⚠️ This repository is no longer being maintained

Storebrand have moved on from Meltano, and we're therefore no longer maintaining this repository.

`map-gpt-embeddings`

Inline mapper for splitting documents and calculating OpenAI embeddings, for purposes of building vectorstore knowledge base usable by GPT and ChatGPT. Split documents into segments, then vectorize.

This version of the mapper is a small modification of the original map-gpt-embeddings mapper, modified to work with the Azure OpenAI API.

Authentication can be either via API key or native Azure authentication either through managed identity or DefaultAzureCredential. If no authentication is provided, the mapper will attempt to use DefaultAzureCredential to authenticate with Azure, otherwise specify either openai_api_key for authentication with a Bearer token, or msi_client_id to authenticate with a specific managed identity.

Built with the Meltano Singer SDK.

Capabilities

stream-maps

Settings

Setting	Required	Default	Description
document_text_property	False	page_content	The name of the property containing the document text.
document_metadata_property	False	None	The name of the property containing the document metadata.
openai_api_key	False	None	OpenAI API key. Optional if `OPENAI_API_KEY` env var is set.
msi_client_id	False	None	Azure Managed Identity for authentication
use_msi	False	0	Use Azure Managed Identity for authentication
api_endpoint	False	https://api.openai.com	Azure OpenAI API Endpoint
deployment_name	False	None	Azure OpenAI Deployment Name
stream_maps	False	None	Config object for stream maps capability. For more information check out Stream Maps.
stream_map_config	False	None	User-defined config values to be used within map expressions.

A full list of supported settings and capabilities is available by running: map-openai-embeddings --about

Configuration

Accepted Config Options

A full list of supported settings and capabilities for this tap is available by running:

map-gpt-embeddings --about

Configure using environment variables

This Singer tap will automatically import any environment variables within the working directory's .env if the --config=ENV is provided, such that config values will be considered if a matching environment variable is set either in the terminal context or in the .env file.

OpenAI Authentication and Authorization

You will need an OpenAI API Key to calculate embeddings using OpenAI's models. Free accounts are rate limited to 60 calls per minute. This is different from ChatGPT Plus account and requires a per-API call billing method established with OpenAI.

Usage

You can easily run map-gpt-embeddings by itself or in a pipeline using Meltano.

Executing the Tap Directly

map-gpt-embeddings --version
map-gpt-embeddings --help
map-gpt-embeddings --config CONFIG --discover > ./catalog.json

Developer Resources

Follow these instructions to contribute to this project.

Initialize your Development Environment

pipx install poetry
poetry install

Create and Run Tests

Create tests within the map_gpt_embeddings/tests subfolder and then run:

poetry run pytest

You can also test the map-gpt-embeddings CLI interface directly using poetry run:

poetry run map-gpt-embeddings --help

Testing with Meltano

Note: This tap will work in any Singer environment and does not require Meltano. Examples here are for convenience and to streamline end-to-end orchestration scenarios.

Next, install Meltano (if you haven't already) and any needed plugins:

# Install meltano
pipx install meltano
# Initialize meltano within this directory
cd map-gpt-embeddings
meltano install

Now you can test and orchestrate using Meltano:

# Test invocation:
meltano invoke map-gpt-embeddings --version
# OR run a test `elt` pipeline:
meltano run tap-smoke-test map-gpt-embeddings target-jsonl

SDK Dev Guide

See the dev guide for more instructions on how to use the SDK to develop your own taps and targets.

storebrand/map-gpt-embeddings