test_moondream.py

from __future__ import annotations

import json
import urllib.request
import base64
import sys

b64 = base64.b64encode(open(r"C:\Users\Ramakrishna\OneDrive\Desktop\pageparse\pageparse\image.png", "rb").read()).decode("utf-8")

prompts = [
    "What does this image say?",
    "Describe this image",
    "Extract all text from this image",
    "Read the text in this screenshot",
    "Transcribe the text",
]
for p in prompts:
    payload = {
        "model": "moondream",
        "prompt": p,
        "images": [b64],
        "stream": False,
        "options": {"temperature": 0.0},
    }
    data = json.dumps(payload).encode("utf-8")
    req = urllib.request.Request(
        "http://localhost:11434/api/generate",
        data=data,
        headers={"Content-Type": "application/json"},
    )
    with urllib.request.urlopen(req, timeout=300) as resp:
        res = json.loads(resp.read().decode("utf-8"))
    print(f"Prompt: {p!r}")
    print(f"  Response: {res.get('response', '')[:200]!r}")
    print(f"  Done: {res.get('done_reason')}, Eval: {res.get('eval_count')}")
    print()