pythia-client 0.2.0

Client to interact with the Pythia API

pythia-client

This is the client make it easier to interact with our internal Pythia API.
This is still a work in progress

Usage

• Create a connection with client object

from pythia_client.client import APIClient
client = APIClient(url="http://localhost:8000", api_key="api-key")

• Query the knowledge base (ask a question)

filters = {
    "operator": "AND",
    "conditions": [
    {
        "field": "meta.source",
        "operator": "==",
        "value": "Salesforce",
        },
    ],
}
query_response = client.query("Hey how are you", filters=filters)
query_response.model_dump()

• Index new files in the knowledge base

index_response = client.upload_files(["path/to/file.pdf"], meta=[{"source": "Salesforce"}])
index_response.model_dump()
# OR
index_response = client.upload_files([("file.pdf", file)], meta=[{"source": "Salesforce"}])
index_response.model_dump()

• List documents in the knowledge base (based on filters)

filters = {
    "operator": "AND",
    "conditions": [
    {
        "field": "meta.source",
        "operator": "==",
        "value": "Salesforce",
        },
    ],
}
list_response = client.get_docs_by_filters(filters=filters)
list_response.model_dump()

• Remove documents from the knowledge base

filters = {
    "operator": "AND",
    "conditions": [
    {
        "field": "meta.source",
        "operator": "==",
        "value": "Salesforce",
        },
    ],
}
remove_response = client.delete_docs_by_filters(filters=filters)
remove_response.model_dump()

• Get a presigned S3 url to the original document

get_doc_response = client.get_file_url("s3fd_sample.pdf", page=2)
s3_url = get_doc_response.url

• Extract JSON data from unstructured test.
  Currently used to showcase the Email2Case feature.

extract_response = client.extract_structured_data("Hello I want to modify the delivery date of order FR73681920")
extract_response.model_dump()