Add support for ingesting Confluence.

[homanp]

Describe the issue

A user should be able to ingest documents from confluence by passing the following payload to the /ingest API endpoint:

{
  "index_name": "string",
  "vector_database": {
    "type": "qdrant",
    "config": {}
  },
  "encoder": {
      "dimensions": 384,
      "model_name": "embed-multilingual-light-v3.0",
      "provider": "cohere"
    },
  "confluence": {
    "api_token": <api_token>,
    "user_email": < user_email >,
    "url":  <url>,
  },
  "webhook_url": "string"
}

Ingestion

After load process has finished we need to loop over the downloaded documents and use the embedding service to chunk, split, generate and save embeddings to the embedding db. The Embedding Service and walkthroughs should be sufficient for this.

Relevant files

Route ingestion on request: https://github.com/superagent-ai/super-rag/blob/main/api/ingest.py#L23
Create a new handler: https://github.com/superagent-ai/super-rag/blob/main/service/ingest.py