Hr/download

aana/core/models/video.py

@@ -133,6 +133,7 @@ class VideoMetadata(BaseModel):
 
     title: str = Field(None, description="The title of the video.")
     description: str = Field(None, description="The description of the video.")
+    duration: float | None  = Field(None, description="The duration of the video.")
     model_config = ConfigDict(
         json_schema_extra={
             "description": "Metadata of a video.",

aana/integrations/external/yt_dlp.py

@@ -1,16 +1,48 @@
 import hashlib
 from pathlib import Path
+from typing_extensions import TypedDict
 
 import yt_dlp
 from yt_dlp.utils import DownloadError
 
 from aana.configs.settings import settings
-from aana.core.models.video import Video, VideoInput
+from aana.core.models.video import Video, VideoInput, VideoMetadata
 from aana.exceptions.io import (
     DownloadException,
 )
 
 
+def get_video_metadata(video_url: str) -> VideoMetadata:
+    """Fetch video's metadata for a url.
+
+    Args:
+        video_url (str): the video input url
+
+    Returns:
+        metadata (VideoMetadata): the metadata of the video
+    """
+
+    ydl_options = {
+        "extract_flat": True,
+        "hls_prefer_native": True,
+        "extractor_args": {"youtube": {"skip": ["hls", "dash"]}},
+    }
+    try:
+        with yt_dlp.YoutubeDL(ydl_options) as ydl:
+            info = ydl.extract_info(video_url, download=False)
+            title = info.get("title", "")
+            description = info.get("description", "")
+            duration = info.get("duration")
+            return VideoMetadata(
+                title=title,
+                description=description,
+                duration= duration,
+            )
+    except DownloadError as e:
+        error_message = e.msg.split(";")[0]
+        raise DownloadException(url=video_url, msg=error_message) from e
+
+
 def download_video(video_input: VideoInput | Video) -> Video:
     """Downloads videos for a VideoInput object.
 

aana/projects/chat_with_video/endpoints.py

@@ -20,7 +20,7 @@
 from aana.exceptions.db import MediaIdAlreadyExistsException, UnfinishedVideoException
 from aana.exceptions.io import VideoTooLongException
 from aana.integrations.external.decord import generate_frames, get_video_duration
-from aana.integrations.external.yt_dlp import download_video
+from aana.integrations.external.yt_dlp import download_video, get_video_metadata
 from aana.processors.remote import run_remote
 from aana.processors.video import extract_audio, generate_combined_timeline
 from aana.projects.chat_with_video.const import (
@@ -154,8 +154,22 @@ async def run(
         if check_media_id_exist(media_id):
             raise MediaIdAlreadyExistsException(table_name="media", media_id=video)
 
+        video_duration = None
+        if video.url is not None:
+            video_metadata = get_video_metadata(video.url)
+            video_duration = video_metadata.duration
+
+        # precheck for max video length before actually download the video if possible
+        if video_duration and video_duration > max_video_len:
+            raise VideoTooLongException(
+                video=video,
+                video_len=video_duration,
+                max_len=max_video_len,
+            )
+
         video_obj: Video = await run_remote(download_video)(video_input=video)
-        video_duration = await run_remote(get_video_duration)(video=video_obj)
+        if video_duration is None:
+            video_duration = await run_remote(get_video_duration)(video=video_obj)
 
         if video_duration > max_video_len:
             raise VideoTooLongException(
@@ -168,7 +182,7 @@ async def run(
         yield {
             "media_id": media_id,
             "metadata": VideoMetadata(
-                title=video_obj.title, description=video_obj.description
+                title=video_obj.title, description=video_obj.description, duration=video_duration
             ),
         }
 

aana/storage/services/video.py

@@ -392,4 +392,5 @@ def load_video_metadata(
         return VideoMetadata(
             title=video_entity.title,
             description=video_entity.description,
+            duration=video_entity.duration
         )