A retrieval augmented generation (RAG) model for completions with the LangChain API, which provides a simple interface for querying the model. An endpoint is provided for indexing data from an Appwrite collection to a Pinecone vector database.
HTML form for interacting with the function.
Index data from an Appwrite collection to a Pinecone vector database. All user defined attributes are indexed as strings and appended to the vector.
Query the model for a completion, including retrieval augmentation. A similarity search is performed using LangChain to retrieve relevant documents for the context.
Parameters
| Name | Description | Location | Type | Sample Value |
|---|---|---|---|---|
| Content-Type | The content type of the request body | Header | application/json |
N/A |
| prompt | Text to prompt the model | Body | String | Write a haiku about Mondays |
Sample 200 Response:
Response from the model.
{
"ok": true,
"completion": "Monday's heavy weight, Dawning with a sigh of grey, Hopeful hearts await."
}Sample 400 Response:
Response when the request body is missing.
{
"ok": false,
"error": "Missing body with a prompt."
}Sample 500 Response:
Response when the model fails to respond.
{
"ok": false,
"error": "Failed to query model."
}| Setting | Value |
|---|---|
| Runtime | Node (18.0) |
| Entrypoint | src/main.js |
| Build Commands | npm install |
| Permissions | any |
| Timeout (Seconds) | 15 |
The ID of the Appwrite database that contains the collection to sync.
| Question | Answer |
|---|---|
| Required | Yes |
| Sample Value | 612a3...5b6c9 |
| Documentation | Appwrite: Databases |
The ID of the collection in the Appwrite database to sync.
| Question | Answer |
|---|---|
| Required | Yes |
| Sample Value | 7c3e8...2a9f1 |
| Documentation | Appwrite: Collections |
The admin API key for Meilisearch. Used to create the index, sync the documents from Appwrite and perform similarity searches.
| Question | Answer |
|---|---|
| Required | Yes |
| Sample Value | masterKey1234 |
| Documentation | Pinecone: API Keys |
The ID of the Pinecone index to sync the documents to.
| Question | Answer |
|---|---|
| Required | Yes |
| Sample Value | index1234 |
| Documentation | Pinecone: Indexes |
API Key for OpenAI API. Used to generate the embeddings for the documents, and query the RAG model.
| Question | Answer |
|---|---|
| Required | Yes |
| Sample Value | searchKey1234 |
| Documentation | OpenAI: API Keys |