Merge main and preserve coverage and history audit checks

Reviewed-on: #17

.gitea/workflows/check.yml

@@ -53,8 +53,11 @@ jobs:
               subprocess.run([python,'-I','-B','-m','paper_atlas','import','arxiv','--help'],cwd=outside,check=True)
               subprocess.run([python,'-I','-B','-m','paper_atlas','maintain','backfill','--help'],cwd=outside,check=True)
               subprocess.run([python,'-I','-B','-m','paper_atlas','audit','public-openreview','--help'],cwd=outside,check=True)
+              subprocess.run([python,'-I','-B','-m','paper_atlas','coverage','--help'],cwd=outside,check=True)
+              subprocess.run([python,'-I','-B','-m','paper_atlas','audit','histories','--help'],cwd=outside,check=True)
               subprocess.run([python,'-I','-B','-c','from paper_atlas.postgres.writer import PostgresStore, WriteConnection'],cwd=outside,check=True)
               smoke = "from paper_atlas.db.store import Store; from paper_atlas.config import DEFAULT_REGISTRY; import json; assert {'COLM','MM','MLSys'} <= {v['acronym'] for v in json.loads(DEFAULT_REGISTRY.read_text())['venues']}; s=Store('installed.sqlite'); s.conn.close()"
               subprocess.run([python,'-I','-B','-c',smoke],cwd=outside,check=True)
               subprocess.run([python,'-I','-B','-m','paper_atlas','query','--db','installed.sqlite','--stats'],cwd=outside,check=True)
+              subprocess.run([python,'-I','-B','-m','paper_atlas','audit','histories','--db','installed.sqlite','--examples','0'],cwd=outside,check=True)
           PY

README.md

@@ -51,7 +51,8 @@ PGSERVICEFILE=/etc/paper-atlas/pg_service.conf paper-atlas query --pg-service fo
 PGSERVICEFILE=/etc/paper-atlas/pg_service.conf paper-atlas query --pg-service foundry-research --paper 123
 
 # 列出全部目标会议，包括尚未入库的项目
-paper-atlas coverage --db /data/paper-index/papers.sqlite \
+PGSERVICEFILE=/etc/paper-atlas/pg_service.conf paper-atlas coverage \
+  --pg-service foundry-research --pg-schema research \
   --out /data/paper-index/collected/venue-coverage-current.json
 
 # 采集与导入分开；这里只采集到独立文件

docs/data-sources.md

@@ -93,12 +93,6 @@ paper-atlas audit public-openreview --db /data/paper-index/papers.sqlite \
 
 输出按会议、届次及 main、position、datasets_benchmarks 分文件。workshop 默认进入单独的 held 清单；有明确 workshop 名称时可通过 `--include-workshops` 单列采集，不合并到主会。Retraction、未映射的轨道或负面标签也保留在 `held-records.jsonl`，不强改为 Reject 或 Withdrawn。当前决定表没有独立 retracted 状态，因此撤回发表记录仍需后续专门建模。
 
-`parser_version=2` 还会将明确负面标签但标题缺失、空白或 null 的记录放入 `held-records.jsonl`，保留原始状态、forum ID、固定来源版本和行号；不制造标题或导入不可识别的论文，也不让这一条缺失阻断其他可用记录。held 文件哈希与各原因数量保存在 inventory；非文本标题、错误身份、冲突届次和哈希不符仍使整批失败。旧年 ICLR 镜像的明确 `Withdraw` 标签归为 `withdrawn`，`Workshop`、`Poster Workshop`、`Oral Workshop` 均不作为拒稿。采集文件与导入主库仍是两个独立步骤。
-
-`parser_version=3` 增加五组已核实身份的 ICLR workshop 标签：2024 ME-FoMo / GenAI4DM 的 Withdrawn Submission、2025 AgenticAI 的 Reject / LMRL 的 Withdrawn Submission，以及 2026 LLM Reasoning 的 Withdrawn Submission。仅在 `--include-workshops` 下输出，并分别保存 `workshop_me_fomo`、`workshop_genai4dm`、`workshop_agenticai`、`workshop_lmrl`、`workshop_llm_reasoning`；默认仍暂存。标签必须与原文件会议、年份和原始 workshop track 一致，错年/错会议使采集失败，错轨道单独暂存，未知标签继续待核查，不用宽泛的 Reject/Withdraw 正则猜测。
-
-[OpenReview 官方 ICLR 活动目录](https://openreview.net/venue?id=ICLR.cc) 确认上述独立 workshop 身份；[ME-FoMo 2024](https://sites.google.com/view/me-fomo2024/call-for-papers)、[GenAI4DM 2024](https://sites.google.com/view/genai4dm-iclr2024)、[AgenticAI 2025](https://iclragenticai.github.io/) 和 [LLM Reasoning 2026](https://sites.google.com/view/iclr-2026-llmreasoning) 的活动网站也提供对应投稿入口。输出在 metadata 保存 `workshop_venueid` 和身份目录证据，决定仍只标为固定镜像的公开分类断言，保留原始标签、forum、来源行号与哈希，不声称核验原始决定签名。2026-11-08 核查时部分原站记录页要求浏览器验证；身份目录和主办方入口可读，不代表逐篇原始决定已可访问，也不外推其他 workshop 的公开政策。采集后仍须经过 PostgreSQL 身份/决定冲突预检、串行导入和独立验收。
-
 该来源明确标记为 `curated_public_status_assertion`：保留源 commit、文件哈希、行号、原标签和 forum ID，不声称已核验原始 PC 签名。只保存标题、摘要、公开作者名、关键词、领域、索引与决定来源，不缓存镜像中的个人资料、邮箱、性别、评审内容或 PDF。所有输出清单均为 partial；盘点中的零条记录仅描述这个固定镜像。
 
 入库前审计核验文件哈希、数量、轨道和来源声明，并用真实 forum ID 对比同届同轨道已有投稿。存在不同决定时阻止整批导入，交由来源和时序核查；另一会议的后续录用属于独立投稿历史，不构成该次拒稿的冲突。同一 forum 的镜像不新增投稿事件，但论文实体的跨平台去重仍须在全量结束后核验。审计通过的报告包含带 SHA-256 的导入队列；应在单写入维护锁下及时串行导入，主库变化后重新审计。

docs/operations.md

@@ -97,13 +97,16 @@ paper-atlas audit arxiv-tags --db /data/paper-index/papers.sqlite \
 采集文件、已入库的来源记录和去重论文实体是不同的计数。来源可能覆盖同一篇论文，同一篇论文也可能有多次投稿；不能把各批文件数量相加作为独立论文总数。实时统计与覆盖报告使用：
 
 ```sh
-paper-atlas query --db /data/paper-index/papers.sqlite --stats
-paper-atlas coverage --db /data/paper-index/papers.sqlite \
+PGSERVICEFILE=/etc/paper-atlas/pg_service.conf paper-atlas query --pg-service foundry-research --pg-schema research --stats
+PGSERVICEFILE=/etc/paper-atlas/pg_service.conf paper-atlas coverage \
+  --pg-service foundry-research --pg-schema research \
   --out /data/paper-index/collected/venue-coverage-current.json
 ```
 
 manifest 和采集器状态清单用于确认文件是否齐全及是否缺摘要；`ingest_runs` 和协调器日志用于确认主库导入是否成功。官方程序、书目目录和公开投稿快照分别保留，未出现在某份清单不等于被拒稿；第三方快照也不能代表平台当前全量状态。代码测试成功不能代替最终全库去重、完整性和逐届覆盖核验。
 
+`coverage` 的所有计数来自同一个只读快照。`latest_observed_status_counts` 按每次投稿的最新观察决定统计，`decision_observation_counts` 保留所有历史观察；重复来源不增加投稿数，缺少决定的投稿计入最新状态 unknown，但不凭空制造 unknown 决定。`source_record_counts` 统计来源在该届次的实际记录与缺摘要数，和 `sources` 中逐清单的预期/导入数量分开；身份补充清单与原始论文集清单可以描述同一条记录。来源缺摘要与论文实体从其他来源获得摘要也分别报告。报告仍将历史覆盖标为待独立审计，不能用某个清单 complete 或没有拒稿记录推断全会议覆盖。
+
 当前代码与合并状态以 [Pull Requests](https://gitea.entroia.com/Zayden-Jung/paper-atlas/pulls) 为准；CI 结果以对应提交的 Actions 检查为准。工作分支交付后等待用户手动合并。合并代码不自动部署、不自动启动采集，也不代表数据批次已完成。
 ## 修正已有 PMLR 投稿的轨道
 

docs/usage.md

@@ -79,22 +79,41 @@ paper-atlas maintain merge --db /data/paper-index/samples/example.sqlite \
 ## 4. 查询与分析前检查
 
 ```sh
-paper-atlas query --db /data/paper-index/papers.sqlite --query '"language model"' --limit 20
-paper-atlas query --db /data/paper-index/papers.sqlite --tag arxiv:category=cs.AI
-paper-atlas query --db /data/paper-index/papers.sqlite --venue ICLR --year 2025 --status rejected
-paper-atlas query --db /data/paper-index/papers.sqlite --paper 123
-paper-atlas coverage --db /data/paper-index/papers.sqlite --out /data/paper-index/collected/coverage.json
+paper-atlas query --pg-service foundry-research --query '"language model"' --limit 20
+paper-atlas query --pg-service foundry-research --tag arxiv:category=cs.AI
+paper-atlas query --pg-service foundry-research --venue ICLR --year 2025 --status rejected
+paper-atlas query --pg-service foundry-research --paper 123
+PGSERVICEFILE=/etc/paper-atlas/pg_service.conf paper-atlas coverage \
+  --pg-service foundry-research --pg-schema research --out /data/paper-index/collected/coverage.json
 ```
 
 `--status rejected` 筛选某次投稿的最新已知决定。它并不表示论文在所有会议均未录用；分析转投情况时查看该论文的完整投稿历史。年份表示会议届次，观察时间和决定日期分别保留。
 
+覆盖报告使用只读一致性快照，标明实际数据库和 schema，并列出零入库目标与注册表外的已入库会议。逐届、逐 track 分别列出最新观察状态的投稿数、全部决定观察数、各来源记录数，以及来源摘要和论文实体摘要的缺失数。摘要缺失指 NULL 或空字符串；有文字不代表内容已经完整核验。同一来源的多个清单 URL 可以重叠，不能把清单计数相加当作来源总量；零条拒稿也不证明该会议没有公开拒稿。SQLite 样本仍可显式使用 `--db`，生产报告应查询当前 PostgreSQL。
+
 例如，查找在 ICLR 有拒稿记录、但在其他已知投稿中已有录用的论文：
 
 ```sh
-paper-atlas query --db /data/paper-index/papers.sqlite \\
+paper-atlas query --pg-service foundry-research \
   --venue ICLR --status rejected --outcome accepted
 ```
 
 `known_acceptances` 列出目前仍为录用状态的各次投稿及会议、年份、track 和决定证据，`publication_outcome` 汇总为 accepted 或 no_known_acceptance。汇总会检查所有会议和 track，不把某次拒稿覆盖到整篇论文；同一次投稿的后续撤稿也会更新该次状态并保留原决定。no_known_acceptance 只表示现有来源中没有查到录用，不能称为“从未录用”。
 
 `--stats` 中 papers 是当前规则下的论文实体数；source_records 是来源记录数；decisions 是决定观察数。一个投稿可能有多个来源和多条观察，不能相加当作论文数。
+
+## 5. 投稿历史关系审计
+
+```sh
+PGSERVICEFILE=/etc/paper-atlas/pg_service.conf paper-atlas audit histories \
+  --pg-service foundry-research --pg-schema research --examples 20 \
+  --out /data/paper-index/collected/history-audit.json
+```
+
+使用一个只读一致性快照，区分论文实体、投稿和决定观察；分别统计同次投稿曾出现负面与录用观察、不同投稿共存负面与当前已知录用、不同会议之间的共存。workshop/main 是不同 track，属于同一会议时不算跨会议；会议别名也不会增加一个会议。当前已知录用使用与 `query --outcome` 相同的“每次投稿最新观察”规则，后续撤稿的该次投稿不再算当前录用，但旧录用决定仍保留。
+
+报告不凭观察时间、届次年份或行顺序断言“拒稿发生在录用之前”。样例保留真实 `decision_date`、`observed_at`、原始标签、证据 URL 和决定来源，供进一步核查；日期缺失仍保持未知。`negative` 统计包含 rejected、desk_rejected 和 withdrawn，其原始状态不互相替换。
+
+总数覆盖快照内全部投稿，样例由 `--examples` 限定为 0–100 篇；每篇最多展示 20 次投稿、每次最多 50 条决定、每条决定最多 100 个来源，并明确标记截断。样例优先保留负面/录用的不同投稿对。四项来源关系检查报告缺失关联和来源论文归属不一致，不自动修复。
+
+这些统计依赖现有论文身份关联，不证明实体合并正确，也不证明原始决定签名、公开拒稿全集、标签类型或最终全库去重已完成。SQLite 独立样本可使用 `--db`；冻结备份不能代表当前 PostgreSQL 的历史情况。

paper_atlas/audits/coverage.py

@@ -1,45 +1,136 @@
-"""Show every target venue, including those with no imported records."""
+"""Describe imported venue coverage without implying a complete census."""
 import argparse
-from paper_atlas.config import DEFAULT_REGISTRY
+from collections import defaultdict
 import json
 from pathlib import Path
 import sqlite3
 
+from paper_atlas.config import DEFAULT_REGISTRY
+from paper_atlas.db.backend import database_arguments, open_read_connection
 from paper_atlas.db.submissions import timestamp
+from paper_atlas.querying import LATEST_DECISION
 
 
-def report(db,registry):
-    registry=json.loads(Path(registry).read_text());targets=registry['venues']
-    names=[row['acronym'].casefold() for row in targets]
-    if len(names)!=len(set(names)):raise ValueError('Duplicate target venue')
-    scope='(v.acronym=? COLLATE NOCASE OR EXISTS(SELECT 1 FROM venue_aliases a WHERE a.venue_id=v.id AND a.alias=? COLLATE NOCASE))'
-    c=sqlite3.connect(Path(db).resolve().as_uri()+'?mode=ro',uri=True);c.row_factory=sqlite3.Row
+STATUSES = ('accepted', 'rejected', 'desk_rejected', 'withdrawn', 'under_review', 'unknown')
+
+
+def _status_counts(connection, *, latest):
+    if latest:
+        query = '''SELECT s.edition_id,COALESCE(d.status,'unknown') AS status,count(*) AS count
+            FROM submissions s LEFT JOIN decisions d ON d.submission_id=s.id AND ''' + LATEST_DECISION
+    else:
+        query = '''SELECT s.edition_id,COALESCE(d.status,'unknown') AS status,count(*) AS count
+            FROM submissions s JOIN decisions d ON d.submission_id=s.id'''
+    counts = defaultdict(dict)
+    for row in connection.execute(query + " GROUP BY s.edition_id,COALESCE(d.status,'unknown')"):
+        counts[row[0]][row[1]] = row[2]
+    return counts
+
+
+def report(db, registry, *, pg_service=None, pg_schema=None):
+    started_at = timestamp()
+    registry = json.loads(Path(registry).read_text())
+    targets = registry['venues']
+    names = [row['acronym'].casefold() for row in targets]
+    if len(names) != len(set(names)):
+        raise ValueError('Duplicate target venue')
+    scope = '''(lower(v.acronym)=lower(?) OR EXISTS(
+        SELECT 1 FROM venue_aliases a WHERE a.venue_id=v.id AND lower(a.alias)=lower(?)))'''
+    connection = open_read_connection(db, pg_service=pg_service, pg_schema=pg_schema)
+    if not pg_service:
+        connection.row_factory = sqlite3.Row
     try:
-        c.execute('BEGIN');venues=[]
+        if pg_service:
+            identity = connection.execute('SELECT current_database(),current_schema()').fetchone()
+            database, schema = identity[0], identity[1]
+        else:
+            database, schema = str(Path(db).resolve()), None
+        latest = _status_counts(connection, latest=True)
+        observations = _status_counts(connection, latest=False)
+        # Aggregate the paper join once for the whole snapshot. Repeating it
+        # for each registry target caused repeated scans on the production DB.
+        edition_counts = {}
+        for row in connection.execute('''SELECT s.edition_id,count(s.id) AS submissions,
+                count(DISTINCT s.paper_id) AS papers,
+                count(DISTINCT CASE WHEN p.abstract IS NULL OR p.abstract='' THEN s.paper_id END)
+                    AS papers_without_abstract_text
+                FROM submissions s LEFT JOIN papers p ON p.id=s.paper_id GROUP BY s.edition_id'''):
+            entry = dict(row)
+            edition_counts[entry.pop('edition_id')] = entry
+        source_counts = defaultdict(list)
+        for row in connection.execute('''SELECT s.edition_id,so.name AS source,
+                count(DISTINCT sr.id) AS source_records,
+                count(DISTINCT CASE WHEN sr.abstract IS NULL OR sr.abstract='' THEN sr.id END)
+                    AS source_records_without_abstract_text
+                FROM submission_sources ss JOIN submissions s ON s.id=ss.submission_id
+                JOIN source_records sr ON sr.id=ss.source_record_id JOIN sources so ON so.id=sr.source_id
+                GROUP BY s.edition_id,so.name ORDER BY s.edition_id,so.name'''):
+            entry = dict(row)
+            source_counts[entry.pop('edition_id')].append(entry)
+        venues, covered_venue_ids = [], set()
         for target in targets:
-            acronym=target['acronym'];params=(acronym,acronym)
-            editions=[]
-            for row in c.execute('SELECT e.id,v.acronym AS canonical_venue,e.year,e.track,count(s.id) AS submissions,count(DISTINCT s.paper_id) AS papers FROM editions e JOIN venues v ON v.id=e.venue_id LEFT JOIN submissions s ON s.edition_id=e.id WHERE '+scope+' GROUP BY e.id ORDER BY e.year,e.track',params):
-                edition=dict(row)
-                edition['sources']=[dict(source) for source in c.execute('SELECT so.name,es.status,es.expected_count,es.imported_count,es.list_url,es.last_success_at,es.last_error FROM edition_sources es JOIN sources so ON so.id=es.source_id WHERE es.edition_id=? ORDER BY so.name',(row['id'],))]
+            acronym = target['acronym']
+            params = (acronym, acronym)
+            editions = []
+            for row in connection.execute('''SELECT e.id,v.id AS venue_id,v.acronym AS canonical_venue,
+                    e.year,e.track FROM editions e JOIN venues v ON v.id=e.venue_id
+                    WHERE ''' + scope + '''
+                    ORDER BY e.year,e.track,v.acronym''', params):
+                edition = dict(row)
+                edition.update(edition_counts.get(row['id'], {
+                    'submissions': 0, 'papers': 0, 'papers_without_abstract_text': 0}))
+                covered_venue_ids.add(edition['venue_id'])
+                edition['latest_observed_status_counts'] = {status: latest[row['id']].get(status, 0) for status in STATUSES}
+                edition['decision_observation_counts'] = {status: observations[row['id']].get(status, 0) for status in STATUSES}
+                edition['source_record_counts'] = source_counts[row['id']]
+                edition['sources'] = [dict(source) for source in connection.execute('''SELECT so.name,
+                    es.status,es.expected_count,es.imported_count,es.list_url,es.last_success_at,es.last_error
+                    FROM edition_sources es JOIN sources so ON so.id=es.source_id
+                    WHERE es.edition_id=? ORDER BY so.name,es.list_url''', (row['id'],))]
                 editions.append(edition)
-            totals=c.execute('SELECT count(s.id),count(DISTINCT s.paper_id) FROM submissions s JOIN editions e ON e.id=s.edition_id JOIN venues v ON v.id=e.venue_id WHERE '+scope,params).fetchone()
-            venues.append({'acronym':acronym,'field':target['field'],'imported_submissions':totals[0],'distinct_papers':totals[1],
-                           'history_status':'partial_history_pending_independent_audit' if totals[0] else 'no_imported_records',
-                           'discovery_sources':target.get('discovery_sources',[]),'editions':editions})
-        return {'checked_at':timestamp(),'target_venue_count':len(targets),'venues_with_imported_records':sum(bool(v['imported_submissions']) for v in venues),
-                'scope_note':registry['scope_note'],'interpretation':'All registry targets are listed. Edition source complete means that source file passed import checks; it does not prove full historical or abstract coverage. Papers can appear in multiple venues/years, so do not sum venue paper counts as a global total.',
-                'venues':venues}
-    finally:c.close()
+            totals = connection.execute('''SELECT count(s.id),count(DISTINCT s.paper_id)
+                FROM submissions s JOIN editions e ON e.id=s.edition_id
+                JOIN venues v ON v.id=e.venue_id WHERE ''' + scope, params).fetchone()
+            venues.append({'acronym': acronym, 'field': target['field'],
+                'imported_submissions': totals[0], 'distinct_papers': totals[1],
+                'history_status': 'partial_history_pending_independent_audit' if totals[0] else 'no_imported_records',
+                'discovery_sources': target.get('discovery_sources', []), 'editions': editions})
+        outside_registry = [dict(row) for row in connection.execute('''SELECT v.id AS venue_id,
+            v.acronym,count(s.id) AS submissions,count(DISTINCT s.paper_id) AS papers
+            FROM submissions s JOIN editions e ON e.id=s.edition_id JOIN venues v ON v.id=e.venue_id
+            GROUP BY v.id,v.acronym ORDER BY v.acronym''') if row[0] not in covered_venue_ids]
+        return {'report_started_at': started_at, 'checked_at': timestamp(),
+            'engine': 'postgresql' if pg_service else 'sqlite', 'database': database, 'schema': schema,
+            'pg_service': pg_service, 'target_venue_count': len(targets),
+            'venues_with_imported_records': sum(bool(v['imported_submissions']) for v in venues),
+            'scope_note': registry['scope_note'],
+            'interpretation': 'All registry targets are listed, including zero imports. '
+                'Latest status counts use the latest observed decision for each attempt; decision observations retain history. '
+                'Source record totals are per source and edition, not per list URL; multiple list entries may describe overlapping records. '
+                'Missing abstract text means NULL or empty string; nonempty text does not prove a complete abstract. '
+                'An edition source marked complete is not a proof of complete conference history or public rejection coverage. '
+                'Zero imported rejections does not establish that no rejections are public. '
+                'Papers can appear in several venues/years; do not sum venue paper counts as a global total.',
+            'venues': venues, 'imported_venues_outside_registry': outside_registry}
+    finally:
+        connection.close()
 
 
 def main(argv=None, *, prog=None):
-    parser=argparse.ArgumentParser(prog=prog);parser.add_argument('--db',required=True)
-    parser.add_argument('--registry',default=str(DEFAULT_REGISTRY));parser.add_argument('--out')
-    args=parser.parse_args(argv);result=report(args.db,args.registry);text=json.dumps(result,ensure_ascii=False,indent=2)
+    parser = argparse.ArgumentParser(prog=prog)
+    database_arguments(parser)
+    parser.add_argument('--registry', default=str(DEFAULT_REGISTRY))
+    parser.add_argument('--out')
+    args = parser.parse_args(argv)
+    result = report(args.db, args.registry, pg_service=args.pg_service, pg_schema=args.pg_schema)
+    text = json.dumps(result, ensure_ascii=False, indent=2)
     if args.out:
-        path=Path(args.out);stage=path.with_suffix('.part');stage.write_text(text+'\n');stage.replace(path)
-    else:print(text)
+        path = Path(args.out)
+        stage = path.with_suffix('.part')
+        stage.write_text(text + '\n')
+        stage.replace(path)
+    else:
+        print(text)
 
 
 if __name__ == '__main__':

paper_atlas/audits/histories.py

@@ -0,0 +1,196 @@
+"""Audit linked submission histories without inferring event chronology."""
+import argparse
+from collections import Counter
+import json
+from pathlib import Path
+import sqlite3
+
+from paper_atlas.db.backend import database_arguments, open_read_connection
+from paper_atlas.db.submissions import timestamp
+from paper_atlas.querying import LATEST_DECISION
+
+
+ATTEMPTS = '''SELECT s.paper_id,s.id AS submission_id,e.venue_id,v.acronym AS venue,e.year,e.track,
+    count(d.id) AS decision_observations,
+    MAX(CASE WHEN d.status IN ('rejected','desk_rejected','withdrawn') THEN 1 ELSE 0 END)
+        AS historical_negative,
+    MAX(CASE WHEN d.status='accepted' THEN 1 ELSE 0 END) AS historical_acceptance,
+    COALESCE(MAX(CASE WHEN ''' + LATEST_DECISION + ''' THEN d.status END),'unknown')
+        AS latest_observed_status
+    FROM submissions s JOIN editions e ON e.id=s.edition_id JOIN venues v ON v.id=e.venue_id
+    LEFT JOIN decisions d ON d.submission_id=s.id
+    GROUP BY s.paper_id,s.id,e.venue_id,v.acronym,e.year,e.track
+    ORDER BY s.paper_id,e.venue_id,s.id'''
+
+
+def _paper_counts(attempts):
+    negative = sum(row['historical_negative'] for row in attempts)
+    accepted = sum(row['latest_observed_status'] == 'accepted' for row in attempts)
+    overlap = sum(row['historical_negative'] and row['latest_observed_status'] == 'accepted'
+                  for row in attempts)
+    venues = {}
+    for row in attempts:
+        counts = venues.setdefault(row['venue_id'], [0, 0])
+        counts[0] += row['historical_negative']
+        counts[1] += row['latest_observed_status'] == 'accepted'
+    return {
+        'papers_with_submissions': 1,
+        'papers_with_multiple_submissions': int(len(attempts) > 1),
+        'papers_with_historical_negative': int(negative > 0),
+        'papers_with_current_known_acceptance': int(accepted > 0),
+        'papers_with_negative_and_acceptance_in_distinct_attempts': int(negative * accepted > overlap),
+        'papers_with_negative_and_acceptance_in_distinct_venues':
+            int(negative * accepted > sum(n * a for n, a in venues.values())),
+    }
+
+
+def _example(connection, attempts):
+    # Keep a negative/accepted pair visible even when a paper has many attempts.
+    accepted = [row for row in attempts if row['latest_observed_status'] == 'accepted']
+    first = accepted[0] if accepted else None
+    other_venue = next((row for row in accepted if row['venue_id'] != first['venue_id']), None)
+    pair = None
+    for negative in attempts:
+        if negative['historical_negative'] and first:
+            candidate = first if first['venue_id'] != negative['venue_id'] else other_venue
+            if candidate:
+                pair = (negative, candidate)
+                break
+    if pair is None:
+        for negative in attempts:
+            if negative['historical_negative'] and first:
+                candidate = first if first['submission_id'] != negative['submission_id'] else (
+                    accepted[1] if len(accepted) > 1 else None)
+                if candidate:
+                    pair = (negative, candidate)
+                    break
+    selected = {}
+    for row in list(pair or ()) + attempts:
+        if len(selected) < 20:
+            selected.setdefault(row['submission_id'], dict(row))
+    for row in selected.values():
+        row['decisions'] = []
+        for decision in connection.execute('''SELECT id,status,raw_value,evidence_url,
+                decision_date,observed_at FROM decisions WHERE submission_id=?
+                ORDER BY observed_at DESC,COALESCE(decision_date,observed_at) DESC,id DESC LIMIT 50''',
+                (row['submission_id'],)):
+            entry = dict(decision)
+            sources = [dict(source) for source in connection.execute('''
+                SELECT so.name,r.external_id,ds.source_decision_id FROM decision_sources ds
+                JOIN source_records r ON r.id=ds.source_record_id JOIN sources so ON so.id=r.source_id
+                WHERE ds.decision_id=? ORDER BY so.name,r.external_id LIMIT 101''', (entry['id'],))]
+            entry['sources'] = sources[:100]
+            entry['sources_truncated'] = len(sources) > 100
+            row['decisions'].append(entry)
+        row['decisions_truncated'] = row['decision_observations'] > len(row['decisions'])
+    paper_id = attempts[0]['paper_id']
+    return {'paper_id': paper_id,
+            'title': connection.execute('SELECT title FROM papers WHERE id=?', (paper_id,)).fetchone()[0],
+            'counts': _paper_counts(attempts), 'attempt_count': len(attempts),
+            'attempts': list(selected.values()), 'attempts_truncated': len(attempts) > len(selected),
+            'chronology': 'Not inferred from observed_at, edition year, or row order.'}
+
+
+def report(db=None, *, pg_service=None, pg_schema=None, examples=20):
+    if not 0 <= examples <= 100:
+        raise ValueError('Examples must be between 0 and 100')
+    started_at = timestamp()
+    connection = open_read_connection(db, pg_service=pg_service, pg_schema=pg_schema)
+    if not pg_service:
+        connection.row_factory = sqlite3.Row
+    try:
+        if pg_service:
+            identity = connection.execute('SELECT current_database(),current_schema()').fetchone()
+            database, schema = identity[0], identity[1]
+        else:
+            database, schema = str(Path(db).resolve()), None
+        names = ['papers_with_submissions', 'papers_with_multiple_submissions',
+                 'papers_with_historical_negative', 'papers_with_current_known_acceptance',
+                 'papers_with_negative_and_acceptance_in_distinct_attempts',
+                 'papers_with_negative_and_acceptance_in_distinct_venues']
+        totals = Counter({key: 0 for key in names})
+        totals.update(submissions=0, decision_observations=0,
+                      attempts_with_historical_negative=0,
+                      attempts_with_historical_negative_and_acceptance=0)
+        latest = Counter()
+        sample_attempts = []
+
+        def consume(rows):
+            if not rows:
+                return
+            counts = _paper_counts(rows)
+            totals.update(counts)
+            if counts['papers_with_negative_and_acceptance_in_distinct_attempts'] and len(sample_attempts) < examples:
+                sample_attempts.append(rows)
+
+        current = []
+        for raw in connection.execute(ATTEMPTS):
+            row = dict(raw)
+            if current and row['paper_id'] != current[0]['paper_id']:
+                consume(current)
+                current = []
+            current.append(row)
+            totals['submissions'] += 1
+            totals['decision_observations'] += row['decision_observations']
+            totals['attempts_with_historical_negative'] += row['historical_negative']
+            totals['attempts_with_historical_negative_and_acceptance'] += int(
+                row['historical_negative'] and row['historical_acceptance'])
+            latest[row['latest_observed_status']] += 1
+        consume(current)
+        # These are relationship checks, not a substitute for the final entity/FK audit.
+        integrity = {}
+        checks = {
+            'submissions_without_source_links': '''SELECT count(*) FROM submissions s
+                WHERE NOT EXISTS(SELECT 1 FROM submission_sources ss WHERE ss.submission_id=s.id)''',
+            'decisions_without_source_links': '''SELECT count(*) FROM decisions d
+                WHERE NOT EXISTS(SELECT 1 FROM decision_sources ds WHERE ds.decision_id=d.id)''',
+            'submission_source_paper_mismatches': '''SELECT count(*) FROM submission_sources ss
+                JOIN submissions s ON s.id=ss.submission_id JOIN source_records r ON r.id=ss.source_record_id
+                WHERE s.paper_id<>r.paper_id''',
+            'decision_source_paper_mismatches': '''SELECT count(*) FROM decision_sources ds
+                JOIN decisions d ON d.id=ds.decision_id JOIN submissions s ON s.id=d.submission_id
+                JOIN source_records r ON r.id=ds.source_record_id WHERE s.paper_id<>r.paper_id''',
+        }
+        for name, query in checks.items():
+            integrity[name] = connection.execute(query).fetchone()[0]
+        return {'report_started_at': started_at, 'checked_at': timestamp(),
+                'engine': 'postgresql' if pg_service else 'sqlite', 'database': database,
+                'schema': schema, 'pg_service': pg_service,
+                'status': 'observed_history_audit_not_final_dedup', 'counts': dict(totals),
+                'latest_attempt_status_counts': dict(sorted(latest.items())),
+                'evidence_relationship_checks': integrity,
+                'examples': [_example(connection, rows) for rows in sample_attempts],
+                'example_limit': examples,
+                'interpretation': [
+                    'Counts use existing canonical paper links, not independently verified real-world identity.',
+                    'Negative includes rejected, desk_rejected and withdrawn observations; these remain distinct decisions.',
+                    'Known acceptance follows the latest observed decision per attempt, matching query --outcome.',
+                    'Distinct attempts/venues require separate submission/venue IDs; aliases do not create new venues.',
+                    'Coexisting evidence does not prove rejection happened before acceptance. observed_at is collection time.',
+                    'No known acceptance is not proof of never accepted. Source signatures and completeness are not established here.',
+                    'Examples are bounded; counts include all attempts and observations in one read-only snapshot.',
+                    'This does not perform entity merging, final deduplication, or label namespace/type validation.',
+                ]}
+    finally:
+        connection.close()
+
+
+def main(argv=None, *, prog=None):
+    parser = argparse.ArgumentParser(prog=prog)
+    database_arguments(parser)
+    parser.add_argument('--examples', type=int, default=20)
+    parser.add_argument('--out')
+    args = parser.parse_args(argv)
+    result = report(args.db, pg_service=args.pg_service, pg_schema=args.pg_schema, examples=args.examples)
+    text = json.dumps(result, ensure_ascii=False, indent=2)
+    if args.out:
+        path = Path(args.out)
+        stage = path.with_suffix('.part')
+        stage.write_text(text + '\n')
+        stage.replace(path)
+    else:
+        print(text)
+
+
+if __name__ == '__main__':
+    main()

paper_atlas/audits/public_openreview.py

@@ -39,7 +39,9 @@ def audit(db, inventory, out, *, pg_service=None, pg_schema=None):
             if json.loads(path.with_suffix('.manifest.json').read_text()) != manifest:
                 raise ValueError('Collected manifest differs from inventory')
             count = 0
-            for line in path.read_text().splitlines():
+            # Unicode paragraph/line separators are valid inside JSON strings;
+            # split physical byte lines, not Unicode text boundaries.
+            for line in path.read_bytes().splitlines():
                 row = json.loads(line)
                 if any(row[key] != manifest[key] for key in ('source', 'venue', 'year', 'track')):
                     raise ValueError('Record edition differs from inventory')

paper_atlas/cli.py

@@ -35,6 +35,7 @@ COMMANDS = {
     ('audit','arxiv'): ('paper_atlas.audits.arxiv', {}, '核验 arXiv 原始编号、重复行与 ZIP 完整性'),
     ('audit','arxiv-tags'): ('paper_atlas.audits.arxiv_tags', {}, '逐记录核验 arXiv 类别及出处完整性'),
     ('audit','public-openreview'): ('paper_atlas.audits.public_openreview', {}, '核对公开负面决定与已有投稿历史'),
+    ('audit','histories'): ('paper_atlas.audits.histories', {}, '只读核对跨会议投稿历史及决定来源关系'),
 }
 
 

paper_atlas/collectors/public_openreview.py

@@ -27,25 +27,11 @@ RAW_ROOT = 'https://raw.githubusercontent.com/papercopilot/openreview/' + REVISI
 TREE_URL = 'https://api.github.com/repos/papercopilot/openreview/git/trees/' + REVISION + '?recursive=1'
 NEGATIVE = {'reject': 'rejected', 'rejected': 'rejected',
             'desk reject': 'desk_rejected', 'desk rejected': 'desk_rejected',
-            'withdraw': 'withdrawn', 'withdrawn': 'withdrawn', 'withdrawal': 'withdrawn'}
+            'withdrawn': 'withdrawn', 'withdrawal': 'withdrawn'}
 TRACKS = {'main': 'main', 'Position': 'position',
           'Datasets & Benchmarks': 'datasets_benchmarks'}
 WORKSHOP_REJECT_LABELS = {"aim-fm workshop @ neurips'24 reject",
                           'coginterp @ neurips 2025 reject'}
-# Exact curated labels, tied to the year and raw workshop track. The official
-# venue directory establishes workshop identity, not an original decision signature.
-ICLR_WORKSHOP_LABELS = {
-    'iclr 2024 workshop mefomo withdrawn submission':
-        (2024, 'workshop_me_fomo', 'withdrawn', 'ME-FoMo'),
-    'iclr 2024 workshop genai4dm withdrawn submission':
-        (2024, 'workshop_genai4dm', 'withdrawn', 'GenAI4DM'),
-    'iclr 2025 workshop agenticai reject':
-        (2025, 'workshop_agenticai', 'rejected', 'AgenticAI'),
-    'iclr 2025 workshop lmrl withdrawn submission':
-        (2025, 'workshop_lmrl', 'withdrawn', 'LMRL'),
-    'iclr 2026 workshop llm reasoning withdrawn submission':
-        (2026, 'workshop_llm_reasoning', 'withdrawn', 'LLM_Reasoning'),
-}
 
 
 def download(url, limit):
@@ -77,20 +63,6 @@ def normalize(row, family, year, provenance, include_workshops=False):
     """Return a normalized assertion or a reason requiring separate review."""
     raw_status = row.get('status', '')
     status = negative_status(raw_status)
-    workshop = ICLR_WORKSHOP_LABELS.get(' '.join(raw_status.casefold().split()))
-    workshop_metadata = {}
-    if workshop:
-        expected_year, expected_track, status, group_name = workshop
-        if family != 'iclr' or year != expected_year:
-            raise ValueError('Workshop status label conflicts with source conference year')
-        if row.get('track') != expected_track:
-            return None, 'conflicting_workshop_track'
-        if not include_workshops:
-            return None, 'separate_workshop'
-        workshop_metadata = {
-            'workshop_venueid': f'ICLR.cc/{year}/Workshop/{group_name}',
-            'workshop_identity_evidence_url': 'https://openreview.net/venue?id=ICLR.cc',
-        }
     if status is None:
         if re.search(r'\bretract\w*\b', raw_status, re.I):
             return None, 'retracted'
@@ -118,9 +90,7 @@ def normalize(row, family, year, provenance, include_workshops=False):
         return None, 'unmapped_venue'
     title = row.get('title')
     abstract = row.get('abstract')
-    if title is None or (isinstance(title, str) and not title.strip()):
-        return None, 'missing_title'
-    if not isinstance(title, str):
+    if not isinstance(title, str) or not title.strip():
         raise ValueError('Public submission lacks a title')
     if abstract is not None and not isinstance(abstract, str):
         raise ValueError('Abstract must be text or null')
@@ -146,8 +116,7 @@ def normalize(row, family, year, provenance, include_workshops=False):
             'venue': venue, 'year': year, 'track': normalized_track,
             'status': status, 'raw_decision': raw_status, 'evidence_url': evidence,
             'observed_at': AS_OF, 'tags': tags,
-            'metadata': {**provenance, **workshop_metadata,
-                         'raw_track': track, 'raw_status': raw_status,
+            'metadata': {**provenance, 'raw_track': track, 'raw_status': raw_status,
                          'evidence_kind': 'curated_public_status_assertion',
                          'original_decision_signature_verified': False,
                          'source_snapshot_upper_bound': AS_OF}}, None
@@ -246,7 +215,7 @@ def _collect_files(out, selected, include_workshops, fetcher):
             name = f'public-openreview-{family}-{year}-{track}.jsonl'
             path = stage / name
             path.write_text(''.join(json.dumps(row, ensure_ascii=False) + '\n' for row in records))
-            manifest = {'source': SOURCE, 'parser_version': 3, 'venue': records[0]['venue'],
+            manifest = {'source': SOURCE, 'parser_version': 1, 'venue': records[0]['venue'],
                         'year': year, 'track': track, 'count': len(records), 'status': 'partial',
                         'file': str(out / name), 'file_sha256': hashlib.sha256(path.read_bytes()).hexdigest(),
                         'list_url': REPOSITORY + '/tree/' + REVISION, 'source_revision': REVISION,

tests/test_coverage_report.py

@@ -7,6 +7,34 @@ from paper_atlas.audits.coverage import report
 from paper_atlas.db.store import Store
 
 
+def add_coverage_evidence(store):
+    paper, record = store.record('official_metadata', 'known', 'https://example.org/known',
+        'Known paper with two source observations', 'Complete canonical abstract', ['Jane Doe', 'John Smith'],
+        {'openreview': 'coverageForum', 'arxiv': '2401.98765'})
+    edition = store.decision(paper, record, 'COLING+LREC', 2024, 'main', 'known',
+        'https://example.org/accept', 'accepted', 'Accept', aliases=['COLING', 'LREC'],
+        identities=[('openreview', 'coverageForum')], observed_at='2025-05-01T00:00:00Z')
+    same, mirror = store.record('bibliography', 'mirror', 'https://example.org/mirror',
+        'Known paper with two source observations', None, ['Jane Doe', 'John Smith'], {'openreview': 'coverageForum'})
+    store.decision(same, mirror, 'COLING+LREC', 2024, 'main', 'mirror', 'https://example.org/older',
+        'rejected', 'Reject', identities=[('openreview', 'coverageForum')],
+        observed_at='2025-05-01T00:00:00Z', decision_date='2025-04-30T00:00:00Z')
+    unknown, unknown_record = store.record('unknown_catalogue', 'unknown', 'https://example.org/unknown',
+        'Indexed work without a public decision', None, [])
+    store.decision(unknown, unknown_record, 'COLING+LREC', 2024, 'main', 'unknown',
+        'https://example.org/unknown', None, '', identities=[('openreview', 'unknownForum')])
+    same, later = store.record('official_metadata', 'later', 'https://example.org/later',
+        'Known paper with two source observations', 'Complete canonical abstract', ['Jane Doe', 'John Smith'],
+        {'openreview': 'coverageNext', 'arxiv': '2401.98765'})
+    store.decision(same, later, 'COLING+LREC', 2025, 'findings', 'later', 'https://example.org/later',
+        'accepted', 'Accept', identities=[('openreview', 'coverageNext')], observed_at='2025-06-01T00:00:00Z')
+    source = store.source('official_metadata', 'https://example.org')
+    for url, status in [('https://example.org/list', 'complete'), ('https://example.org/identity-refresh', 'partial')]:
+        store.conn.execute('''INSERT INTO edition_sources(edition_id,source_id,list_url,status,expected_count,imported_count)
+            VALUES(?,?,?,?,?,?)''', (edition, source, url, status, 1, 1))
+    store.conn.commit()
+
+
 class CoverageTests(unittest.TestCase):
     def setUp(self):
         self.tmp=tempfile.TemporaryDirectory();self.addCleanup(self.tmp.cleanup)
@@ -31,5 +59,46 @@ class CoverageTests(unittest.TestCase):
         self.store.conn.commit();coling=report(self.db,self.registry)['venues'][0]
         self.assertEqual(coling['imported_submissions'],2);self.assertEqual(coling['distinct_papers'],1)
 
+    def test_observations_attempts_and_missing_source_abstracts_stay_separate(self):
+        add_coverage_evidence(self.store)
+        result = report(self.db, self.registry)
+        self.assertEqual(result['engine'], 'sqlite')
+        self.assertEqual(result['database'], str(self.db))
+        coling = result['venues'][0]
+        self.assertEqual((coling['imported_submissions'], coling['distinct_papers']), (3, 2))
+        edition, findings = coling['editions']
+        self.assertEqual((edition['submissions'], edition['papers'], edition['papers_without_abstract_text']), (2, 2, 1))
+        self.assertEqual(edition['latest_observed_status_counts']['accepted'], 1)
+        self.assertEqual(edition['latest_observed_status_counts']['rejected'], 0)
+        self.assertEqual(edition['latest_observed_status_counts']['unknown'], 1)
+        self.assertEqual(edition['decision_observation_counts']['accepted'], 1)
+        self.assertEqual(edition['decision_observation_counts']['rejected'], 1)
+        self.assertEqual(edition['decision_observation_counts']['unknown'], 0)
+        counts = {row['source']: row for row in edition['source_record_counts']}
+        self.assertEqual(counts['bibliography']['source_records_without_abstract_text'], 1)
+        self.assertEqual(counts['official_metadata']['source_records'], 1)
+        self.assertEqual(len(edition['sources']), 2)
+        self.assertEqual(findings['track'], 'findings')
+        self.assertEqual(result['imported_venues_outside_registry'], [])
+
+    def test_missing_sqlite_audit_does_not_create_a_database(self):
+        path = self.root / 'absent.sqlite'
+        with self.assertRaisesRegex(ValueError, 'Existing database'):
+            report(path, self.registry)
+        self.assertFalse(path.exists())
+
+    def test_empty_editions_remain_visible_after_shared_aggregation(self):
+        self.store.conn.execute("INSERT INTO venues(acronym,name,field,venue_type) VALUES('AAAI','AAAI','general_ai','conference')")
+        venue = self.store.conn.execute("SELECT id FROM venues WHERE acronym='AAAI'").fetchone()[0]
+        self.store.conn.execute("INSERT INTO editions(venue_id,year,track) VALUES(?,2026,'main')", (venue,))
+        self.store.conn.commit()
+        aaai = report(self.db, self.registry)['venues'][1]
+        self.assertEqual(aaai['history_status'], 'no_imported_records')
+        self.assertEqual(len(aaai['editions']), 1)
+        edition = aaai['editions'][0]
+        self.assertEqual((edition['submissions'], edition['papers'], edition['papers_without_abstract_text']), (0, 0, 0))
+        self.assertEqual(sum(edition['latest_observed_status_counts'].values()), 0)
+        self.assertEqual(edition['source_record_counts'], [])
+
 
 if __name__=='__main__':unittest.main()

tests/test_histories.py

@@ -0,0 +1,114 @@
+from pathlib import Path
+import tempfile
+import unittest
+
+from paper_atlas.audits.histories import report
+from paper_atlas.db.store import Store
+
+
+def add_history_fixture(store):
+    def event(work, forum, venue, year, track, status, observed):
+        paper, record = store.record('history_fixture', work + '/' + forum,
+            'https://example.org/' + forum, work + ' title\x00 preserved', 'Abstract', ['Jane Doe'],
+            {'arxiv': work})
+        store.decision(paper, record, venue, year, track, forum,
+            'https://example.org/' + forum + '/' + str(status), status, status or '',
+            identities=[('openreview', forum)], observed_at=observed)
+        return paper
+
+    # The negative observation was collected later: observation order is not chronology.
+    event('2501.00001', 'HistoryA01', 'ICLR', 2024, 'main', 'rejected', '2026-02-01T00:00:00Z')
+    event('2501.00001', 'HistoryA02', 'ICML', 2025, 'main', 'accepted', '2026-01-01T00:00:00Z')
+    event('2501.00002', 'HistoryB01', 'ICLR', 2024, 'main', 'rejected', '2026-01-01T00:00:00Z')
+    event('2501.00002', 'HistoryB01', 'ICLR', 2024, 'main', 'accepted', '2026-02-01T00:00:00Z')
+    event('2501.00003', 'HistoryC01', 'ICLR', 2024, 'main', 'desk_rejected', '2026-01-01T00:00:00Z')
+    event('2501.00003', 'HistoryC02', 'ICLR', 2024, 'workshop_example', 'accepted', '2026-02-01T00:00:00Z')
+    event('2501.00004', 'HistoryD01', 'ICML', 2025, 'main', 'accepted', '2026-01-01T00:00:00Z')
+    event('2501.00004', 'HistoryD01', 'ICML', 2025, 'main', 'withdrawn', '2026-02-01T00:00:00Z')
+    event('2501.00004', 'HistoryD02', 'ICLR', 2024, 'main', 'rejected', '2026-01-01T00:00:00Z')
+    event('2501.00005', 'HistoryE01', 'COLM', 2026, 'main', None, '2026-01-01T00:00:00Z')
+    store.conn.commit()
+
+
+class HistoryTests(unittest.TestCase):
+    def setUp(self):
+        self.tmp = tempfile.TemporaryDirectory()
+        self.addCleanup(self.tmp.cleanup)
+        self.db = Path(self.tmp.name) / 'sample.sqlite'
+        self.store = Store(self.db)
+        self.addCleanup(self.store.conn.close)
+
+    def test_distinct_attempts_venues_and_latest_acceptances_are_not_observation_counts(self):
+        add_history_fixture(self.store)
+        result = report(self.db)
+        self.assertEqual(result['counts'], {
+            'papers_with_submissions': 5, 'papers_with_multiple_submissions': 3,
+            'papers_with_historical_negative': 4, 'papers_with_current_known_acceptance': 3,
+            'papers_with_negative_and_acceptance_in_distinct_attempts': 2,
+            'papers_with_negative_and_acceptance_in_distinct_venues': 1,
+            'submissions': 8, 'decision_observations': 9,
+            'attempts_with_historical_negative': 5,
+            'attempts_with_historical_negative_and_acceptance': 2,
+        })
+        self.assertEqual(result['latest_attempt_status_counts'],
+                         {'accepted': 3, 'desk_rejected': 1, 'rejected': 2, 'unknown': 1, 'withdrawn': 1})
+        self.assertEqual(len(result['examples']), 2)
+        first = result['examples'][0]
+        self.assertIn('\x00', first['title'])
+        self.assertEqual({r['venue'] for r in first['attempts']}, {'ICLR', 'ICML'})
+        self.assertTrue(all(r['decisions'][0]['sources'] for r in first['attempts']))
+        self.assertTrue(all(r['decisions'][0]['decision_date'] is None for r in first['attempts']))
+        self.assertIn('Not inferred', first['chronology'])
+        self.assertEqual(result['status'], 'observed_history_audit_not_final_dedup')
+        self.assertTrue(all(v == 0 for v in result['evidence_relationship_checks'].values()))
+
+    def test_missing_and_mislinked_source_evidence_is_reported_without_repair(self):
+        add_history_fixture(self.store)
+        self.store.conn.execute('DELETE FROM decision_sources WHERE decision_id=1')
+        self.store.conn.execute('DELETE FROM submission_sources WHERE submission_id=1')
+        # Record3 belongs to paper2; move it to paper1 to create a real mismatch.
+        self.store.conn.execute('UPDATE source_records SET paper_id=1 WHERE id=3')
+        self.store.conn.commit()
+        result = report(self.db, examples=0)
+        checks = result['evidence_relationship_checks']
+        self.assertEqual(checks['decisions_without_source_links'], 1)
+        self.assertEqual(checks['submissions_without_source_links'], 1)
+        self.assertEqual(checks['submission_source_paper_mismatches'], 1)
+        self.assertEqual(checks['decision_source_paper_mismatches'], 2)
+        self.assertEqual(self.store.conn.execute('SELECT paper_id FROM source_records WHERE id=3').fetchone()[0], 1)
+
+    def test_example_limit_does_not_change_global_counts(self):
+        add_history_fixture(self.store)
+        none, one = report(self.db, examples=0), report(self.db, examples=1)
+        self.assertEqual(none['counts'], one['counts'])
+        self.assertEqual(none['examples'], [])
+        self.assertEqual(len(one['examples']), 1)
+
+    def test_empty_snapshot_and_invalid_limit_do_not_create_missing_database(self):
+        empty = report(self.db)
+        self.assertTrue(all(v == 0 for v in empty['counts'].values()))
+        self.assertEqual(empty['latest_attempt_status_counts'], {})
+        absent = self.db.parent / 'absent.sqlite'
+        with self.assertRaisesRegex(ValueError, 'Examples'):
+            report(absent, examples=-1)
+        with self.assertRaisesRegex(ValueError, 'Existing database'):
+            report(absent)
+        self.assertFalse(absent.exists())
+
+    def test_truncated_example_keeps_the_cross_venue_acceptance_visible(self):
+        paper, record = self.store.record('fixture', 'many', 'https://example.org/many',
+                                          'Many attempts', 'Abstract', [])
+        for index in range(26):
+            accepted = index == 25
+            forum = f'BoundedForum{index}'
+            self.store.decision(paper, record, 'ICML' if accepted else 'ICLR', 2025, 'main', forum,
+                                'https://example.org/' + forum, 'accepted' if accepted else 'rejected',
+                                'Accept' if accepted else 'Reject', identities=[('openreview', forum)])
+        self.store.conn.commit()
+        result = report(self.db)
+        example, = result['examples']
+        self.assertEqual((example['attempt_count'], len(example['attempts'])), (26, 20))
+        self.assertTrue(example['attempts_truncated'])
+        self.assertEqual({row['latest_observed_status'] for row in example['attempts']},
+                         {'accepted', 'rejected'})
+        self.assertEqual(result['counts']['submissions'], 26)

tests/test_postgres.py

@@ -175,6 +175,22 @@ class PostgreSQLIntegrationTests(unittest.TestCase):
         finally:
             reader.close()
 
+    def test_history_audit_matches_sqlite_for_actual_postgresql_snapshot(self):
+        from paper_atlas.audits.histories import report
+        from tests.test_histories import add_history_fixture
+        store = Store(self.db)
+        try:
+            add_history_fixture(store)
+        finally:
+            store.conn.close()
+        expected = report(self.db)
+        self.assertEqual(self.migrate()['status'], 'complete')
+        actual = report(pg_service=self.service, pg_schema=self.target)
+        self.assertEqual(actual['engine'], 'postgresql')
+        self.assertEqual(actual['schema'], self.target)
+        for key in ('counts', 'latest_attempt_status_counts', 'evidence_relationship_checks', 'examples'):
+            self.assertEqual(actual[key], expected[key], key)
+
     def interrupt_after_first_table(self):
         calls = 0
         def interrupt(*args):

tests/test_postgres_ingestion.py

@@ -243,6 +243,50 @@ class PostgreSQLIngestionTests(unittest.TestCase):
         finally:
             reader.close()
 
+    def test_coverage_matches_sqlite_with_history_aliases_and_missing_source_abstracts(self):
+        from paper_atlas.audits.coverage import report
+        from tests.test_coverage_report import add_coverage_evidence
+        registry = self.root / 'venues.json'
+        registry.write_text(json.dumps({'scope_note': 'Candidate scope', 'venues': [
+            {'acronym': 'coling', 'field': 'language'}, {'acronym': 'AAAI', 'field': 'general_ai'}]}))
+        sample = Store(self.root / 'seed.sqlite')
+        try:
+            add_coverage_evidence(sample)
+        finally:
+            sample.conn.close()
+        store = self.store()
+        try:
+            add_coverage_evidence(store)
+        finally:
+            store.conn.close()
+        sqlite_report = report(self.root / 'seed.sqlite', registry)
+        observed = []
+        def reader(*args, **kwargs):
+            connection = open_read_connection(*args, **kwargs)
+            observed.append((connection.execute('SHOW transaction_read_only').fetchone()[0],
+                             connection.execute('SHOW transaction_isolation').fetchone()[0]))
+            return connection
+        with patch('paper_atlas.audits.coverage.open_read_connection', side_effect=reader):
+            actual = report(None, registry, pg_service=self.service, pg_schema=self.schema)
+        self.assertEqual(observed, [('on', 'repeatable read')])
+        self.assertEqual(actual['engine'], 'postgresql')
+        self.assertEqual(actual['schema'], self.schema)
+        self.assertEqual(actual['pg_service'], self.service)
+        def content(result):
+            values = json.loads(json.dumps(result['venues']))
+            for venue in values:
+                for edition in venue['editions']:
+                    for key in ('id', 'venue_id'):
+                        edition.pop(key)
+            return values
+        self.assertEqual(content(actual), content(sqlite_report))
+        self.assertEqual([r['acronym'] for r in actual['imported_venues_outside_registry']], ['ICLR'])
+        edition = actual['venues'][0]['editions'][0]
+        self.assertEqual(edition['latest_observed_status_counts']['accepted'], 1)
+        self.assertEqual(edition['latest_observed_status_counts']['unknown'], 1)
+        self.assertEqual(edition['decision_observation_counts']['rejected'], 1)
+        self.assertEqual(edition['papers_without_abstract_text'], 1)
+
     def test_backfill_merges_evidence_without_reimporting_or_losing_history(self):
         files = []
         for source, external, title in [('official', 'legacy', 'First bibliographic name'), ('openreview_snapshot', 'forum', 'Renamed submission')]:

tests/test_public_openreview.py

@@ -45,12 +45,10 @@ class PublicOpenReviewTests(unittest.TestCase):
 
     def test_only_explicit_decisions_and_verified_scope_become_events(self):
         for value in ('Pending', 'Active', 'Submitted to NeurIPS 2025', 'TBD', '',
-                      'Low score', 'ICML 2025 Retracted by Authors', 'Do not reject',
-                      'Workshop', 'Poster Workshop', 'Oral Workshop'):
+                      'Low score', 'ICML 2025 Retracted by Authors', 'Do not reject'):
             self.assertIsNone(negative_status(value), value)
         self.assertEqual(negative_status('Reject'), 'rejected')
         self.assertEqual(negative_status('Desk rejected'), 'desk_rejected')
-        self.assertEqual(negative_status('Withdraw'), 'withdrawn')
         self.assertEqual(negative_status('NeurIPS 2023 Conference Withdrawn Submission'), 'withdrawn')
         for raw, expected in [('main', 'main'), ('Position', 'position'),
                               ('Datasets & Benchmarks', 'datasets_benchmarks')]:
@@ -100,121 +98,6 @@ class PublicOpenReviewTests(unittest.TestCase):
                 self.assertEqual(list(Path(tmp).glob('*.jsonl')), [])
                 self.assertFalse((Path(tmp) / 'public-openreview-inventory.json').exists())
 
-    def test_missing_title_is_held_without_losing_valid_negative_records(self):
-        with tempfile.TemporaryDirectory() as tmp:
-            result = self.collect(tmp, {'venues/iclr/iclr2017.jsonl': [
-                self.row(id='SJUdkecgx', title=''),
-                self.row(id='NoTitle001', title=None),
-                self.row(id='Whitespace', title='  '),
-                self.row(id='ValidReject'),
-                self.row(id='Withdraw01', status='Withdraw'),
-                self.row(id='Workshop01', status='Workshop'),
-            ]}, families=['iclr'], years=[2017])
-            self.assertEqual(result['input_rows'], 6)
-            self.assertEqual(result['normalized_records'], 2)
-            self.assertEqual(result['source_files'][0]['excluded_or_held'],
-                             {'missing_title': 3, 'not_explicit_negative': 1})
-            held_path = Path(result['held_file'])
-            self.assertEqual(hashlib.sha256(held_path.read_bytes()).hexdigest(),
-                             result['held_file_sha256'])
-            held = [json.loads(line) for line in held_path.read_text().splitlines()]
-            self.assertEqual([row['id'] for row in held], ['SJUdkecgx', 'NoTitle001', 'Whitespace'])
-            self.assertTrue(all(row['reason'] == 'missing_title' and row['status'] == 'Reject'
-                                for row in held))
-            self.assertEqual([row['provenance']['source_line'] for row in held], [1, 2, 3])
-            self.assertNotIn('private@example.org', held_path.read_text())
-            self.assertNotIn('private review', held_path.read_text())
-            manifest = result['files'][0]
-            self.assertEqual((manifest['parser_version'], manifest['status']), (3, 'partial'))
-            rows = [json.loads(line) for line in Path(manifest['file']).read_text().splitlines()]
-            self.assertEqual([(row['identifiers']['openreview'], row['status']) for row in rows],
-                             [('ValidReject', 'rejected'), ('Withdraw01', 'withdrawn')])
-            for row in held:
-                self.assertEqual(row['provenance']['source_revision'], REVISION)
-
-    def test_nontext_title_still_fails_collection(self):
-        with tempfile.TemporaryDirectory() as tmp:
-            with self.assertRaisesRegex(ValueError, 'lacks a title'):
-                self.collect(tmp, {'venues/iclr/iclr2017.jsonl': [self.row(title=['invalid'])]},
-                             families=['iclr'], years=[2017])
-            self.assertFalse((Path(tmp) / 'public-openreview-inventory.json').exists())
-            self.assertEqual(list(Path(tmp).glob('*.jsonl')), [])
-
-    def workshop_cases(self):
-        return [
-            (2024, 'workshop_me_fomo', 'ICLR 2024 Workshop MEFoMo Withdrawn Submission', 'withdrawn', 'ME-FoMo'),
-            (2024, 'workshop_genai4dm', 'ICLR 2024 Workshop GenAI4DM Withdrawn Submission', 'withdrawn', 'GenAI4DM'),
-            (2025, 'workshop_agenticai', 'ICLR 2025 Workshop AgenticAI Reject', 'rejected', 'AgenticAI'),
-            (2025, 'workshop_lmrl', 'ICLR 2025 Workshop LMRL Withdrawn Submission', 'withdrawn', 'LMRL'),
-            (2026, 'workshop_llm_reasoning', 'ICLR 2026 Workshop LLM Reasoning Withdrawn Submission', 'withdrawn', 'LLM_Reasoning'),
-        ]
-
-    def test_iclr_workshop_labels_require_opt_in_and_retain_identity_evidence(self):
-        for year, track, label, status, group in self.workshop_cases():
-            with self.subTest(label=label):
-                raw = self.row(track=track, status=label)
-                self.assertEqual(normalize(raw, 'iclr', year, self.provenance()),
-                                 (None, 'separate_workshop'))
-                row, reason = normalize(raw, 'iclr', year, self.provenance(), True)
-                self.assertIsNone(reason)
-                self.assertEqual((row['venue'], row['year'], row['track'], row['status']),
-                                 ('ICLR', year, track, status))
-                self.assertEqual(row['raw_decision'], label)
-                self.assertEqual(row['metadata']['workshop_venueid'],
-                                 f'ICLR.cc/{year}/Workshop/{group}')
-                self.assertFalse(row['metadata']['original_decision_signature_verified'])
-                self.assertEqual(row['evidence_url'], 'https://example.org/pinned.jsonl#L7')
-                self.assertNotIn('private review', json.dumps(row))
-
-    def test_workshop_label_scope_conflicts_never_become_main_events(self):
-        for year, track, label, _, _ in self.workshop_cases():
-            for enabled in (False, True):
-                for wrong_track in ('main', 'workshop_other'):
-                    self.assertEqual(normalize(self.row(track=wrong_track, status=label),
-                                               'iclr', year, self.provenance(), enabled),
-                                     (None, 'conflicting_workshop_track'))
-                for family, wrong_year in [('icml', year), ('iclr', year + 1)]:
-                    with self.assertRaisesRegex(ValueError, 'conference year'):
-                        normalize(self.row(track=track, status=label), family,
-                                  wrong_year, self.provenance(), enabled)
-        row, reason = normalize(self.row(track='workshop_unknown',
-                                        status='ICLR 2026 Workshop Unknown Reject'),
-                                'iclr', 2026, self.provenance(), True)
-        self.assertIsNone(row)
-        self.assertEqual(reason, 'unmapped_negative_label')
-
-    def test_workshop_collection_and_import_keep_five_tracks_separate_from_main(self):
-        files = {}
-        for index, (year, track, label, _, _) in enumerate(self.workshop_cases()):
-            files.setdefault(f'venues/iclr/iclr{year}.jsonl', []).append(
-                self.row(id=f'Workshop{index:02}', track=track, status=label))
-        files['venues/iclr/iclr2024.jsonl'].append(self.row())
-        with tempfile.TemporaryDirectory() as tmp:
-            root = Path(tmp)
-            default = self.collect(root / 'default', files, families=['iclr'])
-            self.assertEqual(default['normalized_records'], 1)
-            held = [json.loads(line) for line in Path(default['held_file']).read_bytes().split(b'\n') if line]
-            self.assertEqual(len(held), 5)
-            self.assertTrue(all(row['reason'] == 'separate_workshop' for row in held))
-            selected = self.collect(root / 'selected', files, families=['iclr'], include_workshops=True)
-            self.assertEqual(selected['normalized_records'], 6)
-            self.assertEqual(len(selected['files']), 6)
-            self.assertEqual(Path(selected['held_file']).read_bytes(), b'')
-            db = root / 'db.sqlite'
-            for manifest in selected['files']:
-                self.assertEqual(manifest['parser_version'], 3)
-                ingest_file(db, manifest['file'])
-            store = Store(db)
-            try:
-                scopes = set(store.conn.execute('SELECT e.year,e.track,d.status FROM submissions s '
-                                               'JOIN editions e ON e.id=s.edition_id '
-                                               'JOIN decisions d ON d.submission_id=s.id'))
-                expected = {(year, track, status) for year, track, _, status, _ in self.workshop_cases()}
-                self.assertEqual(scopes, expected | {(2024, 'main', 'rejected')})
-                self.assertEqual(store.conn.execute('PRAGMA foreign_key_check').fetchall(), [])
-            finally:
-                store.conn.close()
-
     def test_mirror_reimport_preserves_one_attempt_and_later_acceptance_elsewhere(self):
         with tempfile.TemporaryDirectory() as tmp:
             root = Path(tmp)
@@ -269,6 +152,35 @@ class PublicOpenReviewTests(unittest.TestCase):
             finally:
                 store.conn.close()
 
+    def test_unicode_separators_in_abstract_survive_audit_and_import(self):
+        for separator in ('\u0085', '\u2028', '\u2029'):
+            with self.subTest(separator=repr(separator)), tempfile.TemporaryDirectory() as tmp:
+                root = Path(tmp)
+                abstract = 'First paragraph' + separator + 'Second paragraph'
+                result = self.collect(root / 'source', {
+                    'venues/icml/icml2025.jsonl': [self.row(abstract=abstract)]})
+                path = Path(result['files'][0]['file'])
+                data = path.read_bytes()
+                self.assertIn(separator.encode(), data)
+                self.assertEqual(data.count(b'\n'), 1)
+                db = root / 'db.sqlite'
+                store = Store(db)
+                store.conn.close()
+                inventory = root / 'source/public-openreview-inventory.json'
+                before = audit(db, inventory, root / 'before.json')
+                self.assertEqual(before['status'], 'ready_for_serial_import')
+                self.assertEqual(before['records_checked'], 1)
+                ingest_file(db, path)
+                after = audit(db, inventory, root / 'after.json')
+                self.assertEqual(after['status'], 'ready_for_serial_import')
+                self.assertEqual(after['existing_same_edition_attempts'], 1)
+                store = Store(db)
+                try:
+                    self.assertEqual(store.conn.execute(
+                        'SELECT abstract FROM source_records').fetchone()[0], abstract)
+                finally:
+                    store.conn.close()
+
 
 if __name__ == '__main__':
     unittest.main()