feat: Phase 2 complete — 13 Phases of COBOL type classification and test benchmark

P0.6: gcov infrastructure P1: extract_structure output expansion (11 new feature fields) P2: Confusion group rule engine (8 pairs + contradiction + backtrack) P3: 4-factor confidence calculation + quality gate update P4: 33+2 COBOL program type test samples (22 files, 7 categories) P5: parametrized/ test data generation engine P6: japanese_data.py lookup tables P7-10: Type-specific test suites (~159 parametrized tests) P11: Full classification pipeline (classify_program) + orchestrator integration P12: Documentation (module-interfaces, test-plan v3.0, coverage-matrix) Architecture decisions: - classification_pipeline/ merged to hina/pipeline/ - parametrized/ as independent module - japanese_data.py as root-level file - hina/__all__ only exports classify_program() Co-Authored-By: Claude <noreply@anthropic.com>
2026-06-19 23:51:55 +08:00
parent 63b5284715
commit bc1d56d1a4
129 changed files with 19378 additions and 261 deletions
@@ -0,0 +1,433 @@
+"""Deep coverage tests: HTML report, SEARCH/EVALUATE/PERFORM coverage, locate, index"""
+
+import sys, os
+sys.path.insert(0, os.path.abspath(os.path.join(os.path.dirname(__file__), "..", "..")))
+
+from cobol_testgen.models import BrSeq, CondLeaf
+from cobol_testgen.coverage import (
+    DecisionPoint, LeafStat,
+    mark_coverage, generate_html_report, generate_coverage_index,
+    locate_decision_lines, check_coverage,
+)
+
+
+# ── 1. generate_html_report ──
+
+def test_generate_html_report_full(tmp_path):
+    """Generate full HTML report with known DecisionPoint data — assert table, branch rate, decision points"""
+    dps = [
+        DecisionPoint(id=1, kind="IF", label="A > 100",
+                       branch_names=["T", "F"],
+                       active_branches={"T"},
+                       implied_branches={"T"},
+                       source_line=4),
+        DecisionPoint(id=2, kind="EVALUATE", label="WS-STATUS",
+                       branch_names=["WHEN 1", "WHEN 2", "OTHER"],
+                       active_branches={"WHEN 1"},
+                       implied_branches={"WHEN 1"},
+                       source_line=7),
+    ]
+    leaves = [
+        LeafStat(field="A", op=">", value="100", covered_true=True, covered_false=False),
+        LeafStat(field="B", op="=", value="1", covered_true=False, covered_false=False),
+    ]
+    source_lines = [
+        "       IDENTIFICATION DIVISION.",
+        "       PROGRAM-ID. TESTPGM.",
+        "       PROCEDURE DIVISION.",
+        "           IF A > 100",
+        "               MOVE 1 TO B",
+        "           END-IF.",
+        "           EVALUATE WS-STATUS",
+        "               WHEN 1 ...",
+        "           END-EVALUATE.",
+        "           STOP RUN.",
+    ]
+    outpath = tmp_path / "TESTPGM_coverage.html"
+    generate_html_report(dps, leaves, source_lines, outpath, filename="TESTPGM")
+    html = outpath.read_text(encoding="utf-8")
+
+    # HTML structure
+    assert "<table" in html, "Should contain <table> for decision point list"
+    assert "覆盖率报告" in html, "Should contain report title"
+    assert "TESTPGM" in html, "Should contain program name in title"
+
+    # Branch rate percentage
+    # total=5, covered=2 → 40.0%
+    assert "40.0%" in html or "2/5" in html
+
+    # Coverage section texts
+    assert "决策覆盖率" in html
+    assert "条件覆盖率" in html
+
+    # Decision point list items
+    assert "#1" in html
+    assert "#2" in html
+    assert "IF" in html
+    assert "EVALUATE" in html
+    assert "branch-true" in html
+    assert "branch-false" in html
+
+    # Leaf stats table
+    assert "A" in html
+    assert "B" in html
+
+    # Source lines
+    assert "IF A > 100" in html
+    assert "EVALUATE WS-STATUS" in html
+    assert "hl-green" in html  # IF line is fully covered
+
+
+def test_generate_html_report_no_decision_points(tmp_path):
+    """No decision points → no branch table, no SVG"""
+    outpath = tmp_path / "empty_report.html"
+    generate_html_report([], [], [], outpath, filename="EMPTYPGM")
+    html = outpath.read_text(encoding="utf-8")
+
+    assert "EMPTYPGM" in html
+    # No DP table rows (0个决策点 shown as stat)
+    assert "0个" in html or "0%" in html
+    # Still has the summary section
+    assert "覆盖率概要" in html
+
+
+# ── 2. BrSearch (SEARCH ALL) coverage via _mark_search ──
+
+def test_mark_search_covered_first_branch():
+    """SEARCH ALL DecisionPoint with CondLeaf when_list — first WHEN branch covered"""
+    dp = DecisionPoint(id=1, kind="SEARCH", label="WS-TABLE",
+                       branch_names=["WHEN A > 100", "WHEN B = 50", "AT END"])
+    dp.when_list = [
+        ("A > 100", BrSeq()),
+        ("B = 50", BrSeq()),
+    ]
+    dp.cond_trees = [
+        CondLeaf("A", ">", "100"),
+        CondLeaf("B", "=", "50"),
+    ]
+    dp.has_other = True
+
+    leaf_stats = []
+    branch_paths = [
+        ([("A", ">", "100", True)], []),
+    ]
+    mark_coverage([dp], leaf_stats, branch_paths, [])
+
+    assert "WHEN A > 100" in dp.active_branches
+    assert "AT END" not in dp.active_branches
+    assert "WHEN B = 50" not in dp.active_branches
+
+
+def test_mark_search_covered_at_end():
+    """SEARCH ALL — no WHEN matches → AT END covered"""
+    dp = DecisionPoint(id=1, kind="SEARCH", label="WS-TABLE",
+                       branch_names=["WHEN K > 10", "AT END"])
+    dp.when_list = [
+        ("K > 10", BrSeq()),
+    ]
+    dp.cond_trees = [
+        CondLeaf("K", ">", "10"),
+    ]
+    dp.has_other = True
+
+    leaf_stats = []
+    # K <= 10 → no WHEN matches
+    branch_paths = [
+        ([("K", ">", "10", False)], []),
+    ]
+    mark_coverage([dp], leaf_stats, branch_paths, [])
+
+    assert "AT END" in dp.active_branches
+    assert "WHEN K > 10" not in dp.active_branches
+
+
+def test_mark_search_compound_condition():
+    """SEARCH ALL with compound condition tree"""
+    dp = DecisionPoint(id=1, kind="SEARCH", label="WS-TABLE",
+                       branch_names=["WHEN A>1 AND B<9", "AT END"])
+    dp.when_list = [
+        ("A > 1 AND B < 9", BrSeq()),
+    ]
+    # Build compound tree: CondAnd(CondLeaf("A", ">", "1"), CondLeaf("B", "<", "9"))
+    dp.cond_trees = [
+        type('obj', (object,), {
+            'field': 'dummy', 'op': '=', 'value': '0',
+            '__class__': CondLeaf.__class__,
+        })  # won't be used — tree is CondAnd type
+    ]
+    # Actually use a proper tree
+    from cobol_testgen.models import CondAnd
+    dp.cond_trees = [
+        CondAnd(CondLeaf("A", ">", "1"), CondLeaf("B", "<", "9"))
+    ]
+    dp.has_other = True
+
+    leaf_stats = []
+    branch_paths = [
+        ([("A", ">", "1", True), ("B", "<", "9", True)], []),
+    ]
+    mark_coverage([dp], leaf_stats, branch_paths, [])
+
+    assert "WHEN A>1 AND B<9" in dp.active_branches
+    assert "AT END" not in dp.active_branches
+
+
+# ── 3. BrEval with multiple subjects (ALSO) — _mark_eval ──
+
+def test_mark_eval_simple():
+    """EVALUATE with subject match via constraint field=subject"""
+    dp = DecisionPoint(id=1, kind="EVALUATE", label="WS-STATUS",
+                       branch_names=["WHEN 1", "WHEN 2", "OTHER"])
+    dp.when_list = [
+        ("1", BrSeq()),
+        ("2", BrSeq()),
+    ]
+    leaf_stats = []
+    branch_paths = [
+        ([("WS-STATUS", "=", "1", True)], []),
+    ]
+    mark_coverage([dp], leaf_stats, branch_paths, [])
+
+    assert "WHEN 1" in dp.active_branches
+    assert "WHEN 2" not in dp.active_branches
+    assert "OTHER" not in dp.active_branches
+
+
+def test_mark_eval_other_branch():
+    """EVALUATE — not_in constraint triggers OTHER"""
+    dp = DecisionPoint(id=1, kind="EVALUATE", label="WS-STATUS",
+                       branch_names=["WHEN 1", "WHEN 2", "OTHER"])
+    dp.when_list = [
+        ("1", BrSeq()),
+        ("2", BrSeq()),
+    ]
+    leaf_stats = []
+    branch_paths = [
+        ([("WS-STATUS", "not_in", "", True)], []),
+    ]
+    mark_coverage([dp], leaf_stats, branch_paths, [])
+
+    assert "OTHER" in dp.active_branches
+
+
+def test_mark_eval_true_subject():
+    """EVALUATE TRUE with matched WHEN branch"""
+    dp = DecisionPoint(id=1, kind="EVALUATE", label="TRUE",
+                       branch_names=["WHEN A > 100", "WHEN B = 0", "OTHER"])
+    dp.when_list = [
+        ("A > 100", BrSeq()),
+        ("B = 0", BrSeq()),
+    ]
+    leaf_stats = []
+    branch_paths = [
+        ([("A", ">", "100", True)], []),
+    ]
+    mark_coverage([dp], leaf_stats, branch_paths, [])
+
+    assert "WHEN A > 100" in dp.active_branches
+
+
+# ── 4. BrPerform UNTIL — _mark_perform ──
+
+def test_mark_perform_until_skip():
+    """PERFORM UNTIL condition true → Skip branch active"""
+    dp = DecisionPoint(id=1, kind="PERFORM", label="A > 100",
+                       branch_names=["Enter", "Skip"])
+    # Simulate the "parsed" attribute set by collect_decision_points
+    dp.parsed = ("A", ">", "100")
+
+    leaf_stats = []
+    branch_paths = [
+        ([("A", ">", "100", True)], []),
+    ]
+    mark_coverage([dp], leaf_stats, branch_paths, [])
+
+    assert "Skip" in dp.active_branches
+    assert "Enter" not in dp.active_branches
+
+
+def test_mark_perform_until_enter():
+    """PERFORM UNTIL condition false → Enter branch active"""
+    dp = DecisionPoint(id=1, kind="PERFORM", label="A > 100",
+                       branch_names=["Enter", "Skip"])
+    dp.parsed = ("A", ">", "100")
+
+    leaf_stats = []
+    branch_paths = [
+        ([("A", ">", "100", False)], []),
+    ]
+    mark_coverage([dp], leaf_stats, branch_paths, [])
+
+    assert "Enter" in dp.active_branches
+    assert "Skip" not in dp.active_branches
+
+
+def test_mark_perform_until_compound():
+    """PERFORM UNTIL with compound condition tree"""
+    from cobol_testgen.models import CondAnd
+    leaf_a = CondLeaf("A", ">", "100")
+    leaf_b = CondLeaf("B", "<", "50")
+    dp = DecisionPoint(id=1, kind="PERFORM", label="A > 100 AND B < 50",
+                       branch_names=["Enter", "Skip"])
+    dp.cond_tree = CondAnd(leaf_a, leaf_b)
+    dp.cond_leaves = [leaf_a, leaf_b]
+
+    leaf_stats = []
+    branch_paths = [
+        ([("A", ">", "100", True), ("B", "<", "50", True)], []),
+    ]
+    mark_coverage([dp], leaf_stats, branch_paths, [])
+
+    assert "Skip" in dp.active_branches
+
+
+def test_mark_perform_until_compound_false():
+    """PERFORM UNTIL compound false → Enter active"""
+    from cobol_testgen.models import CondAnd
+    leaf_a = CondLeaf("A", ">", "100")
+    leaf_b = CondLeaf("B", "<", "50")
+    dp = DecisionPoint(id=1, kind="PERFORM", label="A > 100 AND B < 50",
+                       branch_names=["Enter", "Skip"])
+    dp.cond_tree = CondAnd(leaf_a, leaf_b)
+    dp.cond_leaves = [leaf_a, leaf_b]
+
+    leaf_stats = []
+    branch_paths = [
+        ([("A", ">", "100", True), ("B", "<", "50", False)], []),
+    ]
+    mark_coverage([dp], leaf_stats, branch_paths, [])
+
+    assert "Enter" in dp.active_branches
+
+
+# ── 5. locate_decision_lines with real COBOL ──
+
+def test_locate_decision_lines_complex():
+    """Mixed IF/EVALUATE/SEARCH ALL COBOL source → correct line numbers"""
+    source = """       IDENTIFICATION DIVISION.
+       PROGRAM-ID. TESTPGM.
+       DATA DIVISION.
+       WORKING-STORAGE SECTION.
+       01 WS-A PIC 9(4).
+       PROCEDURE DIVISION.
+           IF WS-A > 100
+               MOVE 1 TO B
+           END-IF.
+           EVALUATE WS-A
+               WHEN 1
+                   MOVE 'A' TO B
+               WHEN 2
+                   MOVE 'B' TO B
+               WHEN OTHER
+                   MOVE 'C' TO B
+           END-EVALUATE.
+           SEARCH ALL WS-TABLE
+               AT END DISPLAY 'NOT FOUND'
+               WHEN WS-KEY = 1 DISPLAY 'FOUND'
+           END-SEARCH.
+           STOP RUN.
+       END PROGRAM TESTPGM."""
+
+    dps = [
+        DecisionPoint(id=1, kind="IF", label="WS-A > 100",
+                       branch_names=["T", "F"]),
+        DecisionPoint(id=2, kind="EVALUATE", label="WS-A",
+                       branch_names=["WHEN 1", "WHEN 2", "OTHER"]),
+        # SEARCH kind is not located by _build_search_patterns, expect 0
+        DecisionPoint(id=3, kind="SEARCH", label="WS-TABLE",
+                       branch_names=["WHEN K=1", "AT END"]),
+    ]
+    locate_decision_lines(dps, source)
+
+    assert dps[0].source_line == 7   # IF WS-A > 100
+    assert dps[1].source_line == 10  # EVALUATE WS-A
+    assert dps[2].source_line == 0   # SEARCH not located (no pattern)
+
+
+# ── 6. check_coverage with real-style structure ──
+
+def test_check_coverage_with_structure():
+    """Real-style structure dict with decision_points list and records"""
+    structure = {
+        "total_paragraphs": 5,
+        "total_branches": 10,
+        "decision_points": [
+            {"kind": "IF", "branch_names": ["T", "F"]},
+            {"kind": "EVALUATE", "branch_names": ["W1", "W2", "OTHER"]},
+        ],
+    }
+    test_records = [{"id": 1, "case": "CASE01"}, {"id": 2, "case": "CASE02"}]
+    result = check_coverage(structure, test_records)
+
+    assert isinstance(result, dict)
+    assert result["paragraph_rate"] == 1.0       # has records + paragraphs > 0
+    assert result["branch_rate"] == 0.0          # static analysis limitation
+    assert result["decision_rate"] == 0.0
+    assert result["total_branches"] == 10
+    assert result["total_paragraphs"] == 5
+    assert result["records_count"] == 2
+    assert "gcov" in result["note"]
+
+
+def test_check_coverage_no_records():
+    """No test records → paragraph_rate = 0.0"""
+    structure = {"total_paragraphs": 3, "total_branches": 5, "decision_points": []}
+    result = check_coverage(structure, [])
+    assert result["paragraph_rate"] == 0.0
+    assert result["records_count"] == 0
+
+
+def test_check_coverage_no_paragraphs():
+    """No paragraphs but records exist → paragraph_rate = 0.0"""
+    structure = {"total_paragraphs": 0, "total_branches": 5, "decision_points": []}
+    result = check_coverage(structure, [{"id": 1}])
+    assert result["paragraph_rate"] == 0.0
+
+
+# ── 7. generate_coverage_index with 2 programs ──
+
+def test_generate_coverage_index_two_programs(tmp_path):
+    """Index page with 2 programs → HTML contains both names and SVG ring charts"""
+    programs = [
+        {
+            "name": "PGM001",
+            "detail_relpath": "../PGM001_coverage.html",
+            "total_branches": 5,
+            "covered_branches": 4,
+            "implied_branches": 4,
+            "total_conditions": 6,
+            "covered_conditions": 5,
+        },
+        {
+            "name": "PGM002",
+            "detail_relpath": "../PGM002_coverage.html",
+            "total_branches": 3,
+            "covered_branches": 3,
+            "implied_branches": 3,
+            "total_conditions": 4,
+            "covered_conditions": 4,
+        },
+    ]
+    generate_coverage_index(programs, str(tmp_path))
+
+    index_path = tmp_path / "coverage" / "index.html"
+    assert index_path.exists()
+    html = index_path.read_text(encoding="utf-8")
+
+    # Both program names
+    assert "PGM001" in html
+    assert "PGM002" in html
+
+    # Links to detail pages
+    assert "PGM001_coverage.html" in html
+    assert "PGM002_coverage.html" in html
+
+    # SVG ring chart
+    assert "<svg" in html
+    assert "circle" in html
+    assert "100%" in html or "80.0%"  # PGM002 is 100%, PGM001 is 80%
+
+    # Coverage text
+    assert "覆盖率总览" in html
+    assert "决策覆盖率" in html
+    assert "条件覆盖率" in html