Compare commits

..

13 Commits

Author SHA1 Message Date
codex-bot ae8bcd4c14 Close the SQLite fixture before PostgreSQL migration
Metadata integrity checks / metadata (push) Successful in 9m49s
2026-11-18 22:35:16 +08:00
codex-bot 3acb738aa8 Supply authors in the ECCV integration test seed
Metadata integrity checks / metadata (push) Failing after 6m32s
2026-11-18 22:27:05 +08:00
codex-bot cbf7bab8ca Attach ECCV publication evidence to existing programme attempts
Metadata integrity checks / metadata (push) Failing after 6m0s
2026-11-18 22:19:43 +08:00
Zayden-Jung 48f8304517 Merge pull request '封装 ECCV 2026 Springer 正式出版采集器' (#80) from codex/eccv2026-publications into main
Metadata integrity checks / metadata (push) Successful in 8m36s
Reviewed-on: #80
2026-11-18 10:58:39 +00:00
Zayden-Jung f2b9fe0149 Merge pull request '核对 ECCV 2026 Springer 78 卷与大会节目单' (#79) from codex/eccv2026-publisher-crosswalk into main
Metadata integrity checks / metadata (push) Successful in 13m29s
Reviewed-on: #79
2026-11-18 10:58:32 +00:00
Zayden-Jung 32d99858c1 Merge pull request '记录 IROS 2026 会议录用与期刊报告来源' (#78) from codex/iros2026-source-audit into main
Metadata integrity checks / metadata (push) Successful in 11m10s
Reviewed-on: #78
2026-11-18 10:58:25 +00:00
Zayden-Jung 060d67aba5 Merge pull request '增加 CIKM 2026 节目与出版登记采集命令' (#77) from codex/cikm2026-collector into main
Metadata integrity checks / metadata (push) Successful in 8m59s
Reviewed-on: #77
2026-11-18 10:58:19 +00:00
Zayden-Jung d0a466977c Merge pull request '核对 ICDM 2026 海报范围与跨会历史' (#75) from codex/icdm-presentation-scope into main
Metadata integrity checks / metadata (push) Successful in 8m30s
Reviewed-on: #75
2026-11-18 10:58:13 +00:00
codex-bot 52f9740d99 Record ECCV 2026 publisher and programme reconciliation
Metadata integrity checks / metadata (push) Successful in 7m41s
2026-11-18 18:12:49 +08:00
codex-bot bcb8b2ad7e Record verified IROS arXiv work bindings
Metadata integrity checks / metadata (push) Successful in 11m14s
2026-11-18 17:28:03 +08:00
codex-bot 95ff77d77f Document IROS 2026 publication and presentation sources
Metadata integrity checks / metadata (push) Successful in 7m56s
2026-11-18 17:24:06 +08:00
codex-bot 28805462ee Add reproducible CIKM 2026 source collection
Metadata integrity checks / metadata (push) Successful in 7m34s
2026-11-18 15:49:13 +08:00
codex-bot 2907c3774f Record ICDM poster coverage and cross-venue evidence
Metadata integrity checks / metadata (push) Successful in 6m55s
2026-11-18 15:10:27 +08:00
17 changed files with 2439 additions and 15 deletions
+20
View File
@@ -47,6 +47,26 @@ jobs:
subprocess.run([python,'-B','-m','pip','install','--no-deps',str(wheel)],check=True)
outside = root / 'outside-checkout'
outside.mkdir()
cikm_snapshots = pathlib.Path('/data/paper-index/validation/cikm2026-source-snapshots-v1')
if cikm_snapshots.is_dir():
cikm_output = root / 'installed-cikm2026'
command = [python, '-I', '-B', '-m', 'paper_atlas', 'collect', 'cikm2026-sources']
subprocess.run([*command, '--snapshots', str(cikm_snapshots), '--out', str(cikm_output)], cwd=outside, check=True)
report = json.loads((cikm_output/'collection-report.json').read_text())
assert {name: entry['records'] for name, entry in report['files'].items()} == {
'programme.jsonl': 1008, 'alternate-programme.jsonl': 1005,
'supplemental-programme.jsonl': 73, 'crossref.jsonl': 1081}
assert len(report['registry_outside_doi_programme']) == 73 and not report['database_written']
programme = [json.loads(line) for line in (cikm_output/'programme.jsonl').read_text().splitlines()]
assert all(row['abstract'] and row['author_display_names'] for row in programme)
assert len({row['doi'] for row in programme}) == 1008
broken = root/'tampered-cikm2026'
broken.mkdir()
(broken/'program.html').write_text('tampered')
rejected = subprocess.run([*command, '--snapshots', str(broken), '--out', str(root/'rejected-cikm2026')], cwd=outside, capture_output=True, text=True)
assert rejected.returncode != 0 and 'Snapshot hash mismatch' in rejected.stderr
assert not (root/'rejected-cikm2026').exists()
print('CIKM 2026 full frozen-source CLI check passed', flush=True)
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():
+3 -1
View File
@@ -141,6 +141,8 @@ ECCV 2026 的[暂定节目采集器](docs/eccv2026-program.md)合并重复节目
`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 cikm2026-sources` 核验并分开导出两份官方节目、附属议程和出版登记。见 [CIKM 2026 来源采集与恢复](docs/cikm2026-sources.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)。
@@ -149,4 +151,4 @@ ECCV 2026 的[暂定节目采集器](docs/eccv2026-program.md)合并重复节目
`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)。
ECCV 2026 已归档 Springer 目录和章节可通过 [正式出版采集流程](docs/eccv2026-publications.md)整理为带 DOI 和可复核证据的 JSONL。
ECCV 2026 已归档 Springer 目录和章节可通过 [正式出版采集流程](docs/eccv2026-publications.md)整理为带 DOI 和可复核证据的 JSONL。同届节目投稿可按独立审核计划附加出版证据,保留原投稿 ID 和既往决定。
+22
View File
@@ -0,0 +1,22 @@
# CIKM 2026 来源采集与恢复
使用固定节目和登记快照复现 CIKM 2026 来源盘点。采集命令生成四份独立 JSONL,输入逐文件校验 SHA-256。
```sh
paper-atlas collect cikm2026-sources \
--snapshots /data/paper-index/validation/cikm2026-source-snapshots-v1 \
--out /data/paper-index/collected/cikm2026-sources-v1
```
输入包括两版官方节目 HTML、Crossref 卷登记、五页逐篇登记及分页回执。服务器验证目录保留输入原文。输出目录必须是新目录。先检查 `collection-report.json`,报告在四份 JSONL 写完后发布。
| 文件 | 内容 |
| --- | --- |
| `programme.jsonl` | 带 DOI 的 1,008 条节目记录,包括标题、作者显示名、摘要和五类节目标签 |
| `alternate-programme.jsonl` | 另一版节目 1,005 条记录及各自的字段 |
| `supplemental-programme.jsonl` | 73 条演讲、教程、工作坊及其他附属活动观察 |
| `crossref.jsonl` | 1,081 条 ACM 卷内登记,每条保留登记字段及分页文件位置 |
节目 DOI 指向 Crossref 的出版登记。确认作者和标题差异以及历史投稿绑定后,再为来源观察准备入库队列。报告另列 73 个未出现在带 DOI 节目区的登记,附属活动仍保留各自的活动类型。公开拒稿与撤回决定仍需独立来源核查。
官方原始来源见[带 DOI 的节目](https://cikm2026.diag.uniroma1.it/cikm-program/program.html)、[另一版节目](https://cikm2026.diag.uniroma1.it/cikm-program/CIKM2026_program.html)和 [Crossref 卷登记](https://api.crossref.org/works/10.1145/3799682)。
+2
View File
@@ -204,3 +204,5 @@ IJCAI 联办届次的逐条后续核查见 [2018/2022 分区与目录差集](ijc
- [MM 2026 官方节目来源状态](../mm2026-source-status.md):完整官方日程 JSON、重复展示盘点、28 个来源 track、缺失作者/摘要和四条匿名决定路径的 403 证据;这些不是已导入或公开拒稿全集。
- [2026-11-18 公开负面来源核查](public-negative-source-audit-2026-11-18.md):ResearchArcade 930 条非 ICLR 声明的当前库比对、5 条未绑定声明及原始接口访问证据;不把镜像差集算作新增拒稿。
- [ECCV 2026 出版目录与节目单对账](eccv2026-publisher-crosswalk-2026-11-18.md)记录 Springer 78 卷的 DOI、节目单差异及 PostgreSQL 只读绑定核查。
File diff suppressed because it is too large Load Diff
@@ -0,0 +1,20 @@
# ECCV 2026 出版目录与节目单对账
2026-11-18 核验了 Springer LNCS 17001–17078 的全部 78 卷。会议索引显示 **2,835 条**,逐卷目录合计也有 **2,835 条带 DOI 的章节**,另有 156 项卷首或卷末资料。出版方卷册简介及大会节目单均为 **2,834 条**。两个来源仍有逐条差异,汇总数量相差一条的原因待核。逐卷数字、目录页哈希和只读数据库回执见[机器可读清单](eccv2026-publisher-crosswalk-2026-11-18.json)。[Springer 会议索引](https://link.springer.com/conference/eccv) · [第一卷](https://link.springer.com/book/10.1007/978-3-032-36984-0) · [大会节目单](https://eccv.ecva.net/Conferences/2026/AcceptedPapers)
## 逐篇对账
|核对对象|唯一标题且完整作者顺序一致|唯一标题且目录所显示作者顺序一致|标题对得上、作者有差异|标题待核|
|---|---:|---:|---:|---:|
|全部 2,834 条节目记录|1,332|709|535|258|
|首批已入库的 30 条节目记录|23|3|4|0|
|本次只读核验通过的 248 条既有论文候选|173|72|2|1|
|20 篇有既往拒稿、桌拒或撤回的优先论文|12|7|1|0|
目录作者栏有时只显示前六人。作者差异包括缩写、姓名文字变化与排列变化。原始名单、顺序和来源保留,前缀匹配与作者差异需查章节作者页。节目单与出版目录的摘要、分轨及公开负面决定覆盖仍待分别核查。
## 当前库核对
只读核对了全部 2,835 个出版 DOI,68 个已有论文绑定。对账后另检查 271 个已知论文关联,涵盖 245 个本次候选和 26 个已入库节目记录。其中 14 个 DOI 已指向同一论文,257 个 DOI 尚未绑定。245 个候选内有 10 篇带既往负面决定。
下一批准备以出版 DOI、正式分轨和完整作者证据复核关联。首批 30 条节目记录保留原投稿、决定和来源。确认同届同一投稿后,再补充出版证据,避免重复创建 ECCV 投稿。本次核查未写入数据库。
@@ -0,0 +1,339 @@
{
"sources": {
"poster.html": {
"url": "https://www3.cs.stonybrook.edu/~icdm2026/PosterwDemo/list.htm",
"final_url": "https://www3.cs.stonybrook.edu/~icdm2026/PosterwDemo/list.htm",
"http_status": 200,
"bytes": 132977,
"sha256": "b02083ca7574a188fe876851a700a12c131c262353db557bd43a8320fad5f747",
"at": "2026-11-18T07:06:11.984184+00:00"
},
"schedule.html": {
"url": "https://icdm2026.neu.edu.cn/DetailedSessions/list.htm",
"final_url": "https://icdm2026.neu.edu.cn/DetailedSessions/list.htm",
"http_status": 200,
"bytes": 213830,
"sha256": "c1294b89bae8c2dfcf652d8a25556e733ef42dc3ac3387070abe08419b961b0f",
"at": "2026-11-18T07:06:12.349258+00:00"
},
"deepsync.html": {
"url": "https://dm.cs.univie.ac.at/news/story/deepsync-paper-accepted-at-icdm-2026",
"final_url": "https://dm.cs.univie.ac.at/news/story/deepsync-paper-accepted-at-icdm-2026",
"http_status": 200,
"bytes": 90333,
"sha256": "e243196943434225acec2841b53c77a2b1459184e7c185e1e1cbb055118fe282",
"at": "2026-11-18T07:06:20.561308+00:00"
}
},
"poster_count": 343,
"poster_groups": {
"long-posters": 177,
"short-posters": 166
},
"oral_count": 324,
"shared_count": 324,
"poster_only": {
"DM1022": {
"group": "short-posters",
"floor": "3F",
"area": "Area C",
"slot": "P2-3C-085",
"session": "Applied Track: Real-world AI in finance, healthcare and industry II",
"title": "Beyond Bidding Agents: A Dual-Role Multi-Agent Framework with Dynamic Reserve Pricing for Auto-Bidding Auctions",
"line": 683
},
"DM1189": {
"group": "long-posters",
"floor": "3F",
"area": "Area A",
"slot": "P1-3A-032",
"session": "Heterogeneous data mining: spatiotemporal and representation learning",
"title": "Traj-Transformer: Diffusion Models with Transformer for GPS Trajectory Generation",
"line": 456
},
"DM1234": {
"group": "long-posters",
"floor": "2F",
"area": "Area A",
"slot": "P1-2A-007",
"session": "Foundations, causality and learning theory",
"title": "Prediction-Powered Corpus Mining: Coverage-Valid Statistics from Scarce Labels",
"line": 354
},
"DM1246": {
"group": "short-posters",
"floor": "3F",
"area": "Area B",
"slot": "P2-3B-074",
"session": "LLM-driven data mining and novel applications II",
"title": "Feature Selection for Trajectory Inference",
"line": 672
},
"DM1345": {
"group": "short-posters",
"floor": "3F",
"area": "Area B",
"slot": "P2-3B-060",
"session": "LLM-driven data mining and heterogeneous learning",
"title": "TRACER: An LLM-based Coarse-to-Fine Framework for Adaptive Network Traffic Classification",
"line": 658
},
"DM1408": {
"group": "long-posters",
"floor": "2F",
"area": "Area B",
"slot": "P1-2B-043",
"session": "Applied Track: Industrial, financial and decision intelligence",
"title": "A Generation-to-Application Framework for Semantic ID Collision in Industrial Generative Recommendation",
"line": 390
},
"DM1644": {
"group": "long-posters",
"floor": "2F",
"area": "Area B",
"slot": "P1-2B-072",
"session": "Recommendation, adaptation and spatiotemporal modeling II",
"title": "LQ-Rec: Late Quantization for Sequential Recommendation",
"line": 419
},
"DM1751": {
"group": "short-posters",
"floor": "3F",
"area": "Area A",
"slot": "P2-3A-014",
"session": "ML/DL and statistical methods",
"title": "MissGNN: Missingness-Guided Graph Neural Network for Feature- and Sample-Aware Imputation",
"line": 612
},
"DM1995": {
"group": "long-posters",
"floor": "2F",
"area": "Area B",
"slot": "P1-2B-053",
"session": "Applied Track: Recommendation, security and operational intelligence II",
"title": "Multi-Modal Causal Disentanglement for Multi-Domain Sequential Audio Recommendation",
"line": 400
},
"DM2002": {
"group": "long-posters",
"floor": "3F",
"area": "Area B",
"slot": "P1-3B-048",
"session": "Heterogeneous data mining and systems",
"title": "HiMamba: A Hierarchical State Space Framework for Asymmetric Multimodal Data Mining in Clinical Assessment",
"line": 472
},
"DM2055": {
"group": "short-posters",
"floor": "3F",
"area": "Area A",
"slot": "P2-3A-038",
"session": "Applied Track: Real-world AI in finance, healthcare and industry I",
"title": "CGSBR: Adaptive Cross-Source Gating for Cold-Start Sequential Bundle Recommendation",
"line": 636
},
"DM2140": {
"group": "long-posters",
"floor": "3F",
"area": "Area A",
"slot": "P1-3A-008",
"session": "Machine learning and statistical methods",
"title": "Reliability-Guided Pseudo-Label Learning and Prototype Alignment for Domain-Adaptive Colorectal Segmentation",
"line": 432
},
"DM252": {
"group": "short-posters",
"floor": "2F",
"area": "Area B",
"slot": "P2-2B-046",
"session": "LLM-driven data mining and novel applications I",
"title": "Trust-But-Verify: Poisoning-Resilient Locally Private Graph Learning Protocols",
"line": 572
},
"DM383": {
"group": "long-posters",
"floor": "2F",
"area": "Area A",
"slot": "P1-2A-022",
"session": "Heterogeneous graphs and multimodal intelligence",
"title": "Subgraph Meta-Learning for Direction and Time Aware Transaction Anomaly Detection",
"line": 369
},
"DM463": {
"group": "short-posters",
"floor": "2F",
"area": "Area B",
"slot": "P2-2B-050",
"session": "Modeling, recommendation and LLM methods",
"title": "CaST-POI: Candidate-Conditioned Spatiotemporal Ranking for Next POI Recommendation",
"line": 576
},
"DM483": {
"group": "short-posters",
"floor": "2F",
"area": "Area B",
"slot": "P2-2B-051",
"session": "Modeling, recommendation and LLM methods",
"title": "Causal Masking Lowers Intrinsic Dimension in Full-Vocabulary Sequential Recommendation: Evidence from a Single-Variable Attention Ablation",
"line": 577
},
"DM713": {
"group": "short-posters",
"floor": "2F",
"area": "Area B",
"slot": "P2-2B-060",
"session": "Modeling, recommendation and LLM methods",
"title": "Hint-Conditioned Knowledge Tracing: A Causal-Inspired View of Context-Dependent Hint Usage",
"line": 586
},
"DM736": {
"group": "long-posters",
"floor": "3F",
"area": "Area B",
"slot": "P1-3B-050",
"session": "LLM-driven data mining and retrieval I",
"title": "From Natural-Language Hypotheses to Verified Data-Mining Discoveries",
"line": 474
},
"DM803": {
"group": "short-posters",
"floor": "2F",
"area": "Area B",
"slot": "P2-2B-032",
"session": "Heterogeneous data mining and multimodal learning I",
"title": "Graph-Guided LLM Reasoning for Incident Prediction Over Multi-Source Data Streams",
"line": 558
}
},
"oral_only": [],
"title_differences": [
{
"paper_id": "DM1319",
"poster_title": "FFE-FRCN: A Frequency Feature Enhanced Few-shot Learning Approach for Object Detection in Microscopic Images",
"oral_title": "An FFE-FRCN-Based Data Platform for Environmental Microorganism Image Retrieval and Knowledge Platform"
}
],
"nonpaper_table_rows": [],
"database_written": false,
"scope": "Official presentation metadata. Shared programme IDs identify repeated appearances within this event. Poster-only IDs are scheduling observations pending identity and publication review.",
"pg_readonly": {
"snapshot": {
"at": "2026-11-18 07:08:21.138511+00",
"snapshot": "1095530:1095530:"
},
"report_sha256": "04ad5547342d7e3c2939046f290ecff647878c202c867e47acbdc438d4e10ad4",
"existing_icdm2026_editions": [],
"existing_icdm_program_source": false,
"reviewed_links": [
{
"external_id": "2026/DM1125",
"paper_id": 2783475,
"track": "research",
"source_row_sha256": "4e981605cf471a5ea29bb3ba3a676784cea55f7777a681eb5fe267ff9c83fa4a",
"source_abstract": null,
"work_anchor": {
"scheme": "arxiv",
"value": "2608.29635"
},
"independent_source": "https://arxiv.org/abs/2608.29635",
"existing_history": [
{
"id": 132984,
"venue": "ICLR",
"year": 2026,
"observations": [
{
"id": 134279,
"status": "rejected"
},
{
"id": 539402,
"status": "rejected"
}
]
}
],
"original_icdm_decision_verified": false,
"imported": false
},
{
"external_id": "2026/DM790",
"paper_id": 2485633,
"track": "applied",
"source_row_sha256": "14e03ca3833b267d6f677325f4e4b9295d1fa42cc77be6707b9bca745444888b",
"source_abstract": null,
"work_anchor": {
"scheme": "arxiv",
"value": "2509.19774"
},
"independent_source": "https://arxiv.org/abs/2509.19774",
"existing_history": [
{
"id": 138980,
"venue": "ICLR",
"year": 2026,
"observations": [
{
"id": 545219,
"status": "withdrawn"
}
]
}
],
"original_icdm_decision_verified": false,
"imported": false
},
{
"external_id": "2026/DM1157",
"paper_id": 3286364,
"track": "research",
"source_row_sha256": "22ed027f29a161bc978b3c33e921b42c67d999d520eed49ec46ce5935a6f2f2a",
"source_abstract": null,
"work_anchor": {
"scheme": "openreview_snapshot",
"value": "2026/1dGzRTSLqL"
},
"independent_source": "https://dm.cs.univie.ac.at/news/story/deepsync-paper-accepted-at-icdm-2026",
"existing_history": [
{
"id": 127709,
"venue": "ICLR",
"year": 2026,
"observations": [
{
"id": 129094,
"status": "rejected"
},
{
"id": 534293,
"status": "rejected"
}
]
}
],
"original_icdm_decision_verified": false,
"imported": false
}
],
"held": [
{
"external_id": "2026/DM1791",
"paper_id": 3288747,
"reason": "DomainBridging has name-order/format differences and no arxiv self-accepted link; token-bag retrieval is not author or work identity approval."
},
{
"external_id": "2026/DM1801",
"paper_id": 3288866,
"reason": "LightGFC author lists substantially differ; hold version/work review, no automatic author aliases or merge."
},
{
"external_id": "2026/DM1413",
"paper_id": 3292171,
"reason": "OFS author lists differ including JianchengLv vs JindiLv; not treated as aliases; hold independent version/work verification."
}
]
},
"publisher_catalog_status": "No 2026 IEEE publisher catalogue verified in this check",
"public_negative_availability": "unverified",
"next": "Use the existing oral source denominator by track. Compare the three reviewed candidates with the fresh import baseline and preserve their earlier attempts. Keep the poster-only rows as a separate source inventory, and resolve the title disagreement before linking that poster."
}
@@ -0,0 +1,204 @@
{
"checked_on": "2026-11-18",
"venue": "IROS",
"year": 2026,
"conference_program": {
"url": "https://2026.ieee-iros.org/program/paper_index/",
"status": "unavailable_in_checked_environment",
"paper_level_full_index_verified": false
},
"publisher_registration_check": {
"status": "no_target_container_in_first_25_relevance_results",
"rows_checked": 25,
"observed_container": "2026 IEEE 13th International Conference on Intelligent Systems (IS)",
"complete_registry_search": false
},
"journal_presentations": {
"journal": "IEEE Transactions on Robotics",
"scope": "IROS 2026 conference presentation of journal papers",
"rendered_index_and_embedded_entries_match": true,
"count": 30,
"xplore_document_ids": [
"10989573",
"11251175",
"11488837",
"11268969",
"11098164",
"11219220",
"11304178",
"11267071",
"11304172",
"11455384",
"11397863",
"11288096",
"11544114",
"11534108",
"11488912",
"11535315",
"11342351",
"11353077",
"11429535",
"11364144",
"11427341",
"11427344",
"11455389",
"11488904",
"11535317",
"11554436",
"11495270",
"11575122",
"11488843",
"11581355"
],
"conference_submission_decision_verified": false,
"publication_dois_verified": false
},
"author_reported_conference_acceptance_candidates": [
{
"arxiv": "2607.06499",
"acceptance_evidence": [
"uw-roundup.html",
"uw-faculty.html"
],
"author_count_current": 3,
"work_identity_status": "existing_work_binding_verified_read_only",
"database_paper_id": 2731624,
"ordered_authors_match": true,
"arxiv_versions_stored": [
"v1"
]
},
{
"arxiv": "2505.04999",
"acceptance_evidence": [
"usc-acceptance.html"
],
"author_list_evidence": "usc-clam-project.html",
"author_count_current": 7,
"work_identity_status": "existing_work_binding_verified_read_only",
"database_paper_id": 2381843,
"ordered_authors_match": true,
"arxiv_versions_stored": [
"v1",
"v2"
]
},
{
"acceptance_evidence": [
"ntt-contributed.html"
],
"author_count_current": 6,
"work_identity_status": "stable_identifier_and_track_pending"
}
],
"purdue_scope_note": {
"evidence": "purdue-mars.html",
"reported_combined_count": 10,
"reported_tro_subset": 2
},
"public_rejected_desk_rejected_withdrawn": "not_located_in_checked_public_sources",
"coverage_status": "partial_source_audit",
"database_written": false,
"receipts": [
{
"name": "ieee-tro-list.html",
"url": "https://www.ieee-ras.org/publications/t-ro/t-ro-list-of-conference-papers/",
"final_url": "https://www.ieee-ras.org/publications/t-ro/t-ro-list-of-conference-papers/",
"status": 200,
"bytes": 610713,
"sha256": "2f2bb466446c62c37dc0927f8df45b161a096cbf8ab1aeffc4f43fa5d27e6358"
},
{
"name": "ieee-tro-policy.html",
"url": "https://www.ieee-ras.org/publications/t-ro/",
"final_url": "https://www.ieee-ras.org/publications/t-ro/",
"status": 200,
"bytes": 566001,
"sha256": "37b8877babb62e69d61e8b32193378afad4372117025a139cf70ffc26fcada5c"
},
{
"name": "ntt-contributed.html",
"url": "https://group.ntt/en/topics/2026/09/28/iros2026.html",
"final_url": "https://group.ntt/en/topics/2026/09/28/iros2026.html",
"status": 200,
"bytes": 221293,
"sha256": "ade0a0320a83e1d4504ff5da24c9f495aed581cffa01aa38b1e4cbec331f6dbc"
},
{
"name": "crossref-proceedings-query.json",
"url": "https://api.crossref.org/works?query.container-title=2026+IEEE%2FRSJ+International+Conference+on+Intelligent+Robots+and+Systems+%28IROS%29&filter=prefix%3A10.1109%2Ctype%3Aproceedings-article%2Cfrom-pub-date%3A2026-01-01%2Cuntil-pub-date%3A2026-12-31&rows=25&select=DOI%2Ctitle%2Cevent%2Ccontainer-title%2Cpublisher%2Cpublished%2CURL",
"final_url": "https://api.crossref.org/works?query.container-title=2026+IEEE%2FRSJ+International+Conference+on+Intelligent+Robots+and+Systems+%28IROS%29&filter=prefix%3A10.1109%2Ctype%3Aproceedings-article%2Cfrom-pub-date%3A2026-01-01%2Cuntil-pub-date%3A2026-12-31&rows=25&select=DOI%2Ctitle%2Cevent%2Ccontainer-title%2Cpublisher%2Cpublished%2CURL",
"status": 200,
"bytes": 13634,
"sha256": "415bf105de8006a42d3f2f8e6eb0078eb54c51bd65315296c0735e853431b228"
},
{
"name": "purdue-mars.html",
"url": "https://engineering.purdue.edu/IE/news/2026/she-iros-mars-lab",
"final_url": "https://engineering.purdue.edu/IE/news/2026/she-iros-mars-lab",
"status": 200,
"bytes": 60386,
"sha256": "fae6ba9637ea141ff3fda93643b97ddfb65d6ea3689f1bca3580956272b2d373"
},
{
"name": "uw-acceptance.html",
"url": "https://faculty.washington.edu/chx/post/2026-06-17-iros-mppi/",
"status": 404,
"error": "HTTP Error 404: "
},
{
"name": "usc-acceptance.html",
"url": "https://liralab.usc.edu/news/",
"final_url": "https://liralab.usc.edu/news/",
"status": 200,
"bytes": 25381,
"sha256": "9fa6b272b7995448a4318dd91cc94b93eeb514ddaa92d7df6880b3d93d91fceb"
},
{
"name": "arxiv-2607.06499.html",
"url": "https://arxiv.org/abs/2607.06499",
"final_url": "https://arxiv.org/abs/2607.06499",
"status": 200,
"bytes": 43362,
"sha256": "98fcbcaf39832f3002fa2910f69ca650481bb473cdb450290e5f9ca9245646a2"
},
{
"name": "arxiv-2505.04999.html",
"url": "https://arxiv.org/abs/2505.04999",
"final_url": "https://arxiv.org/abs/2505.04999",
"status": 200,
"bytes": 42506,
"sha256": "26ea0fd9ac8753b911ff97f1aaf29ea495f50b5c1e106b33c1336ea87c03b0df"
},
{
"name": "uw-faculty.html",
"url": "https://www.me.washington.edu/facultyfinder/xu-chen",
"final_url": "https://www.me.washington.edu/facultyfinder/xu-chen",
"status": 200,
"bytes": 99274,
"sha256": "877af368b564ba26a62b1ce4f947b439cceada2490dfb446b8730ba132c72c78"
},
{
"name": "uw-roundup.html",
"url": "https://faculty.washington.edu/chx/post/2026-09-07-new-papers-2026/",
"final_url": "https://faculty.washington.edu/chx/post/2026-09-07-new-papers-2026/",
"status": 200,
"bytes": 180954,
"sha256": "bc9ef19a07ea2875f8cfffe9afd62bcc2998fbe7c4bc8a4440c1bc87bae3856b"
},
{
"name": "usc-clam-project.html",
"url": "https://clamrobot.github.io/",
"final_url": "https://clamrobot.github.io/",
"status": 200,
"bytes": 20098,
"sha256": "3b3ea3f817ce6483807ace9b5a42ccf9835b811837346b63411190316fe2819c"
}
],
"work_identity_review": {
"snapshot_at": "2026-11-18 09:26:04.582744+00",
"evidence_sha256": "5e5cfb57525a369e8c81e1a9ea6edeb4f5e83a33e1961eeead8a014e3d59ad6d",
"existing_work_matches": 2,
"database_written": false
}
}
@@ -0,0 +1,21 @@
# IROS 2026 来源核查
2026-11-18 核查了会议节目、期刊报告名单、作者机构公布的录用信息和 DOI 登记线索。逐项来源、响应哈希与文献定位符见[机器可读清单](iros2026-source-audit-2026-11-18.json)。本次来源核查未写入数据库。
## 期刊论文在会上报告
IEEE RAS 的 T-RO 名单在 IROS 2026 一节列出 30 个不同的 IEEE Xplore 文献编号。页面渲染内容和内嵌数据的编号及顺序完全一致。T-RO 政策允许已录用的期刊文章到 IROS 报告。后续应保留期刊出版与会议报告两种关系,并在确认期刊 DOI 后核对现有论文身份。本清单没有逐条验证 DOI 或会议投稿决定。[T-RO 报告名单](https://www.ieee-ras.org/publications/t-ro/t-ro-list-of-conference-papers/) · [T-RO 报告政策](https://www.ieee-ras.org/publications/t-ro/)
普渡大学的通告将团队的 10 项成果一同介绍,其中 2 项属于 T-RO。这个团队样本可以帮助核查来源口径,暂不用于生成逐篇录用记录。[普渡大学通告](https://engineering.purdue.edu/IE/news/2026/she-iros-mars-lab)
## 作者机构公布的录用线索
- 华盛顿大学的作者通告列出 IROS 2026 待出版论文,标题与 arXiv 2607.06499 对应,三位作者一致。只读查询确认对应论文 2731624,库内标题与三位作者的顺序一致。[作者通告](https://faculty.washington.edu/chx/post/2026-09-07-new-papers-2026/) · [arXiv](https://arxiv.org/abs/2607.06499)
- USC 实验室确认 CLAM 在 IROS 2026 录用。项目页指向 arXiv 2505.04999,当前作者名单为七人。库内 arXiv 保留 v1、v2,现有论文 2381843 的标题和七位作者顺序与当前详情页一致。[实验室通告](https://liralab.usc.edu/news/) · [作者项目页](https://clamrobot.github.io/) · [arXiv](https://arxiv.org/abs/2505.04999)
- NTT 公布一条题名及六位作者,待核实稳定标识和会议 track。[NTT 通告](https://group.ntt/en/topics/2026/09/28/iros2026.html)
这两条工作身份已核实,尚未录入 IROS 录用证据。
## 待补证据
[会议论文索引](https://2026.ieee-iros.org/program/paper_index/)在受检环境未能读取。一次限定 IEEE、2026 年及会议论文类型的 Crossref 相关性查询返回的前 25 条属于另一会议。实际会议论文集 DOI、完整出版清单、正式分轨和逐篇覆盖仍待核查。已检查的公开来源中未找到逐篇拒稿、桌拒或撤回清单,继续保持这些决定的公开可得性待查。
+15
View File
@@ -14,3 +14,18 @@ paper-atlas collect eccv2026-publications \
输出含规范化 JSONL、manifest、采集回执以及所选章节的内容寻址存证。每条保留出版 DOI、卷号、页码、按顺序的完整作者及公开摘要。目录证据用 URL 和哈希连接到输入归档,保存归档以便复核。主会出版观察依据 Springer 正式论文集,原始投稿号、决定日期及签名仍未知。工具仅处理已归档的公开 HTML。
生成的文件待做论文身份复核后方可导入。准备生产批次时核实每个 DOI 与现有 arXiv 标识的绑定,保持两者的来源证据。对已存在的同届节目投稿需先核对投稿身份及轨道,再附加出版来源,避免因直接导入创建重复投稿。具有 ICLR 等既往决定的论文须保留原记录。新的增量导入 manifest 需以当前实际来源覆盖为父版本,留存待核差集和独立验收回执。
## 已有节目投稿的出版证据
先在只读快照上复核节目 ID、章节 DOI、标题、完整有序作者、摘要版本与当前投稿历史。保存计划和原始证据,计划应列出每条出版观察、保留的论文与投稿 ID、节目来源和决定 ID、当前与目标届次及完整历史包的 SHA-256。作者或版本仍有疑点的条目继续留在待核清单。
```sh
paper-atlas maintain eccv-publication-reconciliation \
--pg-service <writer-service> --pg-schema research \
--plan <reviewed-plan.json> --plan-sha256 <reviewed-plan-sha256> \
--springer-previous-count <last-verified-count>
```
命令接收 `readonly_same_submission_reconciliation_plan` 格式的已复核计划。它会在同一事务内核对实时历史和来源覆盖,把原节目投稿归入主会,追加出版来源与观测决定,并分别调整节目与 Springer 的覆盖数。节目主会覆盖保持 partial,因为其目录的主会分母尚未独立核定。随后通过只读连接核验投稿、决定及覆盖,保存计划和回执。相同计划只允许成功提交一次。若执行中失败,先查同一计划的提交标记、覆盖和投稿来源,再决定是否重新执行。
旧节目快照的全量未分类 manifest 无法刷新已有逐条更正的投稿。继续刷新时需按已复核轨道分别整理输出并生成 manifest。
+10
View File
@@ -44,3 +44,13 @@ paper-atlas collect icdm2026-program \
原始来源恢复包为 `/data/paper-index/validation/remaining2026-official-leads-v1-recovery.tar.gz`,SHA-256 为 `0be8c84bb5c591e5afaaa5f95dd5334bf13a9be1671cb942e46e998f502495bb`。独立盘点为 `/data/paper-index/validation/icdm2026-independent-program-inventory-v1.json`,SHA-256 为 `e13eaea0a158dcaafdf5dfb4af7948e4793b0e98041ec87d42b52513acfdc5f6`;本机工作记录目录也保留同名备份。
下次继续先验证源码版本、原始证据和实际输出,再寻找可公开取得的摘要与正式出版记录,核查逐篇版本、稳定标识及旧投稿历史,准备新的只读预检和可审阅入库队列。完成日程规范化不意味着 ICDM2026 整届出版覆盖、公开负面决定普查或最终全库去重完成。
## 2026-11-18 官方展示来源核查
口头日程的 324 个节目编号全部出现在官方海报排期。海报表有 343 个不同编号,其中长演讲组安排 177 项海报,短演讲组安排 166 项。19 个只见于海报表的编号作为独立来源线索登记。两份排期的覆盖范围各自保留,正式出版目录仍待核验。[官方详细日程](https://icdm2026.neu.edu.cn/DetailedSessions/list.htm) · [官方海报表](https://www3.cs.stonybrook.edu/~icdm2026/PosterwDemo/list.htm)
共同编号 DM1319 的标题在两个来源中实质不同。详细日程描述基于 FFE-FRCN 的数据平台,海报表描述 FFE-FRCN 的少样本检测方法。两份标题及来源位置均已留存,该编号的跨来源作品关联待核验。
生产只读快照确认 ICDM 2026 仍无入库届次。节目、作者和独立录用证据支持继续处理 DM1125、DM790、DM1157 三项既有作品关联。DM1125 和 DM1157 各有一次 ICLR 拒稿投稿及两条来源观察,DM790 有一次 ICLR 撤回投稿。三项均尚未导入,节目原始摘要字段缺失。另有三个候选存在作者名称顺序或名单差异,继续暂缓作品关联。[FALCON 作者记录](https://arxiv.org/abs/2608.29635) · [PPGFlowECG 作者记录](https://arxiv.org/abs/2509.19774) · [DeepSynC 大学录用说明](https://dm.cs.univie.ac.at/news/story/deepsync-paper-accepted-at-icdm-2026)
本轮仍未核实公开的 ICDM 2026 拒稿或撤回决定导出。[逐条差集、标题冲突、快照和后续入口](coverage/icdm2026-presentation-scope-2026-11-18.json)
+20 -14
View File
@@ -173,34 +173,40 @@ def find_candidates(connection, records):
records[hit['n']]['title_candidate_ids'].append(hit['paper_id'])
def _named_rows(connection, sql, params):
cursor = connection.execute(sql, params)
columns = [field[0] for field in cursor.description]
return [{name: row[i] for i, name in enumerate(columns)} for row in cursor]
def entity_packet(connection, paper_id):
paper = connection.execute('SELECT * FROM papers WHERE id=?', (paper_id,)).fetchone()
if paper is None:
papers = _named_rows(connection, 'SELECT * FROM papers WHERE id=?', (paper_id,))
if not papers:
raise RuntimeError('Candidate disappeared from read-only snapshot')
sources = [dict(r) for r in connection.execute('''SELECT r.*,s.name AS source
sources = _named_rows(connection, '''SELECT r.*,s.name AS source
FROM source_records r JOIN sources s ON s.id=r.source_id
WHERE r.paper_id=? ORDER BY r.id LIMIT 101''', (paper_id,))]
attempts = [dict(r) for r in connection.execute('''SELECT s.*,v.acronym AS venue,e.year,e.track,e.event_key
WHERE r.paper_id=? ORDER BY r.id LIMIT 101''', (paper_id,))
attempts = _named_rows(connection, '''SELECT s.*,v.acronym AS venue,e.year,e.track,e.event_key
FROM submissions s JOIN editions e ON e.id=s.edition_id JOIN venues v ON v.id=e.venue_id
WHERE s.paper_id=? ORDER BY s.id LIMIT 201''', (paper_id,))]
packet = {'paper': dict(paper), 'identifiers': [dict(r) for r in connection.execute(
'SELECT * FROM identifiers WHERE paper_id=? ORDER BY scheme,value', (paper_id,))],
WHERE s.paper_id=? ORDER BY s.id LIMIT 201''', (paper_id,))
packet = {'paper': papers[0], 'identifiers': _named_rows(connection,
'SELECT * FROM identifiers WHERE paper_id=? ORDER BY scheme,value', (paper_id,)),
'source_records': sources[:100], 'source_records_truncated': len(sources) > 100,
'attempts': attempts[:200], 'attempts_truncated': len(attempts) > 200}
for attempt in packet['attempts']:
observations = [dict(r) for r in connection.execute(
'SELECT * FROM decisions WHERE submission_id=? ORDER BY id LIMIT 201', (attempt['id'],))]
observations = _named_rows(connection,
'SELECT * FROM decisions WHERE submission_id=? ORDER BY id LIMIT 201', (attempt['id'],))
attempt['decisions'], attempt['decisions_truncated'] = observations[:200], len(observations) > 200
provenance = [dict(r) for r in connection.execute('''SELECT ss.*,r.external_id,s.name AS source
provenance = _named_rows(connection, '''SELECT ss.*,r.external_id,s.name AS source
FROM submission_sources ss JOIN source_records r ON r.id=ss.source_record_id
JOIN sources s ON s.id=r.source_id WHERE ss.submission_id=?
ORDER BY ss.source_record_id,ss.source_submission_id LIMIT 101''', (attempt['id'],))]
ORDER BY ss.source_record_id,ss.source_submission_id LIMIT 101''', (attempt['id'],))
attempt['sources'], attempt['sources_truncated'] = provenance[:100], len(provenance) > 100
for decision in attempt['decisions']:
provenance = [dict(r) for r in connection.execute('''SELECT ds.*,r.external_id,s.name AS source
provenance = _named_rows(connection, '''SELECT ds.*,r.external_id,s.name AS source
FROM decision_sources ds JOIN source_records r ON r.id=ds.source_record_id
JOIN sources s ON s.id=r.source_id WHERE ds.decision_id=?
ORDER BY ds.source_record_id,ds.source_decision_id LIMIT 101''', (decision['id'],))]
ORDER BY ds.source_record_id,ds.source_decision_id LIMIT 101''', (decision['id'],))
decision['sources'], decision['sources_truncated'] = provenance[:100], len(provenance) > 100
return packet
+2
View File
@@ -30,6 +30,7 @@ COMMANDS = {
('collect','eccv2026-publications'): ('paper_atlas.collectors.eccv2026_publications', {}, '核验ECCV2026 Springer 正式出版章节'),
('collect','icassp2026-registry'): ('paper_atlas.collectors.icassp2026', {}, '核验 ICASSP 2026 冻结注册论文元数据'),
('collect','icdm2026-program'): ('paper_atlas.collectors.icdm2026', {}, '核验 ICDM2026 冻结日程及明确的研究/应用轨道'),
('collect','cikm2026-sources'): ('paper_atlas.collectors.cikm2026', {}, '核对 CIKM 2026 节目、附属议程与 Crossref 出版登记'),
('collect','sdm2026-proceedings'): ('paper_atlas.collectors.sdm2026', {}, '核验 SDM 2026 注册论文并单独保留公开日程版本'),
('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'),
@@ -48,6 +49,7 @@ COMMANDS = {
('import','arxiv'): ('paper_atlas.ingestion.arxiv', {}, '导入 arXiv 元数据 ZIP'),
('import','records'): ('paper_atlas.ingestion.records', {}, '导入核验过的会议文件'),
('maintain','edition-tracks'): ('paper_atlas.ingestion.edition_tracks', {}, '按逐条来源证据修正既有投稿轨道并保护刷新'),
('maintain','eccv-publication-reconciliation'): ('paper_atlas.ingestion.publication_reconciliation', {}, '向既有 ECCV 2026 投稿附加已复核的正式出版证据'),
('maintain','source-revisions'): ('paper_atlas.ingestion.source_revisions', {}, '按完整来源版本核验并修正出版分轨与覆盖数,保留历史并阻止旧版刷新'),
('maintain','joint-editions'): ('paper_atlas.ingestion.joint_editions', {}, '按已核验计划修正既有联合会议届次归属'),
('maintain','event-identity'): ('paper_atlas.ingestion.event_identity', {}, '按核验计划增加同年会议事件身份'),
+164
View File
@@ -0,0 +1,164 @@
"""Collect CIKM 2026 programme and Crossref metadata from frozen snapshots."""
import argparse
from collections import Counter
from hashlib import sha256
import json
from pathlib import Path
import re
from paper_atlas.io.html import Document, clean
PROGRAM = 'https://cikm2026.diag.uniroma1.it/cikm-program/program.html'
VOLUME = '10.1145/3799682'
CONTAINER = 'Proceedings of the 35th ACM International Conference on Information and Knowledge Management'
TRACKS = {'FULL': 596, 'SHORT': 234, 'APPLIED': 86, 'RESOURCE': 46, 'DEMO': 46}
SUPPLEMENTAL = {'KEYNOTE', 'TUTORIAL', 'INDUSTRY', 'POSTER', 'PHD', 'PHD · REMOTE', 'WORKSHOP', 'ANALYTICUP'}
PAGE_COUNTS = (250, 250, 250, 250, 81)
HASHES = {
'program.html': '3eaa668b6aaf09a671d8a3e19a2889b9ab21afaa7ded013794afa94bf35281d6',
'CIKM2026_program.html': 'e3630bd1247df713b8b618f1d3cd9709e7020b2524bb677f82a10f83badd1732',
'crossref-parent.json': 'f3370418d86ccfcfaeb972073bc5fcd1c1c737caed96570fc83f99db0f7d204a',
'crossref-page-1.json': '16bed29f8ee82d668b124c369b632c5c1de0753bdbec7eb72ecb22ffdda5c710',
'crossref-page-2.json': '54424ca31d99462394c9fde5fecfd7dd16337351d21ede3169f19af54d4f49c5',
'crossref-page-3.json': '2dc31d89f41966614ed66a11d1ebaa4b2a6904dcf92945b7185e72e7e65382dd',
'crossref-page-4.json': 'fbbec2b52650dc99a0251c743f035a6fd95be263b04c55c12134cf8473bdde76',
'crossref-page-5.json': '06151c5f770c1e37e8ea9dabe5210be9c969a14fa3bf9925a2d7730ced47d6e5',
'crossref-pages.json': '8be1f8f05139f7a1231b73ad6430f28d2f625809f1530cd9de654f8b3c68c1ae',
}
def fields(node, css_class):
return [child for child in node.walk() if css_class in child.attrs.get('class', '').split()]
def parse_program(raw, *, require_doi):
document = Document(raw.decode('utf-8'))
papers = {}
for card in fields(document.root, 'paper'):
pid = card.attrs.get('data-pid')
if not pid:
continue
titles = fields(card, 'ptitle')
labels = fields(card, 'badge')
abstracts = fields(card, 'abstxt')
if pid in papers or len(titles) != 1 or len(labels) != 1 or len(abstracts) > 1:
raise ValueError(f'Invalid or repeated programme card {pid}')
title = clean(titles[0].text())
label = clean(labels[0].text())
authors = [clean(author.text()) for author in fields(card, 'au')]
abstract = clean(abstracts[0].text()) if abstracts else None
doi_url = titles[0].attrs.get('href')
doi = doi_url.removeprefix('https://doi.org/') if doi_url else None
if label not in TRACKS or not title or not authors or not all(authors):
raise ValueError(f'Incomplete programme metadata {pid}')
if require_doi and (not doi or not re.fullmatch(re.escape(VOLUME) + r'\.\d+', doi) or not abstract):
raise ValueError(f'Missing programme DOI or abstract {pid}')
papers[pid] = {'programme_id': pid, 'title': title, 'author_display_names': authors,
'abstract': abstract, 'programme_label': label, 'programme_doi_link': doi_url,
'doi': doi if require_doi else None}
return papers
def parse_supplemental(raw):
document = Document(raw.decode('utf-8'))
events = {}
for session in fields(document.root, 'sessblock'):
section = session.attrs.get('id', '')
if not section.startswith(('pre-', 'kn-')):
continue
for card in fields(session, 'paper'):
labels = fields(card, 'badge')
if len(labels) != 1:
raise ValueError(f'Ambiguous supplemental label in {section}')
label = clean(labels[0].text())
if label not in SUPPLEMENTAL:
continue
titles = fields(card, 'ptitle')
abstracts = fields(card, 'abstxt')
locator = card.attrs.get('id') or section
if locator in events or len(titles) != 1 or len(abstracts) > 1:
raise ValueError(f'Ambiguous supplemental entry {locator}')
events[locator] = {'programme_locator': locator, 'programme_section': section,
'programme_label': label, 'title': clean(titles[0].text()),
'author_display_names': [clean(a.text()) for a in fields(card, 'au')],
'abstract': clean(abstracts[0].text()) if abstracts else None}
return events
def parse_registry(snapshots):
volume = json.loads(snapshots['crossref-parent.json'])['message']
if (volume['DOI'], volume['title'], volume['ISBN']) != (VOLUME, [CONTAINER], ['9798400725395']):
raise ValueError('Unexpected Crossref volume')
records = {}
for number, count in enumerate(PAGE_COUNTS, 1):
name = f'crossref-page-{number}.json'
message = json.loads(snapshots[name])['message']
if message['total-results'] != 1081 or len(message['items']) != count:
raise ValueError(f'Unexpected Crossref pagination in {name}')
for item in message['items']:
doi = item['DOI'].lower()
if (doi in records or not re.fullmatch(re.escape(VOLUME) + r'\.\d+', doi)
or item.get('container-title') != [CONTAINER] or item.get('type') != 'proceedings-article'
or item.get('published', {}).get('date-parts') != [[2026, 11, 6]]):
raise ValueError(f'Unexpected Crossref record {doi}')
records[doi] = {'registry': item, 'source_file': name}
return records
def collect(snapshots, output):
snapshots = Path(snapshots)
output = Path(output)
if output.exists() or output.is_symlink():
raise FileExistsError(output)
bodies = {}
for name, expected in HASHES.items():
path = snapshots / name
if path.is_symlink() or not path.is_file() or path.stat().st_size > 8 * 1024**2:
raise ValueError(f'Invalid snapshot file {name}')
raw = path.read_bytes()
if sha256(raw).hexdigest() != expected:
raise ValueError(f'Snapshot hash mismatch for {name}')
bodies[name] = raw
programme = parse_program(bodies['program.html'], require_doi=True)
alternate = parse_program(bodies['CIKM2026_program.html'], require_doi=False)
supplemental = parse_supplemental(bodies['program.html'])
registry = parse_registry(bodies)
programme_dois = {r['doi'] for r in programme.values()}
tracks = Counter(r['programme_label'] for r in programme.values())
if (tracks != TRACKS or len(programme_dois) != len(programme) or programme_dois - registry.keys()
or len(alternate) != 1005 or alternate.keys() - programme.keys()
or len(supplemental) != 73 or len(registry) - len(programme_dois) != 73):
raise ValueError('Unexpected programme or registry coverage')
groups = {
'programme': programme,
'alternate-programme': alternate,
'supplemental-programme': supplemental,
'crossref': registry,
}
output.mkdir(parents=True, exist_ok=False)
files = {}
for name, rows in groups.items():
filename = name + '.jsonl'
payload = ''.join(json.dumps({'source_key': key, **row}, ensure_ascii=False) + '\n'
for key, row in sorted(rows.items())).encode()
(output / filename).write_bytes(payload)
files[filename] = {'records': len(rows), 'sha256': sha256(payload).hexdigest()}
report = {'status': 'collected_pending_identity_review', 'files': files, 'input_sha256': HASHES,
'programme_by_label': dict(tracks), 'programme_dois_in_registry': len(programme_dois),
'registry_outside_doi_programme': sorted(registry.keys() - programme_dois),
'programme_ids_missing_from_alternate': sorted(programme.keys() - alternate.keys()),
'database_written': False, 'public_negative_decisions_verified': False}
partial = output / 'collection-report.json.part'
partial.write_text(json.dumps(report, ensure_ascii=False, indent=2) + '\n', encoding='utf-8')
partial.replace(output / 'collection-report.json')
return report
def main(argv=None, *, prog=None):
parser = argparse.ArgumentParser(prog=prog, description=__doc__)
parser.add_argument('--snapshots', type=Path, required=True)
parser.add_argument('--out', type=Path, required=True)
options = parser.parse_args(argv)
print(json.dumps(collect(options.snapshots, options.out), ensure_ascii=False))
@@ -0,0 +1,222 @@
"""Attach reviewed ECCV proceedings evidence to an existing programme attempt."""
import argparse
from contextlib import closing
from hashlib import sha256
import json
from pathlib import Path
import re
from paper_atlas.audits.source_identity import entity_packet
from paper_atlas.collectors.eccv2026 import SOURCE as PROGRAMME
from paper_atlas.collectors.eccv2026_publications import SOURCE as PUBLISHER
from paper_atlas.db.backend import database_arguments, open_store
from paper_atlas.db.edition_scopes import KEY
from paper_atlas.db.submissions import identities_for_row, timestamp
from paper_atlas.ingestion.edition_tracks import curation
from paper_atlas.ingestion.records import source_record_count
MARKER = 'eccv_publication_reconciliation_v1:'
PUBLISHER_COUNT = 2835
def require(condition, message):
if not condition:
raise ValueError(message)
def packet_digest(packet):
return sha256(json.dumps(packet, sort_keys=True, ensure_ascii=False).encode()).hexdigest()
def validate_plan(plan):
require(plan.get('status') == 'readonly_same_submission_reconciliation_plan'
and plan.get('database_written') is False, 'A reviewed read-only plan is required')
rows = plan['plans']
require(0 < len(rows) <= 1000 and len(rows) == plan['guarded_same_submission_plans'],
'Unexpected approved plan size')
for key in ('doi', 'retained_paper_id', 'retained_submission_id', 'retained_programme_record_id'):
require(len({entry[key] for entry in rows}) == len(rows), 'Repeated reviewed ' + key)
for entry in rows:
doi = entry['doi']
row = entry['source_row']
url = 'https://link.springer.com/chapter/' + doi
evidence = row['metadata']['publisher_chapter_evidence']
require(re.fullmatch(r'10\.1007/978-3-032-\d{5}-[0-9Xx]_[1-9]\d*', doi)
and (row['source'], row['external_id'], row['url'], row['evidence_url']) ==
(PUBLISHER, doi, url, url)
and (row['venue'], row['year'], row['track'], row['status']) ==
('ECCV', 2026, 'main', 'accepted')
and row['identifiers'] == {'doi': doi}
and identities_for_row(row) == [('publication_doi', doi)]
and isinstance(row['abstract'], str) and row['abstract'].strip()
and re.fullmatch(r'[0-9a-f]{64}', evidence['sha256'])
and re.fullmatch(r'[0-9a-f]{64}', entry['before_packet_sha256']),
'Invalid reviewed publisher observation')
return rows
def existing_coverage(conn, edition, source):
rows = conn.execute('''SELECT list_url,status,expected_count,imported_count,cursor,last_success_at
FROM edition_sources WHERE edition_id=? AND source_id=?''', (edition, source)).fetchall()
require(len(rows) <= 1, 'Source coverage is ambiguous')
actual = source_record_count(conn, source, edition)
if not rows:
require(actual == 0, 'Linked records have no coverage listing')
return None
row = rows[0]
require(row[1] == 'partial' and row[3] == actual, 'Coverage differs from linked records')
return tuple(row)
def reconcile_attempt(store, entry, original, target, venue_id, event_key, venue_names, publisher, signature):
conn = store.conn
doi = entry['doi']
canonical_doi = doi.lower()
paper = entry['retained_paper_id']
submission = entry['retained_submission_id']
programme_record = entry['retained_programme_record_id']
row = entry['source_row']
before = entity_packet(conn, paper)
require(packet_digest(before) == entry['before_packet_sha256'], 'Reviewed paper history changed: ' + doi)
require(not before['source_records_truncated'] and not before['attempts_truncated']
and all(not attempt['decisions_truncated'] and not attempt['sources_truncated']
and all(not decision['sources_truncated'] for decision in attempt['decisions'])
for attempt in before['attempts']), 'Reviewed paper history exceeds verification limits')
same_event = [a for a in before['attempts'] if a['venue'].casefold() == 'eccv' and a['year'] == 2026]
require(len(same_event) == 1, 'Paper has multiple or missing ECCV 2026 attempts')
attempt = same_event[0]
require((attempt['id'], attempt['edition_id'], attempt['source_record_id']) ==
(submission, original, programme_record)
and [(d['id'], d['status']) for d in attempt['decisions']] ==
[(entry['retained_programme_decision_id'], 'accepted')]
and [(s['source_record_id'], s['source']) for s in attempt['sources']] ==
[(programme_record, PROGRAMME)], 'Original programme attempt differs')
programme = next(record for record in before['source_records'] if record['id'] == programme_record)
metadata = json.loads(programme['metadata_json'])
require(KEY not in metadata and attempt['source_submission_id'] == programme['external_id'],
'Programme observation already curated or ambiguous')
scope = curation({
'external_id': programme['external_id'], 'venue_names': venue_names, 'from_edition_id': original,
'year': 2026, 'from_track': 'program_unclassified', 'event_key': event_key,
'venue_id': venue_id, 'track': 'main', 'section': 'published ECCV refereed proceedings',
'evidence_url': row['evidence_url'], 'evidence_sha256': row['metadata']['publisher_chapter_evidence']['sha256'],
}, target, signature)
require(not conn.execute('SELECT 1 FROM submission_identifiers WHERE submission_id=?', (submission,)).fetchone()
and not conn.execute("SELECT 1 FROM submission_identifiers WHERE edition_id=? AND scheme='publication_doi' AND value=?",
(target, doi)).fetchone()
and not conn.execute('SELECT 1 FROM source_records WHERE source_id=? AND external_id=?',
(publisher, doi)).fetchone(), 'Publication observation or submission identity already exists')
require(not conn.execute('SELECT 1 FROM identifiers WHERE scheme=? AND value=?',
(PUBLISHER, doi)).fetchone(), 'Publisher identity already exists')
owner = conn.execute("SELECT paper_id FROM identifiers WHERE scheme='doi' AND value=?", (canonical_doi,)).fetchone()
require(owner is None or owner[0] == paper, 'Publication DOI belongs to another paper')
if owner is None:
conn.execute("INSERT INTO identifiers(scheme,value,paper_id) VALUES('doi',?,?)", (canonical_doi, paper))
before['identifiers'].append({'scheme': 'doi', 'value': canonical_doi, 'paper_id': paper})
before['identifiers'].append({'scheme': PUBLISHER, 'value': doi, 'paper_id': paper})
before['identifiers'].sort(key=lambda identifier: (identifier['scheme'], identifier['value']))
require(conn.execute('UPDATE submissions SET edition_id=? WHERE id=? AND edition_id=?',
(target, submission, original)).rowcount == 1, 'Programme attempt changed')
conn.execute("INSERT INTO submission_identifiers VALUES(?,'publication_doi',?,?)", (target, doi, submission))
metadata[KEY] = [scope]
programme['metadata_json'] = json.dumps(metadata, ensure_ascii=False)
conn.execute('UPDATE source_records SET metadata_json=? WHERE id=?',
(programme['metadata_json'], programme_record))
actual_paper, record = store.record(PUBLISHER, doi, row['url'], row['title'], row['abstract'],
row['authors'], row['identifiers'], row['metadata'])
require(actual_paper == paper, 'Unexpected canonical paper')
store.record_tags(paper, record, PUBLISHER, row['tags'])
actual_edition = store.decision(paper, record, 'ECCV', 2026, 'main', doi, row['evidence_url'],
'accepted', row['raw_decision'], identities=identities_for_row(row),
observed_at=row['observed_at'], event_key=event_key)
require(actual_edition == target, 'Unexpected target edition')
after = entity_packet(conn, paper)
published = [r for r in after['source_records'] if r['id'] == record]
require(len(published) == 1 and published[0]['external_id'] == doi
and published[0]['abstract'] == row['abstract'], 'Publication source is missing or changed')
after['source_records'].remove(published[0])
moved = next(a for a in after['attempts'] if a['id'] == submission)
attached = [s for s in moved['sources'] if s['source_record_id'] == record]
observed = [d for d in moved['decisions'] if any(s['source_record_id'] == record for s in d['sources'])]
require(len(attached) == len(observed) == 1 and attached[0]['source_submission_id'] == doi
and observed[0]['status'] == 'accepted' and observed[0]['evidence_url'] == row['evidence_url']
and len(observed[0]['sources']) == 1, 'Publisher evidence did not attach to the existing submission')
moved['sources'].remove(attached[0])
moved['decisions'].remove(observed[0])
attempt.update(edition_id=target, track='main')
require(before == after, 'Existing paper or submission history changed')
def apply(store, plan, signature, previous_count):
rows = validate_plan(plan)
conn = store.conn
marker = MARKER + signature
require(not conn.execute('SELECT 1 FROM schema_migrations WHERE name=?', (marker,)).fetchone(),
'Reconciliation already applied')
edition_rows = conn.execute('''SELECT e.track,e.id,e.venue_id,e.event_key
FROM editions e JOIN venues v ON v.id=e.venue_id WHERE v.acronym='ECCV' AND e.year=2026
AND e.track IN ('main','program_unclassified')''').fetchall()
require(len(edition_rows) == 2 and {r[0] for r in edition_rows} == {'main', 'program_unclassified'},
'Expected two unambiguous reviewed editions')
editions = {r[0]: (r[1], r[2], r[3]) for r in edition_rows}
original, venue, event = editions['program_unclassified']
target, target_venue, target_event = editions['main']
require(venue == target_venue and event == target_event
and all((r['move_from_edition_id'], r['move_to_edition_id']) == (original, target) for r in rows),
'Reviewed edition identity differs')
names = ['ECCV', *(r[0] for r in conn.execute('SELECT alias FROM venue_aliases WHERE venue_id=?', (venue,)))]
sources = {r[0]: r[1] for r in conn.execute('SELECT name,id FROM sources WHERE name IN (?,?)', (PROGRAMME, PUBLISHER))}
programme, publisher = sources[PROGRAMME], sources[PUBLISHER]
old = existing_coverage(conn, original, programme)
main = existing_coverage(conn, target, programme)
published = existing_coverage(conn, target, publisher)
require(old is not None and old[3] >= len(rows) and (main is None or main[0] == old[0])
and published is not None and published[2:4] == (PUBLISHER_COUNT, previous_count)
and previous_count + len(rows) <= PUBLISHER_COUNT, 'Reviewed coverage parent differs')
for entry in rows:
reconcile_attempt(store, entry, original, target, venue, event, names, publisher, signature)
remaining = old[3] - len(rows)
imported = previous_count + len(rows)
main_count = (main[3] if main else 0) + len(rows)
for edition, source, expected in ((original, programme, remaining),
(target, programme, main_count), (target, publisher, imported)):
require(source_record_count(conn, source, edition) == expected, 'Reconciled coverage differs from linked records')
conn.execute('UPDATE edition_sources SET imported_count=? WHERE edition_id=? AND source_id=? AND list_url=?',
(remaining, original, programme, old[0]))
conn.execute('''INSERT INTO edition_sources
(edition_id,source_id,list_url,status,imported_count,cursor,last_success_at) VALUES(?,?,?,'partial',?,?,?)
ON CONFLICT(edition_id,source_id,list_url) DO UPDATE SET imported_count=excluded.imported_count''',
(target, programme, old[0], main_count, old[4], old[5]))
conn.execute('''UPDATE edition_sources SET imported_count=?,last_success_at=?
WHERE edition_id=? AND source_id=? AND list_url=?''', (imported, timestamp(), target, publisher, published[0]))
conn.execute('INSERT INTO schema_migrations(name) VALUES(?)', (marker,))
return {'status': 'applied', 'plan_sha256': signature, 'programme_main': main_count,
'programme_unclassified': remaining, 'springer_main': imported, 'new_papers': 0, 'new_submissions': 0}
def main(argv=None, *, prog=None):
parser = argparse.ArgumentParser(prog=prog)
database_arguments(parser)
parser.add_argument('--plan', type=Path, required=True)
parser.add_argument('--plan-sha256', required=True)
parser.add_argument('--springer-previous-count', type=int, required=True)
args = parser.parse_args(argv)
raw = args.plan.read_bytes()
signature = sha256(raw).hexdigest()
require(signature == args.plan_sha256, 'Plan hash differs')
plan = json.loads(raw)
validate_plan(plan)
store = open_store(args.db, pg_service=args.pg_service, pg_schema=args.pg_schema)
with closing(store.conn) as conn:
try:
if not getattr(conn, 'postgres', False):
conn.commit()
conn.execute('BEGIN IMMEDIATE')
result = apply(store, plan, signature, args.springer_previous_count)
conn.commit()
except BaseException:
conn.rollback()
raise
print(json.dumps(result), flush=True)
+4
View File
@@ -26,6 +26,10 @@ class Result:
def rowcount(self):
return self.cursor.rowcount
@property
def description(self):
return self.cursor.description
class WriteConnection:
postgres = True
+140
View File
@@ -0,0 +1,140 @@
from contextlib import closing, redirect_stdout
from hashlib import sha256
import io
import json
import os
from pathlib import Path
from tempfile import TemporaryDirectory
import unittest
import uuid
from paper_atlas.audits.source_identity import entity_packet
from paper_atlas.cli import main
from paper_atlas.db.edition_scopes import resolve
from paper_atlas.db.store import Store
from paper_atlas.ingestion.publication_reconciliation import PROGRAMME, PUBLISHER, packet_digest
from paper_atlas.postgres.client import ReadConnection, connect
from paper_atlas.postgres.migration import migrate
from paper_atlas.postgres.schema import quote
class PublicationReconciliationTests(unittest.TestCase):
def setUp(self):
directory = TemporaryDirectory()
self.addCleanup(directory.cleanup)
self.root = Path(directory.name)
self.db = self.root / 'db.sqlite'
self.store = Store(self.db)
self.conn = self.store.conn
self.addCleanup(self.conn.close)
self.programme_url = 'https://eccv.ecva.net/Conferences/2026/AcceptedPapers'
self.publisher_url = 'https://link.springer.com/conference/eccv'
programme = self.store.source(PROGRAMME, self.programme_url)
publisher = self.store.source(PUBLISHER, self.publisher_url)
self.entries = [self.programme_attempt(str(i)) for i in range(1, 4)]
title = 'A distinct chapter already integrated in the main track'
paper, record = self.store.record(PUBLISHER, 'prior', self.publisher_url, title, 'Text', [])
self.target = self.store.accepted(paper, record, 'ECCV', 2026, 'main', 'prior', self.publisher_url)
self.original = self.entries[0]['move_from_edition_id']
for entry in self.entries:
entry['move_to_edition_id'] = self.target
self.conn.executemany('''INSERT INTO edition_sources
(edition_id,source_id,list_url,status,expected_count,imported_count) VALUES(?,?,?,'partial',?,?)''',
[(self.original, programme, self.programme_url, 2834, 3),
(self.target, publisher, self.publisher_url, 2835, 1)])
self.conn.commit()
def programme_attempt(self, number):
title = 'A programme observation for continuing research ' + number
url = self.programme_url + '/' + number
authors = ['Ada Example', 'Sam Example']
paper, record = self.store.record(PROGRAMME, number, url, title, 'Programme text', authors)
original = self.store.accepted(paper, record, 'ECCV', 2026, 'program_unclassified', number, url)
other, old_record = self.store.record('openreview', 'test-' + number, url, title, 'Programme text', authors)
self.assertEqual(other, paper)
self.store.decision(paper, old_record, 'ICLR', 2025, 'main', 'test-' + number, url, 'rejected', 'Reject')
packet = entity_packet(self.conn, paper)
attempt = next(row for row in packet['attempts'] if row['source_record_id'] == record)
doi = '10.1007/978-3-032-00000-1_' + number
chapter = 'https://link.springer.com/chapter/' + doi
row = {'source': PUBLISHER, 'external_id': doi, 'url': chapter, 'evidence_url': chapter,
'venue': 'ECCV', 'year': 2026, 'track': 'main', 'status': 'accepted',
'title': title, 'authors': authors, 'abstract': 'Publisher text',
'observed_at': '2026-11-18T00:00:00Z', 'raw_decision': 'Publication observation',
'identifiers': {'doi': doi},
'submission_identifiers': [{'scheme': 'publication_doi', 'value': doi}],
'metadata': {'publisher_chapter_evidence': {'sha256': 'a' * 64}},
'tags': [{'namespace': PUBLISHER + ':format', 'value': 'chapter'}]}
return {'doi': doi, 'retained_paper_id': paper, 'retained_programme_record_id': record,
'retained_submission_id': attempt['id'], 'retained_programme_decision_id': attempt['decisions'][0]['id'],
'move_from_edition_id': original, 'source_row': row, 'before_packet_sha256': packet_digest(packet)}
def run_plan(self, entries, *database):
path = self.root / 'plan.json'
raw = json.dumps({'status': 'readonly_same_submission_reconciliation_plan', 'database_written': False,
'plans': entries, 'guarded_same_submission_plans': len(entries)}).encode()
path.write_bytes(raw)
main(['maintain', 'eccv-publication-reconciliation', *(database or ('--db', str(self.db))),
'--plan', str(path), '--plan-sha256', sha256(raw).hexdigest(), '--springer-previous-count', '1'])
def test_preserves_history_and_updates_all_three_coverage_scopes(self):
entry = self.entries[0]
before = entity_packet(self.conn, entry['retained_paper_id'])
self.run_plan([entry])
packet = entity_packet(self.conn, entry['retained_paper_id'])
self.assertEqual(packet['paper'], before['paper'])
programme, other = packet['attempts']
self.assertEqual((programme['id'], programme['edition_id']), (entry['retained_submission_id'], self.target))
self.assertEqual(other, before['attempts'][1])
self.assertEqual(programme['decisions'][0], before['attempts'][0]['decisions'][0])
self.assertEqual([row['status'] for row in programme['decisions']], ['accepted', 'accepted'])
self.assertEqual({r['source'] for r in programme['sources']}, {PROGRAMME, PUBLISHER})
self.assertEqual(len(packet['source_records']), 3)
self.assertEqual(resolve(self.conn, entry['retained_programme_record_id'], '1',
'ECCV', 2026, 'program_unclassified', None), self.target)
coverage = {(edition, source): (status, expected, imported)
for edition, source, status, expected, imported in self.conn.execute('''SELECT
e.track,s.name,c.status,c.expected_count,c.imported_count FROM edition_sources c
JOIN editions e ON e.id=c.edition_id JOIN sources s ON s.id=c.source_id''')}
self.assertEqual(coverage, {('program_unclassified', PROGRAMME): ('partial', 2834, 2),
('main', PROGRAMME): ('partial', None, 1),
('main', PUBLISHER): ('partial', 2835, 2)})
with self.assertRaisesRegex(ValueError, 'already applied'):
self.run_plan([entry])
def test_changed_history_rolls_back_earlier_observation(self):
first, second = self.entries[:2]
stale = {**second, 'before_packet_sha256': 'b' * 64}
packet = entity_packet(self.conn, first['retained_paper_id'])
with self.assertRaisesRegex(ValueError, 'Reviewed paper history changed'):
self.run_plan([first, stale])
self.assertEqual(entity_packet(self.conn, first['retained_paper_id']), packet)
self.assertIsNone(self.conn.execute(
"SELECT 1 FROM identifiers WHERE scheme='doi' AND value=?", (first['doi'],)).fetchone())
self.assertEqual(self.conn.execute('SELECT count(*) FROM edition_sources').fetchone()[0], 2)
@unittest.skipUnless(os.environ.get('PAPER_ATLAS_TEST_PG_SERVICE'), 'Requires the remote PostgreSQL test service')
def test_postgres_write_verifies_history_from_the_read_backend(self):
service = os.environ['PAPER_ATLAS_TEST_PG_SERVICE']
schema = 'pa_eccv_' + uuid.uuid4().hex[:12]
def cleanup():
with connect(service) as conn:
for name, in conn.execute('SELECT nspname FROM pg_namespace WHERE starts_with(nspname,%s)', (schema,)):
conn.execute('DROP SCHEMA ' + quote(name) + ' CASCADE')
self.addCleanup(cleanup)
self.conn.close()
with redirect_stdout(io.StringIO()):
migrate(self.db, service, self.root / 'migration.json', schema=schema, min_free_gib=0)
entry = self.entries[0]
with closing(ReadConnection(service, schema)) as conn:
before = entity_packet(conn, entry['retained_paper_id'])
reviewed = {**entry, 'before_packet_sha256': packet_digest(before)}
self.run_plan([reviewed], '--pg-service', service, '--pg-schema', schema)
with closing(ReadConnection(service, schema)) as conn:
packet = entity_packet(conn, entry['retained_paper_id'])
self.assertEqual(packet['attempts'][1], before['attempts'][1])
self.assertEqual(packet['attempts'][0]['id'], entry['retained_submission_id'])
self.assertEqual(packet['attempts'][0]['edition_id'], self.target)
self.assertEqual(len(packet['attempts'][0]['decisions']), 2)