Fix: use hi_res strategy + chipper model + GPU for maximum quality

parse_pol177.py

@@ -1,6 +1,7 @@
 #!/usr/bin/env python3
 """
 Parse PDF document using unstructured-api on 192.168.1.103.
+Uses maximum quality settings (hi_res + GPU).
 Saves parsed JSON to /tmp/pol177_parsed.json.
 """
 
@@ -12,17 +13,33 @@ PDF_PATH = "/opt/documents_xk/АОХКСибцем/07 Служба Вице-пр
 API_URL = "http://192.168.1.103:8005/general/v0/general"
 
 def parse_pdf():
-    print("Parsing PDF...")
+    print("Parsing PDF with maximum quality (hi_res + GPU)...")
+    
+    # Maximum quality parameters
+    form_data = {
+        'strategy': 'hi_res',
+        'hi_res_model_name': 'chipper',
+        'pdf_infer_table_structure': 'true',
+        'extract_images': 'true',
+        'extract_image_block_types': '["image", "table"]',
+    }
+    
     with open(PDF_PATH, 'rb') as f:
-        files = {'file': f}
+        # Note: 'files' (plural) - API expects list of UploadFile
-        response = requests.post(API_URL, files=files)
+        files = {'files': f}
+        response = requests.post(API_URL, files=files, data=form_data)
     
     if response.status_code != 200:
         print(f"ERROR: API returned {response.status_code}: {response.text[:500]}")
         sys.exit(1)
     
     result = response.json()
-    elements = result.get('elements', [])
+    
+    # API returns array directly, not {"elements": [...]}
+    if isinstance(result, list):
+        elements = result
+    else:
+        elements = result.get('elements', [])
     
     print(f"Extracted {len(elements)} elements")