fix: site/ → src/ (stdlib conflict — site is built-in Python module)

Dockerfile

@@ -6,8 +6,8 @@ COPY requirements.txt .
 
 RUN pip install --no-cache-dir -r requirements.txt
 
-COPY site /app/site
+COPY src /app/src
 
 EXPOSE 5000
 
-CMD ["gunicorn", "--bind", "0.0.0.0:5000", "--timeout", "300", "--workers", "1", "site.app:app"]
+CMD ["gunicorn", "--bind", "0.0.0.0:5000", "--timeout", "300", "--workers", "1", "src.app:app"]

src/app.py

@@ -2,8 +2,8 @@
 Больше никаких прокси на contracts.kube5s.ru — всё локально.
 """
 from flask import Flask
-from site.config import VERSION, MAX_CONTENT_LENGTH
+from src.config import VERSION, MAX_CONTENT_LENGTH
-from site.routes import register_routes
+from src.routes import register_routes
 
 
 def create_app():
@@ -30,10 +30,10 @@ def create_app():
 
 def _init_db():
     """Создать БД + схему + seed prompts. SQLite — всё в одном файле /tmp."""
-    from site.db.connection import init_db
+    from src.db.connection import init_db
     init_db()
     try:
-        from site.db import prompts as db_prompts
+        from src.db import prompts as db_prompts
         db_prompts.seed_defaults()
     except Exception:
         pass

src/repository.py

@@ -95,77 +95,77 @@ class PgRepository:
     """Реальный доступ к PostgreSQL через существующие db/*.py."""
 
     def insert_document(self, filename, mime_type, original_bytes, batch_id=None, zip_source=None):
-        from site.db import documents
+        from src.db import documents
         return documents.insert(filename, mime_type, original_bytes,
                                 batch_id=batch_id, zip_source=zip_source)
 
     def set_document_parsed(self, doc_id, elements):
-        from site.db import documents
+        from src.db import documents
         documents.set_parsed(doc_id, elements)  # documents.set_parsed уже делает json.dumps
 
     def set_document_error(self, doc_id, error):
-        from site.db import documents
+        from src.db import documents
         documents.set_error(doc_id, error)
 
     def set_classification(self, doc_id, doc_type, own_number=None, parent_number=None,
                            doc_date=None, counterparty=None,
                            classify_raw=None, classify_input=None):
-        from site.db import documents
+        from src.db import documents
         documents.set_classification(doc_id, doc_type, own_number, parent_number,
                                      doc_date, counterparty,
                                      classify_raw=classify_raw, classify_input=classify_input)
 
     def set_classify_garbage(self, doc_id, reason=""):
-        from site.db import documents
+        from src.db import documents
         documents.set_classify_garbage(doc_id, reason)
 
     def set_classify_failed(self, doc_id, error):
-        from site.db import documents
+        from src.db import documents
         documents.set_classify_failed(doc_id, error)
 
     def list_pending(self, batch_id):
-        from site.db import documents
+        from src.db import documents
         return documents.list_pending(batch_id)
 
     def insert_contract(self, number, client=""):
-        from site.db import contracts
+        from src.db import contracts
         c = contracts.insert(number, client)
         return c["id"] if c else ""
 
     def insert_supplement(self, contract_id, doc_id, supp_type):
-        from site.db import supplements
+        from src.db import supplements
         supplements.insert(contract_id, doc_id, supp_type)
 
     def list_supplements(self, contract_id):
-        from site.db import supplements
+        from src.db import supplements
         return supplements.list_by_contract(contract_id)
 
     def get_spec_current(self, contract_id):
-        from site.db import spec_current
+        from src.db import spec_current
         return spec_current.list_by_contract(contract_id)
 
     def get_document(self, doc_id):
-        from site.db import documents
+        from src.db import documents
         return documents.get(doc_id)
 
     def list_by_batch(self, batch_id):
-        from site.db import documents
+        from src.db import documents
         return documents.list_by_batch(batch_id)
 
     def count_by_status(self, batch_id):
-        from site.db import documents
+        from src.db import documents
         return documents.count_by_status(batch_id)
 
     def reset_classify_status(self, batch_id):
-        from site.db import documents
+        from src.db import documents
         documents.reset_classify_status(batch_id)
 
     def set_classify_processing(self, doc_id):
-        from site.db import documents
+        from src.db import documents
         documents.set_classify_processing(doc_id)
 
     def delete_document(self, doc_id):
-        from site.db import documents
+        from src.db import documents
         documents.delete(doc_id)
 
 

src/routes/__init__.py

@@ -2,12 +2,12 @@
 
 
 def register_routes(app):
-    from site.routes.upload_bp import upload_bp
+    from src.routes.upload_bp import upload_bp
-    from site.routes.pipeline_bp import pipeline_bp
+    from src.routes.pipeline_bp import pipeline_bp
-    from site.routes.api_bp import api_bp
+    from src.routes.api_bp import api_bp
-    from site.routes.prompts_bp import prompts_bp
+    from src.routes.prompts_bp import prompts_bp
-    from site.routes.health_bp import health_bp
+    from src.routes.health_bp import health_bp
-    from site.routes.pages_bp import pages_bp
+    from src.routes.pages_bp import pages_bp
 
     app.register_blueprint(upload_bp)
     app.register_blueprint(pipeline_bp)

src/routes/api_bp.py

@@ -1,9 +1,9 @@
 """API blueprint — groups, documents, supplements, sync, cleanup, spec-current, chat."""
 from flask import Blueprint, request, jsonify
-from site.db import documents, supplements, spec_current
+from src.db import documents, supplements, spec_current
-from site.db.connection import execute, query
+from src.db.connection import execute, query
-from site.services.grouping import group_documents, apply_groups
+from src.services.grouping import group_documents, apply_groups
-from site.config import LLM_URL, LLM_KEY, LLM_MODEL
+from src.config import LLM_URL, LLM_KEY, LLM_MODEL
 import httpx
 
 api_bp = Blueprint("api", __name__)
@@ -177,6 +177,6 @@ def chat():
 @api_bp.route("/api/cleanup", methods=["POST"])
 def api_cleanup():
     """Полная очистка: os.remove(DB) + init новой. Данные гарантированно стёрты."""
-    from site.db.connection import cleanup_db
+    from src.db.connection import cleanup_db
     cleanup_db()
     return jsonify(ok=True, message="all data cleaned")

src/routes/health_bp.py

@@ -1,6 +1,6 @@
 """Health probe — обязательно для Штурвала."""
 from flask import Blueprint, jsonify
-from site.config import VERSION
+from src.config import VERSION
 
 health_bp = Blueprint("health", __name__)
 

src/routes/pipeline_bp.py

@@ -1,10 +1,10 @@
 """Pipeline blueprint — SSE-сравнение + classify."""
 import json, re, os, threading
 from flask import Blueprint, request, jsonify, Response, stream_with_context
-from site.services.process import run_pipeline
+from src.services.process import run_pipeline
-from site.services.classify import classify_batch
+from src.services.classify import classify_batch
-from site.llm_prompt import build_prompt
+from src.llm_prompt import build_prompt
-from site.db import documents
+from src.db import documents
 
 pipeline_bp = Blueprint("pipeline", __name__)
 

src/routes/prompts_bp.py

@@ -1,6 +1,6 @@
 """Prompts blueprint — CRUD + activate для версионирования промптов."""
 from flask import Blueprint, request, jsonify
-from site.db import prompts as db_prompts
+from src.db import prompts as db_prompts
 
 prompts_bp = Blueprint("prompts", __name__)
 

src/routes/upload_bp.py

@@ -1,9 +1,9 @@
 """Upload blueprint — загрузка, конвертация, распаковка."""
 import io, os, base64, hashlib, zipfile, tempfile, subprocess
 from flask import Blueprint, request, jsonify, send_file
-from site.services.parse import parse_file
+from src.services.parse import parse_file
-from site.db import documents
+from src.db import documents
-from site.config import MAX_CONTENT_LENGTH
+from src.config import MAX_CONTENT_LENGTH
 
 upload_bp = Blueprint("upload", __name__)
 

src/services/classify.py

@@ -14,8 +14,8 @@ import json, re, os
 from concurrent.futures import ThreadPoolExecutor, as_completed
 
 import httpx
-from site.db import documents as db_docs
+from src.db import documents as db_docs
-from site.llm_prompt import build_classify_prompt
+from src.llm_prompt import build_classify_prompt
 
 log = __import__("logging").getLogger(__name__)
 
@@ -34,7 +34,7 @@ _classify_llm = None
 def _get_classify_client():
     global _classify_llm
     if _classify_llm is None:
-        from site.services.llm_client import HttpxLLMClient
+        from src.services.llm_client import HttpxLLMClient
         _classify_llm = HttpxLLMClient(url=LLM_URL, key=LLM_KEY, model=LLM_MODEL, max_tokens=1000, timeout=60)
     return _classify_llm
 

src/services/grouping.py

@@ -9,9 +9,9 @@ Grouping service — match classified documents into contract groups.
 - apply_groups(): создаёт contracts + supplements с авто-порядком по дате.
 """
 import re
-from site.db import documents as db_docs
+from src.db import documents as db_docs
-from site.db import contracts as db_contracts
+from src.db import contracts as db_contracts
-from site.db import supplements as db_supplements
+from src.db import supplements as db_supplements
 
 
 def normalize_number(num):

src/services/llm.py

@@ -12,7 +12,7 @@ _llm_client = None
 def _get_default_client():
     global _llm_client
     if _llm_client is None:
-        from site.services.llm_client import HttpxLLMClient
+        from src.services.llm_client import HttpxLLMClient
         _llm_client = HttpxLLMClient(url=LLM_URL, key=LLM_KEY, model=LLM_MODEL)
     return _llm_client
 

src/services/process.py

@@ -2,8 +2,8 @@
 Адаптирован из deploy/compare/process.py: callback → generator.
 """
 import json, time
-from site.db import supplements, spec_current, spec_events
+from src.db import supplements, spec_current, spec_events
-from site.services.metrics import check_arithmetic
+from src.services.metrics import check_arithmetic
 
 
 def run_pipeline(contract_id, order_ids, build_prompt_fn):
@@ -31,7 +31,7 @@ def run_pipeline(contract_id, order_ids, build_prompt_fn):
         yield {"type": "error", "message": "Нет распарсенных файлов"}
         return
 
-    from site.services.llm import call_llm
+    from src.services.llm import call_llm
 
     for s in supps:
         sid = s["id"]