v1.0.149: VM parses PDF + stores elements_json in Lucee

deploy/convert_server.py

@@ -346,14 +346,64 @@ class Handler(BaseHTTPRequestHandler):
                 json={"upload_id": upload_id, "contract_id": contract_id},
             )
         if resp.status_code == 200:
+            result = resp.json()
+            doc_id = result.get("DOC_ID", "")
+
+            # Парсинг на VM + сохранить в Lucee
+            parsed = self._parse_file(filename, file_data)
+            if parsed and parsed.get("status") == "parsed" and doc_id:
+                try:
+                    escaped = json.dumps(parsed["elements"], ensure_ascii=False).replace("'", "''")
+                    with httpx.Client(http2=True, timeout=15, verify=False) as c:
+                        c.post(
+                            f"{LUCEE_URL}/api.cfm",
+                            data={
+                                "action": "execute",
+                                "sql": f"UPDATE documents SET elements_json='{escaped}'::jsonb, status='parsed' WHERE id='{doc_id}'"
+                            }
+                        )
+                except Exception:
+                    pass
+
+            result["parsed"] = parsed
             self.send_response(200)
             self._send_cors()
             self.send_header("Content-Type", "application/json; charset=utf-8")
             self.end_headers()
-            self.wfile.write(resp.content)
+            self.wfile.write(json.dumps(result, ensure_ascii=False).encode())
         else:
             self._send_error_cors(502, f"assemble failed: HTTP {resp.status_code}")
 
+    def _parse_file(self, filename, data):
+        """Распарсить файл на VM: PDF→PyPDF2, DOCX/DOC→python-docx, иначе plain text."""
+        import base64
+        ext = filename.rsplit('.', 1)[-1].lower() if '.' in filename else ''
+        try:
+            if ext == 'pdf':
+                from PyPDF2 import PdfReader
+                import io as io_mod
+                reader = PdfReader(io_mod.BytesIO(data))
+                elements = []
+                for page in reader.pages:
+                    text = page.extract_text()
+                    if text:
+                        for line in text.split('\n'):
+                            line = line.strip()
+                            if line:
+                                elements.append({"type": "paragraph", "text": line, "style": ""})
+                return {"status": "parsed", "element_count": len(elements), "elements": elements}
+            elif ext in ('docx', 'doc'):
+                # Для DOCX/DOC — оставляем Lucee (Apache POI работает)
+                return None
+            else:
+                # Plain text
+                text = data.decode('utf-8', errors='ignore')[:5000]
+                lines = [l.strip() for l in text.split('\n') if l.strip()]
+                return {"status": "parsed", "element_count": len(lines),
+                        "elements": [{"type": "paragraph", "text": l, "style": ""} for l in lines]}
+        except Exception as e:
+            return {"status": "error", "error": str(e), "element_count": 0}
+
     def _handle_parse_pdf(self):
         """Распарсить PDF через PyPDF2 (получает base64 тело от Lucee)."""
         import base64