feat: Implement text manager API and database connection

- Added `text_manager.py` to handle the creation of text libraries via FastAPI.
- Introduced database connection management in `conexion.py` using PostgreSQL credentials from environment variables.
- Created abstract base class `EmbedderABC` in `Base_Embedder.py` for embedding models.
- Developed `OpenAIEmbedder` class to generate embeddings using OpenAI's API.
- Implemented `OpenAIEmbedderModel` and repository pattern for managing OpenAI embedders in `Openai_embedder_mmr.py`.
- Established `Biblioteca` class for managing text libraries and their associated notes in `biblioteca.py`.
- Created SQLAlchemy models and mappers for `Biblioteca` and `Nota` in `biblioteca_mmr.py` and `notas_biblioteca_mmr.py`.
- Added functionality for dynamic table generation for notes associated with libraries.
- Included comprehensive methods for adding, retrieving, and managing notes and libraries in their respective repositories.
This commit is contained in:
2025-05-10 17:52:43 +02:00
parent c646bc1fef
commit c47b9474f4
27 changed files with 844 additions and 44 deletions
+107
View File
@@ -0,0 +1,107 @@
import os
from dotenv import load_dotenv
from sqlalchemy import Column, String, Table, MetaData
from pgvector.sqlalchemy import Vector
from sqlalchemy.orm import registry, Session
from src.TextManager.nota import Nota
from src.ConexionSql.Base_conexion import ConexionBase
# ----------------------
# Cargar .env
# ----------------------
from entrypoint import ENV_PATH
load_dotenv(ENV_PATH)
# ----------------------
# REGISTRO DINÁMICO PARA TABLAS
# ----------------------
mapper_registry = registry()
# ----------------------
# FUNCIONES AUXILIARES
# ----------------------
def generar_tabla_nota_para_biblioteca(biblioteca_nombre: str, vector_dim: int, metadata: MetaData):
nombre_tabla = biblioteca_nombre.lower().replace(" ", "_")
tabla = Table(
nombre_tabla,
metadata,
Column("id", String, primary_key=True),
Column("titulo", String, nullable=False),
Column("tags", String),
Column("conexiones", String),
Column("texto", String),
Column("resumen", String),
Column("vector", Vector(vector_dim), nullable=False),
Column("vector_resumen", Vector(vector_dim), nullable=True), # AHORA ES OPCIONAL
)
class NotaModel:
def __init__(self, nota: Nota):
self.id = nota.id
self.titulo = nota.titulo
self.tags = ",".join(nota.tags)
self.conexiones = ",".join(nota.conexiones)
self.texto = nota.texto
self.resumen = nota.resumen
self.vector = nota.vector
self.vector_resumen = getattr(nota, "vector_resumen", None) # AHORA ES OPCIONAL
mapper_registry.map_imperatively(NotaModel, tabla)
return tabla, NotaModel
# ----------------------
# MAPPER
# ----------------------
class NotaMapper:
@staticmethod
def to_dict(nota: Nota) -> dict:
return {
"id": nota.id,
"titulo": nota.titulo,
"tags": ",".join(nota.tags),
"conexiones": ",".join(nota.conexiones),
"texto": nota.texto,
"resumen": nota.resumen,
"vector": nota.vector,
"vector_resumen": nota.vector_resumen
}
@staticmethod
def from_model(model) -> Nota:
return Nota(
id=model.id,
titulo=model.titulo,
tags=model.tags.split(",") if model.tags else [],
conexiones=model.conexiones.split(",") if model.conexiones else [],
texto=model.texto,
resumen=model.resumen,
vector=list(model.vector),
vector_resumen=list(model.vector_resumen) if model.vector_resumen is not None else None
)
# ----------------------
# REPO
# ----------------------
class NotaRepo:
def __init__(self, session: Session, modelo_nota):
self.session = session
self.ModeloNota = modelo_nota
def add(self, nota: Nota) -> str:
model = self.ModeloNota(nota)
self.session.add(model)
self.session.commit()
return model.id
def get_by_id(self, id_: str) -> Nota | None:
model = self.session.get(self.ModeloNota, id_)
return NotaMapper.from_model(model) if model else None
def get_all(self) -> list[Nota]:
models = self.session.query(self.ModeloNota).all()
return [NotaMapper.from_model(m) for m in models]