* Add migration

* Delete legacy from bot
* Clear old models
* Единый http клиент
* РАГ полечен

backend/alembic/versions/003_remove_embeddings_table.py

@@ -0,0 +1,33 @@
+"""Remove unused embeddings table
+
+Revision ID: 003
+Revises: 002
+Create Date: 2024-12-24 12:00:00.000000
+
+"""
+from alembic import op
+import sqlalchemy as sa
+
+
+revision = '003'
+down_revision = '002'
+branch_labels = None
+depends_on = None
+
+
+def upgrade() -> None:
+    op.drop_table('embeddings')
+
+
+def downgrade() -> None:
+    op.create_table(
+        'embeddings',
+        sa.Column('embedding_id', sa.dialects.postgresql.UUID(as_uuid=True), nullable=False),
+        sa.Column('document_id', sa.dialects.postgresql.UUID(as_uuid=True), nullable=False),
+        sa.Column('embedding', sa.dialects.postgresql.JSON(astext_type=sa.Text()), nullable=True),
+        sa.Column('model_version', sa.String(), nullable=True),
+        sa.Column('created_at', sa.DateTime(), nullable=False),
+        sa.ForeignKeyConstraint(['document_id'], ['documents.document_id'], ),
+        sa.PrimaryKeyConstraint('embedding_id')
+    )
+

backend/src/application/use_cases/document_use_cases.py

@@ -8,6 +8,7 @@ from src.domain.repositories.document_repository import IDocumentRepository
 from src.domain.repositories.collection_repository import ICollectionRepository
 from src.domain.repositories.collection_access_repository import ICollectionAccessRepository
 from src.application.services.document_parser_service import DocumentParserService
+from src.application.services.rag_service import RAGService
 from src.shared.exceptions import NotFoundError, ForbiddenError
 
 
@@ -19,12 +20,14 @@ class DocumentUseCases:
         document_repository: IDocumentRepository,
         collection_repository: ICollectionRepository,
         access_repository: ICollectionAccessRepository,
-        parser_service: DocumentParserService
+        parser_service: DocumentParserService,
+        rag_service: Optional[RAGService] = None
     ):
         self.document_repository = document_repository
         self.collection_repository = collection_repository
         self.access_repository = access_repository
         self.parser_service = parser_service
+        self.rag_service = rag_service
     
     async def _check_collection_access(self, user_id: UUID, collection) -> bool:
         """Проверить доступ пользователя к коллекции"""
@@ -64,7 +67,7 @@ class DocumentUseCases:
         filename: str,
         user_id: UUID
     ) -> Document:
-        """Загрузить и распарсить документ"""
+        """Загрузить и распарсить документ, затем автоматически проиндексировать"""
         collection = await self.collection_repository.get_by_id(collection_id)
         if not collection:
             raise NotFoundError(f"Коллекция {collection_id} не найдена")
@@ -81,7 +84,15 @@ class DocumentUseCases:
             content=content,
             metadata={"filename": filename}
         )
-        return await self.document_repository.create(document)
+        document = await self.document_repository.create(document)
+
+        if self.rag_service:
+            try:
+                await self.rag_service.index_document(document)
+            except Exception as e:
+                print(f"Ошибка при автоматической индексации документа {document.document_id}: {e}")
+        
+        return document
     
     async def get_document(self, document_id: UUID) -> Document:
         """Получить документ по ID"""

backend/src/domain/entities/embedding.py

@@ -1,25 +0,0 @@
-"""
-Доменная сущность Embedding
-"""
-from datetime import datetime
-from uuid import UUID, uuid4
-from typing import Any
-
-
-class Embedding:
-    """Эмбеддинг документа"""
-    
-    def __init__(
-        self,
-        document_id: UUID,
-        embedding: list[float] | None = None,
-        model_version: str = "",
-        embedding_id: UUID | None = None,
-        created_at: datetime | None = None
-    ):
-        self.embedding_id = embedding_id or uuid4()
-        self.document_id = document_id
-        self.embedding = embedding or []
-        self.model_version = model_version
-        self.created_at = created_at or datetime.utcnow()
-

backend/src/infrastructure/database/models.py

@@ -53,19 +53,6 @@ class DocumentModel(Base):
     document_metadata = Column("metadata", JSON, nullable=True, default={})
     created_at = Column(DateTime, nullable=False, default=datetime.utcnow)
     collection = relationship("CollectionModel", back_populates="documents")
-    embeddings = relationship("EmbeddingModel", back_populates="document", cascade="all, delete-orphan")
-
-
-class EmbeddingModel(Base):
-    """Модель эмбеддинга (заглушка)"""
-    __tablename__ = "embeddings"
-    
-    embedding_id = Column(UUID(as_uuid=True), primary_key=True, default=uuid.uuid4)
-    document_id = Column(UUID(as_uuid=True), ForeignKey("documents.document_id"), nullable=False)
-    embedding = Column(JSON, nullable=True)
-    model_version = Column(String, nullable=True)
-    created_at = Column(DateTime, nullable=False, default=datetime.utcnow)
-    document = relationship("DocumentModel", back_populates="embeddings")
 
 
 class ConversationModel(Base):

backend/src/presentation/api/v1/rag.py

@@ -1,5 +1,5 @@
 """
-API для RAG: индексация документов и ответы на вопросы
+API для RAG: ответы на вопросы
 """
 from fastapi import APIRouter, status, Request
 from typing import Annotated
@@ -9,30 +9,13 @@ from src.presentation.middleware.auth_middleware import get_current_user
 from src.presentation.schemas.rag_schemas import (
     QuestionRequest,
     RAGAnswer,
-    IndexDocumentRequest,
-    IndexDocumentResponse,
 )
 from src.application.use_cases.rag_use_cases import RAGUseCases
-from src.domain.entities.user import User
 
 
 router = APIRouter(prefix="/rag", tags=["rag"])
 
 
-@router.post("/index", response_model=IndexDocumentResponse, status_code=status.HTTP_200_OK)
-@inject
-async def index_document(
-    body: IndexDocumentRequest,
-    request: Request,
-    user_repo: Annotated[IUserRepository, FromDishka()],
-    use_cases: Annotated[RAGUseCases, FromDishka()],
-):
-    """Индексирование идет через чанкирование, далее эмбеддинг и загрузка в векторную бд"""
-    current_user = await get_current_user(request, user_repo)
-    result = await use_cases.index_document(body.document_id)
-    return IndexDocumentResponse(**result)
-
-
 @router.post("/question", response_model=RAGAnswer, status_code=status.HTTP_200_OK)
 @inject
 async def ask_question(

backend/src/presentation/schemas/rag_schemas.py

@@ -26,10 +26,3 @@ class RAGAnswer(BaseModel):
     usage: dict[str, Any] = {}
 
 
-class IndexDocumentRequest(BaseModel):
-    document_id: UUID
-
-
-class IndexDocumentResponse(BaseModel):
-    chunks_indexed: int
-

backend/src/shared/di_container.py

@@ -152,9 +152,10 @@ class UseCaseProvider(Provider):
         document_repo: IDocumentRepository,
         collection_repo: ICollectionRepository,
         access_repo: ICollectionAccessRepository,
-        parser_service: DocumentParserService
+        parser_service: DocumentParserService,
+        rag_service: RAGService
     ) -> DocumentUseCases:
-        return DocumentUseCases(document_repo, collection_repo, access_repo, parser_service)
+        return DocumentUseCases(document_repo, collection_repo, access_repo, parser_service, rag_service)
 
     @provide(scope=Scope.REQUEST)
     def get_conversation_use_cases(