Skip to content

NLP to Result

The AIP will implement the functionality of getting an SQL form a neutral language sentence.

URL : nibitalkres/{table_name}

Method : POST


Name Type Description
Question string The question you want translated to SQL.
maxResults string /int The max number of results you want to get.
filter Fitler - this is used for adding a filter to the where clause that Niibi is generating. You enter the exact format it will be used in the SQL. The filter will be in relation AND to the conditions in the natural language sentence .

How to use filter

For instance if you ask the question How many companies are B2B With the filter USER_id = 123 you will get the following query:

SELECT COUNT(*) FROM features_company WHERE (LOWER(features_company.business_model) = 'b2b') AND (USER_id = 123) ORDER BY 1 DESC;

Output : JSON

Name Type Description
origQustion string The question you asked.
sql string The translation.
results list List of resuld ecry raw id a dictionary like with column name , value.
JAQL dict JAQL representation of the sql query.
Error string IF Error Acord its reason.


  "question":"Show me b2b companies founded last year",

CURL Request

curl -i -H "Content-type: application/json" -X POST -d "{\"question\":\"Show me b2b companies founded last year\",\"maxResults\":\"10\" }"


    "origQustion": "Show me b2b companies founded last year",
    "sql": "\nSELECT *\nFROM features_company\nWHERE LOWER(features_company.business_model) = 'b2b' AND features_company.founded_at >= '2017-01-01T00:00:00.000-08:00';\n",
    "results": [
            "id": "4",
            "company_id": "7810",
            "name": "NevaBio",
            "website": "None",
            "sector_id": "16",
            "sector_name": "Other",
            "funding_stage_id": "1",
            "funding_stage_name": "Bootstrapped",
            "founded_at": "2017-10-01",
            "proba_prediction_false": "None",
            "proba_prediction_true": "None",
            "label": "None",
            "full_profile": "",
            "business_model": "B2B"