personal memory agent
Add .mp4 to RAW_VIDEO_EXTENSIONS in retention
observe/utils.py already included .mp4 in VIDEO_EXTENSIONS, but retention was out of sync. This change aligns retention with the existing video extension handling.\n\nIt also ensures the is_segment_complete() guard correctly blocks purging segments that contain screen.mp4 when screen.jsonl has not been produced yet.
+16
-5
+1
-1
docs/JOURNAL.md
+1
-1
docs/JOURNAL.md
···
180
180
- `raw_media_days` (integer or null) – Number of days to retain raw media when mode is `"days"`. Required when `raw_media` is `"days"`, ignored otherwise.
181
181
- `per_stream` (object) – Per-stream overrides keyed by stream name. Each entry supports `raw_media` and `raw_media_days`. Omitted fields inherit from the global retention settings.
182
182
183
-
"Raw media" means layer 1 capture files only: audio files (`.flac`, `.opus`, `.ogg`, `.m4a`), video files (`.webm`, `.mov`), and screen diffs (`monitor_*_diff.png`).
183
+
"Raw media" means layer 1 capture files only: audio files (`.flac`, `.opus`, `.ogg`, `.m4a`), video files (`.webm`, `.mov`, `.mp4`), and screen diffs (`monitor_*_diff.png`).
184
184
185
185
All layer 2 and layer 3 content is always preserved regardless of retention policy: transcripts (`audio.jsonl`, `screen.jsonl`), agent outputs (`agents/*.md`), speaker labels (`agents/speaker_labels.json`), facet events (`events/*.jsonl`), entity data, segment metadata (`stream.json`), and search index entries.
186
186
+13
-2
tests/test_retention.py
+13
-2
tests/test_retention.py
···
31
31
assert is_raw_media(p), f"{ext} should be raw media"
32
32
33
33
def test_video_extensions(self, tmp_path):
34
-
for ext in (".webm", ".mov"):
34
+
for ext in (".webm", ".mov", ".mp4"):
35
35
p = tmp_path / f"screen{ext}"
36
36
p.touch()
37
37
assert is_raw_media(p), f"{ext} should be raw media"
···
93
93
*,
94
94
audio=False,
95
95
video=False,
96
+
video_name="screen.webm",
96
97
embeddings=False,
97
98
audio_extract=True,
98
99
screen_extract=True,
···
108
109
if audio:
109
110
(seg / "audio.flac").write_bytes(b"audio")
110
111
if video:
111
-
(seg / "screen.webm").write_bytes(b"video")
112
+
(seg / video_name).write_bytes(b"video")
112
113
if embeddings:
113
114
(seg / "audio.npz").write_bytes(b"npz")
114
115
if audio and audio_extract:
···
144
145
def test_incomplete_missing_screen_extract(self, tmp_path):
145
146
seg = _make_segment(tmp_path, video=True, screen_extract=False)
146
147
assert not is_segment_complete(seg)
148
+
149
+
def test_incomplete_missing_screen_extract_for_mp4(self, tmp_path):
150
+
seg = _make_segment(
151
+
tmp_path, video=True, video_name="screen.mp4", screen_extract=False
152
+
)
153
+
assert not is_segment_complete(seg)
154
+
155
+
def test_complete_mp4_with_screen_extract(self, tmp_path):
156
+
seg = _make_segment(tmp_path, video=True, video_name="screen.mp4")
157
+
assert is_segment_complete(seg)
147
158
148
159
def test_incomplete_missing_speaker_labels(self, tmp_path):
149
160
seg = _make_segment(tmp_path, audio=True, embeddings=True, speaker_labels=False)
+2
-2
think/retention.py
+2
-2
think/retention.py
···
32
32
# ---------------------------------------------------------------------------
33
33
34
34
RAW_AUDIO_EXTENSIONS = frozenset({".flac", ".opus", ".ogg", ".m4a"})
35
-
RAW_VIDEO_EXTENSIONS = frozenset({".webm", ".mov"})
35
+
RAW_VIDEO_EXTENSIONS = frozenset({".webm", ".mov", ".mp4"})
36
36
RAW_MEDIA_EXTENSIONS = RAW_AUDIO_EXTENSIONS | RAW_VIDEO_EXTENSIONS
37
37
38
38
···
40
40
"""Check if a file is raw media (layer 1 capture).
41
41
42
42
Raw media: *.flac, *.opus, *.ogg, *.m4a (audio),
43
-
*.webm, *.mov (video), monitor_*_diff.png (screen diffs).
43
+
*.webm, *.mov, *.mp4 (video), monitor_*_diff.png (screen diffs).
44
44
"""
45
45
if path.suffix.lower() in RAW_MEDIA_EXTENSIONS:
46
46
return True