sourcecode 1.30.19__tar.gz → 1.30.21__tar.gz

{sourcecode-1.30.19 → sourcecode-1.30.21}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: sourcecode
-Version: 1.30.19
+Version: 1.30.21
 Summary: Deterministic codebase context for AI coding agents
 License:                                  Apache License
                                    Version 2.0, January 2004
@@ -221,7 +221,7 @@ Description-Content-Type: text/markdown
 
 **Deterministic, behavior-aware codebase context for AI agents and PR review.**
 
-![Version](https://img.shields.io/badge/version-1.30.19-blue)
+![Version](https://img.shields.io/badge/version-1.30.21-blue)
 ![Python](https://img.shields.io/badge/python-3.10%2B-green)
 
 ---
@@ -257,7 +257,7 @@ pipx install sourcecode
 
 ```bash
 sourcecode version
-# sourcecode 1.30.19
+# sourcecode 1.30.21
 ```
 
 ---

{sourcecode-1.30.19 → sourcecode-1.30.21}/README.md

@@ -2,7 +2,7 @@
 
 **Deterministic, behavior-aware codebase context for AI agents and PR review.**
 
-![Version](https://img.shields.io/badge/version-1.30.19-blue)
+![Version](https://img.shields.io/badge/version-1.30.21-blue)
 ![Python](https://img.shields.io/badge/python-3.10%2B-green)
 
 ---
@@ -38,7 +38,7 @@ pipx install sourcecode
 
 ```bash
 sourcecode version
-# sourcecode 1.30.19
+# sourcecode 1.30.21
 ```
 
 ---

{sourcecode-1.30.19 → sourcecode-1.30.21}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "hatchling.build"
 
 [project]
 name = "sourcecode"
-version = "1.30.19"
+version = "1.30.21"
 description = "Deterministic codebase context for AI coding agents"
 readme = "README.md"
 requires-python = ">=3.9"

{sourcecode-1.30.19 → sourcecode-1.30.21}/src/sourcecode/__init__.py

@@ -1,3 +1,3 @@
 """sourcecode — Deterministic codebase context maps for AI coding agents."""
 
-__version__ = "1.30.19"
+__version__ = "1.30.21"

{sourcecode-1.30.19 → sourcecode-1.30.21}/src/sourcecode/cli.py

@@ -1596,6 +1596,87 @@ def main(
                 typer.echo("✓ copied to clipboard", err=True)
 
 
+# ── prepare-context output helpers ────────────────────────────────────────────
+
+def _make_explanation(reason: str, why: str) -> str:
+    """Merge reason+why into one human string. Drop internal pipe-format jargon."""
+    if why and not why.startswith("artifact_type:"):
+        return why
+    return reason
+
+
+def _serialize_relevant_file(f: Any) -> dict:
+    from dataclasses import asdict as _asdict
+    d = {k: v for k, v in _asdict(f).items() if v != "" and v is not None}
+    reason = d.pop("reason", "") or ""
+    why = d.pop("why", "") or ""
+    explanation = _make_explanation(reason, why)
+    if explanation:
+        d["explanation"] = explanation
+    if reason:
+        d["reason"] = reason
+    return d
+
+
+def _transform_impact_scores(scores: dict) -> dict:
+    """Convert internal impact_score_per_file to human-readable review signals."""
+    _THRESHOLDS = [(0.60, "high"), (0.40, "medium")]
+    _CT_LABELS: dict[str, str] = {
+        "security_change":    "security-sensitive change",
+        "behavioral_change":  "behavioral change",
+        "structural_change":  "structural modification",
+        "configuration_change": "configuration change",
+        "dependency_change":  "dependency update",
+        "ui_change":          "UI change",
+    }
+    result: dict = {}
+    for path, entry in scores.items():
+        raw = entry.get("_rank_score", 0.0)
+        priority = next((p for thresh, p in _THRESHOLDS if raw >= thresh), "low")
+        signals: list[str] = [
+            _CT_LABELS[ct] for ct in entry.get("change_types", []) if ct in _CT_LABELS
+        ]
+        ev = entry.get("evidence", {})
+        callers = ev.get("reverse_edge_count", 0)
+        if callers:
+            signals.append(f"imported by {callers} other file{'s' if callers > 1 else ''}")
+        out_entry: dict = {"review_priority": priority}
+        if signals:
+            out_entry["signals"] = signals
+        result[path] = out_entry
+    return result
+
+
+def _clean_system_impact(si: dict) -> dict:
+    """Remove empty lists/dicts from system_impact — only emit positive signals."""
+    return {k: v for k, v in si.items() if v}
+
+
+def _build_human_summary(output: Any) -> dict:
+    """Human-readable summary block for delta/review-pr output."""
+    concerns: list[str] = []
+    si = getattr(output, "system_impact", {}) or {}
+    for ri in si.get("runtime_impact", []):
+        sig = ri.get("signal", "")
+        if sig:
+            concerns.append(sig)
+    for bc in si.get("behavioral_changes", []):
+        stmt = bc.get("statement", "")
+        if stmt:
+            concerns.append(stmt)
+    cov = getattr(output, "test_coverage_risk", {}) or {}
+    missing = cov.get("changed_files_without_tests", [])
+    if missing:
+        concerns.append(f"{len(missing)} changed file(s) without test coverage detected")
+    summary: dict = {}
+    if getattr(output, "impact_summary", None):
+        summary["description"] = output.impact_summary
+    if concerns:
+        summary["review_concerns"] = concerns[:6]
+    summary["confidence"] = (getattr(output, "confidence", None) or "medium").upper()
+    return summary
+
+
 @app.command("prepare-context")
 def prepare_context_cmd(
     task: Optional[str] = typer.Argument(
@@ -1723,8 +1804,16 @@ def prepare_context_cmd(
     import time as _time
 
     builder = TaskContextBuilder(target)
+    _progress = Progress()
+    _phase = f"analyzing ({task})"
+    if since:
+        _phase += f" since {since}"
+    _progress.start(_phase)
     _t0 = _time.perf_counter()
-    output = builder.build(task, since=since, symptom=symptom)
+    try:
+        output = builder.build(task, since=since, symptom=symptom)
+    finally:
+        _progress.finish()
     _t_total = (_time.perf_counter() - _t0) * 1000
 
     if debug_perf:
@@ -1805,7 +1894,7 @@ def prepare_context_cmd(
         out["confidence"] = output.confidence
     if task != "review-pr" and _task_include("relevant_files"):
         out["relevant_files"] = [
-            {k: v for k, v in asdict(f).items() if v != ""}
+            _serialize_relevant_file(f)
             for f in output.relevant_files
         ]
     if _task_include("key_dependencies") and output.key_dependencies:
@@ -1856,16 +1945,19 @@ def prepare_context_cmd(
             out["affected_modules"] = output.affected_modules
         if output.risk_areas:
             out["risk_areas"] = output.risk_areas
-        if output.why_these_files:
-            out["reasoning"] = output.why_these_files
         if output.change_type:
             out["change_type"] = output.change_type
         if output.system_impact:
-            out["system_impact"] = output.system_impact
+            _si = _clean_system_impact(output.system_impact)
+            if _si:
+                out["system_impact"] = _si
         if output.dependency_graph_summary:
-            out["dependency_graph_summary"] = output.dependency_graph_summary
+            _dgraph = {k: v for k, v in output.dependency_graph_summary.items() if v is not None}
+            if _dgraph:
+                out["dependency_graph_summary"] = _dgraph
         if output.impact_score_per_file:
-            out["impact_score_per_file"] = output.impact_score_per_file
+            out["impact_score_per_file"] = _transform_impact_scores(output.impact_score_per_file)
+        out["summary"] = _build_human_summary(output)
     # review-pr specific fields
     if task == "review-pr":
         if output.error_code:
@@ -1920,8 +2012,7 @@ def prepare_context_cmd(
             out["behavioral_impact"] = output.behavioral_impact
         if output.impact_summary:
             out["impact_summary"] = output.impact_summary
-        if output.why_these_files:
-            out["reasoning"] = output.why_these_files
+        out["summary"] = _build_human_summary(output)
         # git-first scope metadata
         out["scope"] = {
             "source": output.scope_source or "git_diff",
@@ -2105,7 +2196,15 @@ def repo_ir_cmd(
         )
         return
 
-    ir = build_repo_ir(file_list, root, since=since)
+    _ir_phase = f"extracting IR ({len(file_list)} files)"
+    if since:
+        _ir_phase += f" since {since}"
+    _ir_progress = Progress()
+    _ir_progress.start(_ir_phase)
+    try:
+        ir = build_repo_ir(file_list, root, since=since)
+    finally:
+        _ir_progress.finish()
     ir = apply_ir_size_limits(
         ir,
         max_nodes=max_nodes,

{sourcecode-1.30.19 → sourcecode-1.30.21}/src/sourcecode/prepare_context.py

@@ -556,7 +556,7 @@ _ARTIFACT_CHANGE_EFFECT: dict[str, str] = {
     "test":           "test file (no production code modified)",
     "documentation":  "documentation file (no runtime impact)",
     "ide_noise":      "IDE/tooling artifact (no application impact)",
-    "source":         "application source (artifact role requires annotation inspection)",
+    "source":         "application source file (role could not be confirmed from code signals)",
 }
 
 # Maps frontend symptom keywords → backend terms likely to contain the root cause.
@@ -1156,22 +1156,24 @@ class TaskContextBuilder:
                     else "unknown"
                 )
                 _impact_entry = _delta_impact_score_per_file.get(_f) or {}
-                _entry = {
+                _ev_complete = {k: v for k, v in _impact_entry.get("evidence", {}).items() if v}
+                _entry: dict = {
                     "path": _f,
                     "diff_source": _diff_source,
                     "role": _role_obj,
                     "artifact_type": _f_atype,
                     "evidence": _evidence,
                     "change_effect": {
-                        "statement": _ARTIFACT_CHANGE_EFFECT.get(_f_atype, "application source (artifact role requires annotation inspection)"),
+                        "statement": _ARTIFACT_CHANGE_EFFECT.get(_f_atype, "application source file (role could not be confirmed)"),
                         "classification_method": _f_cls.get("confidence", "low"),
                         "epistemic_level": (
                             "STRUCTURAL SIGNAL" if _f_cls.get("confidence", "low") == "high"
                             else "INFERRED (LOW CONFIDENCE)"
                         ),
                     },
-                    "evidence_completeness": _impact_entry.get("evidence", {}),
                 }
+                if _ev_complete:
+                    _entry["evidence_completeness"] = _ev_complete
                 _pr_runtime_changes.append(_entry)
                 # Split committed vs uncommitted — never merged
                 if _f in _committed_set:
@@ -2992,12 +2994,12 @@ class TaskContextBuilder:
 
         # ── Step 6: analysis gaps ──────────────────────────────────────────────
         _bfs_note = (
-            f"BFS multi-hop import propagation repo-wide (propagation_depth={_max_hop})"
+            f"multi-hop import tracing performed (depth {_max_hop})"
             if _max_hop >= 1
-            else "no import-link propagation (no scannable changed files)"
+            else "no downstream dependencies could be verified"
         )
         analysis_gaps: list[str] = [
-            f"Related file expansion: type-aware chain expansion + {_bfs_note} + module/directory heuristics",
+            f"Related file expansion via import analysis and directory proximity ({_bfs_note})",
         ]
         if _expansion_truncated:
             analysis_gaps.insert(0,
@@ -3015,10 +3017,8 @@ class TaskContextBuilder:
         low_confidence = [f for f, cls in classifications.items() if cls.get("confidence") == "low" and not cls["is_noise"]]
         if low_confidence:
             analysis_gaps.append(
-                f"{len(low_confidence)} file(s) classified with low confidence"
-                " (artifact type inferred from extension only)"
-                " — consider adding stem patterns to _classify_changed_file: "
-                + ", ".join(Path(f).name for f in low_confidence[:3])
+                f"{len(low_confidence)} file(s) role inferred from filename only"
+                f" (no code signal found): {', '.join(Path(f).name for f in low_confidence[:3])}"
             )
         if not affected_modules_set and any(not cls["is_noise"] for cls in classifications.values()):
             analysis_gaps.append(

sourcecode-1.30.21/tests/test_output_ux.py

@@ -0,0 +1,235 @@
+"""test_output_ux.py — Verify output UX: no internal jargon, no redundancy, correct structure.
+
+Acceptance criteria (from spec):
+1. impact_score_per_file: no _rank_score, no has_* booleans
+2. relevant_files: no 'why' key; 'explanation' present when meaningful
+3. reasoning block absent from delta/review-pr output
+4. system_impact: no empty arrays
+5. human summary (summary) present in delta/review-pr success output
+6. analysis_gaps: no pipeline jargon ("BFS", "type-aware chain", "import-link propagation")
+7. Output size ≥ 35% smaller than internal-dump baseline for impact_score_per_file
+"""
+from __future__ import annotations
+
+import json
+from pathlib import Path
+from typing import Any
+
+import pytest
+from typer.testing import CliRunner
+
+from sourcecode.cli import app
+from sourcecode import prepare_context as _pc
+
+runner = CliRunner()
+
+FIXTURE = Path(__file__).parent / "fixtures" / "fastapi_app"
+
+
+def _invoke(*args: str) -> Any:
+    return runner.invoke(app, list(args))
+
+
+def _json(result: Any) -> dict:
+    return json.loads(result.output)
+
+
+# ── helpers ───────────────────────────────────────────────────────────────────
+
+def _make_delta_output(monkeypatch, changed: list[str]) -> dict:
+    """Return the parsed JSON output for a delta run with mocked changed files."""
+    monkeypatch.setattr(
+        _pc.TaskContextBuilder,
+        "_resolve_git_root",
+        lambda self: FIXTURE,
+    )
+    result = _invoke("prepare-context", "delta", str(FIXTURE), "--since", "HEAD~1")
+    assert result.exit_code in (0, 1), result.output
+    return _json(result)
+
+
+def _make_review_pr_output(monkeypatch) -> dict:
+    """Return parsed JSON output for review-pr with a mocked changed controller."""
+    changed = ["src/main/java/com/example/UserController.java"]
+    monkeypatch.setattr(
+        _pc.TaskContextBuilder,
+        "_resolve_git_root",
+        lambda self: FIXTURE,
+    )
+    monkeypatch.setattr(
+        _pc.TaskContextBuilder,
+        "_get_pr_scope_files",
+        lambda self, since=None: (changed, "git_diff", changed, []),
+    )
+    result = _invoke("prepare-context", "review-pr", str(FIXTURE))
+    assert result.exit_code == 0, result.output
+    return _json(result)
+
+
+# ── impact_score_per_file structure ──────────────────────────────────────────
+
+class TestImpactScoreOutput:
+    """impact_score_per_file must expose review_priority+signals, not internal scores."""
+
+    def test_no_rank_score_in_output(self, monkeypatch):
+        data = _make_review_pr_output(monkeypatch)
+        for entry in data.get("impact_score_per_file", {}).values():
+            assert "_rank_score" not in entry, f"_rank_score leaked into output: {entry}"
+
+    def test_no_boolean_evidence_spam(self, monkeypatch):
+        data = _make_review_pr_output(monkeypatch)
+        for path, entry in data.get("impact_score_per_file", {}).items():
+            ev = entry.get("evidence", {})
+            for k, v in ev.items():
+                assert v is not False, f"{path}: false evidence key '{k}' should be omitted"
+            assert "has_reverse_edges" not in entry, f"boolean key leaked: {path}"
+            assert "has_route_diff" not in entry, f"boolean key leaked: {path}"
+            assert "has_security_diff" not in entry, f"boolean key leaked: {path}"
+            assert "has_wiring_evidence" not in entry, f"boolean key leaked: {path}"
+
+    def test_review_priority_field_present(self, monkeypatch):
+        data = _make_review_pr_output(monkeypatch)
+        for path, entry in data.get("impact_score_per_file", {}).items():
+            assert "review_priority" in entry, f"review_priority missing for {path}"
+            assert entry["review_priority"] in ("high", "medium", "low"), (
+                f"unexpected priority value for {path}: {entry['review_priority']}"
+            )
+
+    def test_size_reduction_vs_internal(self, monkeypatch):
+        """Transformed impact_score_per_file must be ≥35% smaller in bytes."""
+        from sourcecode.cli import _transform_impact_scores
+        internal = {
+            "src/Foo.java": {
+                "_rank_score": 0.325,
+                "change_types": ["behavioral_change"],
+                "diff_severity": "api_change",
+                "evidence": {
+                    "has_reverse_edges": False,
+                    "reverse_edge_count": 0,
+                    "has_route_diff": True,
+                    "has_security_diff": False,
+                    "has_wiring_evidence": False,
+                },
+            }
+        }
+        before = len(json.dumps(internal))
+        after = len(json.dumps(_transform_impact_scores(internal)))
+        reduction = (before - after) / before
+        assert reduction >= 0.35, (
+            f"Expected ≥35% size reduction, got {reduction:.0%}. "
+            f"before={before}B after={after}B"
+        )
+
+
+# ── relevant_files structure ──────────────────────────────────────────────────
+
+class TestRelevantFilesOutput:
+    """relevant_files must not expose 'why' jargon field; explanation replaces it."""
+
+    def test_no_why_field_in_output(self):
+        result = _invoke("prepare-context", "explain", str(FIXTURE))
+        assert result.exit_code == 0, result.output
+        data = _json(result)
+        for rf in data.get("relevant_files", []):
+            assert "why" not in rf, (
+                f"'why' field leaked into output for {rf.get('path')}: {rf}"
+            )
+
+    def test_explanation_or_reason_present(self):
+        result = _invoke("prepare-context", "explain", str(FIXTURE))
+        assert result.exit_code == 0, result.output
+        data = _json(result)
+        for rf in data.get("relevant_files", []):
+            has_info = "explanation" in rf or "reason" in rf
+            assert has_info, f"No explanation/reason in relevant_file: {rf}"
+
+
+# ── reasoning block removed ───────────────────────────────────────────────────
+
+class TestReasoningBlockRemoved:
+    """delta and review-pr must not emit 'reasoning' block (redundant with relevant_files)."""
+
+    def test_delta_no_reasoning_key(self):
+        result = _invoke("prepare-context", "delta", str(FIXTURE))
+        assert result.exit_code in (0, 1), result.output
+        data = _json(result)
+        if data.get("task") == "delta":
+            assert "reasoning" not in data, (
+                "'reasoning' block still in delta output — should be removed (duplicate)"
+            )
+
+    def test_review_pr_no_reasoning_key(self, monkeypatch):
+        data = _make_review_pr_output(monkeypatch)
+        assert "reasoning" not in data, (
+            "'reasoning' block still in review-pr output — should be removed (duplicate)"
+        )
+
+
+# ── system_impact clean ───────────────────────────────────────────────────────
+
+class TestSystemImpactClean:
+    """system_impact must not contain empty arrays."""
+
+    def test_no_empty_arrays_in_system_impact(self):
+        result = _invoke("prepare-context", "delta", str(FIXTURE))
+        assert result.exit_code in (0, 1), result.output
+        data = _json(result)
+        si = data.get("system_impact", {})
+        for k, v in si.items():
+            assert v != [], f"system_impact['{k}'] is empty list — should be omitted"
+            assert v != {}, f"system_impact['{k}'] is empty dict — should be omitted"
+
+
+# ── human summary present ─────────────────────────────────────────────────────
+
+class TestHumanSummary:
+    """delta and review-pr success output must include a 'summary' block."""
+
+    def test_review_pr_has_summary(self, monkeypatch):
+        data = _make_review_pr_output(monkeypatch)
+        assert "summary" in data, "review-pr output missing 'summary' block"
+        s = data["summary"]
+        assert "confidence" in s, f"summary missing confidence field: {s}"
+        assert s["confidence"] in ("HIGH", "MEDIUM", "LOW"), (
+            f"confidence has unexpected value: {s['confidence']}"
+        )
+
+
+# ── analysis_gaps language ────────────────────────────────────────────────────
+
+class TestAnalysisGapsLanguage:
+    """analysis_gaps must not contain pipeline-internal jargon."""
+
+    _JARGON = [
+        "BFS",
+        "type-aware chain expansion",
+        "import-link propagation",
+        "propagation_depth=",
+        "artifact role requires annotation inspection",
+    ]
+
+    def test_no_jargon_in_analysis_gaps(self):
+        result = _invoke("prepare-context", "delta", str(FIXTURE))
+        assert result.exit_code in (0, 1), result.output
+        data = _json(result)
+        for gap in data.get("gaps", []):
+            for jargon in self._JARGON:
+                assert jargon not in gap, (
+                    f"Internal jargon '{jargon}' found in analysis_gaps: {gap!r}"
+                )
+
+
+# ── dependency_graph_summary clean ───────────────────────────────────────────
+
+class TestDependencyGraphSummaryClean:
+    """dependency_graph_summary must not emit null fields."""
+
+    def test_no_null_propagation_depth(self):
+        result = _invoke("prepare-context", "delta", str(FIXTURE))
+        assert result.exit_code in (0, 1), result.output
+        data = _json(result)
+        dgraph = data.get("dependency_graph_summary", {})
+        for k, v in dgraph.items():
+            assert v is not None, (
+                f"dependency_graph_summary['{k}'] is null — should be omitted"
+            )