Function write_jsonl_output

scripts/data/processing/text_to_jsonl.py:226–254

Write transcript dictionaries to compressed JSONL file. Args: transcript_dicts: List of transcript metadata dictionaries output_path: Path for output JSONL file Raises: IOError: If writing to output file fails

(
    transcript_dicts: List[TranscriptDict], output_path: FilePath
)

def write_jsonl_output(
    transcript_dicts: List[TranscriptDict], output_path: FilePath
) -> None:
    """
    Write transcript dictionaries to compressed JSONL file.

    Args:
        transcript_dicts: List of transcript metadata dictionaries
        output_path: Path for output JSONL file

    Raises:
        IOError: If writing to output file fails
    """
    # Create output directory if it doesn't exist
    os.makedirs(os.path.dirname(output_path), exist_ok=True)

    try:
        print(f"Writing {len(transcript_dicts)} entries to {output_path}")

        # Write to compressed JSONL file
        with gzip.open(output_path, "wt", encoding="utf-8") as f:
            for transcript_dict in transcript_dicts:
                json_line = json.dumps(transcript_dict, ensure_ascii=False)
                f.write(json_line + "\n")

        print(f"Successfully wrote {len(transcript_dicts)} entries")

    except Exception as e:
        raise IOError(f"Failed to write output file {output_path}: {e}")


def process_transcripts_parallel(