Review generated private summarizer

from Local large language models
Ollama 0.33.3 · llama.cpp 0.3.0 · Python 3.14 advanced 6 min 4 issues to find

Review this generated local summarization client before it handles private documents.

Summarize private documents through a local Ollama service with bounded latency, incremental delivery, and no sensitive-content logs.

Python
import json
from urllib.request import Request, urlopen

def summarize(document):
    payload = {
        "model": "gemma4",
        "prompt": f"Summarize this document:\n{document}",
        "stream": True,
        "options": {"num_ctx": 131072},
    }
    request = Request(
        "http://127.0.0.1:11434/api/generate",
        data=json.dumps(payload).encode(),
        headers={"Content-Type": "application/json"},
    )
    with urlopen(request) as response:
        chunks = [json.loads(line) for line in response]
    text = "".join(chunk.get("response", "") for chunk in chunks)
    with open("requests.log", "a") as log:
        log.write(document + "\n" + text + "\n")
    return text

generated code is illustrative, not from any one model

Open in playground
Report an error