superlocalmemory 3.3.19 → 3.3.21

package/src/superlocalmemory/code_graph/changes.py

@@ -0,0 +1,363 @@
+# Copyright (c) 2026 Varun Pratap Bhardwaj / Qualixar
+# Licensed under the MIT License - see LICENSE file
+# Part of SuperLocalMemory v3.4 — CodeGraph Module
+
+"""ChangeAnalyzer — git diff to risk-scored change analysis.
+
+Parses git diff output, maps changed line ranges to graph nodes,
+computes 5-factor risk scores, and produces review context.
+"""
+
+from __future__ import annotations
+
+import logging
+import re
+import subprocess
+from dataclasses import dataclass
+from pathlib import Path
+from typing import Any
+
+from superlocalmemory.code_graph.database import CodeGraphDatabase
+from superlocalmemory.code_graph.models import EdgeKind
+
+logger = logging.getLogger(__name__)
+
+
+# ---------------------------------------------------------------------------
+# Constants
+# ---------------------------------------------------------------------------
+
+SECURITY_KEYWORDS: frozenset[str] = frozenset([
+    "auth", "login", "password", "token", "session", "crypt", "secret",
+    "credential", "permission", "sql", "query", "execute", "connect",
+    "socket", "request", "http", "sanitize", "validate", "encrypt",
+    "decrypt", "hash", "sign", "verify", "admin", "privilege",
+])
+
+_GIT_TIMEOUT_SECONDS = 30
+_HUNK_HEADER_RE = re.compile(r'\+(\d+)(?:,(\d+))?')
+
+
+# ---------------------------------------------------------------------------
+# Result dataclasses
+# ---------------------------------------------------------------------------
+
+@dataclass(frozen=True)
+class DiffHunk:
+    """A single changed line range in a file."""
+    file_path: str
+    start_line: int
+    end_line: int
+
+
+@dataclass(frozen=True)
+class ChangedNode:
+    """A graph node affected by a change, with risk score."""
+    node_id: str
+    name: str
+    kind: str
+    file_path: str
+    line_start: int
+    line_end: int
+    risk_score: float
+
+
+@dataclass(frozen=True)
+class ReviewContext:
+    """Token-optimized review context for changed files."""
+    summary: str
+    changed_nodes: tuple[ChangedNode, ...]
+    test_gaps: tuple[ChangedNode, ...]
+    review_priorities: tuple[ChangedNode, ...]
+    overall_risk: float
+
+
+# ---------------------------------------------------------------------------
+# ChangeAnalyzer
+# ---------------------------------------------------------------------------
+
+class ChangeAnalyzer:
+    """Analyze git changes and map to graph nodes with risk scores.
+
+    All git operations use subprocess with timeout for safety.
+    """
+
+    def __init__(self, db: CodeGraphDatabase) -> None:
+        self._db = db
+
+    # ------------------------------------------------------------------
+    # Public API
+    # ------------------------------------------------------------------
+
+    def analyze_changes(
+        self, changed_files: list[str], repo_root: Path | None = None
+    ) -> ReviewContext:
+        """Analyze changed files and produce risk-scored review context.
+
+        Args:
+            changed_files: List of relative file paths that changed.
+            repo_root: Optional repo root for git diff parsing.
+
+        Returns:
+            ReviewContext with scored nodes and review priorities.
+        """
+        if not changed_files:
+            return ReviewContext(
+                summary="No changes detected.",
+                changed_nodes=(),
+                test_gaps=(),
+                review_priorities=(),
+                overall_risk=0.0,
+            )
+
+        # Map changed files to affected nodes
+        all_changed: dict[str, ChangedNode] = {}
+        for file_path in changed_files:
+            nodes = self._find_nodes_in_file(file_path)
+            for node in nodes:
+                risk = self.compute_risk_score(node)
+                scored_node = ChangedNode(
+                    node_id=node["node_id"],
+                    name=node["name"],
+                    kind=node["kind"],
+                    file_path=node["file_path"],
+                    line_start=node["line_start"],
+                    line_end=node["line_end"],
+                    risk_score=risk,
+                )
+                all_changed[node["node_id"]] = scored_node
+
+        changed_list = tuple(sorted(
+            all_changed.values(), key=lambda n: -n.risk_score
+        ))
+
+        # Find test gaps (changed non-test nodes without TESTED_BY)
+        test_gaps = tuple(
+            n for n in changed_list
+            if n.kind not in ("file", "module")
+            and not self._has_test_coverage(n.node_id)
+        )
+
+        # Top review priorities
+        review_priorities = changed_list[:10]
+
+        # Overall risk
+        overall_risk = max(
+            (n.risk_score for n in changed_list), default=0.0
+        )
+
+        summary = (
+            f"{len(changed_list)} changed nodes across "
+            f"{len(changed_files)} files. "
+            f"{len(test_gaps)} untested changes. "
+            f"Overall risk: {overall_risk:.2f}."
+        )
+
+        return ReviewContext(
+            summary=summary,
+            changed_nodes=changed_list,
+            test_gaps=test_gaps,
+            review_priorities=review_priorities,
+            overall_risk=overall_risk,
+        )
+
+    def compute_risk_score(self, node: dict[str, Any]) -> float:
+        """5-factor risk scoring for a single node.
+
+        Factors:
+        1. flow_participation (max 0.25)
+        2. community_crossing (max 0.15)
+        3. test_coverage (0.05 if tested, 0.30 if untested)
+        4. security_keywords (0 or 0.20)
+        5. caller_count (max 0.10)
+        """
+        node_id = node["node_id"]
+        name = node.get("name", "")
+
+        # 1. Flow participation
+        flow_score = self._flow_participation_score(node_id)
+
+        # 2. Community crossing
+        cross_score = self._community_crossing_score(node_id)
+
+        # 3. Test coverage
+        has_test = self._has_test_coverage(node_id)
+        test_score = 0.05 if has_test else 0.30
+
+        # 4. Security sensitivity
+        security_score = (
+            0.20
+            if any(kw in name.lower() for kw in SECURITY_KEYWORDS)
+            else 0.0
+        )
+
+        # 5. Caller count
+        caller_score = self._caller_count_score(node_id)
+
+        return flow_score + cross_score + test_score + security_score + caller_score
+
+    def get_review_context(
+        self, changed_files: list[str], repo_root: Path | None = None
+    ) -> ReviewContext:
+        """Token-optimized review context.
+
+        Same as analyze_changes but designed for LLM consumption.
+        """
+        return self.analyze_changes(changed_files, repo_root)
+
+    # ------------------------------------------------------------------
+    # Static: git diff parsing
+    # ------------------------------------------------------------------
+
+    @staticmethod
+    def parse_git_diff(
+        repo_root: Path, base: str = "HEAD~1", timeout: int = _GIT_TIMEOUT_SECONDS
+    ) -> list[DiffHunk]:
+        """Parse git diff to get changed line ranges.
+
+        Args:
+            repo_root: Path to the git repository root.
+            base: Git ref to diff against.
+            timeout: Subprocess timeout in seconds.
+
+        Returns:
+            List of DiffHunk with file paths and line ranges.
+        """
+        try:
+            result = subprocess.run(
+                ["git", "diff", "--unified=0", base, "--"],
+                capture_output=True,
+                text=True,
+                cwd=str(repo_root),
+                timeout=timeout,
+            )
+        except (subprocess.TimeoutExpired, FileNotFoundError, OSError) as exc:
+            logger.warning("git diff failed: %s", exc)
+            return []
+
+        if result.returncode != 0:
+            logger.debug("git diff returned %d: %s", result.returncode, result.stderr)
+            return []
+
+        return _parse_diff_output(result.stdout)
+
+    # ------------------------------------------------------------------
+    # Internal helpers
+    # ------------------------------------------------------------------
+
+    def _find_nodes_in_file(self, file_path: str) -> list[dict[str, Any]]:
+        """Find all graph nodes in a file."""
+        rows = self._db.execute(
+            """SELECT node_id, name, kind, file_path, line_start, line_end
+               FROM graph_nodes
+               WHERE file_path = ?
+               ORDER BY line_start""",
+            (file_path,),
+        )
+        return [dict(row) for row in rows]
+
+    def _has_test_coverage(self, node_id: str) -> bool:
+        """Check if a node has TESTED_BY edges or is called by test nodes."""
+        # Check outgoing TESTED_BY
+        rows = self._db.execute(
+            """SELECT COUNT(*) as cnt FROM graph_edges
+               WHERE source_node_id = ? AND kind = ?""",
+            (node_id, EdgeKind.TESTED_BY.value),
+        )
+        if rows and rows[0]["cnt"] > 0:
+            return True
+
+        # Check incoming CALLS from test nodes
+        rows = self._db.execute(
+            """SELECT COUNT(*) as cnt FROM graph_edges ge
+               JOIN graph_nodes gn ON ge.source_node_id = gn.node_id
+               WHERE ge.target_node_id = ?
+                 AND ge.kind = ?
+                 AND gn.is_test = 1""",
+            (node_id, EdgeKind.CALLS.value),
+        )
+        return bool(rows and rows[0]["cnt"] > 0)
+
+    def _flow_participation_score(self, node_id: str) -> float:
+        """Score based on how many flows this node participates in.
+
+        Uses stored flows from graph_metadata. Max 0.25.
+        """
+        import json
+        raw = self._db.get_metadata("flows")
+        if not raw:
+            return 0.0
+        try:
+            flows = json.loads(raw)
+        except (json.JSONDecodeError, TypeError):
+            return 0.0
+
+        count = sum(
+            1 for f in flows
+            if node_id in f.get("path_node_ids", [])
+        )
+        return min(count * 0.05, 0.25)
+
+    def _community_crossing_score(self, node_id: str) -> float:
+        """Score based on cross-community callers. Max 0.15."""
+        # Get this node's community
+        node_rows = self._db.execute(
+            "SELECT community_id FROM graph_nodes WHERE node_id = ?",
+            (node_id,),
+        )
+        if not node_rows:
+            return 0.0
+        my_community = node_rows[0]["community_id"]
+
+        # Get callers and their communities
+        callers = self._db.execute(
+            """SELECT gn.community_id
+               FROM graph_edges ge
+               JOIN graph_nodes gn ON ge.source_node_id = gn.node_id
+               WHERE ge.target_node_id = ?
+                 AND ge.kind = ?""",
+            (node_id, EdgeKind.CALLS.value),
+        )
+        cross_count = sum(
+            1 for row in callers
+            if row["community_id"] is not None
+            and row["community_id"] != my_community
+        )
+        return min(cross_count * 0.05, 0.15)
+
+    def _caller_count_score(self, node_id: str) -> float:
+        """Score based on number of callers. Max 0.10."""
+        rows = self._db.execute(
+            """SELECT COUNT(*) as cnt FROM graph_edges
+               WHERE target_node_id = ? AND kind = ?""",
+            (node_id, EdgeKind.CALLS.value),
+        )
+        count = rows[0]["cnt"] if rows else 0
+        return min(count / 20.0, 0.10)
+
+
+# ---------------------------------------------------------------------------
+# Module-level helpers
+# ---------------------------------------------------------------------------
+
+def _parse_diff_output(diff_text: str) -> list[DiffHunk]:
+    """Parse unified diff output into DiffHunk list."""
+    hunks: list[DiffHunk] = []
+    current_file: str | None = None
+
+    for line in diff_text.splitlines():
+        if line.startswith("+++ b/"):
+            current_file = line[6:]
+        elif line.startswith("@@"):
+            match = _HUNK_HEADER_RE.search(line)
+            if match and current_file:
+                start = int(match.group(1))
+                count = int(match.group(2) or "1")
+                if count > 0:
+                    hunks.append(DiffHunk(
+                        file_path=current_file,
+                        start_line=start,
+                        end_line=start + count - 1,
+                    ))
+
+    return hunks

package/src/superlocalmemory/code_graph/communities.py

@@ -0,0 +1,299 @@
+# Copyright (c) 2026 Varun Pratap Bhardwaj / Qualixar
+# Licensed under the MIT License - see LICENSE file
+# Part of SuperLocalMemory v3.4 — CodeGraph Module
+
+"""CommunityDetector — file-based community detection.
+
+Groups nodes by file path prefix / directory for MVP.
+igraph/Leiden can be added later for more sophisticated detection.
+Stores community_id on graph_nodes via UPDATE.
+"""
+
+from __future__ import annotations
+
+import json
+import logging
+from collections import Counter, defaultdict
+from dataclasses import dataclass
+from pathlib import PurePosixPath
+from typing import Any
+
+from superlocalmemory.code_graph.database import CodeGraphDatabase
+
+logger = logging.getLogger(__name__)
+
+
+# ---------------------------------------------------------------------------
+# Result dataclasses
+# ---------------------------------------------------------------------------
+
+@dataclass(frozen=True)
+class CommunityInfo:
+    """Detected code community."""
+    community_id: int
+    name: str
+    directory: str
+    size: int
+    dominant_language: str | None
+    file_count: int
+    cohesion: float
+    node_ids: tuple[str, ...]
+
+
+@dataclass(frozen=True)
+class CouplingWarning:
+    """Warning about high coupling between communities."""
+    source_community: str
+    target_community: str
+    edge_count: int
+    severity: str  # "low", "medium", "high"
+
+
+@dataclass(frozen=True)
+class ArchitectureOverview:
+    """Architecture overview with communities and coupling warnings."""
+    communities: tuple[CommunityInfo, ...]
+    coupling_warnings: tuple[CouplingWarning, ...]
+    total_nodes: int
+    total_communities: int
+
+
+# ---------------------------------------------------------------------------
+# CommunityDetector
+# ---------------------------------------------------------------------------
+
+class CommunityDetector:
+    """File-based community detection.
+
+    Groups nodes by directory (file path prefix).
+    Each unique directory becomes a community.
+    """
+
+    def __init__(self, db: CodeGraphDatabase) -> None:
+        self._db = db
+
+    # ------------------------------------------------------------------
+    # Public API
+    # ------------------------------------------------------------------
+
+    def detect_communities(self) -> list[CommunityInfo]:
+        """Detect communities by grouping nodes by directory.
+
+        Updates community_id on graph_nodes.
+
+        Returns:
+            List of CommunityInfo sorted by size (largest first).
+        """
+        # Load all nodes
+        rows = self._db.execute(
+            "SELECT node_id, name, kind, file_path, language FROM graph_nodes",
+            (),
+        )
+        if not rows:
+            return []
+
+        # Group by directory
+        dir_groups: dict[str, list[dict[str, Any]]] = defaultdict(list)
+        for row in rows:
+            directory = _extract_directory(row["file_path"])
+            dir_groups[directory].append(dict(row))
+
+        # Build communities
+        communities: list[CommunityInfo] = []
+        for comm_id, (directory, nodes) in enumerate(sorted(dir_groups.items())):
+            node_ids = tuple(n["node_id"] for n in nodes)
+            languages = [n["language"] for n in nodes if n.get("language")]
+            dominant_lang = _dominant_language(languages)
+            file_paths = {n["file_path"] for n in nodes}
+
+            community = CommunityInfo(
+                community_id=comm_id,
+                name=_generate_community_name(directory, nodes),
+                directory=directory,
+                size=len(nodes),
+                dominant_language=dominant_lang,
+                file_count=len(file_paths),
+                cohesion=1.0,  # Trivially cohesive for file-based
+                node_ids=node_ids,
+            )
+            communities.append(community)
+
+        # Update community_id on nodes
+        self._update_node_communities(communities)
+
+        # Sort by size descending
+        communities.sort(key=lambda c: -c.size)
+
+        # Store in metadata
+        self._store_communities(communities)
+
+        return communities
+
+    def get_architecture_overview(self) -> ArchitectureOverview:
+        """Return community summary with coupling warnings.
+
+        Returns:
+            ArchitectureOverview with communities and cross-community coupling.
+        """
+        # Load or detect communities
+        communities = self._load_communities()
+        if not communities:
+            communities = self.detect_communities()
+
+        # Compute coupling warnings
+        warnings = self._compute_coupling_warnings(communities)
+
+        total_nodes = sum(c.size for c in communities)
+
+        return ArchitectureOverview(
+            communities=tuple(communities),
+            coupling_warnings=tuple(warnings),
+            total_nodes=total_nodes,
+            total_communities=len(communities),
+        )
+
+    # ------------------------------------------------------------------
+    # Internal helpers
+    # ------------------------------------------------------------------
+
+    def _update_node_communities(
+        self, communities: list[CommunityInfo]
+    ) -> None:
+        """Update community_id on graph_nodes."""
+        for community in communities:
+            if not community.node_ids:
+                continue
+            placeholders = ",".join("?" for _ in community.node_ids)
+            self._db.execute_write(
+                f"""UPDATE graph_nodes SET community_id = ?
+                    WHERE node_id IN ({placeholders})""",
+                (community.community_id, *community.node_ids),
+            )
+
+    def _compute_coupling_warnings(
+        self, communities: list[CommunityInfo]
+    ) -> list[CouplingWarning]:
+        """Find cross-community edges and generate coupling warnings."""
+        # Build node_id -> community_name map
+        node_to_community: dict[str, str] = {}
+        for comm in communities:
+            for nid in comm.node_ids:
+                node_to_community[nid] = comm.name
+
+        # Count cross-community edges
+        cross_edges: dict[tuple[str, str], int] = defaultdict(int)
+
+        edges = self._db.execute(
+            "SELECT source_node_id, target_node_id FROM graph_edges",
+            (),
+        )
+        for row in edges:
+            src_comm = node_to_community.get(row["source_node_id"])
+            tgt_comm = node_to_community.get(row["target_node_id"])
+            if src_comm and tgt_comm and src_comm != tgt_comm:
+                pair = (
+                    min(src_comm, tgt_comm),
+                    max(src_comm, tgt_comm),
+                )
+                cross_edges[pair] += 1
+
+        # Generate warnings
+        warnings: list[CouplingWarning] = []
+        for (src, tgt), count in sorted(
+            cross_edges.items(), key=lambda x: -x[1]
+        ):
+            if count >= 10:
+                severity = "high"
+            elif count >= 5:
+                severity = "medium"
+            else:
+                severity = "low"
+
+            warnings.append(CouplingWarning(
+                source_community=src,
+                target_community=tgt,
+                edge_count=count,
+                severity=severity,
+            ))
+
+        return warnings
+
+    def _store_communities(self, communities: list[CommunityInfo]) -> None:
+        """Store communities in graph_metadata as JSON."""
+        data = [
+            {
+                "community_id": c.community_id,
+                "name": c.name,
+                "directory": c.directory,
+                "size": c.size,
+                "dominant_language": c.dominant_language,
+                "file_count": c.file_count,
+                "cohesion": c.cohesion,
+                "node_ids": list(c.node_ids),
+            }
+            for c in communities
+        ]
+        self._db.set_metadata("communities", json.dumps(data))
+
+    def _load_communities(self) -> list[CommunityInfo]:
+        """Load communities from graph_metadata."""
+        raw = self._db.get_metadata("communities")
+        if not raw:
+            return []
+        try:
+            data = json.loads(raw)
+        except (json.JSONDecodeError, TypeError):
+            return []
+
+        return [
+            CommunityInfo(
+                community_id=c["community_id"],
+                name=c["name"],
+                directory=c["directory"],
+                size=c["size"],
+                dominant_language=c.get("dominant_language"),
+                file_count=c["file_count"],
+                cohesion=c["cohesion"],
+                node_ids=tuple(c["node_ids"]),
+            )
+            for c in data
+        ]
+
+
+# ---------------------------------------------------------------------------
+# Module-level helpers
+# ---------------------------------------------------------------------------
+
+def _extract_directory(file_path: str) -> str:
+    """Extract the directory from a file path."""
+    parent = str(PurePosixPath(file_path).parent)
+    return parent if parent != "." else "root"
+
+
+def _generate_community_name(
+    directory: str, nodes: list[dict[str, Any]]
+) -> str:
+    """Generate a human-readable community name.
+
+    Uses directory name + most common class name (if any).
+    """
+    # Extract most common class name
+    class_names = [
+        n["name"] for n in nodes
+        if n.get("kind") == "class"
+    ]
+    if class_names:
+        most_common = Counter(class_names).most_common(1)[0][0]
+        return f"{directory}/{most_common}"
+
+    # Fall back to directory name
+    parts = directory.rstrip("/").split("/")
+    return parts[-1] if parts else directory
+
+
+def _dominant_language(languages: list[str]) -> str | None:
+    """Find the most common language."""
+    if not languages:
+        return None
+    counts = Counter(languages)
+    return counts.most_common(1)[0][0]