Skip to content

Conversation

@ankit-v2-3
Copy link
Collaborator

Pull Request

Description:
Add RTStream Audio and Transcript Support, Clip Support

Changes:

  • RTStream Audio and Transcript Support
  • Clip Support
  • Fix upload from gel collections object where going to default collection
  • Add scene_index_id, scene_index_name, metadata in Shot
  • Fix 0 score threshold fix in defaults search
  • Segmentation_type llm in spoken word index

"bot_name": bot_name,
"bot_image_url": bot_image_url,
"meeting_title": meeting_title,
"realtime_stream": realtime_stream,
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Let's remote this for current release.
We will release it separately.
For now only audio support is fine.

Copy link
Collaborator Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Removed

self.time_zone = data.get("time_zone")
self.video_id = data.get("video_id")
self.speaker_timeline = data.get("speaker_timeline")
self.realtime_stream = data.get("realtime_stream")
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Let's remove this as well.

Copy link
Collaborator Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Removed

ankit-v2-3 and others added 20 commits January 7, 2026 13:41
Align collection.search(namespace="rtstream") with RTStream.search()
by removing params not accepted by RTStream.search: stitch, rerank,
rerank_params, and index_platform.
…earch-endpoint

Add rtstream namespace support to Collection.search
- Implemented CaptureSession for pre-streaming session management
- Added create_capture_session to Connection class
- Added generate_session_token for temporal access tokens
- Implemented WebSocketConnection using 'websockets' library for real-time events
- Added connect_websocket factory method to Connection
- Updated _constants.py with new API paths
- Added websockets to requirements.txt
- Rename 'api_url' to 'base_url' in CaptureClient constructor
- Remove 'session_id' from CaptureClient constructor; it is now fetched via 'session_token'
- Implement 'fetch_session_id' to retrieve session ID and 'callback_url' from backend
- Add 'CaptureClient.create()' async factory method for initialization
- Implement structured 'Channels' object with 'mics', 'displays', and 'system_audio' collections
- Rename 'start_capture' to 'start_capture_session' and remove 'callback_url' arg
- internalize 'callback_url' handling from session token response
- Update .gitignore to ignore test_local.py
- Added  and  to  for cleaner indexing API.
- Added  to  to easily retrieve streams.
- Cleaned up  callback/websocket logic.
- CaptureSession.get_rtstream() uses simple name matching for standardized callback names
- Added index_audio() and index_visuals() wrappers to RTStream
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

5 participants