Initial curated RepoPulse source from 2524ccc3c1c27739996dccb2b7992989f88cd26d

backend/app/git_analysis.py

@@ -0,0 +1,104 @@
+from __future__ import annotations
+
+import subprocess
+from collections import Counter
+from datetime import datetime
+from pathlib import Path
+
+from .models import GitSignals
+
+
+def _run(root: Path, *args: str, timeout: int = 8) -> str | None:
+    try:
+        result = subprocess.run(
+            ["git", "-C", str(root), *args],
+            check=True,
+            capture_output=True,
+            text=True,
+            errors="replace",
+            timeout=timeout,
+        )
+    except (OSError, UnicodeError, subprocess.SubprocessError):
+        return None
+    return result.stdout
+
+
+def analyze_git(root: Path) -> tuple[GitSignals, dict[str, int]]:
+    top_level = _run(root, "rev-parse", "--show-toplevel")
+    if not top_level:
+        return GitSignals(), {}
+    try:
+        if Path(top_level.strip()).resolve() != root.resolve():
+            return GitSignals(), {}
+    except (OSError, RuntimeError, ValueError):
+        return GitSignals(), {}
+
+    inside = _run(root, "rev-parse", "--is-inside-work-tree")
+    if not inside or inside.strip() != "true":
+        return GitSignals(), {}
+
+    raw_count = _run(root, "rev-list", "--count", "HEAD")
+    raw_latest = _run(root, "log", "-1", "--format=%aI")
+    raw_authors = _run(root, "log", "--format=%aN%x00%aE", "-n", "2000")
+    raw_files = _run(root, "ls-files", "-z")
+    raw_churn = _run(root, "log", "--numstat", "--format=", "--find-renames=20%", "-n", "200")
+    raw_renames = _run(root, "log", "--name-status", "--format=", "--find-renames=20%", "-n", "200")
+
+    contributors: set[str] = set()
+    if raw_authors:
+        for line in raw_authors.splitlines():
+            if line.strip():
+                contributors.add(line.strip())
+
+    churn: Counter[str] = Counter()
+    renamed_files: set[str] = set()
+    rename_map: dict[str, str] = {}
+    if raw_renames:
+        for line in raw_renames.splitlines():
+            parts = line.split("\t")
+            if len(parts) >= 3 and parts[0].startswith("R"):
+                old_path, new_path = parts[-2:]
+                rename_map[old_path] = new_path
+                renamed_files.add(new_path)
+    if raw_churn:
+        for line in raw_churn.splitlines():
+            parts = line.split("\t", 2)
+            if len(parts) != 3:
+                continue
+            additions, deletions, path = parts
+            churn_path = _rename_target(path) if " => " in path else rename_map.get(path, path)
+            if " => " in path:
+                renamed_files.add(churn_path)
+            if additions.isdigit():
+                churn[churn_path] += int(additions)
+            if deletions.isdigit():
+                churn[churn_path] += int(deletions)
+
+    latest_commit_at = None
+    if raw_latest and raw_latest.strip():
+        try:
+            latest_commit_at = datetime.fromisoformat(raw_latest.strip())
+        except ValueError:
+            latest_commit_at = None
+
+    return (
+        GitSignals(
+            available=True,
+            commit_count=int(raw_count.strip()) if raw_count and raw_count.strip().isdigit() else None,
+            contributor_count=len(contributors) if contributors else None,
+            latest_commit_at=latest_commit_at,
+            tracked_files=len([item for item in raw_files.split("\0") if item]) if raw_files is not None else None,
+            renamed_files=sorted(renamed_files),
+        ),
+        dict(churn),
+    )
+
+
+def _rename_target(path: str) -> str:
+    """Normalize Git's brace-formatted rename path to the new path."""
+    if "{" in path and "}" in path:
+        prefix, remainder = path.split("{", 1)
+        inner, suffix = remainder.split("}", 1)
+        if " => " in inner:
+            return f"{prefix}{inner.split(' => ', 1)[1]}{suffix}"
+    return path.rsplit(" => ", 1)[-1]