Skip to content

WolframAlphaEqual

Edit on GitHub View Code

This simple evaluation function uses the WolframAlpha API to compare two strings. It performs two requests in parallel:

  1. One to check how the user's response is interpreted by WolframAlpha
  2. One to compare the response to the answer, by submitting the following input to the api: res == ans

NOTE: To work, this grading script requires a valid WolframAlpha AppID! This should be stored in the WOLFRAM_APPID env variable.

Inputs

This function doesn't need any parameters, simply two response and answer fields 

{
  "response": "<string>",
  "answer": "<string>"
}

Outputs

{
  "is_correct": "<bool or null>",
  "interp_string": "<string>",
  "raw_comp": "<dict>",
  "raw_interp": "<dict>"
}

is_correct

Extracted from the second WolframAlpha call. More specifically, the "id": "Result" pod's first subpod.plaintext value.

interp_string

Human friendly string which indicates how WolframAlpha interpreted the user response. Extracted from the first WolframAlpha call. Corresponds also to the value of plaintext from the first subpod in the "id" : "Input" pod.

For example, if the user entered 10 kg, this might look like 10 kg (kilograms)

raw_comp

For debugging, passes the raw json result obtained from the second WolframAlpha call (comparison).

raw_interp

For debugging, passes the raw json result obtained from the first WolframAlpha call (interpretation).

Will also return error detailing any issues that were encountered along the way