fix: 遗留清理（枚举同源/边界防御/断言补强）

_AI_USAGE_LOG.md

@@ -36,3 +36,4 @@
 | 2026-08-09 04:20 | Agent 实现 | 里程碑2.5 Task10 实现：data_models 扩展 MixedParagraph/MixedSheet 段落容器，并新增 ExcelParseResult.mixed 字段（默认空列表）。TDD：先追加 3 个测试（字段存在性与默认行为）确认 RED（ImportError: cannot import name 'MixedParagraph'），再在 data_models.py 末尾追加两个 dataclass（顶部 from typing 加 Literal、dataclasses 加 field），excel_parser.py 更新导入并为 ExcelParseResult 加 mixed 字段，确认 GREEN（3 passed），全量 pytest 49 passed（46 baseline + 3）。未修改既有类与字段，仅追加。 | src/genesis/data_models.py, src/genesis/parsers/excel_parser.py, tests/test_excel_parser.py | deepseek-v4-flash-free |
 | 2026-08-09 04:06 | Agent 实现 | 里程碑2.5 Task11 实现：MIXED 完整段落装配（分割→逐段最优解析）。excel_parser.py 新增 MIXED 分支：整表先 forward_fill（合并单元格）再按段切片；对每段 classify_sheet，TABLE 段用 find_header_row(seg)（段内相对，<0 回落 0）并抽出段内子 map seg_fmt_map（物理行 pr→段内 pr-s，修复 formatting_map 坐标错位）传给 extract_table（header_row=header_row）；非表格段归 FREE_TEXT（防二次 MIXED 递归），走 extract_text_blocks + build_free_text_table 构造 kind="free_text"。MixedParagraph 段落追加到 MixedSheet 并进 result.mixed，表/自由文本表进 result.tables。原 else 分支保留为 TABLE 专用，既有 TABLE/FREE_TEXT 路径不变。TDD：先追加 2 测试（段落分割/碎片捕获、表格段不在物理行 0 时的取消线回填）确认 RED（result.mixed 空 + IndexError），再实现确认 GREEN（9 passed），全量 pytest 51 passed（49 baseline + 2）。提交见 git log | src/genesis/parsers/excel_parser.py, tests/test_excel_parser.py | deepseek-v4-flash-free |
 | 2026-08-09 09:40 | 测试验证 | 里程碑2.5 Task12 实现：混合型样本 + 端到端验证。生成 samples/要件定義_混合型.xlsx（機能一覧 表格：表头+3 数据行，空行后接「・改修ポイント」「■対象期間」两条连续碎片文本，无空行）；tests/test_real_samples.py 追加 test_mixed_sample_segments_detected（断言 mixed 段落数==2、kinds==["table","free_text"]、表格段第一批数据行機能ID=="F101"、table.rows==3、碎片段 text 含 改修ポイント/対象期間）。验证：real_samples 全绿 4 passed（3 baseline+1 新增，无 skip）；pytest 全量 52 passed（51 baseline + 1）；实际样本 split_paragraphs 产出两段（0,3）+（5,6）（表格段+碎片段）。提交见 git log | samples/要件定義_混合型.xlsx, tests/test_real_samples.py, _AI_USAGE_LOG.md | deepseek-v4-flash-free |
+| 2026-08-09 | 整体迭代 | 里程碑2.5 Task13 遗留清理：① extraction_method 枚举同源（free_text_extractor→ExtractionMethod.LLM_FROM_FREE_TEXT.value、table_extractor 空/非空两分支→ExtractionMethod.OPENPYXL.value，值不变既有用例兼容）；② table_extractor header_row 越界/负值防御（<0 或 >=len(matrix) 回落 0，keep if not matrix 早返回）；③ merge_fill.forward_fill 完整替换（min_row<1/min_col<1 非法值跳过 + 越界防御）；④ data_models.Provenance.row 语义注释；⑤ 补测试 6 个（sheet_nature MIXED ・/■ 正向 x2、formatting_detector 字体色正向、table_extractor 枚举锁 + forward_fill 防御、free_text_extractor 枚举锁）。TDD：先加测试确认 RED（仅 forward_fill 防御用例失败，其余为已通过回归锁）→ 实现 → GREEN；pytest 全量 58 passed（52 baseline + 6） | src/genesis/parsers/free_text_extractor.py, src/genesis/parsers/table_extractor.py, src/genesis/parsers/merge_fill.py, src/genesis/data_models.py, tests/test_sheet_nature.py, tests/test_formatting_detector.py, tests/test_free_text_extractor.py, tests/test_table_extractor.py, _AI_USAGE_LOG.md | deepseek-v4-flash-free |

src/genesis/data_models.py

@@ -54,7 +54,7 @@ class ExtractionMethod(Enum):
 class Provenance:
     file_name: str
     sheet_name: str
-    row: int
+    row: int          # 数据行号（从 1 起：表格为物理行-表头行；自由文本为块序）
     column: str
     column_header: str
 

src/genesis/parsers/free_text_extractor.py

@@ -2,7 +2,7 @@ from __future__ import annotations
 
 from typing import Any
 
-from genesis.data_models import CellValue, ExcelTable, Provenance, SheetType
+from genesis.data_models import CellValue, ExcelTable, ExtractionMethod, Provenance, SheetType
 
 
 def extract_text_blocks(matrix: list[list[Any]]) -> list[str]:
@@ -45,7 +45,7 @@ def build_free_text_table(
     return ExcelTable(
         name=sheet_name,
         detected_type=detected_type,
-        extraction_method="llm_from_free_text",
+        extraction_method=ExtractionMethod.LLM_FROM_FREE_TEXT.value,
         headers=["text"],
         rows=rows,
     )

src/genesis/parsers/merge_fill.py

@@ -10,7 +10,10 @@ def forward_fill(
     """合并单元格：用左上角主格值填充范围内全部单元格。"""
     out = [list(row) for row in matrix]
     for (min_row, min_col, max_row, max_col) in merged_ranges:
-        if not out or min_row > len(out) or min_col > len(out[min_row - 1]):
+        # 防御：合并范围 1-based，非法（<1）或越界时跳过该范围
+        if not out or min_row < 1 or min_col < 1:
+            continue
+        if min_row > len(out) or min_col > len(out[min_row - 1]):
             continue
         main_value = out[min_row - 1][min_col - 1]
         for r in range(min_row, min(max_row, len(out)) + 1):

src/genesis/parsers/table_extractor.py

@@ -2,7 +2,9 @@ from __future__ import annotations
 
 from typing import Any
 
-from genesis.data_models import CellFormatting, CellValue, ExcelTable, Provenance, SheetType
+from genesis.data_models import (
+    CellFormatting, CellValue, ExcelTable, ExtractionMethod, Provenance, SheetType,
+)
 
 
 def column_letter(index: int) -> str:
@@ -29,8 +31,10 @@ def extract_table(
     if not matrix:
         return ExcelTable(
             name=sheet_name, detected_type=detected_type,
-            extraction_method="openpyxl", headers=[], rows=[],
+            extraction_method=ExtractionMethod.OPENPYXL.value, headers=[], rows=[],
         )
+    if header_row < 0 or header_row >= len(matrix):
+        header_row = 0
     fmt_map = formatting_map or {}
     headers = [str(c) if c is not None else "" for c in matrix[header_row]]
     rows = []
@@ -52,5 +56,5 @@ def extract_table(
         rows.append(row_dict)
     return ExcelTable(
         name=sheet_name, detected_type=detected_type,
-        extraction_method="openpyxl", headers=headers, rows=rows,
+        extraction_method=ExtractionMethod.OPENPYXL.value, headers=headers, rows=rows,
     )

tests/test_formatting_detector.py

@@ -35,6 +35,19 @@ def test_cell_formatting_none_when_plain():
     assert cell_formatting(ws["A1"]) is None
 
 
+def test_cell_formatting_detects_font_color():
+    from openpyxl.styles import Font
+    wb = Workbook()
+    ws = wb.active
+    ws["A1"] = "x"
+    f = copy(ws["A1"].font)
+    f.color = Font(color="FF0000FF").color
+    ws["A1"].font = f
+    fmt = cell_formatting(ws["A1"])
+    assert fmt is not None
+    assert fmt.font_color is not None
+
+
 def test_cell_comment_returns_obj():
     wb = make_wb()
     cm = cell_comment(wb.active["A2"], "f.xlsx")

tests/test_free_text_extractor.py

@@ -1,4 +1,4 @@
-from genesis.data_models import CellValue, ExcelTable, SheetType
+from genesis.data_models import CellValue, ExcelTable, ExtractionMethod, SheetType
 from genesis.parsers.free_text_extractor import build_free_text_table, extract_text_blocks
 
 
@@ -15,3 +15,8 @@ def test_build_free_text_table():
     assert len(table.rows) == 2
     assert isinstance(table.rows[0]["text"], CellValue)
     assert table.rows[0]["text"].value == "A"
+
+
+def test_free_text_uses_enum_value():
+    table = build_free_text_table("メモ", ["A"], "f.xlsx")
+    assert table.extraction_method == ExtractionMethod.LLM_FROM_FREE_TEXT.value

tests/test_sheet_nature.py

@@ -19,4 +19,14 @@ def test_free_text_many_empty_rows():
 def test_header_row_index():
     m = [["機能ID", "名前"], ["1", "田中"]]
     assert find_header_row(m) == 0
-    assert find_header_row([["自由テキスト"]]) == -1
+    assert find_header_row([["自由テキスト"]]) == -1
+
+
+def test_classify_sheet_mixed_with_bullet_line():
+    m = [["ID", "名前"], ["1", "田中"], ["・備考行"]]
+    assert classify_sheet(m) == SheetNature.MIXED
+
+
+def test_classify_sheet_mixed_with_square_line():
+    m = [["ID", "名前"], ["1", "田中"], ["■備考行"]]
+    assert classify_sheet(m) == SheetNature.MIXED

tests/test_table_extractor.py

@@ -1,4 +1,4 @@
-from genesis.data_models import CellValue, ExcelTable, SheetType
+from genesis.data_models import CellValue, ExcelTable, ExtractionMethod, SheetType
 from genesis.parsers.merge_fill import forward_fill
 from genesis.parsers.table_extractor import column_letter, extract_table
 
@@ -37,4 +37,16 @@ def test_extract_table_basic():
     assert isinstance(first["ID"], CellValue)
     assert first["ID"].value == "1"
     assert first["ID"].provenance.sheet_name == "社員一覧"
-    assert first["ID"].provenance.column == "A"
+    assert first["ID"].provenance.column == "A"
+
+
+def test_extract_table_uses_enum_value():
+    matrix = [["ID"], ["1"]]
+    table = extract_table("社員", matrix, "f.xlsx", SheetType.FUNCTION)
+    assert table.extraction_method == ExtractionMethod.OPENPYXL.value
+
+
+def test_forward_fill_ignores_invalid_range():
+    matrix = [["A"], ["B"]]
+    assert forward_fill(matrix, [(0, 1, 2, 1)]) == matrix  # min_row=0 非法 → 不崩溃不改数据
+    assert forward_fill(matrix, [(9, 9, 9, 9)]) == matrix  # 越界 → 不崩溃