Update utils/ingest_video.py

utils/ingest_video.py

@@ -0,0 +1,93 @@
+import zipfile
+import os
+import chromadb
+from chromadb.utils.embedding_functions import OpenCLIPEmbeddingFunction
+from chromadb.utils.data_loaders import ImageLoader
+
+import cv2
+
+path = "mm_vdb2"
+client = chromadb.PersistentClient(path=path)
+
+image_loader = ImageLoader()
+CLIP = OpenCLIPEmbeddingFunction()
+video_collection = client.get_or_create_collection(
+    name='video_collection',
+    embedding_function=CLIP,
+    data_loader=image_loader
+)
+
+def extract_frames(video_folder, output_folder):
+    if not os.path.exists(output_folder):
+        os.makedirs(output_folder)
+
+    for video_filename in os.listdir(video_folder):
+        if video_filename.endswith('.mp4'):
+            video_path = os.path.join(video_folder, video_filename)
+            video_capture = cv2.VideoCapture(video_path)
+            fps = video_capture.get(cv2.CAP_PROP_FPS)
+            frame_count = int(video_capture.get(cv2.CAP_PROP_FRAME_COUNT))
+            duration = frame_count / fps
+
+            output_subfolder = os.path.join(output_folder, os.path.splitext(video_filename)[0])
+            if not os.path.exists(output_subfolder):
+                os.makedirs(output_subfolder)
+
+            success, image = video_capture.read()
+            frame_number = 0
+            while success:
+                if frame_number == 0 or frame_number % int(fps * 5) == 0 or frame_number == frame_count - 1:
+                    frame_time = frame_number / fps
+                    output_frame_filename = os.path.join(output_subfolder, f'frame_{int(frame_time)}.jpg')
+                    cv2.imwrite(output_frame_filename, image)
+
+                success, image = video_capture.read()
+                frame_number += 1
+
+            video_capture.release()
+
+def add_frames_to_chromadb(video_dir, frames_dir):
+    video_frames = {}
+
+    for video_file in os.listdir(video_dir):
+        if video_file.endswith('.mp4'):
+            video_title = video_file[:-4]
+            frame_folder = os.path.join(frames_dir, video_title)
+            if os.path.exists(frame_folder):
+                video_frames[video_title] = [f for f in os.listdir(frame_folder) if f.endswith('.jpg')]
+
+    ids = []
+    uris = []
+    metadatas = []
+
+    for video_title, frames in video_frames.items():
+        video_path = os.path.join(video_dir, f"{video_title}.mp4")
+        for frame in frames:
+            frame_id = f"{frame[:-4]}_{video_title}"
+            frame_path = os.path.join(frames_dir, video_title, frame)
+            ids.append(frame_id)
+            uris.append(frame_path)
+            metadatas.append({'video_uri': video_path})
+
+    video_collection.add(ids=ids, uris=uris, metadatas=metadatas)
+
+def process_video_files(video_paths):
+    frames_output_folder = r"extracted_frames"
+
+    os.makedirs(frames_output_folder, exist_ok=True)
+
+    for video_path in video_paths:
+        video_folder = os.path.dirname(video_path)
+        extract_frames(video_folder, frames_output_folder)
+        add_frames_to_chromadb(video_folder, frames_output_folder)
+
+    return video_collection
+
+# # Example usage
+# video_paths = [
+#     "/path/to/video1.mp4",
+#     "/path/to/video2.mp4",
+#     "/path/to/video3.mp4"
+# ]
+
+# process_video_files(video_paths)