fix: use relative paths in manifest, .graphify_root, and cache for cross-platform portability (fixes #777)

graphify/__main__.py

@@ -3330,7 +3330,7 @@ def _progress(idx: int, total: int, _result: dict) -> None:
                     f"est. cost: ${cost:.4f}"
                 )
             try:
-                _save_manifest(_manifest_files, manifest_path=str(manifest_path), kind="both")
+                _save_manifest(_manifest_files, manifest_path=str(manifest_path), kind="both", root=target)
             except Exception as exc:
                 print(f"[graphify extract] warning: could not write manifest: {exc}", file=sys.stderr)
             if global_merge:
@@ -3419,7 +3419,7 @@ def _progress(idx: int, total: int, _result: dict) -> None:
         }
         analysis_path.write_text(json.dumps(analysis, indent=2), encoding="utf-8")
         try:
-            _save_manifest(_manifest_files, manifest_path=str(manifest_path), kind="both")
+            _save_manifest(_manifest_files, manifest_path=str(manifest_path), kind="both", root=target)
         except Exception as exc:
             print(f"[graphify extract] warning: could not write manifest: {exc}", file=sys.stderr)
 

graphify/cache.py

@@ -159,6 +159,17 @@ def cache_dir(root: Path = Path("."), kind: str = "ast") -> Path:
     return d
 
 
+def _reanchor_source_file(p: str, root: Path) -> str:
+    """If p is a relative path, resolve it against root; otherwise return as-is."""
+    try:
+        candidate = Path(p)
+        if not candidate.is_absolute():
+            return str((root.resolve() / candidate).resolve())
+    except (OSError, ValueError):
+        pass
+    return p
+
+
 def load_cached(path: Path, root: Path = Path("."), kind: str = "ast") -> dict | None:
     """Return cached extraction for this file if hash matches, else None.
 
@@ -176,18 +187,33 @@ def load_cached(path: Path, root: Path = Path("."), kind: str = "ast") -> dict |
     entry = cache_dir(root, kind) / f"{h}.json"
     if entry.exists():
         try:
-            return json.loads(entry.read_text(encoding="utf-8"))
+            result = json.loads(entry.read_text(encoding="utf-8"))
         except (json.JSONDecodeError, OSError):
             return None
     # Migration fallback: check legacy flat cache/ dir for AST entries
-    if kind == "ast":
+    elif kind == "ast":
         legacy = Path(root).resolve() / _GRAPHIFY_OUT / "cache" / f"{h}.json"
         if legacy.exists():
             try:
-                return json.loads(legacy.read_text(encoding="utf-8"))
+                result = json.loads(legacy.read_text(encoding="utf-8"))
             except (json.JSONDecodeError, OSError):
                 return None
-    return None
+        else:
+            return None
+    else:
+        return None
+    # Re-anchor relative source_file paths so cached nodes are consistent with
+    # fresh extractions on this machine (#777)
+    root_resolved = Path(root).resolve()
+    for node in result.get("nodes", []):
+        sf = node.get("source_file")
+        if sf:
+            node["source_file"] = _reanchor_source_file(sf, root_resolved)
+    for edge in result.get("edges", []):
+        sf = edge.get("source_file")
+        if sf:
+            edge["source_file"] = _reanchor_source_file(sf, root_resolved)
+    return result
 
 
 def save_cached(path: Path, result: dict, root: Path = Path("."), kind: str = "ast") -> None:
@@ -203,6 +229,28 @@ def save_cached(path: Path, result: dict, root: Path = Path("."), kind: str = "a
     p = Path(path)
     if not p.is_file():
         return
+
+    # Relativize source_file fields so cached nodes are portable across machines (#777)
+    root_resolved = Path(root).resolve()
+    for node in result.get("nodes", []):
+        sf = node.get("source_file")
+        if sf:
+            try:
+                rel = Path(sf).relative_to(root_resolved).as_posix()
+                if not rel.startswith(".."):
+                    node["source_file"] = rel
+            except ValueError:
+                pass
+    for edge in result.get("edges", []):
+        sf = edge.get("source_file")
+        if sf:
+            try:
+                rel = Path(sf).relative_to(root_resolved).as_posix()
+                if not rel.startswith(".."):
+                    edge["source_file"] = rel
+            except ValueError:
+                pass
+
     h = file_hash(p, root)
     target_dir = cache_dir(root, kind)
     entry = target_dir / f"{h}.json"

graphify/detect.py

@@ -1034,6 +1034,7 @@ def save_manifest(
     manifest_path: str = _MANIFEST_PATH,
     *,
     kind: str = "both",
+    root: Path | None = None,
 ) -> None:
     """Save current file mtimes + content hashes for change detection.
 
@@ -1090,6 +1091,20 @@ def _normalise_entry(entry):
                 # Preserve semantic_hash only when content is unchanged
                 entry["semantic_hash"] = prev.get("semantic_hash", "") if h == prev.get("ast_hash", "") else ""
             manifest[f] = entry
+    # Relativize keys so the manifest is portable across machines (#777)
+    if root is not None:
+        root = Path(root).resolve()
+        relativized: dict[str, dict] = {}
+        for key, val in manifest.items():
+            try:
+                rel = Path(key).relative_to(root).as_posix()
+                if rel.startswith(".."):
+                    relativized[key] = val
+                else:
+                    relativized[rel] = val
+            except ValueError:
+                relativized[key] = val
+        manifest = relativized
     Path(manifest_path).parent.mkdir(parents=True, exist_ok=True)
     Path(manifest_path).write_text(json.dumps(manifest, indent=2), encoding="utf-8")
 
@@ -1129,6 +1144,16 @@ def detect_incremental(
     full = detect(root, follow_symlinks=follow_symlinks, google_workspace=google_workspace, extra_excludes=extra_excludes)
     manifest = load_manifest(manifest_path)
 
+    # Re-anchor relative manifest keys to absolute paths for comparison with detect() output (#777)
+    root_str = str(root.resolve())
+    reanchored: dict[str, dict] = {}
+    for key, val in manifest.items():
+        if Path(key).is_absolute():
+            reanchored[key] = val
+        else:
+            reanchored[str(Path(root_str) / key)] = val
+    manifest = reanchored
+
     if not manifest:
         # No previous run - treat everything as new
         full["incremental"] = True

graphify/watch.py

@@ -452,7 +452,7 @@ def _rebuild_code(
 
         _relativize_source_files(result, project_root)
         out.mkdir(exist_ok=True)
-        (out / ".graphify_root").write_text(str(watch_root), encoding="utf-8")
+        (out / ".graphify_root").write_text(str(watch_path), encoding="utf-8")
 
         if no_cluster:
             # Normalise to "links" key so schema is consistent with the full clustered path.
@@ -482,7 +482,7 @@ def _rebuild_code(
 
             try:
                 from graphify.detect import save_manifest
-                save_manifest(detected["files"], kind="ast")
+                save_manifest(detected["files"], kind="ast", root=watch_root)
             except Exception:
                 pass
 
@@ -520,7 +520,7 @@ def _rebuild_code(
             if same_topology:
                 try:
                     from graphify.detect import save_manifest
-                    save_manifest(detected["files"], kind="ast")
+                    save_manifest(detected["files"], kind="ast", root=watch_root)
                 except Exception:
                     pass
                 flag = out / "needs_update"
@@ -591,7 +591,7 @@ def _rebuild_code(
 
         try:
             from graphify.detect import save_manifest
-            save_manifest(detected["files"], kind="ast")
+            save_manifest(detected["files"], kind="ast", root=watch_root)
         except Exception:
             pass