get_captions_batch¶

Batch retrieval of captions for multiple videos.

Overview¶

Retrieves caption segments for multiple videos at once in a single batch.

Function Signature¶

def get_captions_batch(
    feature_view: str,
    metadata: pd.DataFrame
) -> pd.DataFrame

Description¶

Retrieves captions for multiple videos in a batch. It takes the output of search_metadata() as input and retrieves all captions for each video.

Parameters¶

Parameter	Type	Default	Description
`metadata`	`pd.DataFrame`	required	Return value from the `search_metadata()` function Must include the following columns: `uuid`, `model`, `config_source`, `timestamp`, `segment_ids`

Returns¶

Type: pd.DataFrame

Columns: uuid, model, config_source, timestamp, segment_id, index, start, end, text

Sort Order: Ascending order by uuid, model, config_source, timestamp, and index.

Example¶

# 1. Search metadata
metadata = search_metadata(
    feature_view='caption_summary',
    models=['gpt-4o']
)

# 2. Batch retrieval
all_captions = get_captions_batch(
    feature_view='caption_summary',
    metadata=metadata
)

print(f"Total segments: {len(all_captions)}")

Output:

Total segments: 452

search_metadata - Search metadata
get_captions - Single retrieval