podcast-sponsor-remove

Attempt at identify sponsored segments in audio transcripts and removing them.
Log | Files | Refs

commit 534b7671e2d370bc498fa8d490121434b7a8e17f
parent e4781586464f38f7b3ef1a663b2a50cca6cd2b56
Author: vin <git@vineetk.net>
Date:   Sat, 19 Jul 2025 03:15:43 -0400

output data in jsonl

Diffstat:
Mmain.py | 10++++++----
1 file changed, 6 insertions(+), 4 deletions(-)

diff --git a/main.py b/main.py @@ -3,6 +3,7 @@ import re import sys import backoff import openai +import json from datetime import datetime, timedelta from tqdm import tqdm @@ -263,21 +264,22 @@ def process_file(input_file: str, output_file: str): if ad_timestamps_str: # Split by newlines in case the model returns multiple segments found_timestamps = [ts.strip() for ts in ad_timestamps_str.split('\n') if ts.strip()] - all_ad_timestamps.extend(found_timestamps) + all_ad_timestamps.append({"id": f"{"".join(input_file.split(".")[:-1])}_chunk{i}", f"text": chunk_text_for_api, "target": "\n".join(found_timestamps)}) with open(output_file, "w") as f: - f.write("\n".join(all_ad_timestamps)) + for s in all_ad_timestamps: + f.write(json.dumps(s) + "\n") def main(): paths = sys.argv[1:] for p in paths: - out = "".join(p.split(".")[:-1]) + ".gpt" + out = "".join(p.split(".")[:-1]) + ".jsonl" if os.path.isfile(out): paths.remove(p) print(paths) for p in tqdm(paths): - out = "".join(p.split(".")[:-1]) + ".gpt" + out = "".join(p.split(".")[:-1]) + ".jsonl" process_file(p, out) if __name__ == "__main__":