code-shredding

config.example.yaml

@@ -1,29 +1,15 @@
-# Configuración de Google Cloud Platform
+# Google Cloud Platform
 project_id: "tu-proyecto-gcp"
-location: "us-central1"  # o us-east1, europe-west1, etc.
-bucket: "tu-bucket-nombre"
+location: "us-central1"
 
-# Configuración del índice vectorial
-index:
-  name: "mi-indice-rag"
-  dimensions: 768  # Para text-embedding-005 usa 768
-  machine_type: "e2-standard-2"  # Tipo de máquina para el endpoint
-  approximate_neighbors_count: 150
-  distance_measure_type: "DOT_PRODUCT_DISTANCE"  # O "COSINE_DISTANCE", "EUCLIDEAN_DISTANCE"
+# Embedding model
+agent_embedding_model: "text-embedding-005"
 
-# Configuración de embeddings
-embedder:
-  model_name: "text-embedding-005"
-  task: "RETRIEVAL_DOCUMENT"  # O "RETRIEVAL_QUERY" para queries
-
-# Configuración de LLM para chunking
-llm:
-  model: "gemini-2.0-flash"  # O "gemini-1.5-pro", "gemini-1.5-flash"
-
-# Configuración de chunking
-chunking:
-  strategy: "contextual"  # "recursive", "contextual", "llm"
-  max_chunk_size: 800
-  chunk_overlap: 200  # Solo para LLMChunker
-  merge_related: true  # Solo para LLMChunker
-  extract_images: true  # Solo para LLMChunker
+# Vector index
+index_name: "mi-indice-rag"
+index_dimensions: 768
+index_machine_type: "e2-standard-16"
+index_origin: "gs://tu-bucket/input/"
+index_destination: "gs://tu-bucket/output/"
+index_chunk_limit: 800
+index_distance_measure_type: "DOT_PRODUCT_DISTANCE"