Integrated API

Find Match

POST https://backend.alli.ai/webapi/find_match

This API is for integrated search for fuzzy match, small talk, q&a.

Headers

Name	Type	Description
API-KEY*	string	Your cognitive search API key can be found in your dashboard Settings menu, under the General tab.

Request Body

Name	Type	Description
process_order	array	can be list of "exact_match", "small_talk", "faq", "mrc", default: ["exact_match", "small_talk", "faq"]
exact_match_list	array	exact match list when using "exact_match"
exact_match_threshold	number	default: 0.85
small_talk_threshold	number	default: 0.85
faq_threshold	number	default:0. 90
faq_direct_answer_mdethod	string	can be "adaptive", "simple". default: "adaptive"
faq_direct_answer_adaptive_thrdshold	number	default: 0.4
faq_direct_answer_direct_threshold	string	default: 0.9
faq_hashtags	array	faq filtering hashtags
faq_hashtags_operator	string	can be "and", "or", default: "or"
faq_count	number	default: 5
faq_format	string	can be "draftjs", "html", "text". default: "text"
mrc_threshold	number	default: 0.80
mrc_hashtags	array	mrc filtering hashtags
mrc_hashtags_operator	string	can be "and", "or", default: "or"
mrc_count	string	default: "5"
mrc_format	string	can be "text", "html". default: "text"
mrc_return_po	boolean	default: False
faq_excluding_hashtags	array	faq filter-out hashtags
faq_excluding_hashtags_operator	string	can be "and", "or", default: "or"

200 Successfully retrieved. found: true/false process: matching process name error: error message When process = exact_match result: exact match keyword When process = faq result[].id: Q&A id result[].question: Q&A question result[].answer: Q&A answer

{
  "found": true, 
  "result": [
    {
      "confidence": 0.8660902048311576, 
      "effectiveConfidence": 0.8660902048311576, 
      "question": "QA 게이트(GATE)", 
      "answer": "...", 
      "id": "RkFROjYwZDA3MGU5NWRjZDlkZTliMzFlNTI5MA==", 
      "lastUpdatedDate": "2021-06-21", 
      "createdDate": "2021-06-21", 
      "hashtags": ["실무용어"], 
      "userFeedback": {"positiveCount": 0, "negativeCount": 0}, 
      "agentFeedback": {"positiveCount": 0, "negativeCount": 0}
    }
  ], 
  "process": "faq", 
  "error": null
}

Request Example

Please replace YOUR API KEY with your one in example below, To get the api key, please see getting-api-key section.

$ curl -H "API-KEY: YOUR_API_KEY" \
       -d'{"process_order":["exact_match", "small_talk", "faq"], \
           "exact_match_list": ["aaaa", "bbbb"], \
           "query": "QA", \
           "faq_threshold": 0.85}' \
       "https://backend.alli.ai/webapi/find_match"
{
  "found": true, 
  "result": [
    {
      "confidence": 0.8660902048311576, 
      "effectiveConfidence": 0.8660902048311576, 
      "question": "QA 게이트(GATE)", 
      "answer": "...", 
      "id": "RkFROjYwZDA3MGU5NWRjZDlkZTliMzFlNTI5MA==", 
      "lastUpdatedDate": "2021-06-21", 
      "createdDate": "2021-06-21", 
      "hashtags": ["실무용어"], 
      "userFeedback": {"positiveCount": 0, "negativeCount": 0}, 
      "agentFeedback": {"positiveCount": 0, "negativeCount": 0}
    }
  ], 
  "process": "faq", 
  "error": null,
  "resultId": "RmluZE1hdGNoUmVzdWx0OjYxODhkZGMwMjE1MmZhYWVkNGExMDRmOA=="
}

Match Feedback

POST https://backend.alli.ai/webapi/match_feedback

if there is no positive_feedback, It just returns previous result without feedback.

If there is no feedback_index, it returns all previous result.

Only for 'faq' process can apply feedback. and only 'faq', 'mrc' can get previous result

Headers

Name	Type	Description
API-KEY*	String	Your cognitive search API key can be found in your dashboard Settings menu, under the General tab.

Request Body

Name	Type	Description
result_id*	String	resultId from find_match result
feedback_index	Int	result index
positive_feedback	Boolean	true: positive, false:negative