foo.py

import requests
import base64
import sys
import os
from io import BytesIO
import tempfile

url = "http://llama-cpp.local/v1/chat/completions"
prompt = (
    "Describe this image for use as alt text. "
    "Do not include any preamble, introduction, or phrases like "
    "'This image shows' or 'This appears to be'. "
    "Start directly with the description."
)


def describe_image(image_path, prompt=prompt):
    with open(image_path, "rb") as f:
        image_bytes = f.read()
    b64_image = base64.b64encode(image_bytes).decode("utf-8")

    # adjust mime type if not jpeg (e.g. image/png)
    data_url = f"data:image/jpeg;base64,{b64_image}"

    payload = {
        "model": "Qwen2.5-VL-32B",
        "messages": [
            {
                "role": "user",
                "content": [
                    {"type": "text", "text": prompt},
                    {"type": "image_url", "image_url": {"url": data_url}},
                ],
            }
        ],
        "max_tokens": 64,
    }

    response = requests.post(url, json=payload, timeout=240)
    if not response.ok:
        print("Status:", response.status_code)
        print("Body:", response.text)
    response.raise_for_status()

    return response.json()["choices"][0]["message"]["content"]


def get_image_source(path_or_url):
    if path_or_url.startswith(("http://", "https://")):
        response = requests.get(path_or_url, timeout=10)
        response.raise_for_status()
        suffix = os.path.splitext(path_or_url)[1] or ".jpg"
        tmp = tempfile.NamedTemporaryFile(delete=False, suffix=suffix)
        tmp.write(response.content)
        tmp.close()
        return tmp.name
    else:
        return path_or_url


if __name__ == "__main__":
    if len(sys.argv) > 1:
        image_path = sys.argv[1]
    else:
        print("image path required")
        sys.exit(1)

    image_source = get_image_source(image_path)
    result = describe_image(image_source)
    print(result)