feat: add migrations, test scripts, and utility tools

- Add database migrations (006-028) for face recognition, identity, file_uuid
- Add test scripts for ASR, face, search, processing
- Add portal frontend (Tauri)
- Add config, benchmark, and monitoring utilities
- Add model checkpoints and pretrained model references

test_asr_large_clip.py

@@ -0,0 +1,116 @@
+#!/usr/bin/env python3
+"""Test ASR processor on a 10-minute clip from the large problematic video."""
+
+import sys
+import os
+import subprocess
+import json
+import tempfile
+import time
+import shutil
+
+# Paths
+LARGE_VIDEO = "../test_video/1636719d-c31f-78ac-f1dd-8ab0b0b36c66.mov"
+if not os.path.exists(LARGE_VIDEO):
+    print(f"Large video not found: {LARGE_VIDEO}")
+    sys.exit(1)
+
+print(f"Large video size: {os.path.getsize(LARGE_VIDEO) / (1024**3):.2f} GB")
+
+# Create temporary directory
+temp_dir = tempfile.mkdtemp(prefix="asr_test_")
+clip_path = os.path.join(temp_dir, "clip.mp4")
+output_path = os.path.join(temp_dir, "output.json")
+
+try:
+    # Extract 10-minute clip (600 seconds) starting at 0:00
+    print("Extracting 10-minute clip...")
+    ffmpeg_cmd = [
+        "ffmpeg",
+        "-i",
+        LARGE_VIDEO,
+        "-ss",
+        "0",
+        "-t",
+        "600",  # 10 minutes
+        "-c",
+        "copy",
+        "-y",
+        clip_path,
+    ]
+    result = subprocess.run(ffmpeg_cmd, capture_output=True)
+    if result.returncode != 0:
+        print(f"FFmpeg failed: {result.stderr.decode()}")
+        sys.exit(1)
+    if not os.path.exists(clip_path):
+        print("Clip not created")
+        sys.exit(1)
+    print(
+        f"Clip created: {clip_path} ({os.path.getsize(clip_path) / (1024**2):.1f} MB)"
+    )
+
+    # Run ASR processor on clip with chunked mode forced (set max_direct_duration=300)
+    env = os.environ.copy()
+    env["MOMENTRY_ASR_MAX_DIRECT_DURATION"] = "300"  # 5 minutes, force chunked
+    env["MOMENTRY_ASR_CHUNK_DURATION"] = "120"  # 2-minute chunks for testing
+    env["MOMENTRY_ASR_MODEL_SIZE"] = "tiny"
+    env["MOMENTRY_ASR_COMPUTE_TYPE"] = "int8"
+
+    cmd = [
+        "/opt/homebrew/bin/python3.11",
+        "scripts/asr_processor.py",
+        clip_path,
+        output_path,
+        "--uuid",
+        "test_large",
+    ]
+    print(f"Running ASR processor with forced chunked mode...")
+    print(f"Command: {' '.join(cmd)}")
+
+    start = time.time()
+    proc = subprocess.run(
+        cmd, capture_output=True, text=True, env=env, timeout=900
+    )  # 15 min timeout
+    elapsed = time.time() - start
+
+    print(f"ASR completed in {elapsed:.1f}s")
+    print(f"Return code: {proc.returncode}")
+    if proc.stdout:
+        print(f"STDOUT:\n{proc.stdout}")
+    if proc.stderr:
+        print(f"STDERR:\n{proc.stderr}")
+
+    # Check output
+    if os.path.exists(output_path):
+        with open(output_path, "r") as f:
+            data = json.load(f)
+        segments = data.get("segments", [])
+        print(f"Output contains {len(segments)} segments")
+        print(
+            f"Language: {data.get('language')} (prob {data.get('language_probability')})"
+        )
+        print(f"Processing mode: {data.get('processing_mode', 'unknown')}")
+        if segments:
+            print(f"First segment: {segments[0]}")
+        # Verify timestamps are correct (should be within 0-600s)
+        for seg in segments[:5]:
+            if seg["start"] < 0 or seg["end"] > 600:
+                print(f"WARNING: segment outside clip range: {seg}")
+    else:
+        print("ERROR: Output file not created")
+        sys.exit(1)
+
+except subprocess.TimeoutExpired:
+    print("ERROR: ASR processing timed out after 900 seconds")
+    sys.exit(1)
+except Exception as e:
+    print(f"ERROR: {e}")
+    import traceback
+
+    traceback.print_exc()
+    sys.exit(1)
+finally:
+    # Clean up
+    print(f"Cleaning up {temp_dir}")
+    shutil.rmtree(temp_dir, ignore_errors=True)
+    print("Done.")