Skip to content

Latest commit

 

History

History
154 lines (120 loc) · 6.45 KB

setup.mdx

File metadata and controls

154 lines (120 loc) · 6.45 KB
Raw

RAG API Pipeline - Setup

The rag-api-pipeline CLI offers a setup wizard command that allows you to quickly setup the pipeline settings so you can start creating RAG API pipelines in a few seconds. Below is an overview of the command options:

(venv) user$ rag-api-pipeline setup --help
Usage: rag-api-pipeline setup [OPTIONS]

  Setup wizard to config the pipeline settings prior execution

Options:
  --debug                      enable logging debug level
  --llm-provider [gaia|other]  LLM provider  [default: gaia]
  --help                       Show this message and exit.

The setup wizard is a three stage process that request you to set some configuration parameters required for the pipeline to execute. For each parameter, the wizard suggests a default value/configuration that can be automatically used by hitting the enter key.

Example:

(venv) user$ rag-api-pipeline setup
Which LLM provider you want to use? (gaia, other) [gaia]:
Init pipeline...
(Step 1/3) Setting Pipeline LLM provider settings...
LLM provider API URL [http://127.0.0.1:8080/v1]: https://0xd0275a4f8ef0d4b2dc0a3baf3d405a63bc647340.us.gaianet.network/v1
LLM Provider API connection OK!
Embeddings model Name [Nomic-embed-text-v1.5]:
Embeddings Vector Size [768]:
Pipeline LLM Provider settings OK!
(Step 2/3) Setting API Source settings...
REST API Key:
API Key stored in config/secrets/api_key
(Step 3/3) Setting Vector DB settings...
Do you have a running QdrantDB instance? [y/N]:
Checking Docker settings...
Docker requirements OK!
Starting a QdrantDB instance using Docker....

 Network gaia-rag-api_rag_api_pipeline_network  Creating
 Network gaia-rag-api_rag_api_pipeline_network  Created
 Container gaia-rag-api-qdrant-1  Creating
 Container gaia-rag-api-qdrant-1  Created
 Container gaia-rag-api-qdrant-1  Starting
 Container gaia-rag-api-qdrant-1  Started

Pipeline Setup completed!

            You can use the pipeline by running either rag-api-pipeline run --help
            or docker compose run pipeline --help

You can run the setup command as many times as you want. If the pipeline was already setup, the CLI will prompt you with a confirmation message to cleanup previous configurations:

Init pipeline...
A .env file already exists in config/.env
Do you want to init the pipeline configuration from scratch? [y/N]: y
  1. Set the LLM Provider

By default, the RAG API pipeline is setup to integrate with a Gaia node to use a loaded embeddings model for generating vector embeddings from extracted content. We recommend reading the guide on how to spin up your own Gaia node, or check the documentation on supported LLM Providers.

(venv) user$ rag-api-pipeline setup
Which LLM provider you want to use? (gaia, other) [gaia]:
Init pipeline...

The Next step is to provide the Gaia node Base URI. Under the hood, the CLI validates it can connect to the node API, otherwise it will prompt you the error:

(Step 1/3) Setting Pipeline LLM provider settings...
LLM provider API URL [http://127.0.0.1:8080/v1]:
ERROR: LLM Provider API (@ http://127.0.0.1:8080/v1/models) is down. HTTPConnectionPool(host='127.0.0.1', port=8080): Max retries exceeded with url: /v1/models (Caused by NewConnectionError('<urllib3.connection.HTTPConnection object at 0x110893c90>: Failed to establish a new connection: [Errno 61] Connection refused'))
Try again...
LLM provider API URL [http://127.0.0.1:8080/v1]:

Finally, you'll need to specify the embeddings model name and embeddings vector size:

LLM provider API URL [http://127.0.0.1:8080/v1]: https://0xd0275a4f8ef0d4b2dc0a3baf3d405a63bc647340.us.gaianet.network/v1
LLM Provider API connection OK!
Embeddings model Name [Nomic-embed-text-v1.5]:
Embeddings Vector Size [768]:
Pipeline LLM Provider settings OK!
  1. Set Source API Settings

In this stage, you are required to specify the API Key that will be used to make authenticated requests to your target REST API. This value will be stored in the secrets folder.

(Step 2/3) Setting API Source settings...
REST API Key:
API Key stored in config/secrets/api_key
  1. Set Vector Database Settings

The final stage is in charge of configuring connection to the Vector Database. If you already have a QdrantDB instance running somewhere, you can specify the Server URL. The CLI validates it can connect to the service instance, otherwise it will prompt you the error:

(Step 3/3) Setting Vector DB settings...
Do you have a running QdrantDB instance? [y/N]: y
Enter your Qdrant Server URL [http://127.0.0.1:6333]:
ERROR: QdrantDB (@ http://127.0.0.1:6333) is down. HTTPConnectionPool(host='127.0.0.1', port=6333): Max retries exceeded with url: / (Caused by NewConnectionError('<urllib3.connection.HTTPConnection object at 0x105873910>: Failed to establish a new connection: [Errno 61] Connection refused'))
Try again...
Enter your Qdrant Server URL [http://127.0.0.1:6333]:

If you don't have a running QdrantDB instance, the CLI will attempt to spin up one using Docker. During this process, the CLI validates the host computer has the Docker daemon is installed and running, otherwise it will prompt you the error:

(Step 3/3) Setting Vector DB settings...
Do you have a running QdrantDB instance? [y/N]:
Checking Docker settings...
ERROR: Docker daemon is not running. Please start Docker before proceeding
Do you want to try again? [Y/n]:

If docker is available, it will deploy and start a QdrantDB container using docker-compose (or will omit this step is the container is already running).

Do you have a running QdrantDB instance? [y/N]:
Checking Docker settings...
Docker requirements OK!
Starting a QdrantDB instance using Docker....

 Network gaia-rag-api_rag_api_pipeline_network  Creating
 Network gaia-rag-api_rag_api_pipeline_network  Created
 Container gaia-rag-api-qdrant-1  Creating
 Container gaia-rag-api-qdrant-1  Created
 Container gaia-rag-api-qdrant-1  Starting
 Container gaia-rag-api-qdrant-1  Started

At the end, you'll see a Pipeline Setup completed! message, confirming that all selected settings will be stored in the config/.env file.

Now, you're ready to start using the RAG API pipeline by running either rag-api-pipeline run or docker compose run pipeline command. In the next section, you can find the complete CLI reference for each subcommand or task available in the pipeline.