feat: update Python processors and add utility scripts

- Update ASR, face, OCR, pose processors
- Add release pre-flight check script
- Add synonym generation, chunk processing scripts
- Add face recognition, stamp search utilities

scripts/test_florence2_pipeline.py

@@ -0,0 +1,57 @@
+#!/opt/homebrew/bin/python3.11
+"""
+Test Florence-2 for "Stamps" Detection using Pipeline
+"""
+
+import os
+import cv2
+from transformers import pipeline
+
+UUID = "384b0ff44aaaa1f1"
+VIDEO_PATH = f"output/{UUID}/{UUID}.mp4"
+OUTPUT_DIR = f"output/{UUID}/florence2_results"
+os.makedirs(OUTPUT_DIR, exist_ok=True)
+
+# Frame where "stamp" is heavily discussed
+TIMESTAMP = 6846.0
+
+print(f"📽️ Extracting frame at {TIMESTAMP}s...")
+cap = cv2.VideoCapture(VIDEO_PATH)
+cap.set(cv2.CAP_PROP_POS_MSEC, TIMESTAMP * 1000)
+ret, frame = cap.read()
+cap.release()
+
+if not ret:
+    print("❌ Failed to read frame.")
+    exit()
+
+# Save raw frame
+raw_path = os.path.join(OUTPUT_DIR, f"raw_{int(TIMESTAMP)}.jpg")
+cv2.imwrite(raw_path, frame)
+print(f"💾 Raw frame saved.")
+
+print("🧠 Loading Florence-2 model via pipeline...")
+try:
+    # Using pipeline handles model configuration automatically
+    pipe = pipeline(
+        "image-to-text", model="microsoft/Florence-2-base", trust_remote_code=True
+    )
+
+    print("🔍 Running detection on 'stamp'...")
+    # Florence-2 tasks: '<OPEN_VOCABULARY_DETECTION>', '<CAPTION>', etc.
+    # We want to see if there is a stamp, so let's use caption first to see what it sees.
+
+    result = pipe(raw_path, prompt="<CAPTION>")
+    print(f"📝 Caption Result: {result}")
+
+    # Let's try open vocabulary detection for 'stamp'
+    print("🔍 Running Open Vocabulary Detection for 'stamp'...")
+    result_ood = pipe(
+        raw_path, prompt="<OPEN_VOCABULARY_DETECTION>", text_input="stamp"
+    )
+    print(f"📦 OOD Result: {result_ood}")
+
+except Exception as e:
+    print(f"❌ Error: {e}")
+
+print("🏁 Done.")