Skip to content

Latest commit

 

History

History
42 lines (32 loc) · 1.5 KB

standard-test.md

File metadata and controls

42 lines (32 loc) · 1.5 KB

Standard Test without Ground Truth for Intents

Story

The user wants to test a set of utterances against a trained Watson Assistant instance simply to get the intent and entity predictions.

Prerequisite

There is one and only one column in test_input_file.csv containing the test utterances which is used for testing purpose.

Input file

config.ini (fill in your iam_apikey, url, and one of workspace_id or environment_id at minimum. (Older instances use a url like "https://gateway-wdc.watsonplatform.net/assistant/api")

[ASSISTANT CREDENTIALS]
iam_apikey = <wa iam apikey>
url = https://api.us-east.assistant.watson.cloud.ibm.com
version=2021-06-14

[DEFAULT]
mode = test
workspace_id = 01234567-9ABC-DEF0-1234-56789ABCDEF0
;environment_id=

; optional - defaults shown here
;output_directory = ./data
;test_input_file = ./data/input.csv
;test_output_path = ./data/test-out.csv
;keep_workspace_after_test = no

Sample test_input_file.csv

utterance
utterance 0

Sample out

test_output_path.csv

utterance predicted intent confidence detected entities dialog response
utterance 0 intent 0 0.01 entity 0:value 0;entity 1:value 1 response 0

Note: For Watson Assistant v2, the intent names returned are intent IDs rather than intent names. These IDs can be discovered in the source code of the assistant JSON.