mirror of
https://github.com/docker/docker-agent.git
synced 2026-11-04 18:30:09 +00:00
David Gageot
5c8319add3
Thread the Providers map from the agent config through RuntimeConfig, ManagersBuildConfig, and BuildContext so that custom provider definitions (e.g. a local Ollama instance with a custom base_url) are applied when creating embedding and reranking model providers. Add a NewProvider helper on both ManagersBuildConfig and BuildContext to encapsulate the repeated provider.New + WithGateway + WithProviders pattern. Consolidate duplicate resolveModelConfig logic in builder.go by reusing strategy.ResolveModelConfig. Also fixes missing WithProviders in semantic_embeddings.go chat model creation and missing WithGateway in reranking provider creation. Closes #1701 Assisted-By: docker-agent
41 lines
1.1 KiB
YAML
41 lines
1.1 KiB
YAML
# This example demonstrates using a custom provider for RAG embedding models.
|
|
# For instance, you can use a local Ollama instance or any OpenAI-compatible
|
|
# API endpoint for generating embeddings.
|
|
|
|
providers:
|
|
local-ollama:
|
|
base_url: http://localhost:11434/v1
|
|
|
|
models:
|
|
local-embed:
|
|
provider: local-ollama
|
|
model: nomic-embed-text
|
|
|
|
agents:
|
|
root:
|
|
model: openai/gpt-5-mini
|
|
description: assistant with RAG using custom embedding provider
|
|
instruction: |
|
|
You are a helpful assistant with access to a knowledge base.
|
|
Use the search tool to find relevant information before answering.
|
|
toolsets:
|
|
- type: rag
|
|
ref: knowledge_base
|
|
|
|
rag:
|
|
knowledge_base:
|
|
tool:
|
|
description: search the knowledge base for relevant information
|
|
docs:
|
|
- ./docs
|
|
strategies:
|
|
- type: chunked-embeddings
|
|
embedding_model: local-embed # References the model defined above using the custom provider
|
|
database: ./custom_provider.db
|
|
vector_dimensions: 768
|
|
chunking:
|
|
size: 1000
|
|
overlap: 100
|
|
results:
|
|
limit: 5
|