Compare commits
14 Commits
| Author | SHA1 | Date | |
|---|---|---|---|
| 12f95cff91 | |||
| 9864632774 | |||
| f982851560 | |||
| 88aeb4d2eb | |||
| 83f3ebd26b | |||
| 13536d6268 | |||
| 7e63035570 | |||
| 44f4146f3e | |||
| b0bd9b69dc | |||
| 09a2d4ed20 | |||
| c8db6118d4 | |||
| 1c64971a52 | |||
| ac14df0625 | |||
| 8c147f5e6e |
@@ -19,8 +19,9 @@ jobs:
|
|||||||
# Public repository archive: no owner credentials or runner SSH keys.
|
# Public repository archive: no owner credentials or runner SSH keys.
|
||||||
url = 'https://gitea.entroia.com/Zayden-Jung/paper-atlas/archive/' + sha + '.tar.gz'
|
url = 'https://gitea.entroia.com/Zayden-Jung/paper-atlas/archive/' + sha + '.tar.gz'
|
||||||
with urllib.request.urlopen(url, timeout=60) as response:
|
with urllib.request.urlopen(url, timeout=60) as response:
|
||||||
archive = response.read(16*1024*1024+1)
|
# Full publication-source fixtures retain original bibliographic evidence.
|
||||||
if len(archive)>16*1024*1024:
|
archive = response.read(24*1024*1024+1)
|
||||||
|
if len(archive)>24*1024*1024:
|
||||||
raise SystemExit('Unexpectedly large source archive')
|
raise SystemExit('Unexpectedly large source archive')
|
||||||
with tempfile.TemporaryDirectory(prefix='paper-atlas-ci-') as tmp:
|
with tempfile.TemporaryDirectory(prefix='paper-atlas-ci-') as tmp:
|
||||||
root = pathlib.Path(tmp)
|
root = pathlib.Path(tmp)
|
||||||
@@ -46,6 +47,57 @@ jobs:
|
|||||||
subprocess.run([python,'-B','-m','pip','install','--no-deps',str(wheel)],check=True)
|
subprocess.run([python,'-B','-m','pip','install','--no-deps',str(wheel)],check=True)
|
||||||
outside = root / 'outside-checkout'
|
outside = root / 'outside-checkout'
|
||||||
outside.mkdir()
|
outside.mkdir()
|
||||||
|
mm_fixture = json.loads(gzip.decompress((src/'tests/fixtures/mm2026-source.json.gz').read_bytes()))
|
||||||
|
mm_snapshots = root/'mm2026-snapshots'
|
||||||
|
for name, body in mm_fixture['files'].items():
|
||||||
|
path = mm_snapshots/name
|
||||||
|
path.parent.mkdir(parents=True,exist_ok=True)
|
||||||
|
path.write_bytes(body.encode('utf-8'))
|
||||||
|
subprocess.run([python,'-I','-B','-m','paper_atlas','collect','mm2026-program','--snapshots',str(mm_snapshots),'--out',str(root/'installed-mm2026')],cwd=outside,check=True)
|
||||||
|
mm_report = json.loads((root/'installed-mm2026/collection-report.json').read_text())
|
||||||
|
assert (mm_report['records'],mm_report['program_observations'],mm_report['extra_presentation_observations']) == (2143,4326,2183)
|
||||||
|
assert len(mm_report['files']) == 28 and mm_report['missing_abstract_count'] == 216
|
||||||
|
assert mm_report['session_count'] == 157 and not mm_report['database_written'] and mm_report['network_requests'] == 0 and mm_report['approved_for_import'] == 0
|
||||||
|
mm_rows = {r['paper_id']:r for f in mm_report['files'] for r in map(json.loads,pathlib.Path(f['file']).read_text().splitlines())}
|
||||||
|
import hashlib
|
||||||
|
assert set(mm_rows) == {g['paper_id'] for g in mm_fixture['independent_inventory']['inventory']}
|
||||||
|
for gold in mm_fixture['independent_inventory']['inventory']:
|
||||||
|
row = mm_rows[gold['paper_id']]
|
||||||
|
assert (row['title'],row['authors'],row['affiliations'],row['track']) == (gold['title'],gold['authors'],gold['affiliations'],gold['track_literal'])
|
||||||
|
assert hashlib.sha256(row['abstract'].encode()).hexdigest() == gold['abstract_sha256']
|
||||||
|
assert [c['source_pointer'] for c in row['contexts']] == gold['locations']
|
||||||
|
assert not row['approved_for_import'] and not row['final_decision_verified'] and 'emails' not in row
|
||||||
|
mm_pub_fixture = json.loads(gzip.decompress((src/'tests/fixtures/mm2026-publications.json.gz').read_bytes()))
|
||||||
|
mm_pub_registry = root/'mm2026-publication-registry'
|
||||||
|
for name,body in mm_pub_fixture['files'].items():
|
||||||
|
path=mm_pub_registry/name
|
||||||
|
path.parent.mkdir(parents=True,exist_ok=True)
|
||||||
|
path.write_bytes(body.encode('utf-8'))
|
||||||
|
subprocess.run([python,'-I','-B','-m','paper_atlas','collect','mm2026-publications','--registry-snapshots',str(mm_pub_registry),'--program-snapshots',str(mm_snapshots),'--parent-snapshot',str(mm_pub_registry/'parent.json'),'--out',str(root/'installed-mm2026-publications')],cwd=outside,check=True)
|
||||||
|
mm_pub_report=json.loads((root/'installed-mm2026-publications/collection-report.json').read_text())
|
||||||
|
assert (mm_pub_report['records'],mm_pub_report['unique_registered_dois'],mm_pub_report['abstracts_from_exact_program_correspondence'],mm_pub_report['output_missing_abstract_count'])==(1918,1918,1743,175)
|
||||||
|
assert mm_pub_report['unmatched_program_count']==391 and not mm_pub_report['database_written'] and mm_pub_report['approved_for_import']==0
|
||||||
|
mm_pub_rows={r['external_id']:r for f in mm_pub_report['files'] for r in map(json.loads,pathlib.Path(f['file']).read_text().splitlines())}
|
||||||
|
assert len(mm_pub_rows)==1918
|
||||||
|
mm_format_gold=json.loads((src/'tests/fixtures/mm2026-title-formatting-review.json').read_text())
|
||||||
|
mm_format_overrides={g['doi']:g for g in [*mm_format_gold['gold'],mm_format_gold['held_gold']]}
|
||||||
|
assert mm_pub_report['abstracts_from_reviewed_title_formatting']==26 and mm_pub_report['abstracts_from_literal_program_correspondence']==1717
|
||||||
|
for original_gold in mm_pub_fixture['independent_gold']:
|
||||||
|
gold=mm_format_overrides.get(original_gold['doi'],original_gold)
|
||||||
|
row=mm_pub_rows[gold['doi']]
|
||||||
|
assert (row['title'],row['authors'],row['track'])==(gold['title'],gold['authors'],gold['track'])
|
||||||
|
assert (hashlib.sha256(row['abstract'].encode()).hexdigest() if row['abstract'] is not None else None)==gold['abstract_sha256']
|
||||||
|
assert row['metadata']['program_match_kind']==gold['kind'] and not row['metadata']['scientific_work_identity_approved'] and not row['metadata']['research_type_classified']
|
||||||
|
if service_file.is_file():
|
||||||
|
subprocess.run([python,'-I','-B','-m','paper_atlas','audit','source-identities','--files',*[f['file'] for f in mm_pub_report['files']],'--pg-service','foundry-research','--pg-schema','research','--out',str(root/'installed-mm2026-publication-preflight')],cwd=outside,env=testenv,check=True)
|
||||||
|
mm_pub_preflight=json.loads((root/'installed-mm2026-publication-preflight/postgresql-preflight.json').read_text())
|
||||||
|
assert mm_pub_preflight['records']==1918 and mm_pub_preflight['transaction_read_only'] and not mm_pub_preflight['database_written'] and mm_pub_preflight['approved_for_import']==0
|
||||||
|
subprocess.run([python,'-I','-B','-m','paper_atlas','audit','source-identities','--mm-program-report',str(root/'installed-mm2026/collection-report.json'),'--pg-service','foundry-research','--pg-schema','research','--out',str(root/'installed-mm2026-preflight')],cwd=outside,env=testenv,check=True)
|
||||||
|
mm_preflight = json.loads((root/'installed-mm2026-preflight/postgresql-preflight.json').read_text())
|
||||||
|
assert mm_preflight['records'] == 2143 and len(mm_preflight['inputs']) == 28
|
||||||
|
assert mm_preflight['transaction_read_only'] and not mm_preflight['database_written'] and mm_preflight['approved_for_import'] == 0
|
||||||
|
assert sum(len(r['program_contexts']) for r in mm_preflight['entries']) == 4326
|
||||||
|
assert all(r['status'] == 'unknown' and not r['identifiers'] and not r['contribution_type_verified'] and not r['publication_identity_verified'] and not r['final_decision_verified'] for r in mm_preflight['entries'])
|
||||||
subprocess.run([str(venv/'bin/paper-atlas'),'--help'],cwd=outside,check=True)
|
subprocess.run([str(venv/'bin/paper-atlas'),'--help'],cwd=outside,check=True)
|
||||||
subprocess.run([python,'-I','-B','-m','paper_atlas','collect','cvf','--help'],cwd=outside,check=True)
|
subprocess.run([python,'-I','-B','-m','paper_atlas','collect','cvf','--help'],cwd=outside,check=True)
|
||||||
subprocess.run([python,'-I','-B','-m','paper_atlas','collect','rss2026-proceedings','--help'],cwd=outside,check=True)
|
subprocess.run([python,'-I','-B','-m','paper_atlas','collect','rss2026-proceedings','--help'],cwd=outside,check=True)
|
||||||
@@ -166,6 +218,21 @@ jobs:
|
|||||||
recsys_rows = [json.loads(line) for manifest in recsys_report['files'] for line in pathlib.Path(manifest['file']).read_text().splitlines()]
|
recsys_rows = [json.loads(line) for manifest in recsys_report['files'] for line in pathlib.Path(manifest['file']).read_text().splitlines()]
|
||||||
gold_keys = recsys_fixture['independent_gold'][0].keys()
|
gold_keys = recsys_fixture['independent_gold'][0].keys()
|
||||||
assert [{k:r['metadata']['program_fields'][k] for k in gold_keys} for r in recsys_rows] == recsys_fixture['independent_gold']
|
assert [{k:r['metadata']['program_fields'][k] for k in gold_keys} for r in recsys_rows] == recsys_fixture['independent_gold']
|
||||||
|
ismir_fixture = json.loads(gzip.decompress((src/'tests/fixtures/ismir2026-source.json.gz').read_bytes()))
|
||||||
|
ismir_snapshots = root/'ismir2026-snapshots'
|
||||||
|
ismir_snapshots.mkdir()
|
||||||
|
for name,encoded in ismir_fixture['files'].items():
|
||||||
|
target = ismir_snapshots/name
|
||||||
|
target.parent.mkdir(parents=True,exist_ok=True)
|
||||||
|
target.write_bytes(base64.b64decode(encoded))
|
||||||
|
subprocess.run([python,'-I','-B','-m','paper_atlas','collect','ismir2026-program','--snapshots',str(ismir_snapshots),'--out',str(root/'installed-ismir2026')],cwd=outside,check=True)
|
||||||
|
ismir_report = json.loads((root/'installed-ismir2026/collection-report.json').read_text())
|
||||||
|
assert (ismir_report['records'],ismir_report['missing_abstract_count'],ismir_report['review_optin_yes'],ismir_report['review_optin_no']) == (140,0,101,39)
|
||||||
|
assert ismir_report['literal_title_version_differences'] == 2 and ismir_report['author_sequence_version_differences'] == 3
|
||||||
|
assert not ismir_report['database_written'] and ismir_report['network_requests'] == 0 and ismir_report['approved_for_import'] == 0
|
||||||
|
ismir_rows = [json.loads(line) for line in pathlib.Path(ismir_report['files'][0]['file']).read_text().splitlines()]
|
||||||
|
assert [(r['external_id'],r['title'],r['authors'],r['abstract']) for r in ismir_rows] == [('2026/'+g['program_id'],g['title'],g['authors'],g['abstract']) for g in ismir_fixture['independent_gold']]
|
||||||
|
assert all(not r['identifiers'] and not r['submission_identifiers'] and r['metadata']['paper_type'] is None for r in ismir_rows)
|
||||||
corl_fixture = json.loads(gzip.decompress((src/'tests/fixtures/corl2026-source.json.gz').read_bytes()))
|
corl_fixture = json.loads(gzip.decompress((src/'tests/fixtures/corl2026-source.json.gz').read_bytes()))
|
||||||
corl_official = root/'corl2026-official'
|
corl_official = root/'corl2026-official'
|
||||||
corl_archive = root/'corl2026-archive'
|
corl_archive = root/'corl2026-archive'
|
||||||
|
|||||||
@@ -131,6 +131,8 @@ ICAPS 2026 也提供独立的 `collect icaps2026-registry`,保留 Crossref 卷
|
|||||||
|
|
||||||
ECCV 2026 的[暂定节目采集器](docs/eccv2026-program.md)合并重复节目 ID,并以不可变缓存断点采集公开元数据;保留摘要原文、来源关键词及未分类 track,始终待身份审核,不自动入库。
|
ECCV 2026 的[暂定节目采集器](docs/eccv2026-program.md)合并重复节目 ID,并以不可变缓存断点采集公开元数据;保留摘要原文、来源关键词及未分类 track,始终待身份审核,不自动入库。
|
||||||
|
|
||||||
|
`collect ismir2026-program` 核验 140 项 ISMIR 官方录用节目及原始摘要,保留两处标题、三处作者名单差异和 101 组自愿公开评审文本;不推断原始 OpenReview ID 或拒稿。见 [ISMIR 2026 使用与恢复说明](docs/ismir2026-program.md)。
|
||||||
|
|
||||||
`collect icassp2026-registry` 从冻结的 10 页 Crossref 元数据核验 4,582 条论文,明确排除 7 条非论文及 56 条其他会议记录;保留摘要缺失和单字段作者姓名,不自动融合节目表。见 [ICASSP 2026 采集与恢复说明](docs/icassp2026-registry.md)。
|
`collect icassp2026-registry` 从冻结的 10 页 Crossref 元数据核验 4,582 条论文,明确排除 7 条非论文及 56 条其他会议记录;保留摘要缺失和单字段作者姓名,不自动融合节目表。见 [ICASSP 2026 采集与恢复说明](docs/icassp2026-registry.md)。
|
||||||
|
|
||||||
`collect sdm2026-proceedings` 核验 53 个 Crossref 论文章节,并单独保留 53 项日程观察及 52 页摘要。作者差异、取消展示和摘要缺失分别保留,不自动拼接来源或绑定作品;见 [SDM 2026 使用与恢复说明](docs/sdm2026-proceedings.md)。
|
`collect sdm2026-proceedings` 核验 53 个 Crossref 论文章节,并单独保留 53 项日程观察及 52 页摘要。作者差异、取消展示和摘要缺失分别保留,不自动拼接来源或绑定作品;见 [SDM 2026 使用与恢复说明](docs/sdm2026-proceedings.md)。
|
||||||
@@ -138,3 +140,7 @@ ECCV 2026 的[暂定节目采集器](docs/eccv2026-program.md)合并重复节目
|
|||||||
`collect icdm2026-program` 离线核验 324 条 ICDM2026 官方线下日程,按明确的研究/应用标签输出 241/83 条;展示长短、缺失摘要及未核实出版身份分别保留,不自动入库。见 [ICDM2026 使用与恢复说明](docs/icdm2026-program.md)。
|
`collect icdm2026-program` 离线核验 324 条 ICDM2026 官方线下日程,按明确的研究/应用标签输出 241/83 条;展示长短、缺失摘要及未核实出版身份分别保留,不自动入库。见 [ICDM2026 使用与恢复说明](docs/icdm2026-program.md)。
|
||||||
|
|
||||||
`collect recsys2026-program` 核验九个 track 的 263 项官方录用贡献观察(260 个不同 DOI),保留摘要、原始作者和重复 DOI 的来源版本。见 [RecSys 2026 使用与恢复说明](docs/recsys2026-program.md)。
|
`collect recsys2026-program` 核验九个 track 的 263 项官方录用贡献观察(260 个不同 DOI),保留摘要、原始作者和重复 DOI 的来源版本。见 [RecSys 2026 使用与恢复说明](docs/recsys2026-program.md)。
|
||||||
|
|
||||||
|
`collect mm2026-program` 从已冻结官方快照恢复全部 28 个来源 track 的 2,143 个节目 ID,保留 4,326 次展示的来源位置与场次上下文,单列 216 个摘要缺失项;输出供贡献类型、出版及作品身份核查使用,不是可直接导入的论文/决定队列。见 [MM 2026 使用、来源和恢复说明](docs/mm2026-source-status.md)。
|
||||||
|
|
||||||
|
`collect mm2026-publications` 核验完整论文集的 1,918 个注册出版 DOI,并单独关联节目分轨、展示上下文和摘要。1,743 份节目摘要明确标注来源(其中 26 份来自已复核的标题排版对应),166 项标题/作者版本差异保留为待核;所有节目及未匹配项均保留。注册文章类型不等于研究类型,输出尚需数据库身份/历史审查。见 [MM 2026 出版条目采集与恢复](docs/mm2026-publications.md)。
|
||||||
|
|||||||
@@ -1,6 +1,6 @@
|
|||||||
# 会议覆盖与公开决定核查
|
# 会议覆盖与公开决定核查
|
||||||
|
|
||||||
最新生产进展和剩余工作见 [2026-11-17 补漏状态](backfill-status-2026-11-17.md):固定 278 个会议年份槽位中,270 个有记录、8 个仍为空;这不代表完整覆盖,第 14 步仍未完成。
|
最新生产验收与剩余工作见 [2026-11-17 补漏状态](backfill-status-2026-11-17.md):当日 `14:06:28 UTC` 固定范围快照中,278 个会议年份槽位有 272 个含记录、6 个仍为空;这不代表完整覆盖,第 14 步仍未完成。后续 MM 规范化、只读候选审计与出版核查不等于生产导入。
|
||||||
|
|
||||||
**当前工作范围:2020 年起主动补漏。** 2000 年起已发现资料保留,恢复工作从 [补漏范围与恢复清单(2026-11-10)](backfill-handoff-2026-11-10.md) 开始;下文旧日期快照不改写,较早年份排除不算补齐。
|
**当前工作范围:2020 年起主动补漏。** 2000 年起已发现资料保留,恢复工作从 [补漏范围与恢复清单(2026-11-10)](backfill-handoff-2026-11-10.md) 开始;下文旧日期快照不改写,较早年份排除不算补齐。
|
||||||
|
|
||||||
@@ -72,7 +72,7 @@ PGSERVICEFILE=/etc/paper-atlas/pg_service.conf paper-atlas coverage \
|
|||||||
|
|
||||||
## 优先缺口
|
## 优先缺口
|
||||||
|
|
||||||
- **UAI 2024–2026**:官方均允许作者选择公开拒稿,当前主库拒稿仍为零。原 notes 接口要求 Challenge 验证,公开镜像尚无已核验来源,必须保留这个真实缺口;不能根据 PMLR 录用名单推断拒稿。参见 [2025 投稿指南](https://www.auai.org/uai2025/submission_instructions)、JSON 内其他年份官方证据。
|
- **UAI 2024–2026**:官方均允许作者选择公开拒稿,首份生产快照拒稿为零。原 notes 接口要求 Challenge 验证;新核验镜像的 2022–2025 年 902 个论坛来自录用子集,不能填补公开拒稿缺口。2024 年的元评审也不能直接转成最终决定,见 [2026-11-17 原始来源与回复结构核查](uai-public-negative-source-status-2026-11-17.md)。不能根据 PMLR 或镜像录用名单反推拒稿;政策参见 [2025 投稿指南](https://www.auai.org/uai2025/submission_instructions)、JSON 内其他年份官方证据。
|
||||||
- **EMNLP 2023**:官方允许公开 opt-in 拒稿,目前只导入 9 条。需继续核验实际公开全集。ARR 评审、conference commitment 与最终会议决定分开;[ARR 作者说明](https://aclrollingreview.org/authors) 明确其自愿捐赠数据与 EMNLP 2023 直接投稿发布的范围不同。
|
- **EMNLP 2023**:官方允许公开 opt-in 拒稿,目前只导入 9 条。需继续核验实际公开全集。ARR 评审、conference commitment 与最终会议决定分开;[ARR 作者说明](https://aclrollingreview.org/authors) 明确其自愿捐赠数据与 EMNLP 2023 直接投稿发布的范围不同。
|
||||||
- **ICLR**:固定快照已观察 2017 与 2020–2026;快照之后,2013/2018/2019 的 1,654 条来源记录已完成独立 PostgreSQL 验收(报告时间 `2026-11-07T19:52:28.378Z`),不回写旧快照计数,也不据此宣称逐届全集已齐。快照中的 10,208 次 unknown 不等于 rejected,近期补采与冲突核对仍在进行;评审推荐不能当最终决定。
|
- **ICLR**:固定快照已观察 2017 与 2020–2026;快照之后,2013/2018/2019 的 1,654 条来源记录已完成独立 PostgreSQL 验收(报告时间 `2026-11-07T19:52:28.378Z`),不回写旧快照计数,也不据此宣称逐届全集已齐。快照中的 10,208 次 unknown 不等于 rejected,近期补采与冲突核对仍在进行;评审推荐不能当最终决定。
|
||||||
- **MM / MLSys**:MM 11,200 条书目记录保留 `proceedings_unspecified`;MLSys 的 448 次官方投稿与 313 条 DBLP 记录重叠。2026 MLSys 的研究/工业轨道在 [官方 CFP](https://mlsys.org/Conferences/2026/CallForResearchPapers) 中分开,仍须核对逐篇轨道证据,不能仅凭会议名全部归 main。
|
- **MM / MLSys**:MM 11,200 条书目记录保留 `proceedings_unspecified`;MLSys 的 448 次官方投稿与 313 条 DBLP 记录重叠。2026 MLSys 的研究/工业轨道在 [官方 CFP](https://mlsys.org/Conferences/2026/CallForResearchPapers) 中分开,仍须核对逐篇轨道证据,不能仅凭会议名全部归 main。
|
||||||
|
|||||||
@@ -33,7 +33,7 @@ SDM 新增事件使 SeBA 和 AdamG 同时保有较早 ICLR 拒稿与后来 SDM
|
|||||||
| ECCV | 2,834 个不同初步节目 ID 及公开摘要已独立采集;1,997 个只读候选历史证据包已保存,尚未入库 | 初步日程不等于 Springer 最终论文集;继续作品/版本和出版核查。见 [来源状态](../eccv2026-source-status.md) |
|
| ECCV | 2,834 个不同初步节目 ID 及公开摘要已独立采集;1,997 个只读候选历史证据包已保存,尚未入库 | 初步日程不等于 Springer 最终论文集;继续作品/版本和出版核查。见 [来源状态](../eccv2026-source-status.md) |
|
||||||
| ICDM | 324 项线下节目观察,明确研究241/应用83;演讲长168/短156,摘要均缺失;规范化和80个候选历史包已核验 | 展示长短不能当成论文类型;需审查作品与正式出版版本,尚未导入主库。见 [采集器说明](../icdm2026-program.md) |
|
| ICDM | 324 项线下节目观察,明确研究241/应用83;演讲长168/短156,摘要均缺失;规范化和80个候选历史包已核验 | 展示长短不能当成论文类型;需审查作品与正式出版版本,尚未导入主库。见 [采集器说明](../icdm2026-program.md) |
|
||||||
| ISMIR | 140 项节目和正式名单均已核验,140份摘要,101组自愿公开评审;两处标题、三处作者名单差异保留,78个候选历史包已核验 | 数字 miniconf ID 不是原始 OpenReview ID;评审意见不转成最终拒稿。继续作品身份/历史关联,尚未入库。采集器见 [PR #63](https://gitea.entroia.com/Zayden-Jung/paper-atlas/pulls/63) |
|
| ISMIR | 140 项节目和正式名单均已核验,140份摘要,101组自愿公开评审;两处标题、三处作者名单差异保留,78个候选历史包已核验 | 数字 miniconf ID 不是原始 OpenReview ID;评审意见不转成最终拒稿。继续作品身份/历史关联,尚未入库。采集器见 [PR #63](https://gitea.entroia.com/Zayden-Jung/paper-atlas/pulls/63) |
|
||||||
| MM | 官方10.8MB日程JSON已冻结,4,326次观察对应2,143个节目ID、28个来源track;重复展示核心字段一致,216个ID缺摘要,其中两项主轨同时缺作者 | 不是2,143篇已核验科研论文。先区分各轨贡献与活动、保留全部上下文并单列缺失;尚未规范化/入库。见 [来源与访问证据](../mm2026-source-status.md) |
|
| MM | 全28轨2143个节目ID/4326展示位置已规范化;814个只读候选历史证据包已核验,24篇候选涉及24次负面投稿/37条来源观察;67项原始arXiv DOI中66项注册元数据一致、1项404暂缓 | 不是2143篇已核验科研论文。继续逐轨贡献/出版依据与作品审查,0身份批准、0入库。ISBN查询与父XML均只返回父记录,不作逐篇清单。见 [来源、审计与恢复证据](../mm2026-source-status.md) |
|
||||||
| CIKM | 官方录用/日程入口与五个分轨页面已保存;五个分轨链接实际是CFP/政策,不是逐篇录用名单 | 查找实际节目或出版社完整目录;TBD和没有静态表不能当作没有论文 |
|
| CIKM | 官方录用/日程入口与五个分轨页面已保存;五个分轨链接实际是CFP/政策,不是逐篇录用名单 | 查找实际节目或出版社完整目录;TBD和没有静态表不能当作没有论文 |
|
||||||
| IROS | 官方论文索引入口的原始HTTP403已保存 | 保留访问限制,不重试绕过;寻找可合法读取、能独立核验的出版元数据 |
|
| IROS | 官方论文索引入口的原始HTTP403已保存 | 保留访问限制,不重试绕过;寻找可合法读取、能独立核验的出版元数据 |
|
||||||
|
|
||||||
@@ -43,6 +43,8 @@ SDM 新增事件使 SeBA 和 AdamG 同时保有较早 ICLR 拒稿与后来 SDM
|
|||||||
|
|
||||||
MM 官方群组元数据可匿名读取,但根据其给出的 submission invitation 和 rejected/withdrawn/desk-rejected venue ID,四次限1条的匿名 notes 核查均为403。查询、时间、原文和哈希已保留,没有认证重试或绕过权限。访问限制不等于零条负面记录;详见 [MM来源证据](../mm2026-source-status.md)。ISMIR公开的101组评审来自录用节目,不能记成101条拒稿。
|
MM 官方群组元数据可匿名读取,但根据其给出的 submission invitation 和 rejected/withdrawn/desk-rejected venue ID,四次限1条的匿名 notes 核查均为403。查询、时间、原文和哈希已保留,没有认证重试或绕过权限。访问限制不等于零条负面记录;详见 [MM来源证据](../mm2026-source-status.md)。ISMIR公开的101组评审来自录用节目,不能记成101条拒稿。
|
||||||
|
|
||||||
|
UAI新增固定镜像已核对2022–2025的902个原始论坛及3053个回复的元数据/字段结构。整理版均标录用,2022保存230条明确Accept决定;2024仅有评审、作者回复、元评审,2023/2025未保存回复,不能据此补拒稿。来源限制和恢复见 [逐年原始核查](uai-public-negative-source-status-2026-11-17.md)。这不是UAI没有公开拒稿的结论,缺口保持未补齐。
|
||||||
|
|
||||||
第14步仍需逐会议、年份、track完成公开拒稿/desk-reject/withdrawn来源核查和可得数据采集;继续核查UAI近年自愿公开范围、EMNLP实际覆盖范围及其余已列明缺口。不能从未见录用、评分、评审推荐、缺少摘要或接口不可访问制造负面决定。
|
第14步仍需逐会议、年份、track完成公开拒稿/desk-reject/withdrawn来源核查和可得数据采集;继续核查UAI近年自愿公开范围、EMNLP实际覆盖范围及其余已列明缺口。不能从未见录用、评分、评审推荐、缺少摘要或接口不可访问制造负面决定。
|
||||||
|
|
||||||
之后仍需完成跨会作品/历次投稿与类型标签关联(第15步),所有导入结束后的独立全库去重、标识冲突、近似标题作者、摘要/约束/查询及覆盖审计(第16步),以及长期运行与恢复文档最终整理(第17步)。阶段性来源入库、测试通过、某年度有记录均不能替代这些验收。
|
之后仍需完成跨会作品/历次投稿与类型标签关联(第15步),所有导入结束后的独立全库去重、标识冲突、近似标题作者、摘要/约束/查询及覆盖审计(第16步),以及长期运行与恢复文档最终整理(第17步)。阶段性来源入库、测试通过、某年度有记录均不能替代这些验收。
|
||||||
|
|||||||
@@ -0,0 +1,51 @@
|
|||||||
|
# UAI 公开负面决定来源:2026-11-17 核查
|
||||||
|
|
||||||
|
**公开拒稿缺口仍未补齐,总体仍为 13/17。** 本页保存新来源的实际范围与恢复线索,避免下次把录用子集误当作全部投稿。主动补漏范围仍从2020年开始;本次镜像只含2022–2025,不能替代其他年份或其他会议的核查。
|
||||||
|
|
||||||
|
## 政策与访问边界
|
||||||
|
|
||||||
|
已有逐年核查记录保存在 [覆盖清单](inventory-2026-11-08.json) 的 UAI `policy_checks`:2023–2026 官方说明允许决定后公开录用及作者选择公开的拒稿,默认私密拒稿不属于公开可采集范围。政策并不证明任何镜像完整。
|
||||||
|
|
||||||
|
- [2023 投稿说明](https://www.auai.org/uai2023/submission_instructions)
|
||||||
|
- [2024 投稿说明](https://auai.org/uai2024/submission_instructions)
|
||||||
|
- [2025 投稿说明](https://www.auai.org/uai2025/submission_instructions)
|
||||||
|
- [2026 投稿说明](https://www.auai.org/uai2026/submission_instructions)
|
||||||
|
|
||||||
|
先前匿名 notes 请求遇到 ChallengeRequiredError/403,仍作为访问限制保留。本轮没有认证、重试或绕过这些入口;读取的是整理者公开发布的固定版本镜像。复制的 `readers`、`signatures` 和邀请字段仍只是镜像内的来源断言,不是对当前原始接口可见性或签名真实性的验证。
|
||||||
|
|
||||||
|
## 实际镜像范围
|
||||||
|
|
||||||
|
来源:[latkes/ai-research-index-iclr-openreview](https://huggingface.co/datasets/latkes/ai-research-index-iclr-openreview/tree/4c9e09931a7814d7500630ffd064b74f1c01bcf3),固定提交 `4c9e09931a7814d7500630ffd064b74f1c01bcf3`。目录中存在 `papers/uai-YYYY/papers.jsonl` 与四个 `raw/uaiYYYY.jsonl`,不能仅凭数据集说明宣称拒稿覆盖。
|
||||||
|
|
||||||
|
| 年份 | 整理版及原始论坛数 | 原始回复对象 | 实际回复角色 | 标准结构化决定字段 |
|
||||||
|
| --- | ---: | ---: | --- | --- |
|
||||||
|
| 2022 | 230 | 1059 | Official_Review 829,Decision 230 | Accept (Poster) 194,Accept (Oral) 36 |
|
||||||
|
| 2023 | 243 | 0 | 没有保存回复 | 没有保存决定记录 |
|
||||||
|
| 2024 | 200 | 1994 | Official_Review 905,Rebuttal 889,Meta_Review 200 | 没有单独 Decision 记录或结构化 decision 字段 |
|
||||||
|
| 2025 | 229 | 0 | 没有保存回复 | 没有保存决定记录 |
|
||||||
|
|
||||||
|
整理版902条均被整理者标为 `accepted=true`;原始根记录的 `venue` 为相应会议名称或 oral/poster/spotlight 文本,不能替代逐次决定证据,更不能算作902条已认证的原始录用决定。这不是四届全部投稿或全部录用的完整性证明。2024年还有 `Edit` 邀请,属于同一批回复上的另一个邀请,不能再累加成1994个新回复。
|
||||||
|
|
||||||
|
为排除字段漏读,对含回复的2022、2024快照另做了全部3053个回复的字段名、包装结构、邀请角色和短决定类字段核查;重新读取的正文哈希与首次捕获完全一致。2022的230条明确字段是录用标签。2024仅保存评审、作者回复和元评审,没有可据此新增的结构化拒稿/撤稿决定。没有从元评审文本、评分、推荐或缺少录用推断负面决定,也没有保存评审正文、作者回复正文或论文全文。
|
||||||
|
|
||||||
|
**这只描述上述固定镜像子集,不证明UAI没有公开拒稿。** 本批没有生成负面决定导入队列,也未写数据库。2020、2021、2026及其他可用来源仍需分别核查;某文件未出现在这个镜像目录中不代表全网不存在。
|
||||||
|
|
||||||
|
## 产物与恢复
|
||||||
|
|
||||||
|
服务器目录均相对 `/data/paper-index/collected/`:
|
||||||
|
|
||||||
|
- `uai2022-2025-mirror-metadata-v1`:902条整理版元数据投影、字段盘点与标签计数。
|
||||||
|
- `uai2022-2025-raw-decisions-v1`:902个原始论坛的元数据、明确决定字段及来源正文/逐行哈希。
|
||||||
|
- `uai2022-2024-reply-schema-v2`:3053个回复的元数据、字段名与短标签;评审正文不保存。角色核查包含直接记录和 `note` 包装,未发生静默遗漏。
|
||||||
|
|
||||||
|
源响应正文仅在内存中读取、校验,保存元数据投影与固定来源URL/字节哈希;不是保留了含评审文本的完整原始快照。复核原始源字节时应使用固定提交,并区分投影文件哈希与源正文哈希。恢复包已在本机独立核验全部902个论坛、3053个回复的元数据结构、逐文件哈希及来源关联。
|
||||||
|
|
||||||
|
| 恢复证据 | SHA256 |
|
||||||
|
| --- | --- |
|
||||||
|
| 整理版元数据恢复包 | 55889fb2b9005260cef1001b2cf36e203f3b7e1deeec0291f058b52eeae880e7 |
|
||||||
|
| 原始元数据/回复结构恢复包 | caade7df899dd3e65572d06e2bf2326fad692cc5e1c1e4ab002fdc6d7823743f |
|
||||||
|
| 原始结构独立核验报告 | 5f0ef2948d0db68fe0cd89eb4c8b57a1088b94b53bed98f180d2b96629746e72 |
|
||||||
|
|
||||||
|
服务器恢复包分别为 `/data/paper-index/validation/uai2022-2025-mirror-metadata-recovery-v1.tar.gz` 与 `uai2022-2025-raw-reply-recovery-v1.tar.gz`。本地同名备份位于 `/Users/zzyang/Documents/Codex/2026-11-06/paper-index/`。先检查实际单元、完整manifest和哈希,再恢复;不要因进度文件旧状态或日志安静重跑。
|
||||||
|
|
||||||
|
后续继续寻找能明确覆盖作者公开拒稿选择的来源,保留每次投稿及决定原文标签、原始ID与镜像关系,核对跨来源重叠和后续录用。UAI缺口不关闭,全部43个目标的逐年公开负面决定核查也不缩减;最终全库去重仍为独立必做阶段。
|
||||||
@@ -0,0 +1,50 @@
|
|||||||
|
# ISMIR 2026 官方节目与摘要采集
|
||||||
|
|
||||||
|
这是长期维护项目中的离线采集器,按需处理冻结的官方元数据,不联网、不执行网页脚本、不下载论文 PDF、不连接数据库。输出仍须经过作品身份与投稿历史审查。
|
||||||
|
|
||||||
|
## 使用
|
||||||
|
|
||||||
|
```sh
|
||||||
|
paper-atlas collect ismir2026-program \
|
||||||
|
--snapshots /data/paper-index/collected \
|
||||||
|
--out /data/paper-index/collected/ismir2026-normalized-<new-run>
|
||||||
|
```
|
||||||
|
|
||||||
|
输入父目录需要三个冻结目录:
|
||||||
|
|
||||||
|
- `remaining2026-listings-research-v2`:清单、`ismir-accepted.html`、`ismir-call.html`。
|
||||||
|
- `ismir2026-script-research-v1`:清单及 `papers.js`,用于证明官方程序明确加载的数据文件和 poster 链接模板;只读取脚本原文。
|
||||||
|
- `ismir2026-json-research-v1`:清单及 `papers.json`,提供 140 项完整元数据和公开摘要。
|
||||||
|
|
||||||
|
固定清单和全部输入原始字节均校验 SHA256、HTTP 回执、官方 URL 及带时区的采集时间。已有输出目录不能覆盖;先检查原任务是否已结束,再选择新目录。失败保留 `failed.json`,只有全批成功才发布 `collection-report.json`。
|
||||||
|
|
||||||
|
## 来源与版本
|
||||||
|
|
||||||
|
[正式录用名单](https://ismir2026.ismir.net/accepted-papers)和[官方虚拟平台](https://ismir2026program.ismir.net/papers.html)各有 140 项,经来源内标题对应核查一致;其中 2 处标题有直引号/弯引号文字差异,3 处作者名单有差异,两个来源的原始写法分别保留并标记,不静默修正或拼接作者。
|
||||||
|
|
||||||
|
作者差异涉及节目 ID `60`、`126`、`433`,包含大小写和姓名完整度问题。正式名单的作者机构可能含括号和分号,采集器只在括号外拆分作者;作者原始字符串也保留。来源内名单对应不代表数据库作品身份已经获准关联。
|
||||||
|
|
||||||
|
- 原始 JSON 摘要全文保留,包括数学符号、换行和 Unicode,不作扁平化提取。
|
||||||
|
- 数字 `id`、`forum` 是 miniconf 节目定位符,不视为原始 OpenReview 投稿 ID;不推断 DOI 或 arXiv ID,投稿标识列表保持空。
|
||||||
|
- `accepted` 来自正式录用名单与节目平台,不是原始签名决定。`track=program_unclassified`,不从主题、session 或远程展示方式推断研究方法、workshop 或论文类型。
|
||||||
|
- `ismir_program:keyword` 保存来源非空关键词,规范空白并去除同条记录内完全重复值;原始关键词列表(含空项)保留在元数据。
|
||||||
|
- `ismir_program:presentation` 保存来源展示方式。Remote/In-person 是展示方式,不是研究类型。
|
||||||
|
- poster URL 严格依据冻结的官方 JS 模板生成;本命令不请求 poster 页面。
|
||||||
|
|
||||||
|
## 公开评审与拒稿边界
|
||||||
|
|
||||||
|
原始来源中有 **101 组明确选择公开的评审材料**;其余 39 项 `publish_reviews=No`,评审正文为空。采集器验证此偏好和正文的一致性,原文保存在来源元数据,不把评审意见或 meta-review 中的推荐转换成额外的最终决定。
|
||||||
|
|
||||||
|
这些记录来自录用节目,不能据此当作 101 条公开拒稿记录。公开拒稿导出是否可得仍记为 unknown,需要单独核查。[征稿与评审流程](https://ismir2026.ismir.net/authors/call-for-papers)中的评审规则也不等于具体投稿决定。
|
||||||
|
|
||||||
|
输出包含一份 140 条 JSONL、部分覆盖清单、七个原始来源文件及报告,批准导入数为零。采集成功不等于会议论文/track/公开负面决定覆盖完整,也不替代最后的全库去重审计。
|
||||||
|
|
||||||
|
## 恢复与验证
|
||||||
|
|
||||||
|
`papers.json` SHA256:`864d384cbc928554504e1b39849137bb380143e247a48f2d71563a698cb8e62d`。
|
||||||
|
|
||||||
|
原始 JSON 和 JS 恢复包:服务器 `/data/paper-index/validation/ismir2026-json-research-recovery-v1.tar.gz`,本地 `/Users/zzyang/Documents/Codex/2026-11-06/paper-index/` 同名文件;SHA256 为 `bba7b236cfe4d8b259d6c1230b3fe09a3e9071f5203884c92b016ca0718fb095`。
|
||||||
|
|
||||||
|
正式名单与政策恢复包:同目录 `remaining2026-listings-research-v2-recovery.tar.gz`,SHA256 为 `f5ef7c2e73d46712146e4d6711ad5ad562e16b8b55c0cd75d1041cd0b81142cb`。安全恢复三个源目录、核验文件后可重跑上述命令。
|
||||||
|
|
||||||
|
完整来源和独立 140 项清单在测试夹具中冻结。测试、构建、打包、安装及实际隔离目录 CLI 验证只在用户指定 `ces-control:host` runner 执行。生产规范化后还需独立核验全部记录,并用 PostgreSQL 只读预检检查候选作品和既往投稿;不自动合并来源差异或人物身份。
|
||||||
@@ -0,0 +1,73 @@
|
|||||||
|
# MM 2026 出版条目采集与恢复
|
||||||
|
|
||||||
|
`collect mm2026-publications` 是包内可复用的冻结快照规范化命令,读取完整注册目录和官方节目,不连接数据库、不请求网络、不下载PDF。总体仍为13/17,来源规范化、测试通过或节目匹配都不能替代身份批准、生产入库或最终去重。
|
||||||
|
|
||||||
|
## 使用
|
||||||
|
|
||||||
|
在已配置的服务器环境中,`--registry-snapshots` 是包含 `mm2026-container-registry-v2/` 的父目录;`--program-snapshots` 是包含 `mm2026-program-and-public-probes-v1/` 的父目录。父论文集正文使用独立固定快照,不能从不存在的文章ISBN字段猜测:
|
||||||
|
|
||||||
|
```sh
|
||||||
|
paper-atlas collect mm2026-publications \
|
||||||
|
--registry-snapshots /data/paper-index/collected \
|
||||||
|
--program-snapshots /data/paper-index/collected \
|
||||||
|
--parent-snapshot /data/paper-index/collected/mm2026-publication-metadata-v1/parent.json \
|
||||||
|
--out /data/paper-index/collected/mm2026-publications-native-v3
|
||||||
|
```
|
||||||
|
|
||||||
|
输出目录必须不存在。失败输出保留 `failed.json`,完整 `collection-report.json` 只在全部校验成功后原子生成;不要覆盖输出或因旧进度文件、日志安静重跑。先观察实际单元/报告,再选择新的恢复目录。
|
||||||
|
|
||||||
|
## 注册目录与节目分别保留
|
||||||
|
|
||||||
|
源目录基于 Crossref [官方精确过滤器](https://www.crossref.org/documentation/retrieve-metadata/rest-api/rest-api-filters/) 的 `container-title` 和出版者member320,实际捕获8页、1918个不同DOI。每页请求、时间、原始正文、字节哈希、总量与唯一DOI均核验。父DOI为 `10.1145/3767308`,题名为 *Proceedings of the 34th ACM International Conference on Multimedia*,明确事件为MM2026;文章均注册为 `proceedings-article`。
|
||||||
|
|
||||||
|
ISBN查询和父Unixref XML只返回父项,不能代表逐篇目录。宽泛 `query.container-title` 的检索总量也不能当会议篇数。当前1918是这个明确论文集过滤范围内的注册条目数,不是已经去重的科研作品数、所有MM workshop或全部28个来源track的出版全集。
|
||||||
|
|
||||||
|
每个出版条目都输出,稳定外部标识为真实DOI。身份标识仅为 `doi`,投稿侧标识仅为 `publication_doi`;不推断原始OpenReview/forum/投稿号或arXiv标识。来源名 `mm_registry_program` 明确表示派生的注册/节目材料,不能把其中所有字段都说成Crossref原文。
|
||||||
|
|
||||||
|
原始注册响应及父记录复制到 `evidence/registry/`,完整节目及四个既有403回执复制到 `evidence/program/`,不重新请求受限入口。每项元数据保留注册正文的文件/哈希、数组位置和规范JSON对象哈希;常用书目字段另作投影,完整引用、资助等原始字段仍在原始响应中,未丢弃也不在每条JSONL重复膨胀。
|
||||||
|
|
||||||
|
## 分轨、摘要与版本差异
|
||||||
|
|
||||||
|
按原始题名的 Unicode/字母数字规范化与完整作者多重集合比对,1726 项有唯一节目对应。原始题名没有候选时,再比较经过受限排版投影的题名:只移除平衡、无属性的 `i`、`scp`、`sup`、`sub` 标签,保留上下标内容,HTML 实体只解码一次。未知或不平衡标签、属性、注释及声明不参与这种投影;不使用模糊题名或人物别名。
|
||||||
|
|
||||||
|
26 项新增对应均保留完整作者多重集合一致、唯一节目 ID 以及逐项复核名单,复制节目原文的摘要、track 和展示上下文。原始注册题名、作者和节目字段继续逐字保留,比较文本单独存入 `program_title_formatting_comparison`;这只证明节目材料对应,不批准科学作品合并。原始题名已有作者分歧时,不通过排版回退绕过。其余 122 项原始作者分歧、1 项排版题名下的作者分歧、43 项题名差异,共 166 项仍在 `proceedings_unclassified`。
|
||||||
|
|
||||||
|
| 输出分区 | 注册条目 |
|
||||||
|
| --- | ---: |
|
||||||
|
| main | 1555 |
|
||||||
|
| grand_challenge | 77 |
|
||||||
|
| interactive_art | 17 |
|
||||||
|
| datasets | 41 |
|
||||||
|
| reproducibility | 1 |
|
||||||
|
| bni | 22 |
|
||||||
|
| open_source | 11 |
|
||||||
|
| doctoral_symposium | 2 |
|
||||||
|
| demo_and_video | 23 |
|
||||||
|
| industry_talk | 3 |
|
||||||
|
| proceedings_unclassified | 166 |
|
||||||
|
|
||||||
|
注册摘要 1918 项仍全缺失。1743 项有明确对应的节目摘要:1717 项使用 `abstract_source=official_program_exact_metadata_correspondence`;新增 26 项使用 `official_program_reviewed_title_formatting_correspondence`,两者都逐字复制节目原文。报告分别列出 `abstracts_from_literal_program_correspondence` 和 `abstracts_from_reviewed_title_formatting`;保留的汇总字段 `abstracts_from_exact_program_correspondence` 为 1743,表示两类非模糊对应之和。
|
||||||
|
|
||||||
|
其余 175 项保持摘要缺失,包括 9 项节目本身没有摘要及 166 个版本待核条目。注册原始摘要仍标为缺失,不能把节目内容说成出版者摘要。
|
||||||
|
|
||||||
|
8 个作者仅有单一姓名字段,保留登记值并标记,不猜测 given/family;37 个含排版标记或实体的题名仍保留字面原文和原始标记。排版比较保留上下标字符,不改写原始字段。
|
||||||
|
|
||||||
|
完整2143个节目ID和4326个展示位置仍在匹配证据或391个未匹配节目中,28个原始track均保留。391个节目与166个出版版本待核来自两种观察口径,不能直接相加当成不同论文。Workshop、艺术、行业等贡献不能因为不在这个论文集范围就被当成拒稿、撤稿或不值得保留。
|
||||||
|
|
||||||
|
注册出版支持来源侧 `status=accepted` 的发表观察,`raw_decision` 明确不是原始签署决定;出版日期仅放在注册元数据中,不伪造决定日期。`mm_registry_program:crossref:type` 与 `mm_registry_program:mm_program:track` 分别保存注册书目类型和节目 track;共同的派生来源前缀满足写入器的来源校验,两个标签命名空间仍相互独立,原始平台字段与证据继续保留在各自元数据中,`research_type_classified=false`:书目类型不是研究类型,行业、艺术、demo条目不能默认变成主轨研究论文。
|
||||||
|
|
||||||
|
## 只读数据库预检与导入边界
|
||||||
|
|
||||||
|
每个分区都有带count/SHA256的manifest,coverage始终为partial。规范化完成后,对实际报告中的全部输出文件运行现有 `audit source-identities --files ... --pg-service foundry-research --pg-schema research --out <新目录>`,不能靠已有66个arXiv/DOI候选代表全目录。预检使用只读一致性快照,并保存来源/标识/历次投稿/逐次决定的证据包;缺失或截断必须单列。
|
||||||
|
|
||||||
|
`approved_for_import=0`,每项 `scientific_work_identity_approved=false`。DOI绑定可能有误,精确题名/作者匹配不证明科学作品身份,没有匹配也不证明新作品。进入增量导入前仍须审查已有绑定、科学版本与候选历史,做新鲜保护检查、完整基线/备份,再使用原生写入者。不得覆盖旧拒稿、后续录用、标签、原始元数据或人物姓名版本;既有已消费队列不可重放。
|
||||||
|
|
||||||
|
## 恢复
|
||||||
|
|
||||||
|
完整注册目录及独立节目比对恢复包:`/data/paper-index/validation/mm2026-container-registry-recovery-v2.tar.gz`,SHA256 `ddb05a8134f6437b3fb40250e45ed6350d55431d953812abda3300871b6dd0ac`。其中含注册源目录、父记录 `validation/mm2026-parent.json`、逐项比对和请求方法证据。安全解包后可把该父记录路径传给 `--parent-snapshot`。
|
||||||
|
|
||||||
|
节目原始来源恢复包见 [MM来源说明](mm2026-source-status.md)。本地独立核验备份位于 `/Users/zzyang/Documents/Codex/2026-11-06/paper-index/`。固定回执/正文哈希变化、总量变化、未知注册上下文或节目对应改变都须重新核查,不能放宽为无条件接受。最终全库去重、近似题名作者、标识冲突、摘要、约束、覆盖与查询审计仍是独立必做阶段。
|
||||||
|
|
||||||
|
实际导入回归在隔离的 SQLite 和 PostgreSQL 测试库中运行全部1918条、全部11个输出分区,核验原始字段、1918个书目类型标签、1752个节目标签、publication_doi 身份、partial覆盖,以及已有作品的规范信息和旧拒稿决定不被覆盖。这些测试不会写入生产 research schema;测试通过仍不替代逐条生产身份审阅。
|
||||||
|
|
||||||
|
旧的 `native-v1`、`native-v2` 来源目录及已消费的导入队列不可覆盖或重放。新版本写入新的输出目录,先核对已导入记录的源字段和分轨是否变化;新增来源关联仍须新鲜身份预检、保护基线和独立验收。程序输出的分区预期数量已改变,未来增量 manifest 必须明确保留旧覆盖父值并声明新预期值,不能重置已导入计数。
|
||||||
@@ -73,4 +73,67 @@
|
|||||||
|
|
||||||
服务器恢复包:`/data/paper-index/validation/mm2026-source-discovery-and-inventory-recovery-v1.tar.gz`;本地备份:`/Users/zzyang/Documents/Codex/2026-11-06/paper-index/` 同名文件。先安全解包并核验回执,再做规范化,不能重新抓取 403 入口作为恢复方式。
|
服务器恢复包:`/data/paper-index/validation/mm2026-source-discovery-and-inventory-recovery-v1.tar.gz`;本地备份:`/Users/zzyang/Documents/Codex/2026-11-06/paper-index/` 同名文件。先安全解包并核验回执,再做规范化,不能重新抓取 403 入口作为恢复方式。
|
||||||
|
|
||||||
本批尚未形成生产导入队列。下一步需要可复用采集器保留每个节目 ID 的全部展示上下文、明确贡献类型与来源边界,单列缺失字段,再做作品/出版身份及既往投稿审查。已消费的其他会议导入队列不可重放;最终全库去重和完整审计仍为独立必做阶段。
|
## 可复用采集命令
|
||||||
|
|
||||||
|
在已配置的服务器环境中运行,`--snapshots` 是包含 `mm2026-program-and-public-probes-v1/` 的父目录;恢复时先安全解包上述已核验恢复包。命令只读快照,不请求网络、不连接数据库,不重新访问403入口,也不下载PDF:
|
||||||
|
|
||||||
|
```sh
|
||||||
|
paper-atlas collect mm2026-program \
|
||||||
|
--snapshots /data/paper-index/collected \
|
||||||
|
--out /data/paper-index/collected/mm2026-native-v1
|
||||||
|
```
|
||||||
|
|
||||||
|
输出目录必须不存在;已存在、失败或部分输出均不能覆盖重跑,应先查看报告,另选新的目录。`collection-report.json` 仅在全量验证成功后生成;失败保留 `failed.json`,不冒充完成。
|
||||||
|
|
||||||
|
采集器核验固定回执、正文哈希、URL、抓取时间、完整逐轨数量和字段缺失盘点。28个逐轨 JSONL 保留原始标题、作者、机构、逐字摘要以及同一节目ID的全部展示位置、时间、poster/oral号和展示kind;157个场次及子场次/非论文活动另存 `sessions.jsonl`。外层空track不覆盖 `paper.track`,不同核心版本、未知字段或未处理嵌套需要重新审查,不能静默丢弃或取第一条。公开联系邮箱数组只留在原始证据中,不进入规范记录。`evidence/` 保留原始节目JSON、完整回执及四个既有403响应。
|
||||||
|
|
||||||
|
这些文件的格式为 `mm_program_observation_v1_not_ingestion_record`,所有项 `approved_for_import=false`。节目ID仅为 `mm_program:program_id` 来源定位符,track标签使用 `mm_program:track`;不建立科学作品、原始投稿ID或最终决定,不推断 DOI/arXiv/OpenReview 标识、论文类型、拒稿或撤稿。行业报告、艺术、workshop及其他来源贡献均保留,尚需逐轨核对贡献类型与正式出版依据。
|
||||||
|
|
||||||
|
本批尚未形成生产导入队列。后续做贡献类型、作品/出版身份及既往投稿审查,再生成经批准的增量队列。已消费的其他会议导入队列不可重放;最终全库去重和完整审计仍为独立必做阶段。
|
||||||
|
|
||||||
|
## 直接审计节目身份与既往历史
|
||||||
|
|
||||||
|
全量节目观察可以直接进入只读候选审计,无需伪造成录用论文队列:
|
||||||
|
|
||||||
|
```sh
|
||||||
|
PGSERVICEFILE=/etc/paper-atlas/pg_service.conf \
|
||||||
|
paper-atlas audit source-identities \
|
||||||
|
--mm-program-report /data/paper-index/collected/mm2026-native-v1/collection-report.json \
|
||||||
|
--pg-service foundry-research --pg-schema research \
|
||||||
|
--out /data/paper-index/collected/mm2026-preflight-v1
|
||||||
|
```
|
||||||
|
|
||||||
|
`--mm-program-report` 与常规 `--files` 互斥。审计先核对冻结原文、全部28轨的逐文件哈希/数量,以及每个节目ID的完整字段和展示上下文;缺轨、重复ID、篡改后的标识/决定或上下文均在数据库连接前拒绝。报告中记录的绝对文件路径仅作来源证据,实际读取报告同目录中的对应文件。
|
||||||
|
|
||||||
|
候选审计中的 `status=unknown` 仅为内部查询占位符,不代表原始提交的最终状态。节目ID以 `mm_program` 来源键查询,不转成全局 DOI、arXiv 或 OpenReview 标识;保留原始 `listed_in_official_program` 观察与全部展示位置。所有贡献类型、出版身份、最终决定核验标志均为false;缺作者的两项也保留并标记。
|
||||||
|
|
||||||
|
查询使用 PostgreSQL 只读一致性快照,候选作品的原始来源、标识、历次投稿和逐次决定压缩保存并记录哈希;截断会显式标记。标题/作者候选匹配不是作品身份批准,没有匹配也不证明是新论文。既往拒稿和后续录用不能互相覆盖,评分或评审推荐不能充当最终决定。输出目录必须不存在;遇到已有输出或静默运行,先检查具体单元和报告,不重启重放。
|
||||||
|
|
||||||
|
## 已完成的规范化、只读审计与出版核对
|
||||||
|
|
||||||
|
采集器提交 `c8db611` 和直接节目审计提交 `09a2d4e` 已通过PR #65、#66由用户合并。授权远程runner的535/543项测试、wheel构建安装和安装后全量命令检查通过;没有本机测试或构建。服务器规范化和只读候选报告也已独立验收,并在本机备份、复核全部2143项与814个候选证据包。
|
||||||
|
|
||||||
|
| 只读候选分类 | 来源节目项数 |
|
||||||
|
| --- | ---: |
|
||||||
|
| 标题与完整作者多重集合一致,仍需作品核查 | 660 |
|
||||||
|
| 仅部分作者重叠 | 153 |
|
||||||
|
| 无精确标题候选,不证明是新作品 | 1329 |
|
||||||
|
| 有标题候选但无作者一致性 | 1 |
|
||||||
|
|
||||||
|
全部4326个展示上下文一致,814个证据包均无截断。24篇候选作品涉及24次已有负面投稿、37条负面来源观察;37不能直接算作37次拒稿。所有候选均尚未批准身份或入库,旧负面决定必须在之后关联时保留。
|
||||||
|
|
||||||
|
对已有作品的原始arXiv记录、标题/完整作者多重集合和MM2026语境进一步筛查,得到67个 Main Track 出版DOI核对候选。独立查询其明确给出的DOI及父论文集,共68次单次请求,67次HTTP200、1次404,没有认证或重试。父注册记录为[第34届ACM Multimedia论文集](https://doi.org/10.1145/3767308),ISBN `9798400722134`,出版者ACM,出版年2026。
|
||||||
|
|
||||||
|
其中66项的原始arXiv引用DOI与注册DOI一致,按既有审计规则规范化的标题、完整作者多重集合、出版年、出版者、明确会议名和论文集标题也一致。注册文章未提供ISBN,不能伪称逐篇ISBN已核验;归属依据是文章明确给出的会议和论文集字段,以及独立的原始arXiv语境和DOI。仍需逐项作品/版本审查及新鲜数据库保护检查,**不是66项已入库或最终身份批准**。
|
||||||
|
|
||||||
|
`mfp7067` 对应既有候选作品2749947,原始arXiv提供的 `10.1145/3767308.3836169` 返回404,暂缓。404不是撤稿/拒稿证据,也不据此删除或修正既有标识;需独立核对后再处理。
|
||||||
|
|
||||||
|
完整会议出版名单仍需继续核对。实际ISBN查询只返回父论文集一个条目,父DOI的公开Unixref XML也只含该父DOI、0个 `conference_paper` 节点;都不能代表逐篇名单。按会议名检索得到的 `total-results` 是出版者全年范围内的宽泛检索量,不能当作MM篇数;必须逐项核对明确会议/论文集字段、节目与分轨归属,并保留未匹配项。全部28个来源track继续保留,不能把本批Main Track候选当成会议完整覆盖。
|
||||||
|
|
||||||
|
| 已独立验证的恢复证据 | SHA256 |
|
||||||
|
| --- | --- |
|
||||||
|
| 全轨规范化恢复包 | b119081618b8ce98037dc139be159512f9fdb86a0ccfb127c8b8a1c5722f3d87 |
|
||||||
|
| 只读候选/814证据包恢复包 | 42d0d0404bba04e64e638344b050ad4e7287c3e4c75a777b7069ecc7d7a6ed60 |
|
||||||
|
| 67项DOI及父论文集原始注册回执恢复包 | dd7a0ff76f12070f403cc7352ac54269e454758d807896a086c58c1c1ac7b798 |
|
||||||
|
|
||||||
|
服务器恢复包位于 `/data/paper-index/validation/`,分别名为 `mm2026-native-recovery-v1.tar.gz`、`mm2026-preflight-recovery-v1.tar.gz`、`mm2026-publication-metadata-recovery-v1.tar.gz`;本地同名备份在 `/Users/zzyang/Documents/Codex/2026-11-06/paper-index/`。来源目录为 `collected/mm2026-native-v1`、`mm2026-preflight-v1`、`mm2026-publication-metadata-v1`。先核对完整manifest及实际单元结果,不能用旧进度文件的running字样重启。其他会议已消费的导入队列仍禁止重放;本轮均为来源核查、只读候选审计,没有生产写入。
|
||||||
|
|||||||
@@ -74,6 +74,74 @@ def load_inputs(files):
|
|||||||
return records, inputs
|
return records, inputs
|
||||||
|
|
||||||
|
|
||||||
|
def load_mm_program_inputs(report_path):
|
||||||
|
"""Read every MM observation without turning a program into import records.
|
||||||
|
|
||||||
|
Reconcile the complete output against its frozen raw source before any DB
|
||||||
|
connection. The synthetic unknown status is an audit placeholder only.
|
||||||
|
"""
|
||||||
|
from paper_atlas.collectors import mm2026
|
||||||
|
|
||||||
|
report_path = Path(report_path).absolute()
|
||||||
|
raw_report = mm2026.safe_read(report_path, 256*1024)
|
||||||
|
report = json.loads(raw_report)
|
||||||
|
if (report.get('source') != 'mm_program' or report.get('records') != 2143
|
||||||
|
or report.get('program_observations') != 4326 or len(report.get('files', [])) != 28
|
||||||
|
or report.get('source_body_sha256') != mm2026.BODY_SHA
|
||||||
|
or report.get('source_manifest_sha256') != mm2026.MANIFEST_SHA
|
||||||
|
or report.get('database_written') is not False or report.get('approved_for_import') != 0):
|
||||||
|
raise ValueError('Unreviewed MM observation report')
|
||||||
|
root = report_path.parent
|
||||||
|
raw_source = mm2026.safe_read(root/'evidence/program-data.json', 12*1024**2)
|
||||||
|
raw_receipts = mm2026.safe_read(root/'evidence/manifest.json', 128*1024)
|
||||||
|
if digest(raw_source) != mm2026.BODY_SHA or digest(raw_receipts) != mm2026.MANIFEST_SHA:
|
||||||
|
raise ValueError('MM raw evidence changed')
|
||||||
|
expected_rows, _ = mm2026.parse_program(raw_source)
|
||||||
|
expected = {r['paper_id']:r for r in expected_rows}
|
||||||
|
records, inputs, seen, tracks = [], [], set(), set()
|
||||||
|
for item in report['files']:
|
||||||
|
track = item.get('track_literal')
|
||||||
|
if (track not in mm2026.TRACK_COUNTS or track in tracks
|
||||||
|
or item.get('format') != 'mm_program_observation_v1_not_ingestion_record'
|
||||||
|
or item.get('approved_for_import') != 0):
|
||||||
|
raise ValueError('Duplicate or unreviewed MM track format')
|
||||||
|
tracks.add(track)
|
||||||
|
# Stored absolute paths are provenance, not instructions to read elsewhere.
|
||||||
|
path = root/Path(item['file']).name
|
||||||
|
if path.suffix != '.jsonl':
|
||||||
|
raise ValueError('Invalid MM observation filename')
|
||||||
|
raw = mm2026.safe_read(path, 16*1024**2)
|
||||||
|
lines = raw.splitlines(keepends=True)
|
||||||
|
if digest(raw) != item.get('file_sha256') or len(lines) != item.get('count') or len(lines) != mm2026.TRACK_COUNTS[track]:
|
||||||
|
raise ValueError('MM observation file count/hash mismatch')
|
||||||
|
for number,line in enumerate(lines, 1):
|
||||||
|
row = json.loads(line)
|
||||||
|
pid = row.get('paper_id')
|
||||||
|
if not isinstance(pid,str) or pid in seen or row != expected.get(pid) or row['track'] != track:
|
||||||
|
raise ValueError('MM observation differs from frozen source or repeats an ID')
|
||||||
|
seen.add(pid)
|
||||||
|
records.append({'source':'mm_program','external_id':'2026/'+pid,
|
||||||
|
'title':row['title'],'authors':row['authors'],'venue':'MM','year':2026,'track':track,
|
||||||
|
'status':'unknown','identifiers':{},'file':str(path),'line':number,
|
||||||
|
'raw_line_sha256':digest(line),'source_record_hits':[],
|
||||||
|
'identifier_hits':[],'title_candidate_ids':[],
|
||||||
|
'input_format':'mm_program_observation_v1_not_ingestion_record',
|
||||||
|
'program_id':pid,'program_contexts':row['contexts'],
|
||||||
|
'source_observed_status':row['observed_status'],
|
||||||
|
'missing_source_fields':row['missing_fields'],
|
||||||
|
'publication_identity_verified':False,'final_decision_verified':False,
|
||||||
|
'contribution_type_verified':False})
|
||||||
|
inputs.append({'file':str(path),'file_sha256':digest(raw),'count':len(lines),
|
||||||
|
'manifest_sha256':digest(raw_report),'manifest':{
|
||||||
|
'source':'mm_program','venue':'MM','year':2026,'track':track,'status':'unknown',
|
||||||
|
'format':'mm_program_observation_v1_not_ingestion_record',
|
||||||
|
'scope':'Read-only lookup input only; unknown is not a source decision.',
|
||||||
|
'source_report':str(report_path),'source_body_sha256':digest(raw_source)}})
|
||||||
|
if seen != expected.keys() or tracks != mm2026.TRACK_COUNTS.keys():
|
||||||
|
raise ValueError('Incomplete MM observation input; all28tracks required')
|
||||||
|
return records, inputs
|
||||||
|
|
||||||
|
|
||||||
def _lookup(connection, values, columns, lateral):
|
def _lookup(connection, values, columns, lateral):
|
||||||
"""Bound scalar VALUES allow the existing text codec to encode each key."""
|
"""Bound scalar VALUES allow the existing text codec to encode each key."""
|
||||||
width = len(columns)
|
width = len(columns)
|
||||||
@@ -187,9 +255,11 @@ def _atomic_json(path, value):
|
|||||||
temporary.replace(path)
|
temporary.replace(path)
|
||||||
|
|
||||||
|
|
||||||
def audit(files, out, *, pg_service, pg_schema):
|
def audit(files, out, *, pg_service, pg_schema, mm_program_report=None):
|
||||||
# Reject input corruption before opening a database connection or reserving output.
|
# Reject input corruption before opening a database connection or reserving output.
|
||||||
records, inputs = load_inputs(files)
|
if bool(files) == bool(mm_program_report):
|
||||||
|
raise ValueError('Choose source files or one MM program report')
|
||||||
|
records, inputs = load_mm_program_inputs(mm_program_report) if mm_program_report else load_inputs(files)
|
||||||
out = Path(out)
|
out = Path(out)
|
||||||
out.mkdir(parents=True, exist_ok=False)
|
out.mkdir(parents=True, exist_ok=False)
|
||||||
(out / 'audit.lock').write_text(timestamp() + '\n')
|
(out / 'audit.lock').write_text(timestamp() + '\n')
|
||||||
@@ -229,6 +299,7 @@ def audit(files, out, *, pg_service, pg_schema):
|
|||||||
'Name agreement does not prove person or scientific-work identity.',
|
'Name agreement does not prove person or scientific-work identity.',
|
||||||
'Bound identifiers can be incorrect; review titles, authors and original evidence.',
|
'Bound identifiers can be incorrect; review titles, authors and original evidence.',
|
||||||
'Source/history limits are explicit in compressed entity packets.',
|
'Source/history limits are explicit in compressed entity packets.',
|
||||||
|
'Program observations and unknown audit status do not establish contribution type, publication or a final decision.',
|
||||||
'Not a final whole-database deduplication or coverage audit.']}
|
'Not a final whole-database deduplication or coverage audit.']}
|
||||||
_atomic_json(out / 'postgresql-preflight.json', report)
|
_atomic_json(out / 'postgresql-preflight.json', report)
|
||||||
return report
|
return report
|
||||||
@@ -243,10 +314,14 @@ def audit(files, out, *, pg_service, pg_schema):
|
|||||||
|
|
||||||
def main(argv=None, *, prog=None):
|
def main(argv=None, *, prog=None):
|
||||||
parser = argparse.ArgumentParser(prog=prog, description=__doc__)
|
parser = argparse.ArgumentParser(prog=prog, description=__doc__)
|
||||||
parser.add_argument('--files', nargs='+', required=True, type=Path)
|
inputs = parser.add_mutually_exclusive_group(required=True)
|
||||||
|
inputs.add_argument('--files', nargs='+', type=Path)
|
||||||
|
inputs.add_argument('--mm-program-report', type=Path,
|
||||||
|
help='Read all frozen MM program observations directly; never generate an import queue')
|
||||||
parser.add_argument('--pg-service', required=True)
|
parser.add_argument('--pg-service', required=True)
|
||||||
parser.add_argument('--pg-schema', required=True)
|
parser.add_argument('--pg-schema', required=True)
|
||||||
parser.add_argument('--out', required=True, type=Path)
|
parser.add_argument('--out', required=True, type=Path)
|
||||||
args = parser.parse_args(argv)
|
args = parser.parse_args(argv)
|
||||||
report = audit(args.files, args.out, pg_service=args.pg_service, pg_schema=args.pg_schema)
|
report = audit(args.files, args.out, pg_service=args.pg_service, pg_schema=args.pg_schema,
|
||||||
|
mm_program_report=args.mm_program_report)
|
||||||
print(json.dumps({k: report[k] for k in ('status', 'records', 'kind_counts', 'approved_for_import')}, ensure_ascii=False))
|
print(json.dumps({k: report[k] for k in ('status', 'records', 'kind_counts', 'approved_for_import')}, ensure_ascii=False))
|
||||||
|
|||||||
@@ -23,6 +23,7 @@ COMMANDS = {
|
|||||||
('collect','colm'): ('paper_atlas.collectors.proceedings', {'source':'colm'}, 'COLM 官方论文集'),
|
('collect','colm'): ('paper_atlas.collectors.proceedings', {'source':'colm'}, 'COLM 官方论文集'),
|
||||||
('collect','neurips'): ('paper_atlas.collectors.proceedings', {'source':'neurips'}, 'NeurIPS 官方论文集'),
|
('collect','neurips'): ('paper_atlas.collectors.proceedings', {'source':'neurips'}, 'NeurIPS 官方论文集'),
|
||||||
('collect','mlsys'): ('paper_atlas.collectors.proceedings', {'source':'mlsys'}, 'MLSys 官方历年标题与摘要'),
|
('collect','mlsys'): ('paper_atlas.collectors.proceedings', {'source':'mlsys'}, 'MLSys 官方历年标题与摘要'),
|
||||||
|
('collect','ismir2026-program'): ('paper_atlas.collectors.ismir2026', {}, '核验 ISMIR 2026 公开节目与录用名单并保留原始摘要、来源差异和自愿公开评审'),
|
||||||
('collect','icaps2026-registry'): ('paper_atlas.collectors.icaps_registry2026', {}, '核验 ICAPS 2026 Crossref 注册元数据并保留独立版本'),
|
('collect','icaps2026-registry'): ('paper_atlas.collectors.icaps_registry2026', {}, '核验 ICAPS 2026 Crossref 注册元数据并保留独立版本'),
|
||||||
('collect','eccv2026-program'): ('paper_atlas.collectors.eccv2026', {}, '核验ECCV2026暂定节目并断点采集公开元数据'),
|
('collect','eccv2026-program'): ('paper_atlas.collectors.eccv2026', {}, '核验ECCV2026暂定节目并断点采集公开元数据'),
|
||||||
('collect','icassp2026-registry'): ('paper_atlas.collectors.icassp2026', {}, '核验 ICASSP 2026 冻结注册论文元数据'),
|
('collect','icassp2026-registry'): ('paper_atlas.collectors.icassp2026', {}, '核验 ICASSP 2026 冻结注册论文元数据'),
|
||||||
@@ -31,6 +32,8 @@ COMMANDS = {
|
|||||||
('collect','icaps2026-program'): ('paper_atlas.collectors.icaps2026', {}, '核验 ICAPS 2026 官方录用名单并分开既有论文展示'),
|
('collect','icaps2026-program'): ('paper_atlas.collectors.icaps2026', {}, '核验 ICAPS 2026 官方录用名单并分开既有论文展示'),
|
||||||
('collect','rss2026-proceedings'): ('paper_atlas.collectors.rss2026', {}, '核验 RSS 2026 官方论文集与独立录用表'),
|
('collect','rss2026-proceedings'): ('paper_atlas.collectors.rss2026', {}, '核验 RSS 2026 官方论文集与独立录用表'),
|
||||||
('collect','recsys2026-program'): ('paper_atlas.collectors.recsys2026', {}, '核验 RecSys 2026 公开录用贡献并保留 DOI 来源版本'),
|
('collect','recsys2026-program'): ('paper_atlas.collectors.recsys2026', {}, '核验 RecSys 2026 公开录用贡献并保留 DOI 来源版本'),
|
||||||
|
('collect','mm2026-program'): ('paper_atlas.collectors.mm2026', {}, '恢复 MM 2026 全部来源 track 与重复展示上下文,待贡献和身份核查'),
|
||||||
|
('collect','mm2026-publications'): ('paper_atlas.collectors.mm_publications2026', {}, '核验 MM 2026 全部注册出版条目并分别保留节目分轨、摘要与版本差异'),
|
||||||
('collect','cvf'): ('paper_atlas.collectors.vision', {'source':'cvf'}, 'CVPR / ICCV 官方开放目录'),
|
('collect','cvf'): ('paper_atlas.collectors.vision', {'source':'cvf'}, 'CVPR / ICCV 官方开放目录'),
|
||||||
('collect','ecva'): ('paper_atlas.collectors.vision', {'source':'ecva'}, 'ECCV 官方开放目录'),
|
('collect','ecva'): ('paper_atlas.collectors.vision', {'source':'ecva'}, 'ECCV 官方开放目录'),
|
||||||
('collect','aamas-proceedings'): ('paper_atlas.collectors.aamas', {}, '按官方目录区分 AAMAS 发表形式和摘要'),
|
('collect','aamas-proceedings'): ('paper_atlas.collectors.aamas', {}, '按官方目录区分 AAMAS 发表形式和摘要'),
|
||||||
|
|||||||
@@ -0,0 +1,208 @@
|
|||||||
|
"""Frozen ISMIR 2026 accepted program, abstracts and opt-in review provenance."""
|
||||||
|
import argparse
|
||||||
|
from collections import Counter
|
||||||
|
from datetime import datetime
|
||||||
|
import hashlib
|
||||||
|
import json
|
||||||
|
from pathlib import Path
|
||||||
|
import re
|
||||||
|
import unicodedata
|
||||||
|
|
||||||
|
from paper_atlas.io.html import Document, Node, clean
|
||||||
|
|
||||||
|
SOURCE = 'ismir_program'
|
||||||
|
TRACK = 'program_unclassified'
|
||||||
|
PAPERS = 'https://ismir2026program.ismir.net/papers.html'
|
||||||
|
ACCEPTED = 'https://ismir2026.ismir.net/accepted-papers'
|
||||||
|
SNAPSHOTS = {
|
||||||
|
'remaining2026-listings-research-v2': ('fea7b710587636db06794ee195cd083634c41e9054a217859d9dd0ccdd1afab9',
|
||||||
|
'research_completed_not_coverage_or_import_approval', {
|
||||||
|
'ismir-accepted.html': ACCEPTED,
|
||||||
|
'ismir-call.html': 'https://ismir2026.ismir.net/authors/call-for-papers'}),
|
||||||
|
'ismir2026-script-research-v1': ('5885b4bcfc2fc65374e5dc7a0d74f4eb86a0a36b8689f9fe8edbd10d49a08794',
|
||||||
|
'public_script_research_only', {'papers.js': 'https://ismir2026program.ismir.net/static/js/papers.js'}),
|
||||||
|
'ismir2026-json-research-v1': ('8be3e23277c199720264d940917118572655d18c0f9ac6db1b9cd30a6561b34d',
|
||||||
|
'public_metadata_json_research_only', {'papers.json': 'https://ismir2026program.ismir.net/papers.json'}),
|
||||||
|
}
|
||||||
|
REVIEW_KEYS = {'meta_review','publish_reviews','review1','review2','review3','review4'}
|
||||||
|
|
||||||
|
|
||||||
|
def digest(raw):
|
||||||
|
return hashlib.sha256(raw).hexdigest()
|
||||||
|
|
||||||
|
|
||||||
|
def one(items):
|
||||||
|
if len(items) != 1:
|
||||||
|
raise ValueError('Expected exactly one reviewed source field')
|
||||||
|
return items[0]
|
||||||
|
|
||||||
|
|
||||||
|
def title_key(text):
|
||||||
|
return ''.join(c for c in unicodedata.normalize('NFKD', text).casefold() if c.isalnum())
|
||||||
|
|
||||||
|
|
||||||
|
def load_snapshots(root):
|
||||||
|
root = Path(root)
|
||||||
|
if root.is_symlink():
|
||||||
|
raise ValueError('Unsafe snapshot root')
|
||||||
|
result = {}
|
||||||
|
for name, (expected, status, urls) in SNAPSHOTS.items():
|
||||||
|
directory = root/name; path = directory/'manifest.json'
|
||||||
|
if directory.is_symlink() or path.is_symlink() or path.stat().st_size > 128*1024:
|
||||||
|
raise ValueError('Unsafe snapshot manifest')
|
||||||
|
raw = path.read_bytes()
|
||||||
|
if digest(raw) != expected:
|
||||||
|
raise ValueError('Reviewed manifest hash changed')
|
||||||
|
manifest = json.loads(raw)
|
||||||
|
if manifest['status'] != status or manifest['database_written'] is not False:
|
||||||
|
raise ValueError('Unreviewed snapshot scope')
|
||||||
|
files = {'manifest.json': raw}; receipts = {}
|
||||||
|
for filename, url in urls.items():
|
||||||
|
receipt = one([r for r in manifest['entries'] if r.get('file') == filename])
|
||||||
|
path = directory/filename
|
||||||
|
if path.is_symlink() or path.stat().st_size > 8*1024**2:
|
||||||
|
raise ValueError('Unsafe snapshot source file')
|
||||||
|
body = path.read_bytes()
|
||||||
|
if (receipt['url'] != url or receipt['final_url'] != url or receipt['http_status'] != 200
|
||||||
|
or receipt['bytes'] != len(body) or receipt['sha256'] != digest(body)
|
||||||
|
or datetime.fromisoformat(receipt['at']).utcoffset() is None):
|
||||||
|
raise ValueError('Source receipt/body disagree')
|
||||||
|
files[filename] = body; receipts[filename] = receipt
|
||||||
|
result[name] = (files, receipts)
|
||||||
|
return result
|
||||||
|
|
||||||
|
|
||||||
|
def author_names(literal):
|
||||||
|
# Institution names can contain semicolons; split only outside parentheses.
|
||||||
|
parts = []; depth = 0; start = 0
|
||||||
|
for index, char in enumerate(literal):
|
||||||
|
if char == '(': depth += 1
|
||||||
|
elif char == ')':
|
||||||
|
depth -= 1
|
||||||
|
if depth < 0: raise ValueError('Unbalanced author affiliation')
|
||||||
|
elif char == ';' and depth == 0:
|
||||||
|
parts.append(literal[start:index]); start = index+1
|
||||||
|
if depth:
|
||||||
|
raise ValueError('Unbalanced author affiliation')
|
||||||
|
parts.append(literal[start:])
|
||||||
|
names = [clean(p.split('(', 1)[0]) for p in parts]
|
||||||
|
if any(not name for name in names):
|
||||||
|
raise ValueError('Missing accepted-list author')
|
||||||
|
return names
|
||||||
|
|
||||||
|
|
||||||
|
def accepted_listing(raw):
|
||||||
|
doc = Document(raw.decode('utf-8')); tables = []
|
||||||
|
for table in doc.root.walk():
|
||||||
|
if table.tag != 'table': continue
|
||||||
|
rows = []
|
||||||
|
for row in table.walk():
|
||||||
|
if row.tag == 'tr':
|
||||||
|
rows.append([clean(n.text()) for n in row.children if isinstance(n, Node) and n.tag in ('th','td')])
|
||||||
|
if rows and rows[0] == ['Paper Title', 'Authors']:
|
||||||
|
tables.append(rows)
|
||||||
|
rows = one(tables)[1:]; result = {}
|
||||||
|
if len(rows) != 140:
|
||||||
|
raise ValueError('Incomplete official accepted-list census')
|
||||||
|
for row in rows:
|
||||||
|
if len(row) != 2 or not row[0] or not row[1] or title_key(row[0]) in result:
|
||||||
|
raise ValueError('Missing or ambiguous accepted-list title')
|
||||||
|
result[title_key(row[0])] = {'title': row[0], 'authors_literal': row[1], 'author_names': author_names(row[1])}
|
||||||
|
return result
|
||||||
|
|
||||||
|
|
||||||
|
def program_observations(rows, listing):
|
||||||
|
if not isinstance(rows, list) or len(rows) != 140:
|
||||||
|
raise ValueError('Incomplete public program census')
|
||||||
|
observations = []; seen = set(); used = set()
|
||||||
|
for original in rows:
|
||||||
|
if not isinstance(original, dict) or not isinstance(original.get('content'), dict):
|
||||||
|
raise ValueError('Invalid public program record')
|
||||||
|
pid = original.get('id'); content = original.get('content', {})
|
||||||
|
if not isinstance(pid, str) or not re.fullmatch(r'[1-9][0-9]*', pid) or pid in seen or original.get('forum') != pid:
|
||||||
|
raise ValueError('Missing/repeated numeric program locator, not an OpenReview ID')
|
||||||
|
seen.add(pid); title = content.get('title'); authors = content.get('authors'); abstract = content.get('abstract')
|
||||||
|
if (not isinstance(title, str) or not clean(title) or not isinstance(abstract, str) or not abstract.strip()
|
||||||
|
or not isinstance(authors, list) or not authors or any(not isinstance(a, str) or not clean(a) for a in authors)):
|
||||||
|
raise ValueError('Missing program title, author sequence or abstract')
|
||||||
|
key = title_key(title)
|
||||||
|
if key not in listing or key in used:
|
||||||
|
raise ValueError('Program/accepted-list title membership mismatch')
|
||||||
|
used.add(key); accepted = listing[key]
|
||||||
|
keywords = content.get('keywords'); presentation = content.get('paper_presentation')
|
||||||
|
session = content.get('session'); reviews = original.get('reviews')
|
||||||
|
if (not isinstance(keywords, list) or any(not isinstance(k, str) for k in keywords)
|
||||||
|
or presentation not in ('Remote','In-person','Remote during oral, In-person during nominee')
|
||||||
|
or not isinstance(session, list) or len(session) != 1 or not isinstance(session[0], str) or session[0] != original.get('session')
|
||||||
|
or not re.fullmatch(r'(?:[1-9]|1[0-4])', session[0])):
|
||||||
|
raise ValueError('Unreviewed source keywords, presentation or session')
|
||||||
|
if (not isinstance(reviews, dict) or set(reviews) != REVIEW_KEYS
|
||||||
|
or any(not isinstance(v, str) for v in reviews.values())
|
||||||
|
or reviews['publish_reviews'] not in ('Yes','No')):
|
||||||
|
raise ValueError('Unreviewed opt-in review structure')
|
||||||
|
review_text = [reviews[k] for k in REVIEW_KEYS-{'publish_reviews'}]
|
||||||
|
if (reviews['publish_reviews'] == 'No' and any(review_text)) or (reviews['publish_reviews'] == 'Yes' and not any(review_text)):
|
||||||
|
raise ValueError('Review publication preference/body disagree')
|
||||||
|
flags = []
|
||||||
|
if title != accepted['title']: flags.append('program_and_accepted_title_literal_versions_differ')
|
||||||
|
if authors != accepted['author_names']: flags.append('program_and_accepted_author_sequence_versions_differ')
|
||||||
|
observations.append({'original': original, 'accepted_listing': accepted, 'review_flags': flags})
|
||||||
|
if (used != listing.keys() or sum('program_and_accepted_title_literal_versions_differ' in o['review_flags'] for o in observations) != 2
|
||||||
|
or {o['original']['id'] for o in observations if 'program_and_accepted_author_sequence_versions_differ' in o['review_flags']} != {'60','126','433'}
|
||||||
|
or Counter(o['original']['reviews']['publish_reviews'] for o in observations) != {'Yes':101,'No':39}):
|
||||||
|
raise ValueError('Unreviewed source-version or opt-in census')
|
||||||
|
return observations
|
||||||
|
|
||||||
|
|
||||||
|
def collect(out, *, snapshots):
|
||||||
|
out = Path(out)
|
||||||
|
if out.exists(): raise ValueError('Output exists; inspect previous operation before retrying')
|
||||||
|
out.mkdir(parents=True)
|
||||||
|
try:
|
||||||
|
sources = load_snapshots(snapshots)
|
||||||
|
official, _ = sources['remaining2026-listings-research-v2']; script, _ = sources['ismir2026-script-research-v1']
|
||||||
|
program, receipts = sources['ismir2026-json-research-v1']; receipt = receipts['papers.json']
|
||||||
|
if b"d3.json('papers.json')" not in script['papers.js'] or b'href="poster_${openreview.id}.html"' not in script['papers.js']:
|
||||||
|
raise ValueError('Unreviewed public program data path or poster link template')
|
||||||
|
observations = program_observations(json.loads(program['papers.json']), accepted_listing(official['ismir-accepted.html']))
|
||||||
|
evidence = out/'evidence'; evidence.mkdir()
|
||||||
|
for name, (files, _) in sources.items():
|
||||||
|
directory = evidence/name; directory.mkdir()
|
||||||
|
for filename, body in files.items(): (directory/filename).write_bytes(body)
|
||||||
|
rows = []
|
||||||
|
for observation in observations:
|
||||||
|
original = observation['original']; content = original['content']; pid = original['id']
|
||||||
|
tags = [{'namespace':SOURCE+':presentation','value':content['paper_presentation']}]
|
||||||
|
tags += [{'namespace':SOURCE+':keyword','value':value} for value in dict.fromkeys(clean(k) for k in content['keywords'] if clean(k))]
|
||||||
|
rows.append({'source':SOURCE,'external_id':'2026/'+pid,'url':'https://ismir2026program.ismir.net/poster_'+pid+'.html',
|
||||||
|
'title':content['title'],'authors':content['authors'],'abstract':content['abstract'],
|
||||||
|
'identifiers':{},'submission_identifiers':[],'venue':'ISMIR','year':2026,'track':TRACK,'status':'accepted',
|
||||||
|
'raw_decision':'Present in official ISMIR2026 accepted list and public virtual program; not an original signed decision.',
|
||||||
|
'evidence_url':ACCEPTED,'observed_at':receipt['at'],'tags':tags,
|
||||||
|
'metadata':{'program_entry':original,'accepted_listing':observation['accepted_listing'],'review_flags':observation['review_flags'],
|
||||||
|
'program_receipt':receipt,'source_evidence':{'file':'evidence/ismir2026-json-research-v1/papers.json','sha256':receipt['sha256'],'url':receipt['url']},
|
||||||
|
'original_decision_signature_verified':False,'paper_type':None,
|
||||||
|
'id_scope':'Numeric miniconf id/forum fields are program locators, not original OpenReview submission IDs; no inferred DOI or arxiv ID.',
|
||||||
|
'track_scope':'Unclassified regular accepted-list program; no workshop or research-method class inferred from topic/session/remote presentation.',
|
||||||
|
'review_scope':'101 source-opt-in review sets retained as raw public text;39No contain no review text. Reviewer recommendations/meta-review prose are not separately inferred final negative decisions.',
|
||||||
|
'abstract_transform':'Original JSON abstract string retained verbatim, including math, newlines and Unicode; no markup flattening or PDF extraction.',
|
||||||
|
'program_listing_scope':'Source versions remain separate.2title literal differences and3author-list differences flagged, not silently fused or resolved.',
|
||||||
|
'scope':'Pending work/history review. Poster URL follows frozen official JS template; poster pages not downloaded. No work merge, aliases, negative decision or final coverage/dedup claim.'}})
|
||||||
|
path = out/'ismir-program-2026-unclassified.jsonl';raw=''.join(json.dumps(r,ensure_ascii=False)+'\n' for r in rows).encode();path.write_bytes(raw)
|
||||||
|
manifest={'source':SOURCE,'parser_version':1,'venue':'ISMIR','year':2026,'track':TRACK,'count':140,'file':str(path),
|
||||||
|
'file_sha256':digest(raw),'list_url':ACCEPTED,'status':'partial','missing_abstract_count':0,'source_expected_count':140}
|
||||||
|
path.with_suffix('.manifest.json').write_text(json.dumps(manifest,indent=2)+'\n')
|
||||||
|
report={'status':'collected_pending_identity_review','records':140,'files':[manifest],'missing_abstract_count':0,
|
||||||
|
'literal_title_version_differences':2,'author_sequence_version_differences':3,'review_optin_yes':101,'review_optin_no':39,
|
||||||
|
'database_written':False,'network_requests':0,'pdfs_downloaded':False,'approved_for_import':0,
|
||||||
|
'public_rejected_availability':'unknown_accepted_program_reviews_are_not_a_rejected_export',
|
||||||
|
'scope':'140accepted observations;101opt-in review sets, not101rejected records. Source keyword/presentation tags are namespace-scoped; no original OpenReview ID assumption or source-version fusion.'}
|
||||||
|
part=out/'collection-report.json.part';part.write_text(json.dumps(report,ensure_ascii=False,indent=2)+'\n');part.replace(out/'collection-report.json');return report
|
||||||
|
except Exception as error:
|
||||||
|
(out/'failed.json').write_text(json.dumps({'status':'failed_not_complete','error_type':type(error).__name__,'database_written':False})+'\n')
|
||||||
|
raise
|
||||||
|
|
||||||
|
|
||||||
|
def main(argv=None, *, prog=None):
|
||||||
|
parser=argparse.ArgumentParser(prog=prog,description=__doc__);parser.add_argument('--snapshots',required=True);parser.add_argument('--out',required=True)
|
||||||
|
args=parser.parse_args(argv);print(json.dumps(collect(args.out,snapshots=args.snapshots),ensure_ascii=False))
|
||||||
@@ -0,0 +1,202 @@
|
|||||||
|
"""Restore frozen ACM MM 2026 program observations, not publication decisions."""
|
||||||
|
import argparse
|
||||||
|
from collections import Counter
|
||||||
|
from datetime import datetime
|
||||||
|
import hashlib
|
||||||
|
import json
|
||||||
|
from pathlib import Path
|
||||||
|
import re
|
||||||
|
|
||||||
|
SOURCE = 'mm_program'
|
||||||
|
SNAPSHOT = 'mm2026-program-and-public-probes-v1'
|
||||||
|
URL = 'https://2026.acmmm.org/site/program-data.json'
|
||||||
|
BODY_SHA = '08e2976ecba2d0d535000b0935f41c85076ec674e518a1df2c5843d9bde4d083'
|
||||||
|
MANIFEST_SHA = 'e43ac6cdfc86bac61305553a044dd6148800b5aa41a5d1bf06c486e90d0ed7b8'
|
||||||
|
TRACK_COUNTS = {'MRAC':35,'SUMAC':9,'NeuroMM':10,'IMPACT':8,'RichMediaGAI':10,
|
||||||
|
'MMSports':16,'MUWS':5,'EReL-MIR':7,'GreenMM':10,'DFF':12,'Main Track':1703,
|
||||||
|
'Grand Challenge':97,'Doctoral Symposium':4,'Datasets':50,'Interactive Art':23,
|
||||||
|
'Reproducibility':1,'Demo and Video':27,'Open Source':14,'Industry Talk':7,
|
||||||
|
'BNI':25,'QoEVMA':4,'HUMA':14,'WIPAI':8,'AMI':11,'Mulsemedia':9,
|
||||||
|
'UAVM':11,'MCHM':7,'JAV-CG':6}
|
||||||
|
CORE = ('paper_id','title','authors','affiliations','abstract','track')
|
||||||
|
PRESENTATION = ('kind','start_time','end_time','poster_id','oral_id')
|
||||||
|
|
||||||
|
|
||||||
|
def digest(raw):
|
||||||
|
return hashlib.sha256(raw).hexdigest()
|
||||||
|
|
||||||
|
|
||||||
|
def safe_read(path, limit):
|
||||||
|
if any(p.is_symlink() for p in (path, *path.parents)) or not path.is_file() or path.stat().st_size > limit:
|
||||||
|
raise ValueError('Unsafe snapshot path or size')
|
||||||
|
return path.read_bytes()
|
||||||
|
|
||||||
|
|
||||||
|
def load_snapshot(root):
|
||||||
|
directory = Path(root)/SNAPSHOT
|
||||||
|
raw = safe_read(directory/'manifest.json', 128*1024)
|
||||||
|
if digest(raw) != MANIFEST_SHA:
|
||||||
|
raise ValueError('Reviewed manifest hash changed')
|
||||||
|
manifest = json.loads(raw)
|
||||||
|
if manifest.get('database_written') is not False or len(manifest['entries']) != 5:
|
||||||
|
raise ValueError('Unreviewed manifest scope')
|
||||||
|
evidence = {'manifest.json':raw}
|
||||||
|
for receipt in manifest['entries']:
|
||||||
|
name = receipt['file']
|
||||||
|
if Path(name).name != name or name in evidence:
|
||||||
|
raise ValueError('Unsafe or duplicate receipt name')
|
||||||
|
body = safe_read(directory/name, 12*1024**2)
|
||||||
|
if (receipt['bytes'] != len(body) or receipt['sha256'] != digest(body)
|
||||||
|
or receipt['url'] != receipt['final_url']
|
||||||
|
or not receipt['content_type'].startswith('application/json')
|
||||||
|
or datetime.fromisoformat(receipt['at']).utcoffset() is None):
|
||||||
|
raise ValueError('Snapshot receipt/body disagree')
|
||||||
|
if name == 'program-data.json':
|
||||||
|
if receipt['url'] != URL or receipt['http_status'] != 200 or digest(body) != BODY_SHA:
|
||||||
|
raise ValueError('Wrong program source')
|
||||||
|
elif receipt['http_status'] != 403:
|
||||||
|
raise ValueError('Unreviewed public-access result')
|
||||||
|
evidence[name] = body
|
||||||
|
if 'program-data.json' not in evidence:
|
||||||
|
raise ValueError('Missing program source')
|
||||||
|
return evidence
|
||||||
|
|
||||||
|
|
||||||
|
def without_emails(value):
|
||||||
|
"""Keep public session structure; contact arrays remain in raw evidence only."""
|
||||||
|
if isinstance(value, dict):
|
||||||
|
return {k:without_emails(v) for k,v in value.items() if k != 'emails'}
|
||||||
|
if isinstance(value, list):
|
||||||
|
return [without_emails(v) for v in value]
|
||||||
|
return value
|
||||||
|
|
||||||
|
|
||||||
|
def program_locations(node, pointer=''):
|
||||||
|
"""Independent structural guard against silently skipping a new nesting form."""
|
||||||
|
if isinstance(node, dict):
|
||||||
|
if 'paper_id' in node:
|
||||||
|
yield pointer
|
||||||
|
for key,value in node.items():
|
||||||
|
escaped = key.replace('~','~0').replace('/','~1')
|
||||||
|
yield from program_locations(value, pointer+'/'+escaped)
|
||||||
|
elif isinstance(node, list):
|
||||||
|
for i,value in enumerate(node):
|
||||||
|
yield from program_locations(value, pointer+'/'+str(i))
|
||||||
|
|
||||||
|
|
||||||
|
def parse_program(raw):
|
||||||
|
data = json.loads(raw)
|
||||||
|
if (data.get('generated_at') != '2026-11-14T18:36:26.860677+00:00'
|
||||||
|
or data.get('days') != [1,2,3,4,5] or len(data.get('rooms',[])) != 20
|
||||||
|
or len(data.get('entries',[])) != 157):
|
||||||
|
raise ValueError('Unreviewed program edition or census')
|
||||||
|
records = {}; sessions = []; seen = []
|
||||||
|
|
||||||
|
def observe(paper, pointer, session_pointer, subsession_pointer=None):
|
||||||
|
if not set(CORE) <= paper.keys() or set(paper)-set(CORE)-set(PRESENTATION)-{'emails'}:
|
||||||
|
raise ValueError('Unreviewed program fields')
|
||||||
|
for key in ('paper_id','title','abstract','track'):
|
||||||
|
if not isinstance(paper[key],str):
|
||||||
|
raise ValueError('Wrong program field type')
|
||||||
|
for key in ('authors','affiliations'):
|
||||||
|
if not isinstance(paper[key],list) or any(not isinstance(v,str) for v in paper[key]):
|
||||||
|
raise ValueError('Wrong author or affiliation list')
|
||||||
|
if not paper['paper_id'].strip() or not paper['title'].strip() or paper['track'] not in TRACK_COUNTS:
|
||||||
|
raise ValueError('Missing identity/title or unreviewed source track')
|
||||||
|
core = {k:paper[k] for k in CORE}
|
||||||
|
pid = paper['paper_id']; seen.append(pointer)
|
||||||
|
if pid not in records:
|
||||||
|
missing = [k for k in ('authors','abstract') if not paper[k]]
|
||||||
|
records[pid] = {**core,'source':SOURCE,'year':2026,'venue':'MM',
|
||||||
|
'source_locator':{'namespace':'mm_program:program_id','value':pid},
|
||||||
|
'observed_status':'listed_in_official_program',
|
||||||
|
'publication_identity_verified':False,'final_decision_verified':False,
|
||||||
|
'approved_for_import':False,'missing_fields':missing,
|
||||||
|
'tags':[{'namespace':'mm_program:track','value':paper['track']}],
|
||||||
|
'contexts':[]}
|
||||||
|
elif {k:records[pid][k] for k in CORE} != core:
|
||||||
|
raise ValueError('Conflicting core versions for one program ID; review required')
|
||||||
|
records[pid]['contexts'].append({'source_pointer':pointer,'session_pointer':session_pointer,
|
||||||
|
'subsession_pointer':subsession_pointer,
|
||||||
|
'presentation':{k:paper[k] for k in PRESENTATION if k in paper}})
|
||||||
|
|
||||||
|
for i,entry in enumerate(data['entries']):
|
||||||
|
pointer = '/entries/'+str(i)
|
||||||
|
if entry['track'] != '':
|
||||||
|
raise ValueError('Unreviewed outer track; do not override paper.track')
|
||||||
|
session = without_emails({k:v for k,v in entry.items() if k not in ('papers','subsessions')})
|
||||||
|
session.update(source_pointer=pointer,subsessions=[])
|
||||||
|
for j,paper in enumerate(entry.get('papers') or []):
|
||||||
|
observe(paper,pointer+'/papers/'+str(j),pointer)
|
||||||
|
for j,sub in enumerate(entry.get('subsessions') or []):
|
||||||
|
sp = pointer+'/subsessions/'+str(j)
|
||||||
|
saved = without_emails({k:v for k,v in sub.items() if k != 'items'})
|
||||||
|
saved.update(source_pointer=sp,nonpaper_items=[])
|
||||||
|
for k,item in enumerate(sub.get('items') or []):
|
||||||
|
ip = sp+'/items/'+str(k)
|
||||||
|
if 'paper_id' in item:
|
||||||
|
observe(item,ip,pointer,sp)
|
||||||
|
else:
|
||||||
|
saved['nonpaper_items'].append({'source_pointer':ip,'fields':without_emails(item)})
|
||||||
|
session['subsessions'].append(saved)
|
||||||
|
sessions.append(session)
|
||||||
|
if set(seen) != set(program_locations(data)) or len(set(seen)) != len(seen):
|
||||||
|
raise ValueError('Unhandled or duplicated program location')
|
||||||
|
items = list(records.values())
|
||||||
|
if (len(seen) != 4326 or len(items) != 2143
|
||||||
|
or Counter(r['track'] for r in items) != TRACK_COUNTS
|
||||||
|
or sum(not r['abstract'] for r in items) != 216
|
||||||
|
or {r['paper_id'] for r in items if not r['authors']} != {'mfp4971','mfp7801'}):
|
||||||
|
raise ValueError('Program count or missing-field census changed')
|
||||||
|
return items,sessions
|
||||||
|
|
||||||
|
|
||||||
|
def collect(out, *, snapshots):
|
||||||
|
out = Path(out)
|
||||||
|
if out.exists() or out.is_symlink():
|
||||||
|
raise ValueError('Output exists; use a fresh directory')
|
||||||
|
if any(p.is_symlink() for p in out.parents):
|
||||||
|
raise ValueError('Unsafe output parent')
|
||||||
|
out.mkdir(parents=True)
|
||||||
|
try:
|
||||||
|
evidence = load_snapshot(snapshots)
|
||||||
|
rows,sessions = parse_program(evidence['program-data.json'])
|
||||||
|
(out/'evidence').mkdir()
|
||||||
|
for name,body in evidence.items():
|
||||||
|
(out/'evidence'/name).write_bytes(body)
|
||||||
|
outputs = []
|
||||||
|
for track,count in TRACK_COUNTS.items():
|
||||||
|
slug = re.sub('[^a-z0-9]+','_',track.lower()).strip('_')
|
||||||
|
path = out/('mm-program-2026-'+slug+'.jsonl')
|
||||||
|
raw = ''.join(json.dumps(r,ensure_ascii=False)+'\n' for r in rows if r['track']==track).encode()
|
||||||
|
path.write_bytes(raw)
|
||||||
|
outputs.append({'file':str(path),'file_sha256':digest(raw),'count':count,
|
||||||
|
'track_literal':track,'format':'mm_program_observation_v1_not_ingestion_record',
|
||||||
|
'status':'pending_contribution_and_identity_review','approved_for_import':0})
|
||||||
|
session_raw = ''.join(json.dumps(r,ensure_ascii=False)+'\n' for r in sessions).encode()
|
||||||
|
(out/'sessions.jsonl').write_bytes(session_raw)
|
||||||
|
report = {'status':'collected_pending_contribution_and_identity_review','source':SOURCE,
|
||||||
|
'records':2143,'program_observations':4326,'extra_presentation_observations':2183,
|
||||||
|
'track_counts':TRACK_COUNTS,'files':outputs,'session_count':157,
|
||||||
|
'sessions_file_sha256':digest(session_raw),'missing_abstract_count':216,
|
||||||
|
'missing_author_program_ids':['mfp4971','mfp7801'],
|
||||||
|
'source_body_sha256':BODY_SHA,'source_manifest_sha256':MANIFEST_SHA,
|
||||||
|
'database_written':False,'pdfs_downloaded':False,'network_requests':0,'approved_for_import':0,
|
||||||
|
'public_negative_access':{'anonymous_notes_probes':4,'http_statuses':[403]*4,
|
||||||
|
'negative_decisions_inferred':0,'scope':'Observed endpoints only; not proof of zero public negative decisions.'},
|
||||||
|
'scope':'Program IDs are source locators, not scientific-work or original submission IDs. All28source tracks retained; presentation times/kinds/locators and session context kept separately. No publication status, paper type, DOI, arxiv/OpenReview ID, rejection or withdrawal inferred. Blank fields are not negative decisions. Contribution/publication/work/history review and final whole-database audit remain.'}
|
||||||
|
part = out/'collection-report.json.part'
|
||||||
|
part.write_text(json.dumps(report,ensure_ascii=False,indent=2)+'\n')
|
||||||
|
part.replace(out/'collection-report.json')
|
||||||
|
return report
|
||||||
|
except Exception as error:
|
||||||
|
(out/'failed.json').write_text(json.dumps({'status':'failed_not_complete','error_type':type(error).__name__,'database_written':False})+'\n')
|
||||||
|
raise
|
||||||
|
|
||||||
|
|
||||||
|
def main(argv=None, *, prog=None):
|
||||||
|
parser = argparse.ArgumentParser(prog=prog,description=__doc__)
|
||||||
|
parser.add_argument('--snapshots',required=True)
|
||||||
|
parser.add_argument('--out',required=True)
|
||||||
|
args = parser.parse_args(argv)
|
||||||
|
print(json.dumps(collect(args.out,snapshots=args.snapshots),ensure_ascii=False))
|
||||||
@@ -0,0 +1,336 @@
|
|||||||
|
"""Frozen MM 2026 registered publications with separately attributed program evidence.
|
||||||
|
|
||||||
|
All publications survive normalization. A program metadata match is not a
|
||||||
|
scientific-work identity decision; publisher article type is not research type.
|
||||||
|
"""
|
||||||
|
import argparse
|
||||||
|
from collections import Counter, defaultdict
|
||||||
|
from datetime import datetime
|
||||||
|
import hashlib
|
||||||
|
from html.parser import HTMLParser
|
||||||
|
import json
|
||||||
|
from pathlib import Path
|
||||||
|
import re
|
||||||
|
import unicodedata
|
||||||
|
from urllib.parse import parse_qs, urlsplit
|
||||||
|
|
||||||
|
from paper_atlas.collectors import mm2026 as program
|
||||||
|
|
||||||
|
SOURCE = 'mm_registry_program'
|
||||||
|
SNAPSHOT = 'mm2026-container-registry-v2'
|
||||||
|
MANIFEST_SHA = '128b9d02c72261a66edb9a4537d118bdccab3ed80d49a1e971fefdac399f9855'
|
||||||
|
PARENT_SHA = 'bc3ce7da2914f9ddd198c928b3acb2734a8370add3348c23c5f9f12135381628'
|
||||||
|
PARENT_DOI = '10.1145/3767308'
|
||||||
|
CONTAINER = 'Proceedings of the 34th ACM International Conference on Multimedia'
|
||||||
|
EVENT = "MM '26: The 34th ACM International Conference on Multimedia"
|
||||||
|
TRACKS = {'Main Track':'main','Grand Challenge':'grand_challenge',
|
||||||
|
'Interactive Art':'interactive_art','Datasets':'datasets','Reproducibility':'reproducibility',
|
||||||
|
'BNI':'bni','Open Source':'open_source','Doctoral Symposium':'doctoral_symposium',
|
||||||
|
'Demo and Video':'demo_and_video','Industry Talk':'industry_talk'}
|
||||||
|
BIBLIOGRAPHY = {'DOI','URL','title','original-title','subtitle','short-title','author',
|
||||||
|
'publisher','member','type','container-title','short-container-title','event','abstract',
|
||||||
|
'page','ISBN','isbn-type','published','published-print','published-online','issued',
|
||||||
|
'license','resource','relation','alternative-id'}
|
||||||
|
|
||||||
|
|
||||||
|
def digest(raw):
|
||||||
|
return hashlib.sha256(raw).hexdigest()
|
||||||
|
|
||||||
|
|
||||||
|
def key(text):
|
||||||
|
return ''.join(c for c in unicodedata.normalize('NFKD',text).casefold() if c.isalnum())
|
||||||
|
|
||||||
|
|
||||||
|
def author_names(values):
|
||||||
|
return Counter(key(v) for v in values)
|
||||||
|
|
||||||
|
|
||||||
|
def load_registry(root, parent_path):
|
||||||
|
directory = Path(root)/SNAPSHOT
|
||||||
|
raw_manifest = program.safe_read(directory/'manifest.json', 128*1024)
|
||||||
|
raw_parent = program.safe_read(Path(parent_path), 128*1024)
|
||||||
|
if digest(raw_manifest) != MANIFEST_SHA or digest(raw_parent) != PARENT_SHA:
|
||||||
|
raise ValueError('Reviewed registry or parent hash changed')
|
||||||
|
manifest = json.loads(raw_manifest);parent = json.loads(raw_parent)['message']
|
||||||
|
if (manifest['status'] != 'complete_exact_container_registry_capture_not_work_or_track_approval'
|
||||||
|
or manifest['unique_registry_DOIs'] != 1918 or manifest['expected_registry_items'] != 1918
|
||||||
|
or len(manifest['pages']) != 8 or manifest['database_written'] is not False
|
||||||
|
or manifest['approved_for_import'] != 0 or manifest['parent_body_sha256'] != PARENT_SHA
|
||||||
|
or parent['DOI'] != PARENT_DOI or parent['type'] != 'proceedings'
|
||||||
|
or parent['title'] != [CONTAINER] or parent['event']['name'] != EVENT
|
||||||
|
or parent['member'] != '320' or parent['ISBN'] != ['9798400722134']):
|
||||||
|
raise ValueError('Incomplete reviewed registry or wrong proceedings')
|
||||||
|
evidence = {'manifest.json':raw_manifest,'parent.json':raw_parent};rows=[];seen=set()
|
||||||
|
for number,receipt in enumerate(manifest['pages'],1):
|
||||||
|
name=receipt['file']
|
||||||
|
if name != f'page-{number:03}.json':
|
||||||
|
raise ValueError('Unexpected registry page filename')
|
||||||
|
body=program.safe_read(directory/name,8*1024**2)
|
||||||
|
url=urlsplit(receipt['url']);query=parse_qs(url.query)
|
||||||
|
expected_filter='container-title:'+CONTAINER+',member:320'
|
||||||
|
if (receipt['http_status'] != 200 or receipt['url'] != receipt['final_url']
|
||||||
|
or url.scheme != 'https' or url.netloc != 'api.crossref.org' or url.path != '/works'
|
||||||
|
or set(query) != {'filter','rows','cursor'} or query['filter'] != [expected_filter]
|
||||||
|
or query['rows'] != ['250'] or not query['cursor'][0]
|
||||||
|
or datetime.fromisoformat(receipt['at']).utcoffset() is None
|
||||||
|
or receipt['bytes'] != len(body) or receipt['sha256'] != digest(body)):
|
||||||
|
raise ValueError('Registry receipt and source body disagree')
|
||||||
|
obj=json.loads(body);message=obj['message'];expected=250 if number<8 else 168
|
||||||
|
if (obj['status'] != 'ok' or message['total-results'] != 1918
|
||||||
|
or len(message['items']) != expected or receipt['items'] != expected):
|
||||||
|
raise ValueError('Incomplete registry pagination')
|
||||||
|
evidence[name]=body
|
||||||
|
for position,original in enumerate(message['items']):
|
||||||
|
doi=original.get('DOI')
|
||||||
|
if not isinstance(doi,str) or not re.fullmatch(r'10\.1145/3767308\.[0-9]+',doi) or doi in seen:
|
||||||
|
raise ValueError('Missing, foreign or repeated publication DOI')
|
||||||
|
seen.add(doi)
|
||||||
|
if (original.get('URL') != 'https://doi.org/'+doi or original.get('type') != 'proceedings-article'
|
||||||
|
or original.get('container-title') != [CONTAINER] or original.get('event',{}).get('name') != EVENT
|
||||||
|
or original.get('member') != '320' or original.get('publisher') != 'ACM'
|
||||||
|
or original.get('published') != {'date-parts':[[2026,11,9]]} or original.get('abstract')):
|
||||||
|
raise ValueError('Wrong registered publication context or changed abstract availability')
|
||||||
|
titles=original.get('title');authors=original.get('author')
|
||||||
|
if (not isinstance(titles,list) or len(titles)!=1 or not isinstance(titles[0],str) or not titles[0].strip()
|
||||||
|
or not isinstance(authors,list) or not authors):
|
||||||
|
raise ValueError('Missing publication title or authors')
|
||||||
|
names=[];flags=[]
|
||||||
|
for i,author in enumerate(authors,1):
|
||||||
|
family=author.get('family');given=author.get('given','')
|
||||||
|
if not isinstance(family,str) or not family.strip() or not isinstance(given,str):
|
||||||
|
raise ValueError('Invalid registered author fields')
|
||||||
|
name=' '.join(x.strip() for x in (given,family) if x.strip())
|
||||||
|
if '@' in name or name.casefold()=='anonymous':
|
||||||
|
raise ValueError('Unreviewed author identity')
|
||||||
|
names.append(name)
|
||||||
|
if not given.strip():flags.append({'kind':'single_registered_name_field_not_split','author_position':i})
|
||||||
|
if re.search(r'<[^>]+>|&[A-Za-z#0-9]+;',titles[0]):
|
||||||
|
flags.append({'kind':'literal_title_markup_or_entity_retained_requires_review'})
|
||||||
|
canonical=json.dumps(original,ensure_ascii=False,sort_keys=True,separators=(',',':')).encode()
|
||||||
|
rows.append({'doi':doi,'title':titles[0],'authors':names,'flags':flags,
|
||||||
|
'fields':{k:v for k,v in original.items() if k in BIBLIOGRAPHY},
|
||||||
|
'original_field_names':sorted(original),'original_canonical_json_sha256':digest(canonical),
|
||||||
|
'receipt':receipt,'source_position':position})
|
||||||
|
if len(rows)!=1918:
|
||||||
|
raise ValueError('Incomplete registered publication census')
|
||||||
|
return evidence,rows
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
class _TitleFormatting(HTMLParser):
|
||||||
|
"""Decode text once, removing only balanced, attribute-free reviewed tags."""
|
||||||
|
def __init__(self):
|
||||||
|
super().__init__(convert_charrefs=True)
|
||||||
|
self.parts=[];self.stack=[];self.valid=True
|
||||||
|
|
||||||
|
def handle_data(self, text):
|
||||||
|
self.parts.append(text)
|
||||||
|
|
||||||
|
def handle_starttag(self, tag, attrs):
|
||||||
|
if tag not in {'i','scp','sup','sub'} or attrs:
|
||||||
|
self.valid=False
|
||||||
|
else:
|
||||||
|
self.stack.append(tag)
|
||||||
|
|
||||||
|
def handle_endtag(self, tag):
|
||||||
|
if not self.stack or self.stack[-1]!=tag:
|
||||||
|
self.valid=False
|
||||||
|
else:
|
||||||
|
self.stack.pop()
|
||||||
|
|
||||||
|
def handle_startendtag(self, tag, attrs):
|
||||||
|
self.valid=False
|
||||||
|
|
||||||
|
def handle_comment(self, text):
|
||||||
|
self.valid=False
|
||||||
|
|
||||||
|
def handle_decl(self, text):
|
||||||
|
self.valid=False
|
||||||
|
|
||||||
|
def handle_pi(self, text):
|
||||||
|
self.valid=False
|
||||||
|
|
||||||
|
def unknown_decl(self, text):
|
||||||
|
self.valid=False
|
||||||
|
|
||||||
|
|
||||||
|
def formatting_text(title):
|
||||||
|
# HTMLParser may silently discard a trailing incomplete start tag.
|
||||||
|
# Check lexical coverage before parsing so no title content can disappear.
|
||||||
|
literal=re.sub(r'</?(?:i|scp|sup|sub)\s*>','',title,flags=re.I)
|
||||||
|
if '<' in literal or '>' in literal:
|
||||||
|
return None
|
||||||
|
parser=_TitleFormatting()
|
||||||
|
parser.feed(title);parser.close()
|
||||||
|
text=''.join(parser.parts)
|
||||||
|
return text if parser.valid and not parser.stack and text.strip() and '<' not in text and '>' not in text else None
|
||||||
|
|
||||||
|
|
||||||
|
FORMATTING_MATCHES = {'10.1145/3767308.3834907': 'mfp0053',
|
||||||
|
'10.1145/3767308.3834936': 'mfp0229',
|
||||||
|
'10.1145/3767308.3834946': 'mfp0297',
|
||||||
|
'10.1145/3767308.3834956': 'mfp0339',
|
||||||
|
'10.1145/3767308.3835035': 'mfp0798',
|
||||||
|
'10.1145/3767308.3835055': 'mfp0903',
|
||||||
|
'10.1145/3767308.3835101': 'mfp1175',
|
||||||
|
'10.1145/3767308.3835374': 'mfp2617',
|
||||||
|
'10.1145/3767308.3835424': 'mfp2909',
|
||||||
|
'10.1145/3767308.3835452': 'mfp3048',
|
||||||
|
'10.1145/3767308.3835460': 'mfp3109',
|
||||||
|
'10.1145/3767308.3835544': 'mfp3554',
|
||||||
|
'10.1145/3767308.3835591': 'mfp3783',
|
||||||
|
'10.1145/3767308.3835667': 'mfp4173',
|
||||||
|
'10.1145/3767308.3835763': 'mfp4778',
|
||||||
|
'10.1145/3767308.3835895': 'mfp5537',
|
||||||
|
'10.1145/3767308.3835959': 'mfp5823',
|
||||||
|
'10.1145/3767308.3836215': 'mfp7332',
|
||||||
|
'10.1145/3767308.3836272': 'mfp7635',
|
||||||
|
'10.1145/3767308.3836293': 'mfp7732',
|
||||||
|
'10.1145/3767308.3836447': 'mfp8648',
|
||||||
|
'10.1145/3767308.3836600': 'mfp9815',
|
||||||
|
'10.1145/3767308.3838656': 'mds177',
|
||||||
|
'10.1145/3767308.3839012': 'mde41',
|
||||||
|
'10.1145/3767308.3839024': 'mde15',
|
||||||
|
'10.1145/3767308.3839028': 'mde02'}
|
||||||
|
|
||||||
|
|
||||||
|
def match_program(rows, observations):
|
||||||
|
titles=defaultdict(list);formatted=defaultdict(list)
|
||||||
|
for p in observations:
|
||||||
|
titles[key(p['title'])].append(p)
|
||||||
|
text=formatting_text(p['title'])
|
||||||
|
if text is not None:formatted[key(text)].append(p)
|
||||||
|
matched=set();kinds=Counter()
|
||||||
|
for row in rows:
|
||||||
|
row.pop('program',None);row.pop('formatting_comparison',None)
|
||||||
|
candidates=titles[key(row['title'])]
|
||||||
|
method='literal_title'
|
||||||
|
# A literal-title author disagreement must not be bypassed by fallback.
|
||||||
|
if not candidates:
|
||||||
|
text=formatting_text(row['title'])
|
||||||
|
if text is not None:
|
||||||
|
candidates=formatted[key(text)]
|
||||||
|
if candidates:method='reviewed_title_formatting'
|
||||||
|
exact=[p for p in candidates if author_names(p['authors'])==author_names(row['authors'])]
|
||||||
|
if len(exact)==1:
|
||||||
|
row['kind']=('unique_program_format_title_fullauthor_agreement_not_identity_approval'
|
||||||
|
if method=='reviewed_title_formatting' else 'unique_program_title_fullauthor_agreement_not_identity_approval')
|
||||||
|
row['program']=exact[0];matched.add(exact[0]['paper_id'])
|
||||||
|
if exact[0]['track'] not in TRACKS:
|
||||||
|
raise ValueError('New registered program track requires review')
|
||||||
|
elif len(exact)>1:row['kind']='multiple_program_ids_same_title_authors_hold'
|
||||||
|
elif candidates:
|
||||||
|
row['kind']=('program_format_author_version_requires_review'
|
||||||
|
if method=='reviewed_title_formatting' else 'program_author_version_requires_review')
|
||||||
|
else:row['kind']='no_exact_program_title_requires_review'
|
||||||
|
row['program_title_candidates']=[p['paper_id'] for p in candidates]
|
||||||
|
row['program_exact_candidates']=[p['paper_id'] for p in exact]
|
||||||
|
if method=='reviewed_title_formatting':
|
||||||
|
row['formatting_comparison']={'method':'balanced_attribute_free_i_scp_sup_sub_and_single_entity_decode_then_nfkd_alnum',
|
||||||
|
'registry_comparison_text':formatting_text(row['title']),
|
||||||
|
'program_comparison_texts':{p['paper_id']:formatting_text(p['title']) for p in candidates},
|
||||||
|
'raw_titles_preserved':True,'work_identity_approved':False}
|
||||||
|
kinds[row['kind']]+=1
|
||||||
|
return dict(kinds),[p for p in observations if p['paper_id'] not in matched]
|
||||||
|
|
||||||
|
|
||||||
|
def correlate(rows, observations):
|
||||||
|
kinds,unmatched=match_program(rows,observations)
|
||||||
|
expected={'unique_program_title_fullauthor_agreement_not_identity_approval':1726,
|
||||||
|
'unique_program_format_title_fullauthor_agreement_not_identity_approval':26,
|
||||||
|
'program_author_version_requires_review':122,
|
||||||
|
'program_format_author_version_requires_review':1,
|
||||||
|
'no_exact_program_title_requires_review':43}
|
||||||
|
actual={r['doi']:r['program']['paper_id'] for r in rows
|
||||||
|
if r['kind']=='unique_program_format_title_fullauthor_agreement_not_identity_approval'}
|
||||||
|
if kinds != expected or actual != FORMATTING_MATCHES:
|
||||||
|
raise ValueError('Reviewed program/publication correspondence changed')
|
||||||
|
return kinds,unmatched
|
||||||
|
|
||||||
|
|
||||||
|
def collect(out, *, registry_snapshots, program_snapshots, parent_snapshot):
|
||||||
|
out=Path(out)
|
||||||
|
if out.exists() or out.is_symlink():
|
||||||
|
raise ValueError('Output exists; inspect prior operation, never overwrite')
|
||||||
|
if any(p.is_symlink() for p in out.parents):
|
||||||
|
raise ValueError('Unsafe output parent')
|
||||||
|
out.mkdir(parents=True)
|
||||||
|
try:
|
||||||
|
registry_evidence,publications=load_registry(registry_snapshots,parent_snapshot)
|
||||||
|
program_evidence=program.load_snapshot(program_snapshots)
|
||||||
|
observations,_=program.parse_program(program_evidence['program-data.json'])
|
||||||
|
kinds,unmatched=correlate(publications,observations)
|
||||||
|
for group,files in [('registry',registry_evidence),('program',program_evidence)]:
|
||||||
|
directory=out/'evidence'/group;directory.mkdir(parents=True)
|
||||||
|
for name,body in files.items():(directory/name).write_bytes(body)
|
||||||
|
buckets=defaultdict(list);missing=0;copied=0;known_tracks=Counter()
|
||||||
|
for item in sorted(publications,key=lambda r:r['doi']):
|
||||||
|
p=item.get('program');doi=item['doi'];receipt=item['receipt']
|
||||||
|
track=TRACKS[p['track']] if p else 'proceedings_unclassified'
|
||||||
|
abstract=p['abstract'] if p and p['abstract'] else None
|
||||||
|
missing+=abstract is None;copied+=abstract is not None
|
||||||
|
if p:known_tracks[p['track']]+=1
|
||||||
|
tags=[{'namespace':SOURCE+':crossref:type','value':'proceedings-article'}]
|
||||||
|
if p:tags.append({'namespace':SOURCE+':mm_program:track','value':p['track']})
|
||||||
|
buckets[track].append({'source':SOURCE,'external_id':doi,'url':'https://doi.org/'+doi,
|
||||||
|
'title':item['title'],'authors':item['authors'],'abstract':abstract,
|
||||||
|
'identifiers':{'doi':doi},'submission_identifiers':[{'scheme':'publication_doi','value':doi}],
|
||||||
|
'venue':'MM','year':2026,'track':track,'status':'accepted',
|
||||||
|
'raw_decision':'Registered publication in MM2026 proceedings; not an original signed decision.',
|
||||||
|
'evidence_url':receipt['url'],'observed_at':receipt['at'],'tags':tags,
|
||||||
|
'metadata':{**({'program_title_formatting_comparison':item['formatting_comparison']} if 'formatting_comparison' in item else {}),'registry_fields':item['fields'],'registry_original_field_names':item['original_field_names'],
|
||||||
|
'registry_source_evidence':{'file':'evidence/registry/'+receipt['file'],'sha256':receipt['sha256'],
|
||||||
|
'url':receipt['url'],'json_pointer':'/message/items/'+str(item['source_position']),
|
||||||
|
'canonical_object_json_sha256':item['original_canonical_json_sha256']},
|
||||||
|
'registry_manifest':{'file':'evidence/registry/manifest.json','sha256':MANIFEST_SHA},
|
||||||
|
'parent_evidence':{'file':'evidence/registry/parent.json','sha256':PARENT_SHA},
|
||||||
|
'program_source_evidence':{'file':'evidence/program/program-data.json','sha256':program.BODY_SHA},
|
||||||
|
'program_fields':p,'program_match_kind':item['kind'],
|
||||||
|
'program_title_candidates':item['program_title_candidates'],
|
||||||
|
'program_exact_candidates':item['program_exact_candidates'],
|
||||||
|
'abstract_source':('official_program_reviewed_title_formatting_correspondence' if item.get('formatting_comparison') else 'official_program_exact_metadata_correspondence') if abstract is not None else 'not_available_from_reviewed_sources',
|
||||||
|
'registry_abstract_available':False,'original_decision_signature_verified':False,
|
||||||
|
'scientific_work_identity_approved':False,'research_type_classified':False,
|
||||||
|
'review_flags':item['flags'],
|
||||||
|
'scope':'Derived publication/program metadata with separate original evidence. All registered articles retained, including art/industry/demo; bibliographic type is not research type. Program correlation does not approve work identity or person aliases. Publication dates are not decision dates. DOI is publication evidence, not original OpenReview ID.'}})
|
||||||
|
outputs=[]
|
||||||
|
for track,rows in sorted(buckets.items()):
|
||||||
|
path=out/f'mm-publications-2026-{track}.jsonl'
|
||||||
|
raw=''.join(json.dumps(row,ensure_ascii=False)+'\n' for row in rows).encode();path.write_bytes(raw)
|
||||||
|
manifest={'source':SOURCE,'parser_version':2,'venue':'MM','year':2026,'track':track,'count':len(rows),
|
||||||
|
'file':str(path),'file_sha256':digest(raw),'list_url':'https://doi.org/'+PARENT_DOI,
|
||||||
|
'status':'partial','missing_abstract_count':sum(row['abstract'] is None for row in rows),
|
||||||
|
'scope':'Registered container partition only; not allMMprogram/workshops, independent works or public-negative coverage.'}
|
||||||
|
path.with_suffix('.manifest.json').write_text(json.dumps(manifest,ensure_ascii=False,indent=2)+'\n');outputs.append(manifest)
|
||||||
|
report={'status':'collected_publications_pending_work_identity_review','records':1918,'unique_registered_dois':1918,
|
||||||
|
'files':outputs,'program_ids':2143,'program_observations':4326,'program_match_counts':kinds,
|
||||||
|
'matched_program_track_counts':dict(known_tracks),'unmatched_program':unmatched,
|
||||||
|
'unmatched_program_count':len(unmatched),'registry_missing_abstract_count':1918,
|
||||||
|
'abstracts_from_exact_program_correspondence':copied,'output_missing_abstract_count':missing,
|
||||||
|
'abstracts_from_literal_program_correspondence':copied-26,'abstracts_from_reviewed_title_formatting':26,
|
||||||
|
'registered_single_name_fields':sum(f['kind']=='single_registered_name_field_not_split' for r in publications for f in r['flags']),
|
||||||
|
'registered_title_markup_flags':sum(any(f['kind']=='literal_title_markup_or_entity_retained_requires_review' for f in r['flags']) for r in publications),
|
||||||
|
'database_written':False,'approved_for_import':0,'network_requests':0,'pdfs_downloaded':False,
|
||||||
|
'public_negative_availability':'unknown_no_negative_decisions_inferred',
|
||||||
|
'scope':'1918registeredpublication DOIs, not1918proven distinct scientificworks.166title/authorvariants remainunclassified,391unmatchedprogram items retained.26formattingcorrespondences retainedseparately;rawtitlesunchanged. All28programtracks remain independent census; article type never implies research. Per-work/source/DOI/history review andfinalwholeDBaudit remain.'}
|
||||||
|
if (sum(f['count'] for f in outputs)!=1918 or copied!=1743 or missing!=175 or len(unmatched)!=391):
|
||||||
|
raise ValueError('Reviewed output partition or abstract census changed')
|
||||||
|
part=out/'collection-report.json.part';part.write_text(json.dumps(report,ensure_ascii=False,indent=2)+'\n');part.replace(out/'collection-report.json')
|
||||||
|
return report
|
||||||
|
except Exception as error:
|
||||||
|
(out/'failed.json').write_text(json.dumps({'status':'failed_not_complete','error_type':type(error).__name__,'database_written':False})+'\n')
|
||||||
|
raise
|
||||||
|
|
||||||
|
|
||||||
|
def main(argv=None, *, prog=None):
|
||||||
|
parser=argparse.ArgumentParser(prog=prog,description=__doc__)
|
||||||
|
parser.add_argument('--registry-snapshots',required=True)
|
||||||
|
parser.add_argument('--program-snapshots',required=True)
|
||||||
|
parser.add_argument('--parent-snapshot',required=True)
|
||||||
|
parser.add_argument('--out',required=True)
|
||||||
|
args=parser.parse_args(argv)
|
||||||
|
report=collect(args.out,registry_snapshots=args.registry_snapshots,
|
||||||
|
program_snapshots=args.program_snapshots,parent_snapshot=args.parent_snapshot)
|
||||||
|
print(json.dumps({k:report[k] for k in ('status','records','files','program_match_counts',
|
||||||
|
'abstracts_from_exact_program_correspondence','output_missing_abstract_count',
|
||||||
|
'unmatched_program_count','database_written','approved_for_import')},ensure_ascii=False))
|
||||||
BIN
Binary file not shown.
BIN
Binary file not shown.
Vendored
BIN
Binary file not shown.
+706
@@ -0,0 +1,706 @@
|
|||||||
|
{
|
||||||
|
"status": "reviewed26_presentation_only_title_correspondences_not_work_identity_approval",
|
||||||
|
"gold": [
|
||||||
|
{
|
||||||
|
"doi": "10.1145/3767308.3834907",
|
||||||
|
"title": "H\n <sup>2</sup>\n AL: Hyperbolic Hierarchy-aware Aggregative Learning for Registration-based Few-shot Medical Image Segmentation",
|
||||||
|
"authors": [
|
||||||
|
"Jia Wang",
|
||||||
|
"Jiaming Cai",
|
||||||
|
"Zunying Hu",
|
||||||
|
"Zhanjie Wu",
|
||||||
|
"Jinyuan Liu",
|
||||||
|
"Hua Cheng",
|
||||||
|
"Yun Peng",
|
||||||
|
"Xin Fan"
|
||||||
|
],
|
||||||
|
"kind": "unique_program_format_title_fullauthor_agreement_not_identity_approval",
|
||||||
|
"source_page": "page-004.json",
|
||||||
|
"source_position": 246,
|
||||||
|
"registry_object_sha256": "dd4a627c13ca806e15e59d3ae894c9216d14d33c85bc5ce7362b51fad8c8a692",
|
||||||
|
"program_id": "mfp0053",
|
||||||
|
"track_literal": "Main Track",
|
||||||
|
"track": "main",
|
||||||
|
"abstract_sha256": "516fac88f1b7a2821c1e2b97cc41530f32abbe2514b65b5823427ff3ecdc9a23",
|
||||||
|
"program_context_pointers": [
|
||||||
|
"/entries/129/papers/19",
|
||||||
|
"/entries/129/subsessions/1/items/7"
|
||||||
|
],
|
||||||
|
"single_name_positions": []
|
||||||
|
},
|
||||||
|
{
|
||||||
|
"doi": "10.1145/3767308.3834936",
|
||||||
|
"title": "M\n <sup>3</sup>\n TR: Temporal Retrieval Enhanced Multi-Modal Micro-video Popularity Prediction",
|
||||||
|
"authors": [
|
||||||
|
"Jiacheng Lu",
|
||||||
|
"Weijian Wang",
|
||||||
|
"Mingyuan Xiao",
|
||||||
|
"Yang Hua",
|
||||||
|
"Tao Song",
|
||||||
|
"Bo Peng",
|
||||||
|
"Cheng Hua",
|
||||||
|
"Haibing Guan"
|
||||||
|
],
|
||||||
|
"kind": "unique_program_format_title_fullauthor_agreement_not_identity_approval",
|
||||||
|
"source_page": "page-005.json",
|
||||||
|
"source_position": 25,
|
||||||
|
"registry_object_sha256": "93340f8b923d7cc0d848941107351fb661d95effa35b1869921e7163da36bae4",
|
||||||
|
"program_id": "mfp0229",
|
||||||
|
"track_literal": "Main Track",
|
||||||
|
"track": "main",
|
||||||
|
"abstract_sha256": "6968e6a2d364075020e0b6c95c9076742531681232c40df511a2e8992c6b25f9",
|
||||||
|
"program_context_pointers": [
|
||||||
|
"/entries/109/papers/27",
|
||||||
|
"/entries/109/subsessions/2/items/3"
|
||||||
|
],
|
||||||
|
"single_name_positions": []
|
||||||
|
},
|
||||||
|
{
|
||||||
|
"doi": "10.1145/3767308.3834946",
|
||||||
|
"title": "<i>PAST:</i>\n Prompt-Adaptive Sampling Termination for Efficient Diffusion Model",
|
||||||
|
"authors": [
|
||||||
|
"Renye Yan",
|
||||||
|
"Jikang Cheng",
|
||||||
|
"You Wu",
|
||||||
|
"Wei Peng",
|
||||||
|
"Zongwei Wang",
|
||||||
|
"Ling Liang",
|
||||||
|
"Yimao Cai"
|
||||||
|
],
|
||||||
|
"kind": "unique_program_format_title_fullauthor_agreement_not_identity_approval",
|
||||||
|
"source_page": "page-005.json",
|
||||||
|
"source_position": 35,
|
||||||
|
"registry_object_sha256": "4210191c2871090711820aef9aff06289694380c9d0194ec9b8b1e5e40fad00d",
|
||||||
|
"program_id": "mfp0297",
|
||||||
|
"track_literal": "Main Track",
|
||||||
|
"track": "main",
|
||||||
|
"abstract_sha256": "b47bd7502bf3a8e9e9dc87e3d9552efe928cc455f843d527a4860c430c5e9490",
|
||||||
|
"program_context_pointers": [
|
||||||
|
"/entries/61/papers/40",
|
||||||
|
"/entries/61/subsessions/3/items/4"
|
||||||
|
],
|
||||||
|
"single_name_positions": []
|
||||||
|
},
|
||||||
|
{
|
||||||
|
"doi": "10.1145/3767308.3834956",
|
||||||
|
"title": "<i>SwiftExplorer:</i>\n Training-free diffusion model alignment with Swift Diversity Exploration",
|
||||||
|
"authors": [
|
||||||
|
"Renye Yan",
|
||||||
|
"Jikang Cheng",
|
||||||
|
"You Wu",
|
||||||
|
"Baojin Huang",
|
||||||
|
"Wei Peng",
|
||||||
|
"Zongwei Wang",
|
||||||
|
"Ling Liang",
|
||||||
|
"Yimao Cai"
|
||||||
|
],
|
||||||
|
"kind": "unique_program_format_title_fullauthor_agreement_not_identity_approval",
|
||||||
|
"source_page": "page-006.json",
|
||||||
|
"source_position": 235,
|
||||||
|
"registry_object_sha256": "7f2812ca4072437ecaaa585e85139e7cba0e0d721f9a2c0ac4e9897739faed47",
|
||||||
|
"program_id": "mfp0339",
|
||||||
|
"track_literal": "Main Track",
|
||||||
|
"track": "main",
|
||||||
|
"abstract_sha256": "f7809f57290e44fe291ffdc07ea956e5484aa491390cd385db75d056f95b840a",
|
||||||
|
"program_context_pointers": [
|
||||||
|
"/entries/88/papers/48",
|
||||||
|
"/entries/88/subsessions/4/items/0"
|
||||||
|
],
|
||||||
|
"single_name_positions": []
|
||||||
|
},
|
||||||
|
{
|
||||||
|
"doi": "10.1145/3767308.3835035",
|
||||||
|
"title": "<i>Balancing Efficiency and Efficacy:</i>\n Training-Free Attention-Guided Switching Between Explicit and Latent Thoughts for MLLMs",
|
||||||
|
"authors": [
|
||||||
|
"Haoqian Kang",
|
||||||
|
"Liupeng Li",
|
||||||
|
"Kuofeng Gao",
|
||||||
|
"Jinpeng Wang",
|
||||||
|
"Zhenyu Lu",
|
||||||
|
"Bin Chen",
|
||||||
|
"Ke Chen",
|
||||||
|
"Yaowei Wang"
|
||||||
|
],
|
||||||
|
"kind": "unique_program_format_title_fullauthor_agreement_not_identity_approval",
|
||||||
|
"source_page": "page-004.json",
|
||||||
|
"source_position": 8,
|
||||||
|
"registry_object_sha256": "3d76f692754dc159642eeb1706491631267d437a95c4459674361ebf35f416b9",
|
||||||
|
"program_id": "mfp0798",
|
||||||
|
"track_literal": "Main Track",
|
||||||
|
"track": "main",
|
||||||
|
"abstract_sha256": "477e42fa3d5c8236d6415ff18ba26d799d659b9e2f3f09c1d80bc0971dbca125",
|
||||||
|
"program_context_pointers": [
|
||||||
|
"/entries/88/papers/39",
|
||||||
|
"/entries/88/subsessions/3/items/3"
|
||||||
|
],
|
||||||
|
"single_name_positions": []
|
||||||
|
},
|
||||||
|
{
|
||||||
|
"doi": "10.1145/3767308.3835055",
|
||||||
|
"title": "TM\n <sup>2</sup>\n Q-Det: Boosting Video Object Detection via Text-Guided Spatiotemporal Mamba and Multimodal Query Learning",
|
||||||
|
"authors": [
|
||||||
|
"Qiang Qi",
|
||||||
|
"Wenqi Shang",
|
||||||
|
"Xiao Wang",
|
||||||
|
"Shuyuan Lin"
|
||||||
|
],
|
||||||
|
"kind": "unique_program_format_title_fullauthor_agreement_not_identity_approval",
|
||||||
|
"source_page": "page-004.json",
|
||||||
|
"source_position": 28,
|
||||||
|
"registry_object_sha256": "7e3c3e6acdad588970c681bdb92969f5134086baa2a5a7b9414871135d5ea137",
|
||||||
|
"program_id": "mfp0903",
|
||||||
|
"track_literal": "Main Track",
|
||||||
|
"track": "main",
|
||||||
|
"abstract_sha256": "8549d38e781896079ab04ef6a771a4061a0fe42df2fa5c5c03a037d59cc9f548",
|
||||||
|
"program_context_pointers": [
|
||||||
|
"/entries/129/papers/55",
|
||||||
|
"/entries/129/subsessions/4/items/7"
|
||||||
|
],
|
||||||
|
"single_name_positions": []
|
||||||
|
},
|
||||||
|
{
|
||||||
|
"doi": "10.1145/3767308.3835101",
|
||||||
|
"title": "<scp>ConfAvatar:</scp>\n Confidence-Aware Self-Bootstrapping for Monocular 4D Head Avatar Reconstruction",
|
||||||
|
"authors": [
|
||||||
|
"Zifan Zheng",
|
||||||
|
"Yu Luo",
|
||||||
|
"Yuxin Huang",
|
||||||
|
"Chuancheng Shi",
|
||||||
|
"Mohammad Masbaul Alam Polash",
|
||||||
|
"Zhiyong Wang",
|
||||||
|
"Kun Hu"
|
||||||
|
],
|
||||||
|
"kind": "unique_program_format_title_fullauthor_agreement_not_identity_approval",
|
||||||
|
"source_page": "page-008.json",
|
||||||
|
"source_position": 82,
|
||||||
|
"registry_object_sha256": "b838756bd0b6af036e408936f8f13a6398d83e524ce0282167199eca092bc7d8",
|
||||||
|
"program_id": "mfp1175",
|
||||||
|
"track_literal": "Main Track",
|
||||||
|
"track": "main",
|
||||||
|
"abstract_sha256": "cb8dbe67e16e163e5b136ad3401fc57c827905e9704d7c0a84961bc8f12565f4",
|
||||||
|
"program_context_pointers": [
|
||||||
|
"/entries/109/papers/86",
|
||||||
|
"/entries/109/subsessions/7/items/2"
|
||||||
|
],
|
||||||
|
"single_name_positions": []
|
||||||
|
},
|
||||||
|
{
|
||||||
|
"doi": "10.1145/3767308.3835374",
|
||||||
|
"title": "<scp>DateGBC:</scp>\n Deep Fuzzy Granular-Ball Clustering via Adaptive Topology Evolution for Unsupervised Image Segmentation",
|
||||||
|
"authors": [
|
||||||
|
"Qianpeng Chong",
|
||||||
|
"Jiakun Wen",
|
||||||
|
"Guangyi Wei",
|
||||||
|
"Wenyi Zeng"
|
||||||
|
],
|
||||||
|
"kind": "unique_program_format_title_fullauthor_agreement_not_identity_approval",
|
||||||
|
"source_page": "page-006.json",
|
||||||
|
"source_position": 59,
|
||||||
|
"registry_object_sha256": "a7fe694a04449e841be5181888ae35786d0a4938d87eca1fffb0776d4d350e30",
|
||||||
|
"program_id": "mfp2617",
|
||||||
|
"track_literal": "Main Track",
|
||||||
|
"track": "main",
|
||||||
|
"abstract_sha256": "b57b7026c750aeebb47598860b82980215c9e4e4935f6e8611a31b2fe64f73b9",
|
||||||
|
"program_context_pointers": [
|
||||||
|
"/entries/61/papers/80",
|
||||||
|
"/entries/61/subsessions/6/items/8"
|
||||||
|
],
|
||||||
|
"single_name_positions": []
|
||||||
|
},
|
||||||
|
{
|
||||||
|
"doi": "10.1145/3767308.3835424",
|
||||||
|
"title": "<i>RTCBench</i>\n : A Scalable Benchmarking Framework for Multi-Party Video Conferencing",
|
||||||
|
"authors": [
|
||||||
|
"Evan Ram",
|
||||||
|
"Tuan Tran",
|
||||||
|
"Youngwook Son",
|
||||||
|
"Sangtae Ha"
|
||||||
|
],
|
||||||
|
"kind": "unique_program_format_title_fullauthor_agreement_not_identity_approval",
|
||||||
|
"source_page": "page-005.json",
|
||||||
|
"source_position": 108,
|
||||||
|
"registry_object_sha256": "b21198e0eba35d175bc83292be8c87bca2b80a9c8e01f40b640a4e9efc7ecde5",
|
||||||
|
"program_id": "mfp2909",
|
||||||
|
"track_literal": "Main Track",
|
||||||
|
"track": "main",
|
||||||
|
"abstract_sha256": "2398ec6539c50b14c7ed4fa773277cc860151ab77b5430b49d935f88e90a487e",
|
||||||
|
"program_context_pointers": [
|
||||||
|
"/entries/129/papers/130",
|
||||||
|
"/entries/129/subsessions/8/items/22"
|
||||||
|
],
|
||||||
|
"single_name_positions": []
|
||||||
|
},
|
||||||
|
{
|
||||||
|
"doi": "10.1145/3767308.3835452",
|
||||||
|
"title": "<scp>HeadRouter:</scp>\n Dynamic Head-Weight Routing for Task-Adaptive Audio Token Pruning in Large Audio Language Models",
|
||||||
|
"authors": [
|
||||||
|
"He Peize",
|
||||||
|
"Yaodi Luo",
|
||||||
|
"Xiaoqian Liu",
|
||||||
|
"Xuyang Liu",
|
||||||
|
"Jiahang Deng",
|
||||||
|
"Yaosong Du",
|
||||||
|
"Li Bangyu",
|
||||||
|
"Xiyan Gui",
|
||||||
|
"Yuxuan Chen",
|
||||||
|
"Linfeng Zhang"
|
||||||
|
],
|
||||||
|
"kind": "unique_program_format_title_fullauthor_agreement_not_identity_approval",
|
||||||
|
"source_page": "page-004.json",
|
||||||
|
"source_position": 167,
|
||||||
|
"registry_object_sha256": "ee31b12972ea5844a50c6c63a6fc3430875c3388a4bf261f8e76cab31112d351",
|
||||||
|
"program_id": "mfp3048",
|
||||||
|
"track_literal": "Main Track",
|
||||||
|
"track": "main",
|
||||||
|
"abstract_sha256": "c455de7c27f43d0fb49990ead5fca795334c0b613eafa2b3c6150eb147d9c3fa",
|
||||||
|
"program_context_pointers": [
|
||||||
|
"/entries/46/papers/78",
|
||||||
|
"/entries/46/subsessions/6/items/6"
|
||||||
|
],
|
||||||
|
"single_name_positions": []
|
||||||
|
},
|
||||||
|
{
|
||||||
|
"doi": "10.1145/3767308.3835460",
|
||||||
|
"title": "<i>CoAnchor:</i>\n Robust Collaborative Perception under Spatio-Temporal Misalignment via Object-Level Anchors",
|
||||||
|
"authors": [
|
||||||
|
"Chi Li",
|
||||||
|
"Rui Lin",
|
||||||
|
"Aobo Ji",
|
||||||
|
"Dongzhu Xu"
|
||||||
|
],
|
||||||
|
"kind": "unique_program_format_title_fullauthor_agreement_not_identity_approval",
|
||||||
|
"source_page": "page-004.json",
|
||||||
|
"source_position": 175,
|
||||||
|
"registry_object_sha256": "7c41b0dcb65ad27a1f0961f095889ccc73044776250ec816cf95fc25c30da7f1",
|
||||||
|
"program_id": "mfp3109",
|
||||||
|
"track_literal": "Main Track",
|
||||||
|
"track": "main",
|
||||||
|
"abstract_sha256": "5f3d97775d83528d6f41fb2fc481d1848cac7c3f9de1d469039450f9b2ecbff1",
|
||||||
|
"program_context_pointers": [
|
||||||
|
"/entries/109/papers/121",
|
||||||
|
"/entries/109/subsessions/8/items/13"
|
||||||
|
],
|
||||||
|
"single_name_positions": []
|
||||||
|
},
|
||||||
|
{
|
||||||
|
"doi": "10.1145/3767308.3835544",
|
||||||
|
"title": "<i>OC-GS:</i>\n Object-Centric Representation and Demand Alignment for Immersive 3D Gaussian Streaming",
|
||||||
|
"authors": [
|
||||||
|
"Boyan Li",
|
||||||
|
"Ruijun Mao",
|
||||||
|
"Zhicheng Liang",
|
||||||
|
"Dayou Zhang",
|
||||||
|
"Fangxin Wang"
|
||||||
|
],
|
||||||
|
"kind": "unique_program_format_title_fullauthor_agreement_not_identity_approval",
|
||||||
|
"source_page": "page-006.json",
|
||||||
|
"source_position": 129,
|
||||||
|
"registry_object_sha256": "afc0c7db304251511eef5e223f027dc6314dd70895955ff1a1cb0a461dbfdbe9",
|
||||||
|
"program_id": "mfp3554",
|
||||||
|
"track_literal": "Main Track",
|
||||||
|
"track": "main",
|
||||||
|
"abstract_sha256": "00b2ffe1073feede8447bac439ac7842c1ba0841cb6ee38ec1b5e06c3a3dc22e",
|
||||||
|
"program_context_pointers": [
|
||||||
|
"/entries/46/papers/153",
|
||||||
|
"/entries/46/subsessions/9/items/21"
|
||||||
|
],
|
||||||
|
"single_name_positions": []
|
||||||
|
},
|
||||||
|
{
|
||||||
|
"doi": "10.1145/3767308.3835591",
|
||||||
|
"title": "DP\n <sup>2</sup>\n -NeRF: Dual-Prior Probabilistic Neural Radiance Fields for Event-Guided Motion Deblurring",
|
||||||
|
"authors": [
|
||||||
|
"Jiafeng Huang",
|
||||||
|
"Wenzhen Jia",
|
||||||
|
"Shilong Dong",
|
||||||
|
"Shengjie Zhao"
|
||||||
|
],
|
||||||
|
"kind": "unique_program_format_title_fullauthor_agreement_not_identity_approval",
|
||||||
|
"source_page": "page-007.json",
|
||||||
|
"source_position": 97,
|
||||||
|
"registry_object_sha256": "145025e784ec251247f8cd86d0dbee1749666ec71f35566c20e41edd0a2347ce",
|
||||||
|
"program_id": "mfp3783",
|
||||||
|
"track_literal": "Main Track",
|
||||||
|
"track": "main",
|
||||||
|
"abstract_sha256": "0b2d267736cfe506de666618820e2934c23395bd25dab63c5a347710445a361c",
|
||||||
|
"program_context_pointers": [
|
||||||
|
"/entries/33/papers/158",
|
||||||
|
"/entries/33/subsessions/10/items/2"
|
||||||
|
],
|
||||||
|
"single_name_positions": []
|
||||||
|
},
|
||||||
|
{
|
||||||
|
"doi": "10.1145/3767308.3835667",
|
||||||
|
"title": "MMCoT\n <sup>3</sup>\n : Empowering 3D Scene Understanding with 3D-Text Interleaved Multi-Modal Chain-of-Thought Reasoning",
|
||||||
|
"authors": [
|
||||||
|
"Yongqi Shan",
|
||||||
|
"Yifan Wang",
|
||||||
|
"Yunzhi Zhuge",
|
||||||
|
"Lijun Wang",
|
||||||
|
"Huchuan Lu"
|
||||||
|
],
|
||||||
|
"kind": "unique_program_format_title_fullauthor_agreement_not_identity_approval",
|
||||||
|
"source_page": "page-005.json",
|
||||||
|
"source_position": 209,
|
||||||
|
"registry_object_sha256": "8b5ad8d891692e441b6e210304dcd35103835fd35d086c860c7611bf5385a6d9",
|
||||||
|
"program_id": "mfp4173",
|
||||||
|
"track_literal": "Main Track",
|
||||||
|
"track": "main",
|
||||||
|
"abstract_sha256": "4924721a2c3328e6bca661654569bca86d3fa69aa5f6c9e3719f827c5f0d7639",
|
||||||
|
"program_context_pointers": [
|
||||||
|
"/entries/88/papers/114",
|
||||||
|
"/entries/88/subsessions/8/items/6"
|
||||||
|
],
|
||||||
|
"single_name_positions": []
|
||||||
|
},
|
||||||
|
{
|
||||||
|
"doi": "10.1145/3767308.3835763",
|
||||||
|
"title": "<scp>AudioKV:</scp>\n KV Cache Eviction in Efficient Large Audio Language Models",
|
||||||
|
"authors": [
|
||||||
|
"Yuxuan Wang",
|
||||||
|
"He Peize",
|
||||||
|
"Xiyan Gui",
|
||||||
|
"Xiaoqian Liu",
|
||||||
|
"Junhao He",
|
||||||
|
"Xuyang Liu",
|
||||||
|
"Xuming Hu",
|
||||||
|
"Linfeng Zhang"
|
||||||
|
],
|
||||||
|
"kind": "unique_program_format_title_fullauthor_agreement_not_identity_approval",
|
||||||
|
"source_page": "page-006.json",
|
||||||
|
"source_position": 192,
|
||||||
|
"registry_object_sha256": "70054b0a5a2586fd08e75494eee08a26498730554601a5fe2e4efddcaad309de",
|
||||||
|
"program_id": "mfp4778",
|
||||||
|
"track_literal": "Main Track",
|
||||||
|
"track": "main",
|
||||||
|
"abstract_sha256": "274b360da553dfc7afdb9072505c2957cec4213e096373e2edbf58236a3fc3a2",
|
||||||
|
"program_context_pointers": [
|
||||||
|
"/entries/129/papers/100",
|
||||||
|
"/entries/129/subsessions/7/items/16"
|
||||||
|
],
|
||||||
|
"single_name_positions": []
|
||||||
|
},
|
||||||
|
{
|
||||||
|
"doi": "10.1145/3767308.3835895",
|
||||||
|
"title": "F\n <sup>2</sup>\n -Adapter: A Focus-and-Forget Framework for Generalizable CLIP-based Face Forgery Detection",
|
||||||
|
"authors": [
|
||||||
|
"Zan Gao",
|
||||||
|
"Xin Duan",
|
||||||
|
"Nian Hu",
|
||||||
|
"Hao Jia",
|
||||||
|
"Yibo Zhao",
|
||||||
|
"Cong Liu"
|
||||||
|
],
|
||||||
|
"kind": "unique_program_format_title_fullauthor_agreement_not_identity_approval",
|
||||||
|
"source_page": "page-001.json",
|
||||||
|
"source_position": 39,
|
||||||
|
"registry_object_sha256": "16bdbcc4398df4d9556c14d4d4203127349070669f5d0e1c4bf87756e241b698",
|
||||||
|
"program_id": "mfp5537",
|
||||||
|
"track_literal": "Main Track",
|
||||||
|
"track": "main",
|
||||||
|
"abstract_sha256": "b3ef626b66d93ac37a73caf15a4cf9ff92241209c80ee66be5575e6b56f9c899",
|
||||||
|
"program_context_pointers": [
|
||||||
|
"/entries/88/papers/184",
|
||||||
|
"/entries/88/subsessions/11/items/6"
|
||||||
|
],
|
||||||
|
"single_name_positions": []
|
||||||
|
},
|
||||||
|
{
|
||||||
|
"doi": "10.1145/3767308.3835959",
|
||||||
|
"title": "UPR-VAD:\n <i>U</i>\n ncertainty-guided Signal\n <i>P</i>\n urification and\n <i>R</i>\n egularization for Weakly-Supervised\n <i>V</i>\n ideo\n <i>A</i>\n nomaly\n <i>D</i>\n etection",
|
||||||
|
"authors": [
|
||||||
|
"Weiliang Huang",
|
||||||
|
"Weichao Cai",
|
||||||
|
"Bob Zhang",
|
||||||
|
"Chao Huang",
|
||||||
|
"Biao Xue",
|
||||||
|
"Rongxin Zhang"
|
||||||
|
],
|
||||||
|
"kind": "unique_program_format_title_fullauthor_agreement_not_identity_approval",
|
||||||
|
"source_page": "page-007.json",
|
||||||
|
"source_position": 21,
|
||||||
|
"registry_object_sha256": "e96f9e9d67cce31a9235bdc7dc5dba551676f3d36ba72654ae3d5a98ae8d5174",
|
||||||
|
"program_id": "mfp5823",
|
||||||
|
"track_literal": "Main Track",
|
||||||
|
"track": "main",
|
||||||
|
"abstract_sha256": "d8469bcb68f8cc19070f3bf9a866c3377ba27fe6445b5bf9337b7599a6881592",
|
||||||
|
"program_context_pointers": [
|
||||||
|
"/entries/88/papers/66",
|
||||||
|
"/entries/88/subsessions/5/items/6"
|
||||||
|
],
|
||||||
|
"single_name_positions": []
|
||||||
|
},
|
||||||
|
{
|
||||||
|
"doi": "10.1145/3767308.3836215",
|
||||||
|
"title": "BLM\n <sub>1</sub>\n : A Boundless Large Model for Cross-Space, Cross-Task, and Cross-Embodiment Learning",
|
||||||
|
"authors": [
|
||||||
|
"Wentao Tan",
|
||||||
|
"Bowen Wang",
|
||||||
|
"Heng Zhi",
|
||||||
|
"Chenyv Liu",
|
||||||
|
"Zhe Li",
|
||||||
|
"Jian Liu",
|
||||||
|
"Zengrong Lin",
|
||||||
|
"Yukun Dai",
|
||||||
|
"Yipeng Chen",
|
||||||
|
"Wenjie Yang",
|
||||||
|
"Enci Xie",
|
||||||
|
"Hao Xue",
|
||||||
|
"Baixu Ji",
|
||||||
|
"Chen Xu",
|
||||||
|
"Zhibin Wang",
|
||||||
|
"Tianshi Wang",
|
||||||
|
"Lei Zhu",
|
||||||
|
"Heng Tao Shen"
|
||||||
|
],
|
||||||
|
"kind": "unique_program_format_title_fullauthor_agreement_not_identity_approval",
|
||||||
|
"source_page": "page-003.json",
|
||||||
|
"source_position": 46,
|
||||||
|
"registry_object_sha256": "b4b51e5d0a4c0a330d01e7638c0bcc73871d750312081bafbd3b8a10529153f2",
|
||||||
|
"program_id": "mfp7332",
|
||||||
|
"track_literal": "Main Track",
|
||||||
|
"track": "main",
|
||||||
|
"abstract_sha256": "aa458905209ad562117d72dd33d4d61e000b0d9efaa55dd9672919f40e6cc2fd",
|
||||||
|
"program_context_pointers": [
|
||||||
|
"/entries/33/papers/13",
|
||||||
|
"/entries/33/subsessions/1/items/1",
|
||||||
|
"/entries/47/papers/0"
|
||||||
|
],
|
||||||
|
"single_name_positions": []
|
||||||
|
},
|
||||||
|
{
|
||||||
|
"doi": "10.1145/3767308.3836272",
|
||||||
|
"title": "SEE: Structure-aware Exploring & Exploiting for Long-horizon GUI Agent Trajectory Synthesis",
|
||||||
|
"authors": [
|
||||||
|
"Zhuohang Fan",
|
||||||
|
"Beichen Zhang",
|
||||||
|
"Yuanfa Li",
|
||||||
|
"Changqiao Wu",
|
||||||
|
"Wei Liu",
|
||||||
|
"Jian Luan",
|
||||||
|
"Weigang Zhang"
|
||||||
|
],
|
||||||
|
"kind": "unique_program_format_title_fullauthor_agreement_not_identity_approval",
|
||||||
|
"source_page": "page-002.json",
|
||||||
|
"source_position": 127,
|
||||||
|
"registry_object_sha256": "66eebbdef9e2d07fe08f3cb41cd577095f282f1256bdb73fe1e0d4c5d6e55d51",
|
||||||
|
"program_id": "mfp7635",
|
||||||
|
"track_literal": "Main Track",
|
||||||
|
"track": "main",
|
||||||
|
"abstract_sha256": "df11d727337f397bab72977bcdce4246b388ef0c8e1198ac47fdc32e68a3aff4",
|
||||||
|
"program_context_pointers": [
|
||||||
|
"/entries/88/papers/131",
|
||||||
|
"/entries/88/subsessions/8/items/23"
|
||||||
|
],
|
||||||
|
"single_name_positions": []
|
||||||
|
},
|
||||||
|
{
|
||||||
|
"doi": "10.1145/3767308.3836293",
|
||||||
|
"title": "Video\n <sup>*</sup>\n Bench: Benchmarking Fine-grained Video Perception with \"Think with Videos\"",
|
||||||
|
"authors": [
|
||||||
|
"Ruizhe Chen",
|
||||||
|
"Zhiting Fan",
|
||||||
|
"Yang Shi",
|
||||||
|
"Enxin Song",
|
||||||
|
"Wenhao Chai",
|
||||||
|
"Tongkun Guan",
|
||||||
|
"Songtao Jiang",
|
||||||
|
"Ruilin Luo",
|
||||||
|
"Yuanxing Zhang",
|
||||||
|
"Shuai Bai",
|
||||||
|
"Sibo Song",
|
||||||
|
"Zhibo Yang",
|
||||||
|
"Zuozhu Liu"
|
||||||
|
],
|
||||||
|
"kind": "unique_program_format_title_fullauthor_agreement_not_identity_approval",
|
||||||
|
"source_page": "page-002.json",
|
||||||
|
"source_position": 148,
|
||||||
|
"registry_object_sha256": "7332426edb3d15499f175fb98c093b78df21ea95fdd9cfea33da5e925cbec171",
|
||||||
|
"program_id": "mfp7732",
|
||||||
|
"track_literal": "Main Track",
|
||||||
|
"track": "main",
|
||||||
|
"abstract_sha256": "f6e4df110fe85d7f6d34eb712cd338443c2e34e924393a4e65e07a0c7b152c9e",
|
||||||
|
"program_context_pointers": [
|
||||||
|
"/entries/46/papers/198",
|
||||||
|
"/entries/46/subsessions/11/items/20"
|
||||||
|
],
|
||||||
|
"single_name_positions": []
|
||||||
|
},
|
||||||
|
{
|
||||||
|
"doi": "10.1145/3767308.3836447",
|
||||||
|
"title": "<i>Insert In Style</i>\n : A Zero-Shot Generative Framework for Harmonious Cross-Domain Object Composition",
|
||||||
|
"authors": [
|
||||||
|
"Raghu Vamsi Chittersu",
|
||||||
|
"Yuvraj Singh Rathore",
|
||||||
|
"Pranav Adlinge",
|
||||||
|
"Eswar Sai Pashaula",
|
||||||
|
"Kunal Swami"
|
||||||
|
],
|
||||||
|
"kind": "unique_program_format_title_fullauthor_agreement_not_identity_approval",
|
||||||
|
"source_page": "page-002.json",
|
||||||
|
"source_position": 183,
|
||||||
|
"registry_object_sha256": "92b2c094b96cebbace0150805b65b0a1bc3ca99ed42124bcb8de442a81094e1b",
|
||||||
|
"program_id": "mfp8648",
|
||||||
|
"track_literal": "Main Track",
|
||||||
|
"track": "main",
|
||||||
|
"abstract_sha256": "39d3f21f774010980f8b30f2fd800b5243ea7ddc6e091b4d8769de3dee622273",
|
||||||
|
"program_context_pointers": [
|
||||||
|
"/entries/107/papers/4",
|
||||||
|
"/entries/117/papers/30",
|
||||||
|
"/entries/117/subsessions/2/items/6"
|
||||||
|
],
|
||||||
|
"single_name_positions": []
|
||||||
|
},
|
||||||
|
{
|
||||||
|
"doi": "10.1145/3767308.3836600",
|
||||||
|
"title": "<scp>EEGaze</scp>\n : Multimodal Attention Prediction in Visual Search via EEG-Based Saliency Refinement",
|
||||||
|
"authors": [
|
||||||
|
"Mansi Sharma",
|
||||||
|
"Dinar Orazgaliyev",
|
||||||
|
"Antonio Krüger",
|
||||||
|
"Philipp Müller"
|
||||||
|
],
|
||||||
|
"kind": "unique_program_format_title_fullauthor_agreement_not_identity_approval",
|
||||||
|
"source_page": "page-003.json",
|
||||||
|
"source_position": 140,
|
||||||
|
"registry_object_sha256": "30b99b234aa2b944d6890119b2524d3b3c3c425bde22f20e813b9517b8f8536b",
|
||||||
|
"program_id": "mfp9815",
|
||||||
|
"track_literal": "Main Track",
|
||||||
|
"track": "main",
|
||||||
|
"abstract_sha256": "47dcb456aee600663d0df59429c1745b335dbae9bdef2dd1063265429fff49fb",
|
||||||
|
"program_context_pointers": [
|
||||||
|
"/entries/46/papers/219",
|
||||||
|
"/entries/46/subsessions/12/items/17"
|
||||||
|
],
|
||||||
|
"single_name_positions": []
|
||||||
|
},
|
||||||
|
{
|
||||||
|
"doi": "10.1145/3767308.3838656",
|
||||||
|
"title": "<scp>MedMKG:</scp>\n Benchmarking Medical Knowledge Exploitation with Multimodal Knowledge Graph",
|
||||||
|
"authors": [
|
||||||
|
"Xiaochen Wang",
|
||||||
|
"Yuan Zhong",
|
||||||
|
"Lingwei Zhang",
|
||||||
|
"Lisong Dai",
|
||||||
|
"Ting Wang",
|
||||||
|
"Fenglong Ma"
|
||||||
|
],
|
||||||
|
"kind": "unique_program_format_title_fullauthor_agreement_not_identity_approval",
|
||||||
|
"source_page": "page-003.json",
|
||||||
|
"source_position": 225,
|
||||||
|
"registry_object_sha256": "b910b66541ca65127646d35d0cd5737a56ae2e0f311a79526005dec75fca739f",
|
||||||
|
"program_id": "mds177",
|
||||||
|
"track_literal": "Datasets",
|
||||||
|
"track": "datasets",
|
||||||
|
"abstract_sha256": "5bee3b5e9ff60690509d1f34953f3dcc185c50861e112d2c4a3baf5673f6b945",
|
||||||
|
"program_context_pointers": [
|
||||||
|
"/entries/46/papers/23",
|
||||||
|
"/entries/46/subsessions/1/items/11"
|
||||||
|
],
|
||||||
|
"single_name_positions": []
|
||||||
|
},
|
||||||
|
{
|
||||||
|
"doi": "10.1145/3767308.3839012",
|
||||||
|
"title": "<i>EULALens</i>\n : SEE What YOU Agree!",
|
||||||
|
"authors": [
|
||||||
|
"Abhijeet Narang",
|
||||||
|
"Hagar Elbatanouny",
|
||||||
|
"Vanishri Shekhawat",
|
||||||
|
"Hasan Al-Nashash",
|
||||||
|
"Abhinav Dhall",
|
||||||
|
"Usman Tariq"
|
||||||
|
],
|
||||||
|
"kind": "unique_program_format_title_fullauthor_agreement_not_identity_approval",
|
||||||
|
"source_page": "page-007.json",
|
||||||
|
"source_position": 238,
|
||||||
|
"registry_object_sha256": "81dd29dafd6228f51ae2f32e2bb713272a88df78c78347b9170b54f83db00c84",
|
||||||
|
"program_id": "mde41",
|
||||||
|
"track_literal": "Demo and Video",
|
||||||
|
"track": "demo_and_video",
|
||||||
|
"abstract_sha256": "9acbb264b8019916aad0845aeae4eb5a9c94903ac5bca3cfb612c857306f6d55",
|
||||||
|
"program_context_pointers": [
|
||||||
|
"/entries/94/papers/7"
|
||||||
|
],
|
||||||
|
"single_name_positions": []
|
||||||
|
},
|
||||||
|
{
|
||||||
|
"doi": "10.1145/3767308.3839024",
|
||||||
|
"title": "<scp>ThreatVerse:</scp>\n A Closed-Loop World Model Simulator for Autonomous Driving Safety Evaluation",
|
||||||
|
"authors": [
|
||||||
|
"Jiangfan Liu",
|
||||||
|
"Zexuan Cui",
|
||||||
|
"Tianyuan Zhang",
|
||||||
|
"Zonglei Jing",
|
||||||
|
"Yaoyuan Zhang",
|
||||||
|
"Jiakai Wang",
|
||||||
|
"Jinyang Guo",
|
||||||
|
"Yuqing Ma",
|
||||||
|
"Lu Sheng",
|
||||||
|
"Aishan Liu",
|
||||||
|
"Xianglong Liu"
|
||||||
|
],
|
||||||
|
"kind": "unique_program_format_title_fullauthor_agreement_not_identity_approval",
|
||||||
|
"source_page": "page-008.json",
|
||||||
|
"source_position": 0,
|
||||||
|
"registry_object_sha256": "d0571f4562b30e4e7fb2e3aa826068ac14a540d1ae70e8cf0b11fd73c67f26ee",
|
||||||
|
"program_id": "mde15",
|
||||||
|
"track_literal": "Demo and Video",
|
||||||
|
"track": "demo_and_video",
|
||||||
|
"abstract_sha256": "02fc9b31d2b7667ea13f00c74f923a375373042171aaa1ff1f2010ee9697211a",
|
||||||
|
"program_context_pointers": [
|
||||||
|
"/entries/49/papers/4"
|
||||||
|
],
|
||||||
|
"single_name_positions": []
|
||||||
|
},
|
||||||
|
{
|
||||||
|
"doi": "10.1145/3767308.3839028",
|
||||||
|
"title": "<scp>FinAlogy:</scp>\n A Visual Analogy Retrieval System for Financial K-Line Analysis",
|
||||||
|
"authors": [
|
||||||
|
"Xiaoyan Yu",
|
||||||
|
"Ziya Zhao",
|
||||||
|
"Yunshan Ma",
|
||||||
|
"Yifan Wei",
|
||||||
|
"Dianbo Sui",
|
||||||
|
"Tat-Seng Chua"
|
||||||
|
],
|
||||||
|
"kind": "unique_program_format_title_fullauthor_agreement_not_identity_approval",
|
||||||
|
"source_page": "page-008.json",
|
||||||
|
"source_position": 4,
|
||||||
|
"registry_object_sha256": "b18c8e2821873a59318f6f09d1f453dd02195ccedad5c8e0ce1da60ac07466fb",
|
||||||
|
"program_id": "mde02",
|
||||||
|
"track_literal": "Demo and Video",
|
||||||
|
"track": "demo_and_video",
|
||||||
|
"abstract_sha256": "22f978dd02c05439663c5724f25bb997098fd85f5558d01d836711252d119931",
|
||||||
|
"program_context_pointers": [
|
||||||
|
"/entries/49/papers/0"
|
||||||
|
],
|
||||||
|
"single_name_positions": []
|
||||||
|
}
|
||||||
|
],
|
||||||
|
"author_disagreement_hold": {
|
||||||
|
"doi": "10.1145/3767308.3836308",
|
||||||
|
"program_id": "mfp7801"
|
||||||
|
},
|
||||||
|
"scope": "Onlytitlecomparisonchanges;originalregistryandprogramfixturesremainunchanged;sourceABandtrackcomefromtheexplicitprogramID.",
|
||||||
|
"held_gold": {
|
||||||
|
"doi": "10.1145/3767308.3836308",
|
||||||
|
"title": "Dynamic Late Fusion Multiple Kernel\n <i>k</i>\n -means Clustering based on Adaptive Sample Selection and Alignment",
|
||||||
|
"authors": [
|
||||||
|
"Zhirong Tian",
|
||||||
|
"Yi Zhang",
|
||||||
|
"Weixuan Liang",
|
||||||
|
"Miaomiao Li",
|
||||||
|
"Xinwang Liu"
|
||||||
|
],
|
||||||
|
"kind": "program_format_author_version_requires_review",
|
||||||
|
"source_page": "page-001.json",
|
||||||
|
"source_position": 82,
|
||||||
|
"registry_object_sha256": "26a3c1715d276b5fbe988596e3115091086158f6ef12600a36e5206785fc9dfb",
|
||||||
|
"program_id": null,
|
||||||
|
"track_literal": null,
|
||||||
|
"track": "proceedings_unclassified",
|
||||||
|
"abstract_sha256": null,
|
||||||
|
"program_context_pointers": [],
|
||||||
|
"single_name_positions": []
|
||||||
|
}
|
||||||
|
}
|
||||||
@@ -0,0 +1,92 @@
|
|||||||
|
import base64
|
||||||
|
import copy
|
||||||
|
import gzip
|
||||||
|
import hashlib
|
||||||
|
import json
|
||||||
|
from pathlib import Path
|
||||||
|
import tempfile
|
||||||
|
import unittest
|
||||||
|
|
||||||
|
from paper_atlas.collectors import ismir2026 as c
|
||||||
|
|
||||||
|
|
||||||
|
class ISMIR2026Test(unittest.TestCase):
|
||||||
|
@classmethod
|
||||||
|
def setUpClass(cls):
|
||||||
|
cls.fixture=json.loads(gzip.decompress((Path(__file__).parent/'fixtures/ismir2026-source.json.gz').read_bytes()))
|
||||||
|
cls.files={name:base64.b64decode(raw) for name,raw in cls.fixture['files'].items()}
|
||||||
|
cls.program=json.loads(cls.files['ismir2026-json-research-v1/papers.json'])
|
||||||
|
cls.accepted=cls.files['remaining2026-listings-research-v2/ismir-accepted.html']
|
||||||
|
|
||||||
|
def materialize(self,root):
|
||||||
|
for name,raw in self.files.items():
|
||||||
|
p=root/name;p.parent.mkdir(parents=True,exist_ok=True);p.write_bytes(raw)
|
||||||
|
|
||||||
|
def test140_source_versions_and_review_optins_against_independent_gold(self):
|
||||||
|
listing=c.accepted_listing(self.accepted);observations=c.program_observations(self.program,listing)
|
||||||
|
self.assertEqual(len(observations),140)
|
||||||
|
for observation,gold in zip(observations,self.fixture['independent_gold']):
|
||||||
|
r=observation['original'];v=r['content'];official=observation['accepted_listing']
|
||||||
|
self.assertEqual((r['id'],r['forum'],v['title'],v['authors'],v['abstract']),(gold['program_id'],gold['forum_literal'],gold['title'],gold['authors'],gold['abstract']))
|
||||||
|
self.assertEqual(official,{'title':gold['official_title'],'authors_literal':gold['official_authors_literal'],'author_names':gold['official_author_names']})
|
||||||
|
self.assertEqual(v['session'],gold['session_literal']);self.assertEqual(v['paper_presentation'],gold['presentation_literal']);self.assertEqual(v['keywords'],gold['keywords'])
|
||||||
|
self.assertEqual(sum('program_and_accepted_title_literal_versions_differ' in r['review_flags'] for r in observations),2)
|
||||||
|
self.assertEqual({r['original']['id'] for r in observations if 'program_and_accepted_author_sequence_versions_differ' in r['review_flags']},{'60','126','433'})
|
||||||
|
self.assertEqual(sum(r['original']['reviews']['publish_reviews']=='Yes' for r in observations),101)
|
||||||
|
|
||||||
|
def test_full_collection_keeps_abstract_bytes_review_text_tags_and_no_inferred_ids(self):
|
||||||
|
with tempfile.TemporaryDirectory() as tmp:
|
||||||
|
root=Path(tmp);src=root/'src';self.materialize(src);out=root/'out';report=c.collect(out,snapshots=src)
|
||||||
|
self.assertEqual((report['records'],report['missing_abstract_count'],report['review_optin_yes'],report['review_optin_no']),(140,0,101,39))
|
||||||
|
self.assertFalse(report['database_written']);self.assertFalse(report['pdfs_downloaded']);self.assertEqual(report['network_requests'],0);self.assertEqual(report['approved_for_import'],0)
|
||||||
|
self.assertEqual(report['public_rejected_availability'],'unknown_accepted_program_reviews_are_not_a_rejected_export')
|
||||||
|
manifest,=report['files'];raw=Path(manifest['file']).read_bytes();self.assertEqual(hashlib.sha256(raw).hexdigest(),manifest['file_sha256']);self.assertEqual(manifest['status'],'partial')
|
||||||
|
rows=[json.loads(l) for l in raw.splitlines()];self.assertEqual(len(rows),140)
|
||||||
|
for row,original,gold in zip(rows,self.program,self.fixture['independent_gold']):
|
||||||
|
self.assertEqual(row['metadata']['program_entry'],original);self.assertEqual(row['abstract'],gold['abstract']);self.assertEqual(row['title'],gold['title']);self.assertEqual(row['authors'],gold['authors'])
|
||||||
|
self.assertEqual(row['external_id'],'2026/'+gold['program_id']);self.assertEqual(row['identifiers'],{});self.assertEqual(row['submission_identifiers'],[])
|
||||||
|
self.assertEqual(row['status'],'accepted');self.assertEqual(row['track'],'program_unclassified');self.assertIsNone(row['metadata']['paper_type']);self.assertFalse(row['metadata']['original_decision_signature_verified'])
|
||||||
|
tags=[{'namespace':'ismir_program:presentation','value':gold['presentation_literal']}]+[{'namespace':'ismir_program:keyword','value':v} for v in dict.fromkeys(' '.join(k.split()) for k in gold['keywords'] if k.strip())]
|
||||||
|
self.assertEqual(row['tags'],tags);self.assertTrue(all(t['value'] for t in row['tags']))
|
||||||
|
if original['reviews']['publish_reviews']=='No':self.assertFalse(any(v for k,v in row['metadata']['program_entry']['reviews'].items() if k!='publish_reviews'))
|
||||||
|
for name,body in self.files.items():self.assertEqual((out/'evidence'/name).read_bytes(),body)
|
||||||
|
with self.assertRaisesRegex(ValueError,'Output exists'):c.collect(out,snapshots=src)
|
||||||
|
|
||||||
|
def test_duplicate_and_malformed_locator_or_incomplete_census_rejected(self):
|
||||||
|
listing=c.accepted_listing(self.accepted)
|
||||||
|
for kind in ('repeat','forum','non_numeric','missing_row'):
|
||||||
|
rows=copy.deepcopy(self.program)
|
||||||
|
if kind=='repeat':rows[1]['id']=rows[0]['id'];rows[1]['forum']=rows[0]['forum']
|
||||||
|
if kind=='forum':rows[0]['forum']='OriginalOpenReviewID'
|
||||||
|
if kind=='non_numeric':rows[0]['id']='abc';rows[0]['forum']='abc'
|
||||||
|
if kind=='missing_row':rows.pop()
|
||||||
|
with self.subTest(kind=kind),self.assertRaises(ValueError):c.program_observations(rows,listing)
|
||||||
|
|
||||||
|
def test_empty_fields_foreign_title_and_session_disagree_rejected(self):
|
||||||
|
listing=c.accepted_listing(self.accepted)
|
||||||
|
for field,value in [('authors',[]),('abstract',''),('title','A foreign paper'),('session',['99']),('keywords',[42])]:
|
||||||
|
rows=copy.deepcopy(self.program);rows[0]['content'][field]=value
|
||||||
|
with self.subTest(field=field),self.assertRaises(ValueError):c.program_observations(rows,listing)
|
||||||
|
|
||||||
|
def test_source_optin_no_cannot_publish_review_text(self):
|
||||||
|
rows=copy.deepcopy(self.program);r=next(r for r in rows if r['reviews']['publish_reviews']=='No');r['reviews']['meta_review']='Do not infer a negative decision'
|
||||||
|
with self.assertRaisesRegex(ValueError,'preference'):c.program_observations(rows,c.accepted_listing(self.accepted))
|
||||||
|
|
||||||
|
def test_parenthesized_affiliation_semicolons_do_not_split_people(self):
|
||||||
|
self.assertEqual(c.author_names('A Person (University A; Lab (X)); B Person (Institute B)'),['A Person','B Person'])
|
||||||
|
for raw in ['A Person (Broken','; A Person (Lab)','A Person )']:
|
||||||
|
with self.assertRaises(ValueError):c.author_names(raw)
|
||||||
|
|
||||||
|
def test_ambiguous_official_title_is_not_fused(self):
|
||||||
|
source=self.accepted.replace(b'FusID: Modality-Fused Semantic IDs for Generative Music Recommendation',b'Off-Manifold Robustness in Synthesizer Inversion with Joint Distribution Flow Matching',1)
|
||||||
|
with self.assertRaisesRegex(ValueError,'ambiguous'):c.accepted_listing(source)
|
||||||
|
|
||||||
|
def test_tamper_missing_file_and_symlink_fail_without_success_report(self):
|
||||||
|
for kind in ('manifest','body','missing','symlink'):
|
||||||
|
with tempfile.TemporaryDirectory() as tmp:
|
||||||
|
root=Path(tmp);src=root/'src';self.materialize(src);path=src/'ismir2026-json-research-v1'/('manifest.json' if kind=='manifest' else 'papers.json')
|
||||||
|
if kind=='missing':path.unlink()
|
||||||
|
elif kind=='symlink':other=root/'actual.json';path.rename(other);path.symlink_to(other)
|
||||||
|
else:path.write_bytes(path.read_bytes()+b' ')
|
||||||
|
with self.subTest(kind=kind),self.assertRaises((ValueError,FileNotFoundError)):c.collect(root/'out',snapshots=src)
|
||||||
|
self.assertTrue((root/'out/failed.json').exists());self.assertFalse((root/'out/collection-report.json').exists())
|
||||||
@@ -0,0 +1,146 @@
|
|||||||
|
import copy
|
||||||
|
import gzip
|
||||||
|
import hashlib
|
||||||
|
import json
|
||||||
|
from pathlib import Path
|
||||||
|
import tempfile
|
||||||
|
import unittest
|
||||||
|
|
||||||
|
from paper_atlas.collectors import mm2026 as c
|
||||||
|
|
||||||
|
|
||||||
|
class MM2026Test(unittest.TestCase):
|
||||||
|
@classmethod
|
||||||
|
def setUpClass(cls):
|
||||||
|
cls.fixture = json.loads(gzip.decompress((Path(__file__).parent/'fixtures/mm2026-source.json.gz').read_bytes()))
|
||||||
|
cls.files = {name:body.encode('utf-8') for name,body in cls.fixture['files'].items()}
|
||||||
|
cls.raw = cls.files[c.SNAPSHOT+'/program-data.json']
|
||||||
|
cls.data = json.loads(cls.raw)
|
||||||
|
cls.gold = cls.fixture['independent_inventory']
|
||||||
|
|
||||||
|
def materialize(self, root):
|
||||||
|
for name,body in self.files.items():
|
||||||
|
path = root/name; path.parent.mkdir(parents=True,exist_ok=True); path.write_bytes(body)
|
||||||
|
|
||||||
|
def parse_changed(self, data):
|
||||||
|
return c.parse_program(json.dumps(data,ensure_ascii=False).encode())
|
||||||
|
|
||||||
|
def test_all2143_ids_and4326_locations_against_independent_inventory(self):
|
||||||
|
rows,sessions = c.parse_program(self.raw)
|
||||||
|
self.assertEqual(len(rows),2143); self.assertEqual(len(sessions),157)
|
||||||
|
self.assertEqual(sum(len(r['contexts']) for r in rows),4326)
|
||||||
|
actual = {r['paper_id']:r for r in rows}
|
||||||
|
self.assertEqual(set(actual),{g['paper_id'] for g in self.gold['inventory']})
|
||||||
|
for gold in self.gold['inventory']:
|
||||||
|
r = actual[gold['paper_id']]
|
||||||
|
self.assertEqual((r['title'],r['authors'],r['track'],r['affiliations']),
|
||||||
|
(gold['title'],gold['authors'],gold['track_literal'],gold['affiliations']))
|
||||||
|
self.assertEqual(hashlib.sha256(r['abstract'].encode()).hexdigest(),gold['abstract_sha256'])
|
||||||
|
self.assertEqual([x['source_pointer'] for x in r['contexts']],gold['locations'])
|
||||||
|
|
||||||
|
def test_every_presentation_and_session_context_is_preserved(self):
|
||||||
|
rows,sessions = c.parse_program(self.raw)
|
||||||
|
for r in rows:
|
||||||
|
for context in r['contexts']:
|
||||||
|
node = self.data
|
||||||
|
for key in context['source_pointer'].split('/')[1:]:
|
||||||
|
node = node[int(key)] if isinstance(node,list) else node[key]
|
||||||
|
self.assertEqual(context['presentation'],{k:node[k] for k in c.PRESENTATION if k in node})
|
||||||
|
self.assertEqual(r['paper_id'],node['paper_id'])
|
||||||
|
for i,session in enumerate(sessions):
|
||||||
|
original = self.data['entries'][i]
|
||||||
|
self.assertEqual(session['source_pointer'],'/entries/'+str(i))
|
||||||
|
self.assertEqual(session['track'],'')
|
||||||
|
for key in original.keys()-{'papers','subsessions'}:
|
||||||
|
self.assertEqual(session[key],original[key])
|
||||||
|
for j,sub in enumerate(original.get('subsessions') or []):
|
||||||
|
saved = session['subsessions'][j]
|
||||||
|
self.assertEqual(saved['title'],sub['title'])
|
||||||
|
expected = [x for x in sub.get('items') or [] if 'paper_id' not in x]
|
||||||
|
self.assertEqual([x['fields'] for x in saved['nonpaper_items']],expected)
|
||||||
|
|
||||||
|
def test_all_tracks_retained_without_publication_or_negative_inference(self):
|
||||||
|
rows,_ = c.parse_program(self.raw)
|
||||||
|
self.assertEqual({r['track'] for r in rows},set(self.gold['unique_id_track_counts']))
|
||||||
|
for r in rows:
|
||||||
|
self.assertFalse(r['approved_for_import']); self.assertFalse(r['final_decision_verified'])
|
||||||
|
self.assertFalse(r['publication_identity_verified'])
|
||||||
|
self.assertEqual(r['source_locator'],{'namespace':'mm_program:program_id','value':r['paper_id']})
|
||||||
|
self.assertEqual(r['tags'],[{'namespace':'mm_program:track','value':r['track']}])
|
||||||
|
self.assertEqual(r['observed_status'],'listed_in_official_program')
|
||||||
|
self.assertTrue({'status','identifiers','submission_identifiers','paper_type','emails'}.isdisjoint(r))
|
||||||
|
missing = [r for r in rows if not r['authors']]
|
||||||
|
self.assertEqual({r['paper_id'] for r in missing},{'mfp4971','mfp7801'})
|
||||||
|
self.assertTrue(all(r['missing_fields']==['authors','abstract'] for r in missing))
|
||||||
|
self.assertEqual(sum(not r['abstract'] for r in rows),216)
|
||||||
|
|
||||||
|
def test_collect_evidence_hashes_and_no_contact_arrays_in_normalized_output(self):
|
||||||
|
with tempfile.TemporaryDirectory() as tmp:
|
||||||
|
root=Path(tmp);self.materialize(root/'src')
|
||||||
|
report=c.collect(root/'out',snapshots=root/'src')
|
||||||
|
self.assertEqual((report['records'],report['program_observations'],report['extra_presentation_observations']),(2143,4326,2183))
|
||||||
|
self.assertFalse(report['database_written']);self.assertFalse(report['pdfs_downloaded'])
|
||||||
|
self.assertEqual(report['network_requests'],0);self.assertEqual(report['approved_for_import'],0)
|
||||||
|
self.assertEqual(report['public_negative_access']['http_statuses'],[403]*4)
|
||||||
|
self.assertEqual(report['track_counts'],self.gold['unique_id_track_counts'])
|
||||||
|
for name,body in self.files.items():
|
||||||
|
self.assertEqual((root/'out/evidence'/Path(name).name).read_bytes(),body)
|
||||||
|
n=0
|
||||||
|
for item in report['files']:
|
||||||
|
raw=Path(item['file']).read_bytes();self.assertEqual(hashlib.sha256(raw).hexdigest(),item['file_sha256'])
|
||||||
|
self.assertNotIn(b'"emails"',raw);self.assertEqual(len(raw.splitlines()),item['count']);n+=item['count']
|
||||||
|
self.assertEqual(item['format'],'mm_program_observation_v1_not_ingestion_record')
|
||||||
|
self.assertEqual(n,2143)
|
||||||
|
raw=(root/'out/sessions.jsonl').read_bytes()
|
||||||
|
self.assertEqual(hashlib.sha256(raw).hexdigest(),report['sessions_file_sha256']);self.assertNotIn(b'"emails"',raw)
|
||||||
|
with self.assertRaisesRegex(ValueError,'Output exists'):c.collect(root/'out',snapshots=root/'src')
|
||||||
|
|
||||||
|
def test_conflicting_core_version_is_not_silently_merged(self):
|
||||||
|
data=copy.deepcopy(self.data)
|
||||||
|
first=data['entries'][0]['papers'][0]
|
||||||
|
data['entries'][0]['papers'].append({**first,'abstract':'conflicting abstract'})
|
||||||
|
with self.assertRaisesRegex(ValueError,'Conflicting core'):self.parse_changed(data)
|
||||||
|
|
||||||
|
def test_unknown_nesting_cannot_silently_lose_observations(self):
|
||||||
|
data=copy.deepcopy(self.data)
|
||||||
|
moved=data['entries'][0]['papers'].pop()
|
||||||
|
data['entries'][0]['unreviewed_bucket']=[moved]
|
||||||
|
with self.assertRaisesRegex(ValueError,'Unhandled'):self.parse_changed(data)
|
||||||
|
|
||||||
|
def test_track_drift_unknown_fields_or_missing_title_require_review(self):
|
||||||
|
for kind in ('outer_track','paper_track','field','title','authors'):
|
||||||
|
data=copy.deepcopy(self.data);paper=data['entries'][0]['papers'][0]
|
||||||
|
if kind=='outer_track':data['entries'][0]['track']='Main Track'
|
||||||
|
elif kind=='paper_track':paper['track']='New Track'
|
||||||
|
elif kind=='field':paper['doi']='unreviewed'
|
||||||
|
elif kind=='title':paper['title']=''
|
||||||
|
else:paper['authors']='Unparsed author string'
|
||||||
|
with self.subTest(kind=kind),self.assertRaises(ValueError):self.parse_changed(data)
|
||||||
|
|
||||||
|
def test_edition_or_census_change_requires_review(self):
|
||||||
|
for kind in ('edition','count'):
|
||||||
|
data=copy.deepcopy(self.data)
|
||||||
|
if kind=='edition':data['generated_at']='2025-11-14T18:36:26.860677+00:00'
|
||||||
|
else:data['entries'][0]['papers'].pop()
|
||||||
|
with self.subTest(kind=kind),self.assertRaises(ValueError):self.parse_changed(data)
|
||||||
|
|
||||||
|
def test_corruption_missing_evidence_and_symlink_fail_without_completed_report(self):
|
||||||
|
for kind in ('manifest','program','probe','missing','symlink'):
|
||||||
|
with tempfile.TemporaryDirectory() as tmp:
|
||||||
|
root=Path(tmp);self.materialize(root/'src');directory=root/'src'/c.SNAPSHOT
|
||||||
|
name='manifest.json' if kind=='manifest' else 'openreview-rejected-probe.json' if kind=='probe' else 'program-data.json'
|
||||||
|
path=directory/name
|
||||||
|
if kind=='missing':path.unlink()
|
||||||
|
elif kind=='symlink':
|
||||||
|
original=directory/'original.json';path.rename(original);path.symlink_to(original)
|
||||||
|
else:path.write_bytes(path.read_bytes()+b' ')
|
||||||
|
with self.subTest(kind=kind),self.assertRaises((ValueError,FileNotFoundError)):c.collect(root/'out',snapshots=root/'src')
|
||||||
|
self.assertTrue((root/'out/failed.json').exists());self.assertFalse((root/'out/collection-report.json').exists())
|
||||||
|
|
||||||
|
def test_literal_abstract_math_and_nonpaper_session_events_retained(self):
|
||||||
|
rows,sessions=c.parse_program(self.raw)
|
||||||
|
self.assertTrue(any('\\' in r['abstract'] for r in rows))
|
||||||
|
self.assertTrue(any('\n' in r['abstract'] for r in rows))
|
||||||
|
self.assertTrue(any(s['type']=='coffee_break' for s in sessions))
|
||||||
|
self.assertTrue(any(s['type']=='industry_talk' for s in sessions))
|
||||||
|
self.assertEqual(c.without_emails({'emails':['contact'],'items':[{'emails':['contact'],'title':'talk'}]}),{'items':[{'title':'talk'}]})
|
||||||
@@ -0,0 +1,273 @@
|
|||||||
|
from contextlib import redirect_stdout
|
||||||
|
import io
|
||||||
|
import os
|
||||||
|
import uuid
|
||||||
|
import gzip
|
||||||
|
import hashlib
|
||||||
|
import json
|
||||||
|
from pathlib import Path
|
||||||
|
import re
|
||||||
|
import tempfile
|
||||||
|
import unittest
|
||||||
|
|
||||||
|
from paper_atlas.collectors import mm_publications2026 as c
|
||||||
|
|
||||||
|
|
||||||
|
class MMPublications2026Tests(unittest.TestCase):
|
||||||
|
@classmethod
|
||||||
|
def setUpClass(cls):
|
||||||
|
fixtures=Path(__file__).parent/'fixtures'
|
||||||
|
cls.fixture=json.loads(gzip.decompress((fixtures/'mm2026-publications.json.gz').read_bytes()))
|
||||||
|
cls.program_fixture=json.loads(gzip.decompress((fixtures/'mm2026-source.json.gz').read_bytes()))
|
||||||
|
cls.format_gold=json.loads((fixtures/'mm2026-title-formatting-review.json').read_text())
|
||||||
|
overrides={g['doi']:g for g in [*cls.format_gold['gold'],cls.format_gold['held_gold']]}
|
||||||
|
cls.gold=[overrides.get(g['doi'],g) for g in cls.fixture['independent_gold']]
|
||||||
|
cls.temp=tempfile.TemporaryDirectory();cls.root=Path(cls.temp.name)
|
||||||
|
cls.materialize(cls.root/'registry',cls.fixture['files'])
|
||||||
|
cls.materialize(cls.root/'program',cls.program_fixture['files'])
|
||||||
|
cls.out=cls.root/'out'
|
||||||
|
cls.report=c.collect(cls.out,registry_snapshots=cls.root/'registry',program_snapshots=cls.root/'program',parent_snapshot=cls.root/'registry/parent.json')
|
||||||
|
cls.rows=[json.loads(line) for f in cls.report['files'] for line in Path(f['file']).read_bytes().splitlines()]
|
||||||
|
cls.by_doi={r['external_id']:r for r in cls.rows}
|
||||||
|
|
||||||
|
@classmethod
|
||||||
|
def tearDownClass(cls):
|
||||||
|
cls.temp.cleanup()
|
||||||
|
|
||||||
|
@staticmethod
|
||||||
|
def materialize(root,files):
|
||||||
|
for name,body in files.items():
|
||||||
|
path=root/name;path.parent.mkdir(parents=True,exist_ok=True);path.write_bytes(body.encode('utf-8'))
|
||||||
|
|
||||||
|
def test_all1918_publications_against_independent_registry_program_gold(self):
|
||||||
|
self.assertEqual(len(self.rows),1918);self.assertEqual(len(self.by_doi),1918)
|
||||||
|
self.assertEqual(set(self.by_doi),{g['doi'] for g in self.gold})
|
||||||
|
for g in self.gold:
|
||||||
|
r=self.by_doi[g['doi']];m=r['metadata'];p=m['program_fields']
|
||||||
|
self.assertEqual((r['title'],r['authors'],r['track']),(g['title'],g['authors'],g['track']))
|
||||||
|
self.assertEqual(m['program_match_kind'],g['kind'])
|
||||||
|
self.assertEqual(m['registry_source_evidence']['json_pointer'],'/message/items/'+str(g['source_position']))
|
||||||
|
self.assertEqual(m['registry_source_evidence']['file'],'evidence/registry/'+g['source_page'])
|
||||||
|
self.assertEqual(m['registry_source_evidence']['canonical_object_json_sha256'],g['registry_object_sha256'])
|
||||||
|
self.assertEqual(p['paper_id'] if p else None,g['program_id'])
|
||||||
|
self.assertEqual([x['source_pointer'] for x in p['contexts']] if p else [],g['program_context_pointers'])
|
||||||
|
|
||||||
|
def test_abstract_source_is_separate_and_verbatim_not_fabricated_registry_abstract(self):
|
||||||
|
count=0
|
||||||
|
for g in self.gold:
|
||||||
|
r=self.by_doi[g['doi']];m=r['metadata']
|
||||||
|
self.assertFalse(m['registry_abstract_available']);self.assertFalse(m['registry_fields'].get('abstract'))
|
||||||
|
if g['abstract_sha256'] is not None:
|
||||||
|
count+=1;self.assertEqual(hashlib.sha256(r['abstract'].encode()).hexdigest(),g['abstract_sha256'])
|
||||||
|
self.assertEqual(r['abstract'],m['program_fields']['abstract'])
|
||||||
|
self.assertEqual(m['abstract_source'],'official_program_reviewed_title_formatting_correspondence' if g['kind'].startswith('unique_program_format_') else 'official_program_exact_metadata_correspondence')
|
||||||
|
else:
|
||||||
|
self.assertIsNone(r['abstract']);self.assertEqual(m['abstract_source'],'not_available_from_reviewed_sources')
|
||||||
|
self.assertEqual(count,1743);self.assertEqual(self.report['output_missing_abstract_count'],175)
|
||||||
|
|
||||||
|
def test_all_program_tracks_and_4326_contexts_survive_in_matches_or_unmatched(self):
|
||||||
|
programs={r['metadata']['program_fields']['paper_id']:r['metadata']['program_fields'] for r in self.rows if r['metadata']['program_fields']}
|
||||||
|
for p in self.report['unmatched_program']:
|
||||||
|
self.assertNotIn(p['paper_id'],programs);programs[p['paper_id']]=p
|
||||||
|
self.assertEqual(len(programs),2143);self.assertEqual(self.report['unmatched_program_count'],391)
|
||||||
|
gold=self.program_fixture['independent_inventory']['inventory']
|
||||||
|
self.assertEqual(set(programs),{g['paper_id'] for g in gold})
|
||||||
|
for g in gold:
|
||||||
|
p=programs[g['paper_id']]
|
||||||
|
self.assertEqual((p['title'],p['authors'],p['track']),(g['title'],g['authors'],g['track_literal']))
|
||||||
|
self.assertEqual([x['source_pointer'] for x in p['contexts']],g['locations'])
|
||||||
|
self.assertEqual(sum(len(p['contexts']) for p in programs.values()),4326)
|
||||||
|
self.assertEqual(len({p['track'] for p in programs.values()}),28)
|
||||||
|
|
||||||
|
def test_no_person_alias_original_submission_decision_or_research_type_inference(self):
|
||||||
|
for r in self.rows:
|
||||||
|
doi=r['external_id'];m=r['metadata']
|
||||||
|
self.assertEqual(r['identifiers'],{'doi':doi})
|
||||||
|
self.assertEqual(r['submission_identifiers'],[{'scheme':'publication_doi','value':doi}])
|
||||||
|
self.assertEqual((r['source'],r['venue'],r['year']),(c.SOURCE,'MM',2026))
|
||||||
|
self.assertEqual(r['status'],'accepted');self.assertIn('not an original signed decision',r['raw_decision'])
|
||||||
|
self.assertFalse(m['original_decision_signature_verified']);self.assertFalse(m['research_type_classified'])
|
||||||
|
self.assertFalse(m['scientific_work_identity_approved'])
|
||||||
|
self.assertTrue({'decided_at','decision_date','source_decision_id','paper_type'}.isdisjoint(r))
|
||||||
|
self.assertIn({'namespace':c.SOURCE+':crossref:type','value':'proceedings-article'},r['tags'])
|
||||||
|
self.assertTrue({x['namespace'] for x in r['tags']} <= {c.SOURCE+':crossref:type',c.SOURCE+':mm_program:track'})
|
||||||
|
for track in ('interactive_art','industry_talk','demo_and_video'):
|
||||||
|
self.assertTrue(any(r['track']==track for r in self.rows))
|
||||||
|
variants=[r for r in self.rows if r['track']=='proceedings_unclassified']
|
||||||
|
self.assertEqual(len(variants),166)
|
||||||
|
self.assertTrue(all(r['metadata']['program_fields'] is None and r['abstract'] is None for r in variants))
|
||||||
|
|
||||||
|
def test_original_bibliography_evidence_kept_without_duplicating_reference_payload(self):
|
||||||
|
authors_by_page={}
|
||||||
|
for name,body in self.fixture['files'].items():
|
||||||
|
actual=self.out/'evidence/registry'/('parent.json' if name=='parent.json' else Path(name).name)
|
||||||
|
self.assertEqual(actual.read_bytes(),body.encode())
|
||||||
|
if re.fullmatch(r'page-[0-9]+\.json',Path(name).name):
|
||||||
|
authors_by_page[str(actual.relative_to(self.out))]=[row['author'] for row in json.loads(body)['message']['items']]
|
||||||
|
for name,body in self.program_fixture['files'].items():
|
||||||
|
self.assertEqual((self.out/'evidence/program'/Path(name).name).read_bytes(),body.encode())
|
||||||
|
for r in self.rows:
|
||||||
|
m=r['metadata'];self.assertNotIn('reference',m['registry_fields'])
|
||||||
|
self.assertEqual(m['registry_fields']['title'],[r['title']])
|
||||||
|
self.assertEqual(m['registry_fields']['author'],authors_by_page[m['registry_source_evidence']['file']][int(m['registry_source_evidence']['json_pointer'].rsplit('/',1)[-1])])
|
||||||
|
for f in self.report['files']:
|
||||||
|
raw=Path(f['file']).read_bytes();self.assertEqual(hashlib.sha256(raw).hexdigest(),f['file_sha256'])
|
||||||
|
self.assertEqual(len(raw.splitlines()),f['count']);self.assertEqual(f['status'],'partial')
|
||||||
|
|
||||||
|
def test_eight_single_name_fields_preserved_without_splitting(self):
|
||||||
|
n=0
|
||||||
|
for g in self.gold:
|
||||||
|
r=self.by_doi[g['doi']];flags=r['metadata']['review_flags']
|
||||||
|
actual=[x['author_position'] for x in flags if x['kind']=='single_registered_name_field_not_split']
|
||||||
|
self.assertEqual(actual,g['single_name_positions']);n+=len(actual)
|
||||||
|
for i in actual:
|
||||||
|
self.assertEqual(r['authors'][i-1],r['metadata']['registry_fields']['author'][i-1]['family'].strip())
|
||||||
|
self.assertEqual(n,8)
|
||||||
|
|
||||||
|
def test_structured_title_markup_retained_and_flagged_without_flattening(self):
|
||||||
|
count=0
|
||||||
|
for g in self.gold:
|
||||||
|
if re.search(r'<[^>]+>|&[A-Za-z#0-9]+;',g['title']):
|
||||||
|
count+=1;r=self.by_doi[g['doi']]
|
||||||
|
self.assertEqual(r['title'],g['title'])
|
||||||
|
self.assertIn({'kind':'literal_title_markup_or_entity_retained_requires_review'},r['metadata']['review_flags'])
|
||||||
|
self.assertGreater(count,0);self.assertEqual(count,self.report['registered_title_markup_flags'])
|
||||||
|
|
||||||
|
|
||||||
|
def test_formatting_gold_preserves_raw_fields_and_copies_exact_program_evidence(self):
|
||||||
|
self.assertEqual(len(self.format_gold['gold']),26)
|
||||||
|
for gold in self.format_gold['gold']:
|
||||||
|
row=self.by_doi[gold['doi']];m=row['metadata'];p=m['program_fields']
|
||||||
|
self.assertEqual((row['title'],row['authors']),(gold['title'],gold['authors']))
|
||||||
|
self.assertEqual((p['paper_id'],row['track']),(gold['program_id'],gold['track']))
|
||||||
|
self.assertEqual(hashlib.sha256(row['abstract'].encode()).hexdigest(),gold['abstract_sha256'])
|
||||||
|
comparison=m['program_title_formatting_comparison']
|
||||||
|
self.assertTrue(comparison['raw_titles_preserved']);self.assertFalse(comparison['work_identity_approved'])
|
||||||
|
self.assertEqual(set(comparison['program_comparison_texts']),{p['paper_id']})
|
||||||
|
self.assertFalse(m['scientific_work_identity_approved'])
|
||||||
|
held=self.by_doi[self.format_gold['author_disagreement_hold']['doi']]
|
||||||
|
self.assertEqual(held['metadata']['program_title_candidates'],['mfp7801'])
|
||||||
|
self.assertEqual(held['metadata']['program_exact_candidates'],[])
|
||||||
|
self.assertEqual(held['metadata']['program_match_kind'],'program_format_author_version_requires_review')
|
||||||
|
self.assertIsNone(held['abstract']);self.assertEqual(held['track'],'proceedings_unclassified')
|
||||||
|
|
||||||
|
def test_reviewed_formatting_keeps_superscript_subscript_and_decodes_entities_once(self):
|
||||||
|
self.assertEqual(c.formatting_text('<i>A</i> H<sup>2</sup> B<sub>1</sub> & C'),'A H2 B1 & C')
|
||||||
|
self.assertEqual(c.formatting_text('A &amp; B'),'A & B')
|
||||||
|
|
||||||
|
def test_unknown_attributes_unbalanced_or_encoded_markup_not_discarded(self):
|
||||||
|
for title in ['A <model>B</model>','A <i class="x">B</i>','A <i>B','A </i>B',
|
||||||
|
'A <sup/>B','A <!--hidden-->B','<!DOCTYPE html>A','A <i>B</i>',
|
||||||
|
'A <sup','A <i>B</sup>','<![CDATA[A]]>']:
|
||||||
|
with self.subTest(title=title):self.assertIsNone(c.formatting_text(title))
|
||||||
|
|
||||||
|
def test_ambiguous_program_and_author_multiset_differences_remain_held(self):
|
||||||
|
def observation(pid,authors):
|
||||||
|
return {'paper_id':pid,'title':'M3 Model','authors':authors,'track':'Main Track'}
|
||||||
|
rows=[{'title':'M<sup>3</sup> Model','authors':['A','B']}]
|
||||||
|
kinds,_=c.match_program(rows,[observation('p1',['A','B']),observation('p2',['B','A'])])
|
||||||
|
self.assertEqual(kinds,{'multiple_program_ids_same_title_authors_hold':1})
|
||||||
|
self.assertNotIn('program',rows[0])
|
||||||
|
rows=[{'title':'M<sup>3</sup> Model','authors':['A','A','B']}]
|
||||||
|
c.match_program(rows,[observation('p1',['A','B'])])
|
||||||
|
self.assertEqual(rows[0]['kind'],'program_format_author_version_requires_review')
|
||||||
|
self.assertNotIn('program',rows[0])
|
||||||
|
|
||||||
|
def test_literal_title_author_disagreement_does_not_use_formatting_bypass(self):
|
||||||
|
rows=[{'title':'M<sup>3</sup> Model','authors':['A','B']}]
|
||||||
|
observations=[{'paper_id':'literal','title':'M<sup>3</sup> Model','authors':['Other'],'track':'Main Track'},
|
||||||
|
{'paper_id':'formatted','title':'M3 Model','authors':['A','B'],'track':'Main Track'}]
|
||||||
|
c.match_program(rows,observations)
|
||||||
|
self.assertEqual(rows[0]['kind'],'program_author_version_requires_review')
|
||||||
|
self.assertEqual(rows[0]['program_title_candidates'],['literal']);self.assertNotIn('program',rows[0])
|
||||||
|
|
||||||
|
|
||||||
|
def test_corrupt_registry_parent_or_program_keeps_failure_not_complete_output(self):
|
||||||
|
for kind in ('registry_manifest','registry_page','parent','program'):
|
||||||
|
with tempfile.TemporaryDirectory() as tmp:
|
||||||
|
root=Path(tmp);self.materialize(root/'registry',self.fixture['files']);self.materialize(root/'program',self.program_fixture['files'])
|
||||||
|
path={'registry_manifest':root/'registry'/c.SNAPSHOT/'manifest.json','registry_page':root/'registry'/c.SNAPSHOT/'page-001.json','parent':root/'registry/parent.json','program':root/'program'/c.program.SNAPSHOT/'program-data.json'}[kind]
|
||||||
|
path.write_bytes(path.read_bytes()+b' ')
|
||||||
|
with self.subTest(kind=kind),self.assertRaises(ValueError):
|
||||||
|
c.collect(root/'out',registry_snapshots=root/'registry',program_snapshots=root/'program',parent_snapshot=root/'registry/parent.json')
|
||||||
|
self.assertTrue((root/'out/failed.json').exists());self.assertFalse((root/'out/collection-report.json').exists())
|
||||||
|
|
||||||
|
def test_symlink_source_rejected_and_no_previous_output_overwrite(self):
|
||||||
|
with tempfile.TemporaryDirectory() as tmp:
|
||||||
|
root=Path(tmp);self.materialize(root/'registry',self.fixture['files'])
|
||||||
|
path=root/'registry/parent.json';original=root/'registry/actual.json';path.rename(original);path.symlink_to(original)
|
||||||
|
with self.assertRaisesRegex(ValueError,'Unsafe'):
|
||||||
|
c.collect(root/'out',registry_snapshots=root/'registry',program_snapshots=self.root/'program',parent_snapshot=path)
|
||||||
|
with self.assertRaisesRegex(ValueError,'Output exists'):
|
||||||
|
c.collect(self.out,registry_snapshots=self.root/'registry',program_snapshots=self.root/'program',parent_snapshot=self.root/'registry/parent.json')
|
||||||
|
|
||||||
|
def test_complete_report_retains_source_partitions_not_import_or_final_dedup_claim(self):
|
||||||
|
self.assertEqual(self.report['program_match_counts'],{'unique_program_title_fullauthor_agreement_not_identity_approval':1726,'unique_program_format_title_fullauthor_agreement_not_identity_approval':26,'program_author_version_requires_review':122,'program_format_author_version_requires_review':1,'no_exact_program_title_requires_review':43})
|
||||||
|
self.assertEqual(self.report['registry_missing_abstract_count'],1918)
|
||||||
|
self.assertEqual(self.report['abstracts_from_exact_program_correspondence'],1743)
|
||||||
|
self.assertEqual(self.report['abstracts_from_literal_program_correspondence'],1717)
|
||||||
|
self.assertEqual(self.report['abstracts_from_reviewed_title_formatting'],26)
|
||||||
|
self.assertEqual(self.report['unmatched_program_count'],391)
|
||||||
|
self.assertEqual(sum(f['count'] for f in self.report['files']),1918)
|
||||||
|
self.assertEqual(self.report['approved_for_import'],0);self.assertFalse(self.report['database_written'])
|
||||||
|
self.assertFalse(self.report['pdfs_downloaded']);self.assertEqual(self.report['network_requests'],0)
|
||||||
|
self.assertIn('unknown',self.report['public_negative_availability'])
|
||||||
|
|
||||||
|
|
||||||
|
def assert_native_ingestion(self, pg_service=None):
|
||||||
|
# An isolated synthetic database exercises the writer contract, not a
|
||||||
|
# production identity approval. All source rows/track partitions run.
|
||||||
|
from paper_atlas.db.store import Store
|
||||||
|
from paper_atlas.db.backend import open_store
|
||||||
|
from paper_atlas.ingestion.records import ingest_file
|
||||||
|
from paper_atlas.postgres.client import connect
|
||||||
|
from paper_atlas.postgres.migration import migrate
|
||||||
|
from paper_atlas.postgres.schema import quote
|
||||||
|
with tempfile.TemporaryDirectory() as tmp:
|
||||||
|
root=Path(tmp);db=root/'seed.sqlite';seed=self.rows[0]
|
||||||
|
store=Store(db)
|
||||||
|
paper,record=store.record('arxiv','synthetic-mm-anchor','https://example.org/arxiv',
|
||||||
|
seed['title'],'Existing canonical abstract',seed['authors'],seed['identifiers'])
|
||||||
|
store.decision(paper,record,'ICLR',2025,'main','synthetic-old-forum',
|
||||||
|
'https://example.org/decision','rejected','Reject')
|
||||||
|
before=tuple(store.conn.execute('SELECT * FROM papers WHERE id=?',(paper,)).fetchone())
|
||||||
|
store.conn.commit();store.conn.close()
|
||||||
|
schema='pa_mm_pub_'+uuid.uuid4().hex[:12] if pg_service else None
|
||||||
|
try:
|
||||||
|
with redirect_stdout(io.StringIO()):
|
||||||
|
if pg_service:migrate(db,pg_service,root/'migration.json',schema=schema,min_free_gib=0)
|
||||||
|
for f in self.report['files']:
|
||||||
|
ingest_file(None if pg_service else db,f['file'],pg_service=pg_service,pg_schema=schema)
|
||||||
|
store=open_store(None if pg_service else db,pg_service=pg_service,pg_schema=schema)
|
||||||
|
try:
|
||||||
|
conn=store.conn
|
||||||
|
self.assertEqual(tuple(conn.execute('SELECT * FROM papers WHERE id=?',(paper,)).fetchone()),before)
|
||||||
|
self.assertEqual(conn.execute('SELECT status FROM decisions WHERE submission_id=1').fetchone()[0],'rejected')
|
||||||
|
actual={r[0]:r for r in conn.execute("SELECT sr.external_id,sr.paper_id,sr.title,sr.abstract,sr.authors_json,sr.metadata_json,sr.id FROM source_records sr JOIN sources s ON s.id=sr.source_id WHERE s.name=?",(c.SOURCE,))}
|
||||||
|
self.assertEqual(set(actual),set(self.by_doi))
|
||||||
|
self.assertEqual(actual[seed['external_id']][1],paper)
|
||||||
|
for doi,row in self.by_doi.items():
|
||||||
|
got=actual[doi];self.assertEqual((got[2],got[3],json.loads(got[4])),(row['title'],row['abstract'],row['authors']))
|
||||||
|
meta=json.loads(got[5]);self.assertEqual(meta.pop('record_identifiers'),{'doi':doi,c.SOURCE:doi});self.assertEqual(meta,row['metadata'])
|
||||||
|
tags=conn.execute("SELECT t.namespace,t.value,count(*) FROM paper_tags pt JOIN tags t ON t.id=pt.tag_id GROUP BY t.namespace,t.value").fetchall()
|
||||||
|
self.assertEqual(sum(r[2] for r in tags if r[0]==c.SOURCE+':crossref:type'),1918)
|
||||||
|
self.assertEqual(sum(r[2] for r in tags if r[0]==c.SOURCE+':mm_program:track'),1752)
|
||||||
|
self.assertEqual({r[0] for r in tags},{c.SOURCE+':crossref:type',c.SOURCE+':mm_program:track'})
|
||||||
|
decisions=conn.execute("SELECT d.status,d.decision_date,count(*) FROM decisions d JOIN submissions s ON s.id=d.submission_id JOIN source_records r ON r.id=s.source_record_id JOIN sources src ON src.id=r.source_id WHERE src.name=? GROUP BY d.status,d.decision_date",(c.SOURCE,)).fetchall()
|
||||||
|
self.assertEqual([tuple(r) for r in decisions],[('accepted',None,1918)])
|
||||||
|
self.assertEqual(conn.execute("SELECT count(*) FROM submission_identifiers WHERE scheme='publication_doi'").fetchone()[0],1918)
|
||||||
|
coverage=conn.execute("SELECT es.status,sum(es.imported_count) FROM edition_sources es JOIN sources s ON s.id=es.source_id WHERE s.name=? GROUP BY es.status",(c.SOURCE,)).fetchall()
|
||||||
|
self.assertEqual([tuple(r) for r in coverage],[('partial',1918)])
|
||||||
|
finally:store.conn.close()
|
||||||
|
finally:
|
||||||
|
if pg_service:
|
||||||
|
with connect(pg_service) as pg:
|
||||||
|
for name, in pg.execute('SELECT nspname FROM pg_namespace WHERE starts_with(nspname,%s)',(schema,)).fetchall():
|
||||||
|
pg.execute('DROP SCHEMA '+quote(name)+' CASCADE')
|
||||||
|
|
||||||
|
def test_all1918_native_sqlite_source_tags_and_prior_rejection_survive(self):
|
||||||
|
self.assert_native_ingestion()
|
||||||
|
|
||||||
|
@unittest.skipUnless(os.environ.get('PAPER_ATLAS_TEST_PG_SERVICE'),'Requires approved remote PostgreSQL test service')
|
||||||
|
def test_all1918_native_postgres_source_tags_and_prior_rejection_survive(self):
|
||||||
|
self.assert_native_ingestion(os.environ['PAPER_ATLAS_TEST_PG_SERVICE'])
|
||||||
@@ -0,0 +1,101 @@
|
|||||||
|
import gzip
|
||||||
|
import hashlib
|
||||||
|
import json
|
||||||
|
from pathlib import Path
|
||||||
|
import tempfile
|
||||||
|
import unittest
|
||||||
|
from unittest.mock import patch
|
||||||
|
|
||||||
|
from paper_atlas.collectors import mm2026
|
||||||
|
from paper_atlas.audits.source_identity import audit, load_inputs, load_mm_program_inputs
|
||||||
|
|
||||||
|
|
||||||
|
class MMProgramIdentityInputTests(unittest.TestCase):
|
||||||
|
@classmethod
|
||||||
|
def setUpClass(cls):
|
||||||
|
cls.fixture = json.loads(gzip.decompress((Path(__file__).parent/'fixtures/mm2026-source.json.gz').read_bytes()))
|
||||||
|
|
||||||
|
def setUp(self):
|
||||||
|
self.tmp = tempfile.TemporaryDirectory()
|
||||||
|
self.addCleanup(self.tmp.cleanup)
|
||||||
|
self.root = Path(self.tmp.name)
|
||||||
|
src = self.root/'source'
|
||||||
|
for name,body in self.fixture['files'].items():
|
||||||
|
path=src/name;path.parent.mkdir(parents=True,exist_ok=True);path.write_bytes(body.encode())
|
||||||
|
mm2026.collect(self.root/'program', snapshots=src)
|
||||||
|
self.report_path = self.root/'program/collection-report.json'
|
||||||
|
|
||||||
|
def change_report(self, change):
|
||||||
|
report=json.loads(self.report_path.read_text());change(report)
|
||||||
|
self.report_path.write_text(json.dumps(report))
|
||||||
|
|
||||||
|
def assert_rejected_before_database(self):
|
||||||
|
with patch('paper_atlas.audits.source_identity.ReadConnection') as connection:
|
||||||
|
with self.assertRaises(ValueError):
|
||||||
|
audit(None,self.root/'audit',pg_service='unused',pg_schema='research',mm_program_report=self.report_path)
|
||||||
|
connection.assert_not_called()
|
||||||
|
self.assertFalse((self.root/'audit').exists())
|
||||||
|
|
||||||
|
def test_all2143_programs_are_unknown_readonly_inputs_not_import_records(self):
|
||||||
|
rows,inputs=load_mm_program_inputs(self.report_path)
|
||||||
|
self.assertEqual(len(rows),2143);self.assertEqual(len(inputs),28)
|
||||||
|
self.assertEqual(sum(len(r['program_contexts']) for r in rows),4326)
|
||||||
|
actual={r['program_id']:r for r in rows}
|
||||||
|
for gold in self.fixture['independent_inventory']['inventory']:
|
||||||
|
r=actual[gold['paper_id']]
|
||||||
|
self.assertEqual((r['title'],r['authors'],r['track']),
|
||||||
|
(gold['title'],gold['authors'],gold['track_literal']))
|
||||||
|
self.assertEqual(r['external_id'],'2026/'+gold['paper_id'])
|
||||||
|
self.assertEqual(r['status'],'unknown');self.assertEqual(r['identifiers'],{})
|
||||||
|
self.assertEqual(r['source_observed_status'],'listed_in_official_program')
|
||||||
|
self.assertFalse(r['publication_identity_verified']);self.assertFalse(r['final_decision_verified'])
|
||||||
|
self.assertFalse(r['contribution_type_verified'])
|
||||||
|
self.assertEqual([c['source_pointer'] for c in r['program_contexts']],gold['locations'])
|
||||||
|
self.assertEqual({r['program_id'] for r in rows if not r['authors']},{'mfp4971','mfp7801'})
|
||||||
|
self.assertTrue(all('unknown is not a source decision' in i['manifest']['scope'] for i in inputs))
|
||||||
|
self.assertFalse(any(self.root.rglob('*.manifest.json')))
|
||||||
|
|
||||||
|
def test_untrusted_absolute_file_paths_are_not_followed(self):
|
||||||
|
self.change_report(lambda r:[f.update(file='/untrusted/'+Path(f['file']).name) for f in r['files']])
|
||||||
|
rows,inputs=load_mm_program_inputs(self.report_path)
|
||||||
|
self.assertEqual(len(rows),2143)
|
||||||
|
self.assertTrue(all(Path(i['file']).parent==self.report_path.parent for i in inputs))
|
||||||
|
|
||||||
|
def test_program_files_are_not_silently_accepted_by_regular_ingestion_input_loader(self):
|
||||||
|
report=json.loads(self.report_path.read_text())
|
||||||
|
with self.assertRaises(FileNotFoundError):load_inputs([report['files'][0]['file']])
|
||||||
|
|
||||||
|
def test_corrupt_jsonl_fails_before_connection_or_output(self):
|
||||||
|
report=json.loads(self.report_path.read_text());path=Path(report['files'][0]['file'])
|
||||||
|
path.write_bytes(path.read_bytes()+b' ')
|
||||||
|
self.assert_rejected_before_database()
|
||||||
|
|
||||||
|
def test_rehashed_forged_identifiers_decisions_and_contexts_rejected(self):
|
||||||
|
original=self.report_path.read_bytes();report=json.loads(original);path=Path(report['files'][0]['file']);raw=path.read_bytes()
|
||||||
|
for field,value in [('identifiers',{'openreview':'invented'}),('status','accepted'),('contexts',[])]:
|
||||||
|
path.write_bytes(raw);self.report_path.write_bytes(original)
|
||||||
|
lines=raw.splitlines();row=json.loads(lines[0]);row[field]=value;lines[0]=json.dumps(row).encode()
|
||||||
|
changed=b'\n'.join(lines)+b'\n';path.write_bytes(changed)
|
||||||
|
self.change_report(lambda r:r['files'][0].update(file_sha256=hashlib.sha256(changed).hexdigest()))
|
||||||
|
with self.subTest(field=field):self.assert_rejected_before_database()
|
||||||
|
|
||||||
|
def test_missing_or_repeated_track_rejected(self):
|
||||||
|
original=self.report_path.read_bytes()
|
||||||
|
for kind in ('missing','repeated'):
|
||||||
|
self.report_path.write_bytes(original)
|
||||||
|
if kind=='missing':self.change_report(lambda r:r['files'].pop())
|
||||||
|
else:self.change_report(lambda r:r['files'].__setitem__(1,r['files'][0]))
|
||||||
|
with self.subTest(kind=kind):self.assert_rejected_before_database()
|
||||||
|
|
||||||
|
def test_raw_source_corruption_or_symlink_rejected(self):
|
||||||
|
path=self.report_path.parent/'evidence/program-data.json';raw=path.read_bytes()
|
||||||
|
path.write_bytes(raw+b' ');self.assert_rejected_before_database()
|
||||||
|
path.write_bytes(raw);saved=path.with_name('original.json');path.rename(saved);path.symlink_to(saved)
|
||||||
|
self.assert_rejected_before_database()
|
||||||
|
|
||||||
|
def test_mixed_input_modes_rejected_before_connection(self):
|
||||||
|
with patch('paper_atlas.audits.source_identity.ReadConnection') as connection:
|
||||||
|
for files,program in [([],None),([Path('unused')],self.report_path)]:
|
||||||
|
with self.assertRaisesRegex(ValueError,'Choose source files'):
|
||||||
|
audit(files,self.root/'out',pg_service='unused',pg_schema='research',mm_program_report=program)
|
||||||
|
connection.assert_not_called()
|
||||||
Reference in New Issue
Block a user