Begin

main.py

@@ -0,0 +1,45 @@
+from pathlib import Path
+import base64
+import requests
+
+OLLAMA_URL = "http://localhost:11434/api/chat"
+MODEL = "llama3.2-vision"
+IMAGE_EXTS = {".png", ".jpg", ".jpeg", ".webp", ".bmp", ".gif"}
+
+PROMPT = """
+Extract all readable text from this image.
+Return only the extracted text.
+Do not describe the image.
+If nothing is readable, return an empty string.
+""".strip()
+
+def img_to_b64(path: Path) -> str:
+    return base64.b64encode(path.read_bytes()).decode("utf-8")
+
+def ocr_image(path: Path) -> str:
+    payload = {
+        "model": MODEL,
+        "stream": False,
+        "messages": [
+            {
+                "role": "user",
+                "content": PROMPT,
+                "images": [img_to_b64(path)]
+            }
+        ]
+    }
+    r = requests.post(OLLAMA_URL, json=payload, timeout=300)
+    r.raise_for_status()
+    data = r.json()
+    return data.get("message", {}).get("content", "").strip()
+
+def process_folder(folder: str):
+    folder_path = Path(folder)
+    for img in folder_path.iterdir():
+        if img.is_file() and img.suffix.lower() in IMAGE_EXTS:
+            print(f"Processing: {img.name}")
+            text = ocr_image(img)
+            img.with_suffix(".txt").write_text(text, encoding="utf-8")
+
+if __name__ == "__main__":
+    process_folder("kepek")