podcast-sponsor-remove

commit 950c077309aff3036bdd2ab618983089faef41de
parent 2db3b781c15b45bc3f0771ad2ef30d6d88352adb
Author: vin <git@vineetk.net>
Date:   Sat, 19 Jul 2025 12:36:46 -0400

save output in data/, not where podcasts are

Diffstat:
M
main.py
 | 
18
++++++++++++++----

1 file changed, 14 insertions(+), 4 deletions(-)
diff --git a/main.py b/main.py
@@ -14,6 +14,8 @@ client = openai.OpenAI(
 )
 
 MODEL = "deepseek/deepseek-chat-v3-0324"
+#MODEL = "deepseek/deepseek-chat-v3-0324:free"
+
 SYSTEM_PROMPT_TEMPLATE = """
 You are an expert podcast content analyzer. Your task is to identify and extract pre-recorded dynamic advertising segments from podcast transcripts. These ads typically have a distinct tone shift, often using more direct, persuasive language, and frequently include calls to action, product mentions, or specific brand names.
 
@@ -258,25 +260,33 @@ def process_file(input_file: str, output_file: str):
         if ad_timestamps_str.startswith("[") or "no output" in ad_timestamps_str.lower() or "no ad" in ad_timestamps_str.lower():
             ad_timestamps_str = ""
 
+
+        promptid = os.path.dirname(input_file).split('/')[-1] + "_"
+        promptid += os.path.splitext(os.path.basename(output_file))[0].replace(" ", "") + f"_chunk{i}"
+
         # 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.append({"id": f"{"".join(input_file.split(".")[:-1])}_chunk{i}", f"text": "Instruction: Identify the timestamp range of the pre-recorded ad in the following transcript. Output only the HH:MM:SS.mmm-HH:MM:SS.mmm range.\n\nTranscript:\n" + user_prompt, "target": "\n".join(found_timestamps)})
+        all_ad_timestamps.append({"id": promptid, f"text": "Instruction: Identify the timestamp range of the pre-recorded ad in the following transcript. Output only the HH:MM:SS.mmm-HH:MM:SS.mmm range.\n\nTranscript:\n" + user_prompt, "target": "\n".join(found_timestamps)})
 
     with open(output_file, "w") as f:
         for s in all_ad_timestamps:
             f.write(json.dumps(s) + "\n")
 
 def main():
-    os.makedirs("data/")
+    os.makedirs("data/", exist_ok=True)
     paths = sys.argv[1:]
     for p in paths:
-        out = "data/" + os.path.splitext(os.path.basename(p))[0] + ".jsonl"
+        out = "data/"
+        out += os.path.dirname(p).split('/')[-1] + "_"
+        out += os.path.splitext(os.path.basename(p))[0] + ".jsonl"
         if os.path.isfile(out):
             paths.remove(p)
     print(paths)
 
     for p in tqdm(paths):
-        out = "data/" + os.path.splitext(os.path.basename(p))[0] + ".jsonl"
+        out = "data/"
+        out += os.path.dirname(p).split('/')[-1] + "_"
+        out += os.path.splitext(os.path.basename(p))[0] + ".jsonl"
         process_file(p, out)
 
 if __name__ == "__main__":