Merge pull request #2134 from fatkobra/fix/2118-convo-dry-run-mined-state

fix(convos): honor mined state during dry runs

mempalace/convo_miner.py

@@ -19,6 +19,7 @@ from datetime import datetime
 from collections import defaultdict
 from typing import Optional
 
+from .backends import PalaceNotFoundError
 from .collision_scan import assert_no_collisions
 from .ids import ID_RECIPE, make_convo_drawer_id, make_convo_sentinel_id
 from .normalize import normalize_conversations
@@ -830,6 +831,27 @@ def _normalize_convo_conversations(
     return conversations
 
 
+def _open_convo_collection(
+    palace_path: str,
+    *,
+    dry_run: bool,
+):
+    """Open the conversation collection without creating it during dry-run."""
+    if not dry_run:
+        return get_collection(palace_path)
+
+    try:
+        return get_collection(
+            palace_path,
+            create=False,
+            read_only=True,
+        )
+    except PalaceNotFoundError:
+        # A missing palace or uninitialized collection represents empty
+        # prior state to a dry-run. Do not create either one.
+        return None
+
+
 def _mine_convos_impl(
     convo_dir: str,
     palace_path: str,
@@ -871,7 +893,10 @@ def _mine_convos_impl(
         print("  DRY RUN — nothing will be filed")
     print(f"{'-' * 55}\n")
 
-    collection = get_collection(palace_path) if not dry_run else None
+    collection = _open_convo_collection(
+        palace_path,
+        dry_run=dry_run,
+    )
 
     # Bulk pre-fetch already-mined source_file -> stored mtime in one
     # paginated pass instead of `len(files)` separate WHERE-source_file
@@ -880,7 +905,7 @@ def _mine_convos_impl(
     # prefetch_mined_set() does the same decisions in a single scan; loop
     # body becomes an O(1) dict lookup + a cheap local mtime comparison.
     mined_mtimes: dict = (
-        prefetch_mined_set(collection, extract_mode=extract_mode) if not dry_run else {}
+        prefetch_mined_set(collection, extract_mode=extract_mode) if collection is not None else {}
     )
     # content_hash -> source_file for transcripts already filed. Repeated
     # exports from Claude/ChatGPT commonly land under a new filename each
@@ -888,7 +913,9 @@ def _mine_convos_impl(
     # source_file-keyed skip above ("mined_mtimes") never recognizes them —
     # this catches the same conversation reappearing at a new path.
     mined_content_hashes: dict = (
-        prefetch_content_hashes(collection, extract_mode=extract_mode) if not dry_run else {}
+        prefetch_content_hashes(collection, extract_mode=extract_mode)
+        if collection is not None
+        else {}
     )
 
     total_drawers = 0
@@ -909,7 +936,7 @@ def _mine_convos_impl(
         # Falling through re-mines: _file_chunks_locked purges this
         # source_file's stale drawers before inserting fresh ones, so this
         # never leaves duplicates behind.
-        if not dry_run and _is_unchanged_since_last_mine(source_file, mined_mtimes):
+        if _is_unchanged_since_last_mine(source_file, mined_mtimes):
             files_skipped += 1
             continue
 

tests/test_convo_miner.py

@@ -834,3 +834,119 @@ def test_register_file_sentinel_includes_source_mtime():
         assert abs(mined[str(tiny_file)] - os.path.getmtime(tiny_file)) < 0.001
     finally:
         shutil.rmtree(tmpdir, ignore_errors=True)
+
+
+def _write_dry_run_transcript(path: Path) -> None:
+    path.write_text(
+        "> What is the plan?\n"
+        "Start with the schema, then the API.\n\n"
+        "> Are there any risks?\n"
+        "Migration ordering is the main one.\n\n"
+        "> What comes next?\n"
+        "Run focused tests before the full suite.\n",
+        encoding="utf-8",
+    )
+
+
+def test_mine_convos_dry_run_skips_unchanged_mined_file(
+    tmp_path,
+    capsys,
+    monkeypatch,
+):
+    monkeypatch.setenv("HOME", str(tmp_path))
+
+    convo_dir = tmp_path / "convos"
+    convo_dir.mkdir()
+    transcript = convo_dir / "session.txt"
+    _write_dry_run_transcript(transcript)
+    palace_path = str(tmp_path / "palace")
+
+    mine_convos(
+        str(convo_dir),
+        palace_path,
+        wing="original",
+    )
+    capsys.readouterr()
+
+    mine_convos(
+        str(convo_dir),
+        palace_path,
+        wing="target",
+        dry_run=True,
+    )
+    output = capsys.readouterr().out
+
+    assert "[DRY RUN] session.txt" not in output
+    assert "Files processed: 0" in output
+    assert "Files skipped (already filed): 1" in output
+    assert "Drawers filed: 0" in output
+
+
+def test_mine_convos_dry_run_keeps_modified_file_as_work(
+    tmp_path,
+    capsys,
+    monkeypatch,
+):
+    monkeypatch.setenv("HOME", str(tmp_path))
+
+    convo_dir = tmp_path / "convos"
+    convo_dir.mkdir()
+    transcript = convo_dir / "session.txt"
+    _write_dry_run_transcript(transcript)
+    palace_path = str(tmp_path / "palace")
+
+    mine_convos(
+        str(convo_dir),
+        palace_path,
+        wing="original",
+    )
+    capsys.readouterr()
+
+    transcript.write_text(
+        transcript.read_text(encoding="utf-8")
+        + "\n> Did the plan change?\n"
+        + "Yes, add a migration rollback test.\n",
+        encoding="utf-8",
+    )
+
+    future = time.time() + 60
+    os.utime(transcript, (future, future))
+
+    mine_convos(
+        str(convo_dir),
+        palace_path,
+        wing="target",
+        dry_run=True,
+    )
+    output = capsys.readouterr().out
+
+    assert "[DRY RUN] session.txt" in output
+    assert "Files processed: 1" in output
+    assert "Files skipped (already filed): 0" in output
+
+
+def test_mine_convos_dry_run_missing_palace_does_not_create_it(
+    tmp_path,
+    capsys,
+    monkeypatch,
+):
+    monkeypatch.setenv("HOME", str(tmp_path))
+
+    convo_dir = tmp_path / "convos"
+    convo_dir.mkdir()
+    transcript = convo_dir / "session.txt"
+    _write_dry_run_transcript(transcript)
+    palace_path = tmp_path / "palace"
+
+    mine_convos(
+        str(convo_dir),
+        str(palace_path),
+        wing="target",
+        dry_run=True,
+    )
+    output = capsys.readouterr().out
+
+    assert "[DRY RUN] session.txt" in output
+    assert "Files processed: 1" in output
+    assert "Files skipped (already filed): 0" in output
+    assert not palace_path.exists()