code2llm 0.5.172__tar.gz → 0.5.174__tar.gz

{code2llm-0.5.172/code2llm.egg-info → code2llm-0.5.174}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: code2llm
-Version: 0.5.172
+Version: 0.5.174
 Summary: High-performance Python code flow analysis with optimized TOON format - CFG, DFG, call graphs, and intelligent code queries
 Home-page: https://github.com/wronai/stts
 Author: Tom Sapletta
@@ -67,13 +67,13 @@ Dynamic: requires-python
 
 ## AI Cost Tracking
 
-![PyPI](https://img.shields.io/badge/pypi-costs-blue) ![Version](https://img.shields.io/badge/version-0.5.172-blue) ![Python](https://img.shields.io/badge/python-3.9+-blue) ![License](https://img.shields.io/badge/license-Apache--2.0-green)
-![AI Cost](https://img.shields.io/badge/AI%20Cost-$17.33-orange) ![Human Time](https://img.shields.io/badge/Human%20Time-92.8h-blue) ![Model](https://img.shields.io/badge/Model-openrouter%2Fdeep%2Fdeep--v4--pro-lightgrey)
+![PyPI](https://img.shields.io/badge/pypi-costs-blue) ![Version](https://img.shields.io/badge/version-0.5.174-blue) ![Python](https://img.shields.io/badge/python-3.9+-blue) ![License](https://img.shields.io/badge/license-Apache--2.0-green)
+![AI Cost](https://img.shields.io/badge/AI%20Cost-$3.80-orange) ![Human Time](https://img.shields.io/badge/Human%20Time-102.3h-blue) ![Model](https://img.shields.io/badge/Model-openrouter%2Fdeep%2Fdeep--v4--pro-lightgrey)
 
-- 🤖 **LLM usage:** $17.3260 (246 commits)
-- 👤 **Human dev:** ~$9280 (92.8h @ $100/h, 30min dedup)
+- 🤖 **LLM usage:** $3.8031 (260 commits)
+- 👤 **Human dev:** ~$10235 (102.3h @ $100/h, 30min dedup)
 
-Generated on 2026-07-06 using [openrouter/deep/deep-v4-pro](https://openrouter.ai/deep/deep-v4-pro)
+Generated on 2026-07-13 using [openrouter/deep/deep-v4-pro](https://openrouter.ai/deep/deep-v4-pro)
 
 ---
 

{code2llm-0.5.172 → code2llm-0.5.174}/README.md

@@ -2,13 +2,13 @@
 
 ## AI Cost Tracking
 
-![PyPI](https://img.shields.io/badge/pypi-costs-blue) ![Version](https://img.shields.io/badge/version-0.5.172-blue) ![Python](https://img.shields.io/badge/python-3.9+-blue) ![License](https://img.shields.io/badge/license-Apache--2.0-green)
-![AI Cost](https://img.shields.io/badge/AI%20Cost-$17.33-orange) ![Human Time](https://img.shields.io/badge/Human%20Time-92.8h-blue) ![Model](https://img.shields.io/badge/Model-openrouter%2Fdeep%2Fdeep--v4--pro-lightgrey)
+![PyPI](https://img.shields.io/badge/pypi-costs-blue) ![Version](https://img.shields.io/badge/version-0.5.174-blue) ![Python](https://img.shields.io/badge/python-3.9+-blue) ![License](https://img.shields.io/badge/license-Apache--2.0-green)
+![AI Cost](https://img.shields.io/badge/AI%20Cost-$3.80-orange) ![Human Time](https://img.shields.io/badge/Human%20Time-102.3h-blue) ![Model](https://img.shields.io/badge/Model-openrouter%2Fdeep%2Fdeep--v4--pro-lightgrey)
 
-- 🤖 **LLM usage:** $17.3260 (246 commits)
-- 👤 **Human dev:** ~$9280 (92.8h @ $100/h, 30min dedup)
+- 🤖 **LLM usage:** $3.8031 (260 commits)
+- 👤 **Human dev:** ~$10235 (102.3h @ $100/h, 30min dedup)
 
-Generated on 2026-07-06 using [openrouter/deep/deep-v4-pro](https://openrouter.ai/deep/deep-v4-pro)
+Generated on 2026-07-13 using [openrouter/deep/deep-v4-pro](https://openrouter.ai/deep/deep-v4-pro)
 
 ---
 

code2llm-0.5.174/VERSION

@@ -0,0 +1 @@
+0.5.174

{code2llm-0.5.172 → code2llm-0.5.174}/code2llm/__init__.py

@@ -8,7 +8,7 @@ Includes NLP Processing Pipeline for query normalization, intent matching,
 and entity resolution with multilingual support.
 """
 
-__version__ = "0.5.172"
+__version__ = "0.5.174"
 __author__ = "STTS Project"
 
 # Core analysis components (lightweight, always needed)

{code2llm-0.5.172 → code2llm-0.5.174}/code2llm/analysis/smells.py

@@ -131,18 +131,34 @@ class SmellDetector:
         return smells
 
     def _detect_data_clumps(self) -> List[CodeSmell]:
-        """Detect 3+ variables frequently passed together."""
+        """Detect 3+ variables frequently passed together.
+
+        Grouped by (file, argument-set), not argument-set alone: a common
+        constructor shape like `(self, project_root, config)` is a
+        conventional pattern reused independently by unrelated classes
+        across unrelated files (any two classes that each take a project
+        root and a config object) -- that's coincidence, not evidence the
+        two classes should share a bundled parameter object. Grouping by
+        bare argument-set alone falsely flagged such coincidences as data
+        clumps (reproduced: `wup.anomaly_detector.AnomalyDetector.__init__`
+        and `wup.testql_monitor.TestQLMonitor.__init__` -- two unrelated
+        classes in unrelated files -- flagged as clumped on `project_root,
+        self, config`). A genuine data clump -- args that should become one
+        object -- is a same-file (usually same-class) phenomenon; scoping
+        by file preserves that signal while dropping cross-file
+        coincidental-signature noise.
+        """
         smells = []
         # Simplified: find functions with same 3+ arguments
-        arg_sets = {}  # frozenset(args) -> List[func_names]
+        arg_sets = {}  # (file, frozenset(args)) -> List[func_names]
         for func_name, func_info in self.result.functions.items():
             if len(func_info.args) >= 3:
-                args = frozenset(func_info.args)
-                if args not in arg_sets:
-                    arg_sets[args] = []
-                arg_sets[args].append(func_name)
+                key = (func_info.file, frozenset(func_info.args))
+                if key not in arg_sets:
+                    arg_sets[key] = []
+                arg_sets[key].append(func_name)
 
-        for args, funcs in arg_sets.items():
+        for (file, args), funcs in arg_sets.items():
             if len(funcs) >= 2:
                 for func_name in funcs:
                     func_info = self.result.functions[func_name]

{code2llm-0.5.172 → code2llm-0.5.174}/code2llm/exporters/flow_constants.py

@@ -39,6 +39,24 @@ EXCLUDE_PATTERNS = {
     ".envs",
 }
 
+# Paths where duplicate class names are intentional (twin sites, examples, contracts).
+# Excluded from duplicate *detection* only — other metrics still apply.
+INTENTIONAL_DUP_COPY_MARKERS: tuple[str, ...] = (
+    "pc1/",
+    "pc2/",
+    "examples/",
+    "-contract-",
+)
+
+
+@lru_cache(maxsize=4096)
+def is_intentional_duplicate_copy(path: str) -> bool:
+    """True when *path* is an intentional parallel copy (not actionable duplication)."""
+    if not path:
+        return False
+    norm = path.replace("\\", "/").lower()
+    return any(marker in norm for marker in INTENTIONAL_DUP_COPY_MARKERS)
+
 
 @lru_cache(maxsize=4096)
 def is_excluded_path(path: str) -> bool:

{code2llm-0.5.172 → code2llm-0.5.174}/code2llm/exporters/planfile_tickets.py

@@ -256,9 +256,51 @@ def _god_module_tickets(result: AnalysisResult) -> list[PlanfileTicketSuggestion
 
 
 def _duplicate_class_tickets(result: AnalysisResult) -> list[PlanfileTicketSuggestion]:
-    """Yield deduplication tickets for class pairs with ≥60% method-name overlap."""
+    """Yield one deduplication ticket per duplicate *class name group* (not per pair)."""
+    from collections import defaultdict
+
+    from code2llm.exporters.flow_constants import is_intentional_duplicate_copy
+
+    pairs = _duplicate_class_pairs(result)
+    groups: dict[str, set[str]] = defaultdict(set)
+    for left_name, left_file, right_file in pairs:
+        groups[left_name].add(left_file)
+        groups[left_name].add(right_file)
+
     tickets: list[PlanfileTicketSuggestion] = []
-    classes = list(result.classes.values())
+    for class_name, files in sorted(groups.items()):
+        rel_files = tuple(sorted(files))
+        if all(is_intentional_duplicate_copy(f) for f in rel_files):
+            continue
+        tickets.append(
+            PlanfileTicketSuggestion(
+                signal="code2llm_dup",
+                title=f"Consolidate duplicate class {class_name}",
+                description=(
+                    f"code2llm found `{class_name}` in {len(rel_files)} files: "
+                    f"{', '.join(rel_files)}.\n\n"
+                    "Extract a shared base/helper or merge the duplicated behavior, then "
+                    "re-run code2llm to confirm the duplicate signal is gone."
+                ),
+                priority="high",
+                labels=("llm-ready", "code2llm", "duplication", "refactor"),
+                files=rel_files,
+                dedupe_key=f"code2llm:dup:{class_name}",
+            ),
+        )
+    return tickets
+
+
+def _duplicate_class_pairs(result: AnalysisResult) -> list[tuple[str, str, str]]:
+    """Return (class_name, file_a, file_b) for each similar class pair."""
+    from code2llm.exporters.flow_constants import is_intentional_duplicate_copy
+
+    pairs: list[tuple[str, str, str]] = []
+    classes = [
+        c
+        for c in result.classes.values()
+        if not is_intentional_duplicate_copy(c.file)
+    ]
     for index, left in enumerate(classes):
         left_methods = _method_names(left)
         if len(left_methods) < 3:
@@ -275,23 +317,8 @@ def _duplicate_class_tickets(result: AnalysisResult) -> list[PlanfileTicketSugge
                 continue
             left_file = _rel_path(left.file, result.project_path)
             right_file = _rel_path(right.file, result.project_path)
-            tickets.append(
-                PlanfileTicketSuggestion(
-                    signal="code2llm_dup",
-                    title=f"Consolidate duplicate class {left.name}",
-                    description=(
-                        f"code2llm found similar `{left.name}` classes in `{left_file}` "
-                        f"and `{right_file}` with {len(overlap)} overlapping methods.\n\n"
-                        "Extract a shared base/helper or merge the duplicated behavior, then "
-                        "re-run code2llm to confirm the duplicate signal is gone."
-                    ),
-                    priority="high",
-                    labels=("llm-ready", "code2llm", "duplication", "refactor"),
-                    files=(left_file, right_file),
-                    dedupe_key=f"code2llm:dup:{left.name}:{left_file}:{right_file}",
-                ),
-            )
-    return tickets
+            pairs.append((left.name, left_file, right_file))
+    return pairs
 
 
 def _smell_tickets(result: AnalysisResult) -> list[PlanfileTicketSuggestion]:

{code2llm-0.5.172 → code2llm-0.5.174}/code2llm/exporters/toon/_render_section_helpers.py

@@ -54,7 +54,8 @@ def _build_refactor_steps(ctx: Dict[str, Any]) -> List[str]:
     """Build list of refactoring step strings from health, duplicates, and cycles."""
     steps: List[str] = []
     if ctx["duplicates"]:
-        steps.append(f"rm duplicates  (-{len(ctx['duplicates'])} dup classes)")
+        group_count = len({d["class_name"] for d in ctx["duplicates"]})
+        steps.append(f"rm duplicates  (-{group_count} dup groups)")
     for gi in [h for h in ctx["health"] if h["code"] == "GOD"]:
         steps.append(f"split {gi['message'].split('=')[0].strip()}  (god module)")
     cc_issues = [h for h in ctx["health"] if h["code"] == "CC"]

{code2llm-0.5.172 → code2llm-0.5.174}/code2llm/exporters/toon/helpers.py

@@ -5,7 +5,7 @@ from pathlib import Path
 from typing import Any, Dict, Set
 
 from code2llm.core.models import AnalysisResult, FunctionInfo
-from code2llm.exporters.flow_constants import is_excluded_path
+from code2llm.exporters.flow_constants import is_excluded_path, is_intentional_duplicate_copy
 
 # Re-export is_excluded_path from flow_constants to eliminate duplication
 

{code2llm-0.5.172 → code2llm-0.5.174}/code2llm/exporters/toon/metrics_duplicates.py

@@ -7,7 +7,7 @@ from typing import Any, Dict, List, Optional, Set
 
 from code2llm.core.models import AnalysisResult, ClassInfo
 
-from .helpers import _is_excluded, _rel_path
+from .helpers import _is_excluded, _rel_path, is_intentional_duplicate_copy
 
 
 class DuplicatesMetricsComputer:
@@ -22,7 +22,9 @@ class DuplicatesMetricsComputer:
         dupes: List[Dict[str, Any]] = []
         # Filter out excluded classes first
         class_list = [
-            (q, c) for q, c in result.classes.items() if not _is_excluded(c.file)
+            (q, c)
+            for q, c in result.classes.items()
+            if not _is_excluded(c.file) and not is_intentional_duplicate_copy(c.file)
         ]
 
         for i, (qa, ca) in enumerate(class_list):

{code2llm-0.5.172 → code2llm-0.5.174}/code2llm/exporters/toon/metrics_health.py

@@ -31,21 +31,23 @@ class HealthMetricsComputer:
         return issues[:20]  # MAX_HEALTH_ISSUES
 
     def _check_duplicates_health(self, ctx, issues):
-        """Check for duplicate classes."""
-        ndups = len(ctx["duplicates"])
+        """Check for duplicate classes (report merged name-groups, not pair count)."""
+        dupes = ctx["duplicates"]
+        group_names = {d.get("class_name") for d in dupes if d.get("class_name")}
+        ndups = len(group_names)
         if ndups > 0:
             dup_lines = sum(
                 ctx["files"].get(d.get("fileA_abs", ""), {}).get("lines", 0)
-                for d in ctx["duplicates"]
+                for d in dupes
             )
             issues.append(
                 {
                     "severity": "red",
                     "code": "DUP",
-                    "message": f"{ndups} classes duplicated ({dup_lines}L wasted)"
+                    "message": f"{ndups} duplicate class groups ({dup_lines}L wasted)"
                     if dup_lines
-                    else f"{ndups} classes duplicated",
-                    "impact": f"-{ndups} dup classes",
+                    else f"{ndups} duplicate class groups",
+                    "impact": f"-{ndups} dup groups",
                 }
             )
 

{code2llm-0.5.172 → code2llm-0.5.174}/code2llm/exporters/toon/renderer.py

@@ -41,7 +41,7 @@ class ToonRenderer:
         all_cc = [f["cc"] for f in ctx["func_metrics"]]
         avg_cc = round(sum(all_cc) / len(all_cc), 1) if all_cc else 0.0
         critical = len([f for f in ctx["func_metrics"] if f["cc"] >= CC_CRITICAL])
-        ndups = len(ctx["duplicates"])
+        ndups = len({d["class_name"] for d in ctx["duplicates"]})
         ncycles = len(ctx["cycles"])
         lang_label = _detect_language_label(result)
         return [

{code2llm-0.5.172 → code2llm-0.5.174}/code2llm/nlp/__init__.py

@@ -4,7 +4,7 @@ Provides query normalization, intent matching, and entity resolution
 with multilingual support and fuzzy matching.
 """
 
-__version__ = "0.5.172"
+__version__ = "0.5.174"
 
 from .pipeline import NLPPipeline
 from .normalization import QueryNormalizer

{code2llm-0.5.172 → code2llm-0.5.174/code2llm.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: code2llm
-Version: 0.5.172
+Version: 0.5.174
 Summary: High-performance Python code flow analysis with optimized TOON format - CFG, DFG, call graphs, and intelligent code queries
 Home-page: https://github.com/wronai/stts
 Author: Tom Sapletta
@@ -67,13 +67,13 @@ Dynamic: requires-python
 
 ## AI Cost Tracking
 
-![PyPI](https://img.shields.io/badge/pypi-costs-blue) ![Version](https://img.shields.io/badge/version-0.5.172-blue) ![Python](https://img.shields.io/badge/python-3.9+-blue) ![License](https://img.shields.io/badge/license-Apache--2.0-green)
-![AI Cost](https://img.shields.io/badge/AI%20Cost-$17.33-orange) ![Human Time](https://img.shields.io/badge/Human%20Time-92.8h-blue) ![Model](https://img.shields.io/badge/Model-openrouter%2Fdeep%2Fdeep--v4--pro-lightgrey)
+![PyPI](https://img.shields.io/badge/pypi-costs-blue) ![Version](https://img.shields.io/badge/version-0.5.174-blue) ![Python](https://img.shields.io/badge/python-3.9+-blue) ![License](https://img.shields.io/badge/license-Apache--2.0-green)
+![AI Cost](https://img.shields.io/badge/AI%20Cost-$3.80-orange) ![Human Time](https://img.shields.io/badge/Human%20Time-102.3h-blue) ![Model](https://img.shields.io/badge/Model-openrouter%2Fdeep%2Fdeep--v4--pro-lightgrey)
 
-- 🤖 **LLM usage:** $17.3260 (246 commits)
-- 👤 **Human dev:** ~$9280 (92.8h @ $100/h, 30min dedup)
+- 🤖 **LLM usage:** $3.8031 (260 commits)
+- 👤 **Human dev:** ~$10235 (102.3h @ $100/h, 30min dedup)
 
-Generated on 2026-07-06 using [openrouter/deep/deep-v4-pro](https://openrouter.ai/deep/deep-v4-pro)
+Generated on 2026-07-13 using [openrouter/deep/deep-v4-pro](https://openrouter.ai/deep/deep-v4-pro)
 
 ---
 

{code2llm-0.5.172 → code2llm-0.5.174}/code2llm.egg-info/SOURCES.txt

@@ -186,6 +186,7 @@ tests/test_cache_invalidation_e2e.py
 tests/test_calls_toon_export.py
 tests/test_declarative_collection.py
 tests/test_deep_analysis.py
+tests/test_dup_groups.py
 tests/test_edge_cases.py
 tests/test_export_cache_flags.py
 tests/test_file_analyzer_tagging.py
@@ -202,5 +203,6 @@ tests/test_project_toon_export.py
 tests/test_prompt_engine.py
 tests/test_prompt_txt.py
 tests/test_refactoring_engine.py
+tests/test_smells_data_clumps.py
 tests/test_smells_shotgun_surgery.py
 tests/test_toon_v2.py

{code2llm-0.5.172 → code2llm-0.5.174}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "code2llm"
-version = "0.5.172"
+version = "0.5.174"
 description = "High-performance Python code flow analysis with optimized TOON format - CFG, DFG, call graphs, and intelligent code queries"
 readme = "README.md"
 requires-python = ">=3.10"

code2llm-0.5.174/tests/test_dup_groups.py

@@ -0,0 +1,80 @@
+"""Regression: duplicate class groups vs pairwise pairs (IFURI-245)."""
+
+from __future__ import annotations
+
+from pathlib import Path
+
+from code2llm.core.models import AnalysisResult, ClassInfo
+from code2llm.exporters.flow_constants import is_intentional_duplicate_copy
+from code2llm.exporters.planfile_tickets import collect_planfile_tickets
+from code2llm.exporters.toon.metrics_duplicates import DuplicatesMetricsComputer
+from code2llm.exporters.toon.metrics_health import HealthMetricsComputer
+
+
+def _worker(file: Path, pkg: str) -> ClassInfo:
+    methods = [f"{pkg}.Worker.load", f"{pkg}.Worker.save", f"{pkg}.Worker.run"]
+    return ClassInfo(
+        name="Worker",
+        qualified_name=f"{pkg}.Worker",
+        file=str(file),
+        line=1,
+        methods=methods,
+    )
+
+
+def test_intentional_copy_paths_are_detected() -> None:
+    assert is_intentional_duplicate_copy("/repo/pc1/net-user-pl/handler.py")
+    assert is_intentional_duplicate_copy("/repo/examples/demo/foo.py")
+    assert is_intentional_duplicate_copy("/repo/urirun-contract-windowpair/x.py")
+    assert not is_intentional_duplicate_copy("/repo/app/src/handler.py")
+
+
+def test_duplicate_groups_not_pair_count(tmp_path: Path) -> None:
+    project = tmp_path
+    result = AnalysisResult(project_path=str(project))
+    for index in range(4):
+        pkg = f"pkg{index}"
+        path = project / f"{pkg}.py"
+        path.write_text("# stub\n", encoding="utf-8")
+        result.classes[f"{pkg}.Worker"] = _worker(path, pkg)
+
+    dupes = DuplicatesMetricsComputer(str(project)).detect_duplicates(result)
+    assert len(dupes) == 6  # C(4,2) pairs for same class name
+
+    ctx = {
+        "duplicates": dupes,
+        "files": {str(project / f"pkg{i}.py"): {"lines": 10} for i in range(4)},
+        "result": result,
+        "health": [],
+        "func_metrics": [],
+    }
+    issues = HealthMetricsComputer().compute_health(ctx)
+    dup_issue = next(i for i in issues if i["code"] == "DUP")
+    assert "1 duplicate class groups" in dup_issue["message"]
+
+
+def test_planfile_dup_ticket_uses_stable_group_dedupe_key(tmp_path: Path) -> None:
+    project = tmp_path
+    result = AnalysisResult(project_path=str(project))
+    for pkg in ("one", "two", "three"):
+        path = project / f"{pkg}.py"
+        path.write_text("# stub\n", encoding="utf-8")
+        result.classes[f"pkg.{pkg}.Worker"] = _worker(path, f"pkg.{pkg}")
+
+    dup_tickets = [t for t in collect_planfile_tickets(result) if t.signal == "code2llm_dup"]
+    assert len(dup_tickets) == 1
+    assert dup_tickets[0].dedupe_key == "code2llm:dup:Worker"
+    assert len(dup_tickets[0].files) == 3
+
+
+def test_examples_dir_excluded_from_duplicate_pairs(tmp_path: Path) -> None:
+    project = tmp_path
+    result = AnalysisResult(project_path=str(project))
+    for pkg in ("app/main", "examples/demo"):
+        path = project / f"{pkg}.py"
+        path.parent.mkdir(parents=True, exist_ok=True)
+        path.write_text("# stub\n", encoding="utf-8")
+        result.classes[f"{pkg}.Worker"] = _worker(path, pkg.replace("/", "."))
+
+    dupes = DuplicatesMetricsComputer(str(project)).detect_duplicates(result)
+    assert dupes == []

code2llm-0.5.174/tests/test_smells_data_clumps.py

@@ -0,0 +1,83 @@
+"""Tests for SmellDetector._detect_data_clumps.
+
+2026-07-06 incident (same root cause as the Shotgun Surgery fix): grouping
+functions by their argument-name set alone, ignoring file, meant a common
+constructor shape like `(self, project_root, config)` -- conventionally
+reused independently by unrelated classes across unrelated files -- falsely
+triggered "Data Clump". Reproduced live:
+`wup.anomaly_detector.AnomalyDetector.__init__` and
+`wup.testql_monitor.TestQLMonitor.__init__` are two unrelated classes in
+unrelated files, flagged as clumped on `project_root, self, config` purely
+because both happen to take a project root and a config object.
+"""
+
+from __future__ import annotations
+
+from code2llm.analysis.smells import SmellDetector
+from code2llm.core.models import AnalysisResult, FunctionInfo
+
+
+def _function(name: str, file: str, args: list[str], line: int = 1) -> FunctionInfo:
+    return FunctionInfo(
+        name=name, qualified_name=f"{file}::{name}", file=file, line=line, args=args
+    )
+
+
+def test_same_args_across_unrelated_files_is_not_a_data_clump():
+    """The core regression: two unrelated classes in unrelated files each
+    taking (self, project_root, config) must NOT be flagged."""
+    result = AnalysisResult()
+    result.functions["a::AnomalyDetector.__init__"] = _function(
+        "AnomalyDetector.__init__", "wup/anomaly_detector.py", ["self", "project_root", "config"]
+    )
+    result.functions["b::TestQLMonitor.__init__"] = _function(
+        "TestQLMonitor.__init__", "wup/testql_monitor.py", ["self", "project_root", "config"]
+    )
+
+    smells = SmellDetector(result)._detect_data_clumps()
+
+    assert smells == []
+
+
+def test_genuine_same_file_data_clump_still_detected():
+    """Two methods in the SAME file sharing a 3+ arg signature is the real
+    smell and must still be caught."""
+    result = AnalysisResult()
+    file = "core/session.py"
+    result.functions["core.session::start"] = _function(
+        "start", file, ["self", "user_id", "token", "scope"]
+    )
+    result.functions["core.session::refresh"] = _function(
+        "refresh", file, ["self", "user_id", "token", "scope"]
+    )
+
+    smells = SmellDetector(result)._detect_data_clumps()
+
+    assert len(smells) == 2
+    assert all(s.name == "Data Clump: self, user_id, token, scope"
+               or set(s.context["clump"]) == {"self", "user_id", "token", "scope"}
+               for s in smells)
+    assert all(s.file == file for s in smells)
+
+
+def test_single_function_with_matching_args_in_other_file_not_flagged():
+    """Only one function per file matches the arg-set -- no clump within
+    either file, and cross-file coincidence must not count."""
+    result = AnalysisResult()
+    result.functions["a::f"] = _function("f", "a.py", ["self", "x", "y", "z"])
+    result.functions["b::g"] = _function("g", "b.py", ["self", "x", "y", "z"])
+
+    smells = SmellDetector(result)._detect_data_clumps()
+
+    assert smells == []
+
+
+def test_fewer_than_three_args_never_counted():
+    result = AnalysisResult()
+    file = "core/session.py"
+    result.functions["core.session::start"] = _function("start", file, ["self", "user_id"])
+    result.functions["core.session::refresh"] = _function("refresh", file, ["self", "user_id"])
+
+    smells = SmellDetector(result)._detect_data_clumps()
+
+    assert smells == []

code2llm-0.5.172/VERSION

@@ -1 +0,0 @@
-0.5.172