NeonAI API Proxy

Proxies API calls to consolidate usage to a single point and allow for caching of API data.

Request Format

API requests should be in the form of a dictionary. The service requested should be specified as service and the remaining data will be passed to the handle_query method of the requested service as kwargs.

Example Wolfram|Alpha Request:

{
 "service": "wolfram_alpha",
 "query": "how far away is Rome?",
 "api": "simple",
 "units": "metric",
 "ip": "64.34.186.120"
}

Response Format

Responses will be returned as dictionaries. Responses should contain the following:

• status_code - Usually contains the HTTP status code from the requested API, -1 should be used to specify any other errors
• content - Usually contains the HTTP content (bytes) from the requested API, but may include a string message for errors.
• encoding = Usually contains the HTTP content encoding if content is the byte representation of a string, may be None

Docker Configuration

When running this as a docker container, the XDG_CONFIG_HOME envvar is set to /config. A configuration file at /config/neon/diana.yaml is required and should look like:

MQ:
  port: <MQ Port>
  server: <MQ Hostname or IP>
  users:
    neon_api_connector:
      password: <neon_api user's password>
      user: neon_api
keys:
  api_services:
    alpha_vantage:
      api_key: <Alpha Vantage Key>
    open_weather_map:
      api_key: <OWM Key>
    wolfram_alpha:
      api_key: <Wolfram|Alpha Key>

For example, if your configuration resides in ~/.config:

export CONFIG_PATH="/home/${USER}/.config"
docker run -v ${CONFIG_PATH}:/config neon_api_proxy

Note: If connecting to a local MQ server, you may need to specify --network host