fix: 第27轮搜索审计修复 — 跨域日期污染/部分日期范围/标记组路径等7项

backend/app/services/pubmed_query_parser.py

@@ -464,6 +464,13 @@ class PubmedQueryParser:
             if t.field in _DATE_RANGE_FIELDS:
                 result._top_level_date_fields.add(t.field)
 
+        # R27: also track top-level date range markers (excluded from _ungrouped by _is_range_end filter)
+        for t in terms:
+            if getattr(t, '_is_range_end', False) and t.group_id < 0:
+                _tag = t.field.replace("__RANGE_", "").replace("__", "")
+                if _tag in _DATE_RANGE_FIELDS:
+                    result._top_level_date_fields.add(_tag)
+
         # P2-1: Handle unconsumed tokens (e.g., orphan text after RPAREN)
         if self.pos < len(self.tokens) - 1:
             for t in self.tokens[self.pos:-1]:  # exclude EOF token
@@ -935,8 +942,8 @@ class PubmedQueryParser:
             return []
         token = self.advance()
         text = token.value.strip('"') if token.type == TokenType.QUOTED else token.value
-        # R21: skip empty quoted text ""[TI]
-        if not text:
+        # R21: skip empty quoted text ""[TI]; R27: also skip whitespace-only
+        if not text or not text.strip():
             return []
         is_exact = (token.type == TokenType.QUOTED)
         field = None
@@ -1013,6 +1020,11 @@ class PubmedQueryParser:
                 except IndexError:
                     pass
             # P12: validate date values — non-numeric garbage falls back to plain text
+            # R27: expand partial dates YYYY-MM in ranges before validation
+            if _PARTIAL_DATE_RE.match(start_val):
+                start_val, _ = _expand_partial_date(start_val)
+            if _PARTIAL_DATE_RE.match(end_val):
+                _, end_val = _expand_partial_date(end_val)
             _valid_date = lambda s: _validate_date_str(s)
             if not _valid_date(start_val) or not _valid_date(end_val):
                 txt = f"{start_val}:{end_val}[{field}]"
@@ -1021,71 +1033,74 @@ class PubmedQueryParser:
             _start_is_year = start_val.isdigit() and len(start_val) == 4
             _end_is_year = end_val.isdigit() and len(end_val) == 4
             # Year-only range (e.g., 2024:2026[EDAT])
-            if _start_is_year and _end_is_year:
-                if negated:
-                    # R26: store in _neg_single_dates instead of main fields
-                    result._neg_single_dates.setdefault(field, []).append(
-                        (f"{start_val}-01-01", f"{end_val}-12-31")
-                    )
-                else:
-                    try:
-                        if yr_from_attr:
-                            curr_f = getattr(result, yr_from_attr)
-                            new_f = int(start_val)
-                            setattr(result, yr_from_attr, max(curr_f, new_f) if curr_f is not None else new_f)
-                            curr_t = getattr(result, yr_to_attr)
-                            new_t = int(end_val)
-                            setattr(result, yr_to_attr, min(curr_t, new_t) if curr_t is not None else new_t)
-                        else:
-                            # For non-DP date fields: convert year to full date for consistency
-                            curr_f = getattr(result, date_attr)
-                            new_f = f"{start_val}-01-01"
-                            setattr(result, date_attr, max(curr_f, new_f) if curr_f is not None else new_f)
-                            curr_t = getattr(result, date_attr_to)
-                            new_t = f"{end_val}-12-31"
-                            setattr(result, date_attr_to, min(curr_t, new_t) if curr_t is not None else new_t)
-                    except (ValueError, TypeError):
-                        pass
-            elif _start_is_year and not _end_is_year:
-                # Mixed: start is year, end is full date (e.g., 2024:2024-12-01[EDAT])
-                if negated:
-                    result._neg_single_dates.setdefault(field, []).append(
-                        (f"{start_val}-01-01", end_val)
-                    )
-                else:
-                    if yr_from_attr:
+            # R27: only set global attributes for top-level ranges (depth == 0)
+            # Ranges inside groups are handled via markers in the engine's group path.
+            if self._depth == 0:
+                if _start_is_year and _end_is_year:
+                    if negated:
+                        # R26: store in _neg_single_dates instead of main fields
+                        result._neg_single_dates.setdefault(field, []).append(
+                            (f"{start_val}-01-01", f"{end_val}-12-31")
+                        )
+                    else:
                         try:
-                            curr = getattr(result, yr_from_attr)
-                            v = int(start_val)
-                            setattr(result, yr_from_attr, max(curr, v) if curr is not None else v)
+                            if yr_from_attr:
+                                curr_f = getattr(result, yr_from_attr)
+                                new_f = int(start_val)
+                                setattr(result, yr_from_attr, max(curr_f, new_f) if curr_f is not None else new_f)
+                                curr_t = getattr(result, yr_to_attr)
+                                new_t = int(end_val)
+                                setattr(result, yr_to_attr, min(curr_t, new_t) if curr_t is not None else new_t)
+                            else:
+                                # For non-DP date fields: convert year to full date for consistency
+                                curr_f = getattr(result, date_attr)
+                                new_f = f"{start_val}-01-01"
+                                setattr(result, date_attr, max(curr_f, new_f) if curr_f is not None else new_f)
+                                curr_t = getattr(result, date_attr_to)
+                                new_t = f"{end_val}-12-31"
+                                setattr(result, date_attr_to, min(curr_t, new_t) if curr_t is not None else new_t)
                         except (ValueError, TypeError):
                             pass
-                    curr_f = getattr(result, date_attr)
-                    new_f = f"{start_val}-01-01"
-                    setattr(result, date_attr, max(curr_f, new_f) if curr_f is not None else new_f)
-                    curr_t = getattr(result, date_attr_to)
-                    setattr(result, date_attr_to, min(curr_t, end_val) if curr_t is not None else end_val)
-            elif not _start_is_year and _end_is_year:
-                # Mixed: start is full date, end is year (e.g., 2024-01-01:2026[EDAT])
-                if negated:
-                    result._neg_single_dates.setdefault(field, []).append(
-                        (start_val, f"{end_val}-12-31")
-                    )
+                elif _start_is_year and not _end_is_year:
+                    # Mixed: start is year, end is full date (e.g., 2024:2024-12-01[EDAT])
+                    if negated:
+                        result._neg_single_dates.setdefault(field, []).append(
+                            (f"{start_val}-01-01", end_val)
+                        )
+                    else:
+                        if yr_from_attr:
+                            try:
+                                curr = getattr(result, yr_from_attr)
+                                v = int(start_val)
+                                setattr(result, yr_from_attr, max(curr, v) if curr is not None else v)
+                            except (ValueError, TypeError):
+                                pass
+                        curr_f = getattr(result, date_attr)
+                        new_f = f"{start_val}-01-01"
+                        setattr(result, date_attr, max(curr_f, new_f) if curr_f is not None else new_f)
+                        curr_t = getattr(result, date_attr_to)
+                        setattr(result, date_attr_to, min(curr_t, end_val) if curr_t is not None else end_val)
+                elif not _start_is_year and _end_is_year:
+                    # Mixed: start is full date, end is year (e.g., 2024-01-01:2026[EDAT])
+                    if negated:
+                        result._neg_single_dates.setdefault(field, []).append(
+                            (start_val, f"{end_val}-12-31")
+                        )
+                    else:
+                        curr_f = getattr(result, date_attr)
+                        setattr(result, date_attr, max(curr_f, start_val) if curr_f is not None else start_val)
+                        curr_t = getattr(result, date_attr_to)
+                        v = f"{end_val}-12-31"
+                        setattr(result, date_attr_to, min(curr_t, v) if curr_t is not None else v)
                 else:
-                    curr_f = getattr(result, date_attr)
-                    setattr(result, date_attr, max(curr_f, start_val) if curr_f is not None else start_val)
-                    curr_t = getattr(result, date_attr_to)
-                    v = f"{end_val}-12-31"
-                    setattr(result, date_attr_to, min(curr_t, v) if curr_t is not None else v)
-            else:
-                # Full date range (e.g., 2024-01-01:2024-12-31[EDAT])
-                if negated:
-                    result._neg_single_dates.setdefault(field, []).append((start_val, end_val))
-                else:
-                    curr_f = getattr(result, date_attr)
-                    setattr(result, date_attr, max(curr_f, start_val) if curr_f is not None else start_val)
-                    curr_t = getattr(result, date_attr_to)
-                    setattr(result, date_attr_to, min(curr_t, end_val) if curr_t is not None else end_val)
+                    # Full date range (e.g., 2024-01-01:2024-12-31[EDAT])
+                    if negated:
+                        result._neg_single_dates.setdefault(field, []).append((start_val, end_val))
+                    else:
+                        curr_f = getattr(result, date_attr)
+                        setattr(result, date_attr, max(curr_f, start_val) if curr_f is not None else start_val)
+                        curr_t = getattr(result, date_attr_to)
+                        setattr(result, date_attr_to, min(curr_t, end_val) if curr_t is not None else end_val)
             marker = Term(f"{start_val}:{end_val}", field=marker_field, is_not=False)
             marker._is_range_end = True
             result._date_range_markers.append(marker)

backend/app/services/search_engine.py

@@ -1248,15 +1248,25 @@ class AdvancedSearchEngine:
                 # R23-3: track date fields referenced in negated groups (De Morgan fix)
                 _neg_group_date_fields: set[str] = set()
                 for t in group:
-                    # P15: __RANGE_* markers are side-effect-only
+                    # R27: process date markers for ALL groups (follows text term pattern)
                     if getattr(t, '_is_range_end', False):
-                        if _negated:
+                        _marker_cond = AdvancedSearchEngine._build_marker_condition(t)
+                        if _negated and _marker_cond is not None:
                             # R23-3: identify which date field this marker references
                             _mf = t.field or ""
                             if _mf.startswith("__RANGE_") and _mf.endswith("__"):
                                 _tag = _mf[8:-2]
                                 if _tag in ("DP", "EDAT", "CRDT", "MHDA", "LR", "DCOM", "DEP"):
                                     _neg_group_date_fields.add(_tag)
+                            # R27: negated group → all go to g_neg
+                            # is_not=True means inner NOT: not_(cond), outer NOT wraps: not_(not_(cond)) = cond
+                            g_neg.append(not_(_marker_cond) if getattr(t, 'is_not', False) else _marker_cond)
+                        elif _marker_cond is not None:
+                            # Non-negated group: is_not → negated range; otherwise → positive
+                            if getattr(t, 'is_not', False):
+                                g_neg.append(not_(_marker_cond))
+                            else:
+                                g_pos.append(_marker_cond)
                         continue
                     cond = await AdvancedSearchEngine._single_term_condition(db, t)
                     if cond is not None:
@@ -1297,12 +1307,9 @@ class AdvancedSearchEngine:
                         else:
                             g_pos.append(child_combined)
 
-                # R23-3: include date conditions inside negated group scope for correct De Morgan
+                # R27: date conditions are now built directly from markers in the group loop above.
+                # Only track handled fields for top-level skip logic.
                 if _negated and _neg_group_date_fields:
-                    for _tag in _neg_group_date_fields:
-                        _date_cond = AdvancedSearchEngine._build_date_cond_from_pp(pp, _tag)
-                        if _date_cond is not None:
-                            g_neg.append(_date_cond)
                     _handled_neg_group_date_fields.update(_neg_group_date_fields)
 
                 gop = (pp.group_operators[idx]
@@ -1342,11 +1349,15 @@ class AdvancedSearchEngine:
                     if isinstance(c, UnaryExpression) and c.modifier == _sa_ops.inv:
                         return True
                     # R26: NULL-safe NOT: or_(not_(inner), col.is_(None))
+                    # R27: verify second clause is indeed is_(None) to avoid false positives
                     try:
                         if hasattr(c, 'operator') and c.operator is _sa_ops.or_:
                             clauses = list(getattr(c, 'clauses', ()))
                             if len(clauses) >= 2 and isinstance(clauses[0], UnaryExpression) and clauses[0].modifier == _sa_ops.inv:
-                                return True
+                                # Verify remaining clause is an IS NULL check
+                                for _cl in clauses[1:]:
+                                    if hasattr(_cl, 'operator') and _cl.operator in (_sa_ops.is_, _sa_ops.isnot):
+                                        return True
                     except Exception:
                         pass
                     return False
@@ -1406,7 +1417,9 @@ class AdvancedSearchEngine:
                         except ValueError:
                             pass
                     if neg_conds:
-                        conditions.append(not_(and_(*neg_conds) if len(neg_conds) > 1 else neg_conds[0]))
+                        # R27: NULL-safe NOT — rows with NULL pub_date should not be excluded
+                        _range_cond = and_(*neg_conds) if len(neg_conds) > 1 else neg_conds[0]
+                        conditions.append(or_(not_(_range_cond), GlobalLiterature.pub_date.is_(None)))
             elif dp_conds:
                 cond = and_(*dp_conds) if len(dp_conds) > 1 else dp_conds[0]
                 conditions.append(not_(cond) if dp_negated else cond)
@@ -1462,7 +1475,9 @@ class AdvancedSearchEngine:
                         except ValueError:
                             pass
                     if neg_conds:
-                        conditions.append(not_(and_(*neg_conds) if len(neg_conds) > 1 else neg_conds[0]))
+                        # R27: NULL-safe NOT — rows with NULL column should not be excluded
+                        _range_cond = and_(*neg_conds) if len(neg_conds) > 1 else neg_conds[0]
+                        conditions.append(or_(not_(_range_cond), col.is_(None)))
             elif field_conds:
                 cond = and_(*field_conds) if len(field_conds) > 1 else field_conds[0]
                 negated = field_tag in getattr(pp, 'negated_date_ranges', set())
@@ -2054,3 +2069,68 @@ class AdvancedSearchEngine:
             except ValueError:
                 pass
         return and_(*conds) if conds else None
+
+    @staticmethod
+    def _build_marker_condition(marker_term) -> object | None:
+        """Build SQL condition from a __RANGE_* marker term for group-internal date ranges.
+
+        R27: used by the group processing path to build date conditions from markers,
+        replacing the old approach that read conflated global attributes.
+        """
+        _mf = marker_term.field or ""
+        if not _mf.startswith("__RANGE_") or not _mf.endswith("__"):
+            return None
+        _tag = _mf[8:-2]
+        if _tag not in ("DP", "EDAT", "CRDT", "MHDA", "LR", "DCOM", "DEP"):
+            return None
+
+        parts = marker_term.text.split(":", 1)
+        if len(parts) != 2:
+            return None
+        _from_str, _to_str = parts
+
+        from datetime import date as _dt_date
+
+        if _tag == "DP":
+            conds = []
+            if _from_str.isdigit() and len(_from_str) == 4:
+                conds.append(GlobalLiterature.pub_year >= int(_from_str))
+            else:
+                try:
+                    conds.append(GlobalLiterature.pub_date >= _dt_date.fromisoformat(_from_str))
+                except ValueError:
+                    pass
+            if _to_str.isdigit() and len(_to_str) == 4:
+                conds.append(GlobalLiterature.pub_year <= int(_to_str))
+            else:
+                try:
+                    conds.append(GlobalLiterature.pub_date <= _dt_date.fromisoformat(_to_str))
+                except ValueError:
+                    pass
+            return and_(*conds) if conds else None
+
+        _DATE_COL_MAP = {
+            "EDAT": GlobalLiterature.entrez_date,
+            "CRDT": GlobalLiterature.create_date,
+            "MHDA": GlobalLiterature.meshed_date,
+            "LR": GlobalLiterature.pubmed_revised,
+            "DCOM": GlobalLiterature.date_completed,
+            "DEP": GlobalLiterature.pub_date,
+        }
+        col = _DATE_COL_MAP.get(_tag)
+        if col is None:
+            return None
+        conds = []
+        if _from_str.isdigit() and len(_from_str) == 4:
+            _from_str = f"{_from_str}-01-01"
+        try:
+            conds.append(col >= _dt_date.fromisoformat(_from_str))
+        except ValueError:
+            pass
+        if _to_str.isdigit() and len(_to_str) == 4:
+            _to_str = f"{_to_str}-12-31"
+        try:
+            conds.append(col <= _dt_date.fromisoformat(_to_str))
+        except ValueError:
+            pass
+        return and_(*conds) if conds else None

docs/13-搜索修复全记录.md

@@ -2014,3 +2014,84 @@ def _is_negated_cond(c):
 | Bug 3 | LOW | Facet cache 从未写入 p2+（`total=0` 回退） |
 | Bug 4 | LOW | 无效日期字段标签静默降级为全字段搜索 |
 | Bug 5 | LOW | 否定组内冗余日期条件（单个年份 + 全范围） |
+
+---
+
+# 第27轮审计修复 (R27)
+
+## 背景
+
+第 27 轮由 3 个并行审计 agent 覆盖：解析器、引擎、前端/集成。发现 2 个 CRITICAL 引擎-解析器交互 bug、1 个 MEDIUM 解析器 bug，以及若干 LOW 问题。
+
+## 修复清单
+
+### R27-1 (CRITICAL): 日期范围属性跨作用域污染
+
+**文件**：[pubmed_query_parser.py:1024-1088](backend/app/services/pubmed_query_parser.py#L1024) + [search_engine.py:1300-1306](backend/app/services/search_engine.py#L1300)
+
+**根因**：`_parse_range` 对所有非否定日期范围都用 `max/min` intersect 更新全局 `edat_from/edat_to` 属性。当同一日期字段同时出现在顶层和括号组内（如 `2000:2010[EDAT] NOT (cancer AND 2005:2006[EDAT])`），顶层范围被组内范围错误缩小：`edat_from` 从 `2000-01-01` 变成 `2005-01-01`，`edat_to` 从 `2010-12-31` 变成 `2006-12-31`。加上 `_top_level_date_fields` 未追踪 `__RANGE_*` 标记，导致 `_neg_only` 判定跳过整个日期节。
+
+**修复**：
+- `_parse_range` 增加 `if self._depth == 0` 守卫——仅顶层范围更新全局属性，组内范围不再污染
+- `parse()` 循环 `terms` 追踪 `__RANGE_*` 标记的 `_top_level_date_fields`
+- 新增 `_build_marker_condition()` 静态方法——从标记内嵌的 `start_val:end_val` 文本重建 SQL 条件
+- 组循环中处理标记：非否定组 → `g_pos.append`，否定组 → `g_neg.append`
+- 移除 `_build_date_cond_from_pp` 在否定组中的调用（已被标记直接处理取代）
+
+### R27-2 (MEDIUM): 范围语法中的部分日期 YYYY-MM 静默降级
+
+**文件**：[pubmed_query_parser.py:1015-1017](backend/app/services/pubmed_query_parser.py#L1015)
+
+**根因**：`_validate_date_str` 只接受 `YYYY` 或 `YYYY-MM-DD`。`2024-01:2024-06[DP]` 中两个端点无法通过验证，整个范围降级为纯文本。单值部分日期（`2024-01[DP]`）因 R26 的 `_expand_partial_date` 调用路径已正确处理，但 `_parse_range` 缺少相同调用。
+
+**修复**：在 `_parse_range` 的验证前添加 `_PARTIAL_DATE_RE` 匹配 + `_expand_partial_date` 展开：
+
+```python
+if _PARTIAL_DATE_RE.match(start_val):
+    start_val, _ = _expand_partial_date(start_val)
+if _PARTIAL_DATE_RE.match(end_val):
+    _, end_val = _expand_partial_date(end_val)
+```
+
+### R27-3 (MEDIUM): `_is_negated_cond` NULL-safe 模式验证不完整
+
+**文件**：[search_engine.py:1345-1351](backend/app/services/search_engine.py#L1345)
+
+**根因**：NULL-safe NOT 检测只检查 `or_` 的第一个子句是 `not_(...)`，没有验证第二个子句是 `is_(None)`。潜在误报：`or_(not_(X), Y)` 会被错误识别为否定。
+
+**修复**：增加对第二子句的 `is_(None)` / `isnot(None)` 验证。
+
+### R27-4 (LOW): `_neg_single_dates` NOT 条件缺少 NULL 安全
+
+**文件**：[search_engine.py:1421, 1477](backend/app/services/search_engine.py#L1421)
+
+**根因**：`not_(and_(col >= from, col <= to))` — 当 `col` 为 NULL 时，`NULL >= date` → NULL，`NOT(NULL)` → NULL（WHERE 中为假）。NULL 日期行的记录被错误排除。
+
+**修复**：使用 `or_(not_(_range_cond), col.is_(None))`。
+
+### R27-5 (LOW): 空白字符引号文本不跳过
+
+**文件**：[pubmed_query_parser.py:939](backend/app/services/pubmed_query_parser.py#L939)
+
+**根因**：`"  "[TI]` — `text = "  "`（两个空格），`if not text` 为 False，创建搜索条件包含双空格。
+
+**修复**：改为 `if not text or not text.strip(): return []`。
+
+## 前端审计发现（R27 未修复）
+
+| 编号 | 严重度 | 描述 | 文件 |
+|------|--------|------|------|
+| C-1 | CRITICAL | 非 keyset 排序（best_match/relevance）第 2 页起总分页消失 | SearchView.vue:278 |
+| C-2 | CRITICAL | `#N` 引用在 SearchView 中不解析 | SearchView.vue:865 |
+| H-1 | HIGH | offset 分页第 2 页起 total 显示为 0 | SearchView.vue:278,353 |
+| H-2 | HIGH | 错误时 goToPage 不回退页码 | SearchView.vue:369-380 |
+| M-1 | MEDIUM | 空查询时年份直方图与"未找到"同时显示 | SearchView.vue:625 |
+| M-3 | MEDIUM | URL 同时存储 `date_preset` 和绝对日期 | SearchView.vue:512-521 |
+| L-3 | LOW | 空查询无筛选时返回全库年份分布 | search_engine.py:622-641 |
+
+## 验证
+
+- ✅ **110 search tests passed**（parser + search engine + integration）
+- 解析器 36 测试全部通过
+- 5 个新增 R27 正确性检查通过
+- 全量套件中仅外部服务连接失败（httpx.ConnectError），与改动无关