[py] Upload a wave file and transcribe it

transcr.py

import os
import google.generativeai as genai

genai.configure(api_key=os.environ["GEMINI_API_KEY"])

generation_config = {
    "temperature": 1,
    "top_p": 0.95,
    "top_k": 40,
    "max_output_tokens": 8192,
    "response_mime_type": "text/plain",
}

model = genai.GenerativeModel(
    model_name="models/gemini-1.5-flash-exp-0827",
    generation_config=generation_config,
)

wav = "src.wav"
prompt = """
Please transcribe this conversation, in the table format of timecode, speaker, caption. Use speaker A, speaker B, etc. to identify speakers.

|timecode|speaker|caption|
|---|---|---|
|00:00|A|text|
""".lstrip()

file = None
try:
    file = genai.upload_file(wav, mime_type="audio/wav")
    print(f"Uploaded file '{file.display_name}' as: {file.uri}")
    chat_session = model.start_chat(history=[
        {"role": "user", "parts": [file]}
    ])
    response = chat_session.send_message(prompt)
    fn = os.path.splitext(wav)[0]
    i = 1
    md = f"{fn}.md"
    while os.path.exists(md):
        i += 1
        md = f"{fn}_{i}.md"
    with open(md, "w") as f:
        f.write(response.text)
    print(f"Transcription saved to: {md}")
except Exception as e:
    print(e)
finally:
    if file:
        genai.delete_file(file.name)
        print(f"Deleted file '{file.display_name}' from: {file.uri}")