v2.0.5: интеграция с convert-service (HTTP вместо subprocess libreoffice)

site/config.py

@@ -9,3 +9,4 @@ LLM_MODEL = os.getenv("LLM_MODEL", "gpt-oss-120b")
 
 MAX_CONTENT_LENGTH = 200 * 1024 * 1024  # 200 MB
 API_KEY = os.getenv("API_KEY", "")
+CONVERT_SERVICE_URL = os.getenv("CONVERT_SERVICE_URL", "http://containerk8s.df36c8af-1a95-4623-b551-0d37b731ccca.svc.cluster.local:5000")

site/routes/upload_bp.py

@@ -1,9 +1,10 @@
 """Upload blueprint — загрузка, конвертация, распаковка."""
-import io, os, base64, hashlib, zipfile, tempfile, subprocess
+import io, os, base64, hashlib, zipfile
+import httpx
 from flask import Blueprint, request, jsonify, send_file
 from services.parse import parse_file
 from db import documents
-from config import MAX_CONTENT_LENGTH
+import config
 
 upload_bp = Blueprint("upload", __name__)
 
@@ -70,38 +71,28 @@ def upload():
 
 @upload_bp.route("/convert-doc", methods=["POST"])
 def convert_doc():
-    """.doc → .docx через libreoffice (без сохранения на диск)."""
+    """.doc → .docx через внешний libreoffice-сервис (HTTP)."""
     f = request.files.get("files")
     if not f:
         return jsonify(ok=False, error="no file"), 400
 
-    data = f.read()
-    doc_path = None
-    tmpdir = None
     try:
-        with tempfile.NamedTemporaryFile(suffix=".doc", delete=False) as tmp:
-            tmp.write(data)
-            doc_path = tmp.name
-        tmpdir = tempfile.mkdtemp()
-        subprocess.run(
-            ["libreoffice", "--headless", "--convert-to", "docx", "--outdir", tmpdir, doc_path],
-            timeout=30, capture_output=True,
+        resp = httpx.post(
+            config.CONVERT_SERVICE_URL + "/convert",
+            files={"file": (f.filename, f.read(), f.content_type or "application/msword")},
+            timeout=120,
         )
-        docx_files = [x for x in os.listdir(tmpdir) if x.endswith(".docx")]
-        if docx_files:
-            with open(os.path.join(tmpdir, docx_files[0]), "rb") as out:
-                return send_file(
-                    io.BytesIO(out.read()),
-                    mimetype="application/vnd.openxmlformats-officedocument.wordprocessingml.document",
-                )
-        return jsonify(ok=False, error="conversion produced no output"), 500
-    finally:
-        if doc_path and os.path.exists(doc_path):
-            os.unlink(doc_path)
-        if tmpdir and os.path.exists(tmpdir):
-            for x in os.listdir(tmpdir):
-                os.unlink(os.path.join(tmpdir, x))
-            os.rmdir(tmpdir)
+        if resp.status_code == 200:
+            return send_file(
+                io.BytesIO(resp.content),
+                mimetype="application/vnd.openxmlformats-officedocument.wordprocessingml.document",
+            )
+        data = resp.json()
+        return jsonify(ok=False, error=data.get("error", "conversion failed")), 500
+    except httpx.TimeoutException:
+        return jsonify(ok=False, error="conversion timeout"), 500
+    except Exception as e:
+        return jsonify(ok=False, error=str(e)), 500
 
 
 @upload_bp.route("/unzip-upload", methods=["POST"])