Data sources

Data sources

Overview

Data sources let you call third-party APIs. Sperta provides pre-built connectors such as ComplyAdvantage and CRS, as well as HTTP connectors for any JSON APIs.

Connectors can be shared by all workflows under the same environment. Before you delete a connector, please make sure no workflows are using it.

Data source steps use the connection provided by connectors and configure the rest of the request parameters.

icon
There is a timeout of 5 seconds for data sources.

Pre-built connectors

To add a pre-built connector, just add the connector name and credentials.

image

To create the data source step, you need to create an entity feature as an input feature first. Take a look at Entity featuresEntity features to learn more. For some products, you can also configure some options, such as Inquiry type and Scoring model in the following example:

image

Make sure to roll out the data source step. Then, you can use the features of the data source in subsequent stages. Use the sources namespace to access the features, e.g., sources.equifax_soft_pull_vantage.credit_score. Please refer to Data Source Features to find out which features are available.

HTTP connectors

HTTP connectors let you connect to any data vendors or internal services that have a JSON API.

icon
The response of JSON APIs needs to be a JSON object instead of a JSON array.

To add a HTTP connector, specify the name, base URL, authentication method, and headers. You can securely store API keys and passwords in the headers since they’re encrypted when stored. You can also configure TLS settings if you’re connecting to an API that requires mTLS.

image

When creating the data source, you can add the method, path, and request body. The path is an expression that evaluates to a String, and it can include query parameters too. The body is an expression that evaluates to a Map. Just like other expressions, you can use any available features in the path and body.

image

Sperta doesn’t know the response schema of HTTP data sources. As a result, you need to parse the response in order to use its fields as features in subsequent stages:

  • response.body contains the HTTP response body as a JSON. In the previous example, you can access Socure’s email risk score like this: sources.socure_risk.response.body.emailRisk.score.
  • response.status_code contains the HTTP response status code as an Integer. For example, sources.socure_risk.response.status_code.

Sperta doesn’t add the response from HTTP data sources to the workflow execution results, but you can build formulas that parse the response to get the features you care about.

Data source features

ComplyAdvantage

ProductFeature IDTypeDescription

Searches

match_status
String

Possible values are unknown, no_match, potential_match, false_positive, true_positive, true_positive_approve, true_positive_reject. See this help article to learn more.

risk_level
String

Possible values are low, medium, high, unknown

total_hits
Integer

The total number of hits

total_matches
Integer

The total number of matches

total_blacklist_hits
Integer

The total number of blacklist hits

match_aml_types
List<String>

The list of aml types associated with all the matches

match_sources
List<String>

The list of sources associated with all the matches

CRS

ProductFeature IDTypeDescription

Equifax Consumer Credit Report

credit_score
Integer

The FICO 9 Score or VantageScore 4.0

reason_codes
List<String>

The adverse action codes. Please reach out to support to get the list of codes

Experian Consumer Credit Report

credit_score
Integer

The FICO 9 Score or VantageScore 4.0

reason_codes
List<String>

The adverse action codes. Please reach out to support to get the list of codes

TransUnion Consumer Credit Report

credit_score
Integer

The FICO 9 Score or VantageScore 4.0

reason_codes
List<String>

The adverse action codes. Please reach out to support to get the list of codes

SentiLink

ProductFeature IDTypeDescription

Scores

abuse_score
Double

Abuse score

abuse_score_reason_codes
List<String>

The reason codes for the abuse score. See SentiLink documentation to learn more.

first_party_synthetic_score
Double

First party synthetic score

first_party_synthetic_score_reason_codes
List<String>

The reason codes for the first party synthetic score. See SentiLink documentation to learn more.

third_party_synthetic_score
Double

Third party synthetic score

third_party_synthetic_score_reason_codes
List<String>

The reason codes for the third party synthetic score. See SentiLink documentation to learn more.

id_theft_score
Double

ID theft score

id_theft_score_reason_codes
List<String>

The reason codes for the id theft score

Insights

insights_codes
List<String>

A list of insight codes. See SentiLink documentation to learn more.

name_match_type
String

Possible values are exact, fuzzy, no_match

dob_match_type
String

Possible values are exact, fuzzy, no_match

ssn_match_type
String

Possible values are exact, fuzzy, no_match

address_match_type
String

Possible values are exact, fuzzy, no_match