piscada-foresight 0.6.3

Library for Piscada Foresight; Access knowledge-graph and timeseries data.

Piscada Foresight

Access knowledge-graph and timeseries data.

Overview

This library provides access to the knowledge-graph and timeseries data in the Foresight platform. It implements a Transport using HTTPX to be used with the GQL GraphQL client. It provides the following modules:

• data module: Read timeseries values as Pandas DataFrames or Series

  • get_value(): Get latest value before a given time
  • get_values(): Get values between start and end time
  • get_all_values(): Extract all values from a graph query response

• domains module: Access domain definitions and trait hierarchies

  • Domain class: Contains traits and relationships for a domain
    • get_trait_by_id(): Gets a trait by its ID
  • get_domains(): Retrieve list of all available domains
  • get_trait_by_id(): Get a trait from any domain by its ID string
  • get_parent_traits(): Get all parent traits for a given trait

• http module: OAuth2 authenticated transport

  • ForesightHTTPXTransport class: HTTPX transport with OAuth2 authentication
    • connect(): Establish authenticated connection

Installation

pip install piscada-foresight

You will need access to a Piscada Foresight instance. The library uses the OAuth2 Authorization Code Flow with Proof Key for Code Exchange (PKCE) to act on behalf of your user. After an initial interactive login, the library persists the session state in $HOME/.<client_id>_state and will use that to authenticate non-interactive the next time.

Usage

Local usage

When building the QueryManager singleton, the transport (redirection) is built with the ForesightHTTPXTransport class. This class is a subclass of the HTTPXTransport class from the gql library. The ForesightHTTPXTransport class THAT is responsible for handling the OAuth2 authentication and token refreshing.

There are 3 different transport mechanisms available (interactive, non-interactive and production_ready:

• Interactive (local development) : Specify the domain in the QueryManager creation. Specify the client_id in the QueryManager creation. (default is "foresight-lib-py")

from datetime import datetime, timedelta, timezone

from piscada_foresight.data import get_values
from piscada_foresight.queries_templates.query_manager import QueryManager

domain = "foresight.piscada.cloud"
query_manager = QueryManager(domain=domain) #client_id default is "foresight-lib-py" or query_manager = QueryManager(domain=domain, client_id="<client_id>")

# Retrieve timeseries values for two specific entites:
get_values(
  query_manager,
  entity_ids=["ENTITY_ID", "ENTITY_ID2"],
  start=datetime.now(tz=timezone.utc) - timedelta(hours=8),
)

# Retrieve aggregated timeseries values for two specific entites:
get_values(
  query_manager,
  entity_ids=["ENTITY_ID", "ENTITY_ID2"],
  start=datetime.now(tz=timezone.utc) - timedelta(hours=8),
  interval="1h",
  aggregation_functions=["min", "max", "avg", "count", "last"],
)

• Non-interactive (client_secrets are passed directly): Specify the client_id and the client_secret in the QueryManager creation. QueryManager(domain=domain, client_id="foresight-lib-py", client_secret="")

from datetime import datetime, timedelta, timezone

from piscada_foresight.data import get_values
from piscada_foresight.queries_templates.query_manager import QueryManager

domain = "foresight.piscada.cloud"
query_manager = QueryManager(domain=domain, client_id="<client_id>", client_secret="<client_secret>")

• Production : to handle authorization, catch the headers from the request and pass them to the query_manager transport.

from piscada_foresight.queries_templates.query_manager import QueryManager
from fastapi import Request
from starlette.datastructures import Headers
...

domain = "foresight.piscada.cloud"
#build your query manager object
query_manager = QueryManager(domain=domain)

#Catcht the request
@count_and_duration_app.api_route("/graphql", methods=["GET", "POST"])
async def graphql_server(request: Request):
    # Get the headers       
    headers: Headers = request.headers
    dict_headers = dict(headers)
    
    #sort/filter the headers that you need to keep 
    keys_to_keep = ["x-auth-request-email",
                    "X-Auth-Request-Email",
                    "X-Auth-Request-User",
                    "x-auth-request-user",
                    "X-Auth-Request-Groups",
                    "x-auth-request-groups",
                    "X-Forwarded-Host",
                    "x-forwarded-host",
                    ] 
    filtered_headers = {key: dict_headers[key] for key in keys_to_keep if key in dict_headers}
    query_manager.update_transport(headers=filtered_headers)

Note: Not all interval values are accepted (e.g. 30m work but 45m won't work)

Contributing

Contributions are welcome! You can contact us at [e-mail].

Support

If you have any questions, issues, or suggestions, please contact us at [e-mail].

Copyright

© Piscada AS 2024