API Integration Instructions

GiniMachine API documentation is split into three sections: authentication, endpoints section and sdk usage samples.

All requests are made to endpoints beginning:

https://gcredit.ginimachine.com/

All requests must be secure, i.e. https, not http.

The API endpoint bodies support two content types JSON and tab-delimited CSV. It’s specified at the end of the endpoint URL.

Authentication

In order to make a single API call to GiniMachine, you will need to attach an access token to the request headers.

Important!

For every http request pass the value of the access token as an Authorization header with type API Key.

API Endpoints

When integrating with the model, it is crucial to ensure that the structure of the request JSON file aligns with the structure of the model itself.

This means that if the dataset used for modeling consists of 50 columns, the request body should contain 49 columns. The "ok" column should not be included as it serves as an outcome column that the model will predict.

It is important for the column names to match the JSON representation. For instance, if the training dataset has a column named "age_on_application" the JSON should also include this column with the exact same name, including case sensitivity.

Here you could find our Youtube video guide explaining the API integration with GiniMachine.

Predict JSON, single - application or batch scoring

Query params

Response status codes

  • Status: 200 - OK
  • Status: 201 - Created
  • Status: 400 - Bad Request
    The operation was not completed;
    Limit exceeded
  • Status: 401 - Unauthorized
  • Status: 403 - Forbidden
  • Status: 404 - Model with id not Found

Response body example

Predict JSON with interpretation for each application, single - application or batch scoring
Important!

The process uses additional heuristics to calculate the effect of each feature on the prediction. So, the inference time can vary in this case.

Query params

Response status codes

  • Status: 200 - OK
  • Status: 201 - Created
  • Status: 400 - Bad Request
    The operation was not completed;
    Limit exceeded
  • Status: 401 - Unauthorized
  • Status: 403 - Forbidden
  • Status: 404 - Model with id not Found

Response body example

Predict CSV, single - application or batch scoring

Query params

Response status codes

  • Status: 200 - OK
  • Status: 201 - Created
  • Status: 400 - Bad Request
    The operation was not completed;
    Limit exceeded
  • Status: 401 - Unauthorized
  • Status: 403 - Forbidden
  • Status: 404 - Model with id not Found

Predict CSV with interpretation for each application, single - application or batch scoring

Query params

Response status codes

  • Status: 200 - OK
  • Status: 201 - Created
  • Status: 400 - Bad Request
    The operation was not completed;
    Limit exceeded
  • Status: 401 - Unauthorized
  • Status: 403 - Forbidden
  • Status: 404 - Model with id not Found

Usage and SDK Samples

Curl SDK 


curl --location --request POST 'https://gcredit.ginimachine.com.ginimachine.com/api/v1/scoring/predict.json?modelId=4' \
--header 'Content-Type: application/json' \
--data-raw '[{
    "amount": "30000",
    "source_type": "OFFLINE",
    "remittance_type": "CASH",
    "term": 21,
    "income": 120000,
    "occupation_type": "SERVICES",
    "pension_contributions": true,
    "criminal_record": false,
    "marital_status": "SINGLE",
    "add_info_fraud": false,
    "add_info_inadequate": false,
    "add_info_poor_appearance": false,
    "credit_product_id": "1",
    "verificator_id": "1",
    "partner_id": "1",
    "add_income_pension": true,
    "add_info_call_before_visit": true,
    "had_arrears": true,
    "had_credits_before": true,
    "has_active_credit": true,
    "monthly_payment": 1200,
    "pt_income": 0,
    "banned": false
}]
'

JavaScript SDK 


var settings = {
  "url": "https://gcredit.ginimachine.com.ginimachine.com/api/v1/scoring/predict.json?modelId=4",
  "method": "POST",
  "timeout": 0,
  "headers": {
    "Content-Type": "application/json"
  },
  "data": JSON.stringify([
  {
      "amount": "30000",
      "source_type": "OFFLINE",
      "remittance_type": "CASH",
      "term": 21,
      "income": 120000,
      "occupation_type": "SERVICES",
      "pension_contributions": true,
      "criminal_record": false,
      "marital_status": "SINGLE",
      "add_info_fraud": false,
      "add_info_inadequate": false,
      "add_info_poor_appearance": false,
      "credit_product_id": "1",
      "verificator_id": "1",
      "partner_id": "1",
      "add_income_pension": true,
      "add_info_call_before_visit": true,
      "had_arrears": true,
      "had_credits_before": true,
      "has_active_credit": true,
      "monthly_payment": 1200,
      "pt_income": 0,
      "banned": false
  }
  ]),
};

$.ajax(settings).done(function (response) {
  console.log(response);
});

C# SDK 


var client = new RestClient("https://gcredit.ginimachine.com.ginimachine.com/api/v1/scoring/predict.json?modelId=4");
client.Timeout = -1;
var request = new RestRequest(Method.POST);
request.AddHeader("Content-Type", "application/json");
var body = @"[{" + "\n" +
@"  ""amount"" : ""30000""," + "\n" +
@"  ""source_type"" : ""OFFLINE""," + "\n" +
@"  ""remittance_type"" : ""CASH""," + "\n" +
@"  ""term"" : ""21""," + "\n" +
@"  ""income"" : ""120000""," + "\n" +
@"  ""occupation_type"" : ""SERVICES""," + "\n" +
@"  ""pension_contributions"" : ""true""," + "\n" +
@"  ""criminal_record"" : ""false""," + "\n" +
@"  ""marital_status"" : ""SINGLE""," + "\n" +
@"  ""add_info_fraud"" : ""false""," + "\n" +
@"  ""add_info_poor_appearance"" : ""false""," + "\n" +
@"  ""credit_product_id"" : ""1""," + "\n" +
@"  ""verificator_id"" : ""1""," + "\n" +
@"  ""partner_id"" : ""1""" + "\n" +
@"  ""add_income_pension"" : ""true""" + "\n" +
@"  ""add_info_call_before_visit"" : ""true""" + "\n" +
@"  ""had_arrears"" : ""true""" + "\n" +
@"  ""had_credits_before"" : ""true""" + "\n" +
@"  ""has_active_credit"" : ""true""" + "\n" +
@"  ""monthly_payment"" : ""1200""" + "\n" +
@"  ""pt_income"" : ""0""" + "\n" +
@"  ""banned"" : ""false""" + "\n" +
@"}]" + "\n" +
@"";t.AddParameter("application/json", body,  ParameterType.RequestBody);
IRestResponse response = client.Execute(request);
Console.WriteLine(response.Content);

Python SDK 


import requests
import json

url = "https://gcredit.ginimachine.com.ginimachine.com/api/v1/scoring/predict.json?modelId=4"

payload = json.dumps([
{
    "amount": "30000",
    "source_type": "OFFLINE",
    "remittance_type": "CASH",
    "term": 21,
    "income": 120000,
    "occupation_type": "SERVICES",
    "pension_contributions": true,
    "criminal_record": false,
    "marital_status": "SINGLE",
    "add_info_fraud": false,
    "add_info_inadequate": false,
    "add_info_poor_appearance": false,
    "credit_product_id": "1",
    "verificator_id": "1",
    "partner_id": "1",
    "add_income_pension": true,
    "add_info_call_before_visit": true,
    "had_arrears": true,
    "had_credits_before": true,
    "has_active_credit": true,
    "monthly_payment": 1200,
    "pt_income": 0,
    "banned": false
}
])
headers = {
  'Content-Type': 'application/json'
}

response = requests.request("POST", url, headers=headers, data=payload)

print(response.text)

#With empty columns list

url = "https://gcredit.ginimachine.com.ginimachine.com/api/v1/scoring/predict.json?modelId=4"

payload = json.dumps([
{
    "amount": "30000",
    "source_type": "OFFLINE",
    "remittance_type": "CASH",
    "term": 21,
    "income": 120000,
    "occupation_type": "SERVICES",
    "pension_contributions": true,
    "criminal_record": false,
    "marital_status": "SINGLE",
    "add_info_fraud": false,
    "add_info_inadequate": false,
    "add_info_poor_appearance": false,
    "credit_product_id": "1",
    "verificator_id": "1",
    "partner_id": "1",
    "add_income_pension": true,
    "add_info_call_before_visit": true,
    "had_arrears": true,
    "had_credits_before": true,
    "has_active_credit": true,
    "monthly_payment": 1200,
    "pt_income": 0,
    "banned": false
}
])
headers = {
  'Content-Type': 'application/json'
}

response = requests.request("POST", url, headers=headers, data=payload)

print(response.text)

Got a question?
Contact us: support@ginimachine.com