Rate-limit backup_if_protected to one folder per day via content hash

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

Author: safishamsi

graphify/export.py

@@ -1,5 +1,6 @@
 # write graph to HTML, JSON, SVG, GraphML, Obsidian vault, and Neo4j Cypher
 from __future__ import annotations
+import hashlib
 import html as _html
 import json
 import math
@@ -62,10 +63,16 @@ def backup_if_protected(out_dir: Path) -> "Path | None":
     reason = "+".join(filter(None, ["semantic" if is_semantic else "", "curated" if is_curated else ""]))
     today = date.today().isoformat()
     backup_dir = out / today
-    suffix = 2
-    while backup_dir.exists():
-        backup_dir = out / f"{today}_{suffix}"
-        suffix += 1
+    graph_src = out / "graph.json"
+
+    # Skip re-copying if today's backup already has identical graph.json content.
+    # If content differs (graph changed since the last backup today), overwrite
+    # the backup in place — one folder per day, always the latest pre-overwrite state.
+    if backup_dir.exists() and (backup_dir / "graph.json").exists():
+        src_hash = hashlib.sha256(graph_src.read_bytes()).hexdigest()
+        bak_hash = hashlib.sha256((backup_dir / "graph.json").read_bytes()).hexdigest()
+        if src_hash == bak_hash:
+            return backup_dir  # identical content, nothing to do
 
     try:
         backup_dir.mkdir(parents=True, exist_ok=True)

tests/test_export.py

@@ -228,17 +228,30 @@ def test_backup_default_labels_only(tmp_path):
     assert backup_if_protected(tmp_path) is None
 
 
-def test_backup_same_day_collision(tmp_path):
-    """Second backup on same day gets _2 suffix."""
+def test_backup_same_day_no_accumulation(tmp_path):
+    """Same content on same day returns existing backup dir without re-copying."""
     from graphify.export import backup_if_protected
     from datetime import date
     (tmp_path / "graph.json").write_text('{"nodes":[],"links":[]}')
     (tmp_path / ".graphify_semantic_marker").write_text("{}")
     b1 = backup_if_protected(tmp_path)
     b2 = backup_if_protected(tmp_path)
     assert b1 is not None and b2 is not None
-    assert b1 != b2
-    assert b2.name == f"{date.today().isoformat()}_2"
+    assert b1 == b2  # same dir, no _2 accumulation
+    assert b1.name == date.today().isoformat()
+
+
+def test_backup_same_day_changed_content(tmp_path):
+    """Changed graph.json on same day overwrites the existing backup in place."""
+    from graphify.export import backup_if_protected
+    from datetime import date
+    (tmp_path / "graph.json").write_text('{"nodes":[],"links":[]}')
+    (tmp_path / ".graphify_semantic_marker").write_text("{}")
+    b1 = backup_if_protected(tmp_path)
+    (tmp_path / "graph.json").write_text('{"nodes":[{"id":"x"}],"links":[]}')
+    b2 = backup_if_protected(tmp_path)
+    assert b1 == b2  # still one folder per day
+    assert (b2 / "graph.json").read_text() == '{"nodes":[{"id":"x"}],"links":[]}'
 
 
 def test_backup_env_disable(tmp_path, monkeypatch):