personal memory agent
Fix tilde expansion in sol import and add text dry-run segmentation
Expand ~ in media path with os.path.expanduser() to fix FileNotFoundError
when paths are shell-quoted. Add LLM segmentation preview to text/PDF
dry-run output, paralleling audio's segment/key display. Update help
text to reflect audio/text (not video) support.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
+46
-2
+46
-2
think/importers/cli.py
+46
-2
think/importers/cli.py
···
161
161
global _callosum, _message_queue, _import_id, _current_stage, _start_time
162
162
global _stage_start_time, _stages_run, _status_thread, _status_running
163
163
164
-
parser = argparse.ArgumentParser(description="Chunk a media file into the journal")
165
-
parser.add_argument("media", nargs="?", help="Path to video or audio file")
164
+
parser = argparse.ArgumentParser(description="Import a media file into the journal")
165
+
parser.add_argument("media", nargs="?", help="Path to audio or text file")
166
166
parser.add_argument(
167
167
"--timestamp", help="Timestamp YYYYMMDD_HHMMSS for journal entry"
168
168
)
···
237
237
238
238
if not args.media:
239
239
parser.error("the following arguments are required: media")
240
+
241
+
args.media = os.path.expanduser(args.media)
240
242
241
243
# Track detection result for metadata
242
244
detection_result = None
···
306
308
307
309
ext = os.path.splitext(args.media)[1].lower()
308
310
if ext in {".txt", ".md", ".pdf"}:
311
+
from think.detect_transcript import detect_transcript_segment
312
+
from think.importers.text import _time_to_seconds
313
+
309
314
text = _read_transcript(args.media)
310
315
chars = len(text)
311
316
lines = text.count("\n") + (1 if text and not text.endswith("\n") else 0)
312
317
print()
313
318
print(f" Content: {chars:,} characters, {lines:,} lines")
319
+
320
+
# Run segmentation to preview what would be created
321
+
start_time = base_dt.strftime("%H:%M:%S")
322
+
if args.verbose:
323
+
print()
324
+
print(" Segmenting transcript...")
325
+
segments = detect_transcript_segment(text, start_time)
326
+
if segments:
327
+
keys = []
328
+
for idx, (start_at, seg_text) in enumerate(segments):
329
+
time_part = start_at.replace(":", "")
330
+
start_seconds = _time_to_seconds(start_at)
331
+
if idx + 1 < len(segments):
332
+
next_start_at, _ = segments[idx + 1]
333
+
duration = _time_to_seconds(next_start_at) - start_seconds
334
+
else:
335
+
duration = 5
336
+
duration = max(1, duration)
337
+
key = f"{time_part}_{duration}"
338
+
keys.append(key)
339
+
340
+
print(f" Segments: {len(segments)}")
341
+
print(f" Keys: {', '.join(keys)}")
342
+
343
+
if args.verbose:
344
+
print()
345
+
for idx, (key, (start_at, seg_text)) in enumerate(
346
+
zip(keys, segments), 1
347
+
):
348
+
seg_lines = seg_text.count("\n") + (
349
+
1 if seg_text and not seg_text.endswith("\n") else 0
350
+
)
351
+
print(
352
+
f" Segment {idx}: {key} "
353
+
f"({seg_lines:,} lines, {len(seg_text):,} chars)"
354
+
)
355
+
else:
356
+
print()
357
+
print(" Segments: segmentation failed")
314
358
else:
315
359
duration = _get_audio_duration(args.media)
316
360
if duration is not None: