test_llava2.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")

prompt = "Read all text in this image. Return ONLY the exact text you see, preserving all words, punctuation and line breaks. Do NOT add any explanation or commentary."

payload = {
    "model": "llava:7b",
    "prompt": prompt,
    "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"},
)
sys.stderr.write("Sending request to llava:7b...\n")
with urllib.request.urlopen(req, timeout=600) as resp:
    res = json.loads(resp.read().decode("utf-8"))
sys.stderr.write("Done.\n")
resp_text = res.get("response", "")
print(resp_text)
print(f"\n--- Stats: done={res.get('done_reason')}, eval_count={res.get('eval_count')}, duration_ns={res.get('total_duration')}", file=sys.stderr)