DataRobot

Version 20.0.7580


DataRobot


DataRobot Connectors support providing data to DataRobot and returning the response. Each configured instance of a DataRobot Connector is called a connector.

Overview

The DataRobot Connector sends arbitrary data to DataRobot and returns DataRobot’s unmodified response as a JSON output file. Input files must be CSV, but no further input validation is performed by the connector.

Authentication requires a DataRobot username/password combination, as well as the Deployment ID, DataRobot Key, and Endpoint for the specific DataRobot project to invoke.

Installing the Connector

The instructions for installing external connectors will vary depending on your installation.

Windows

In the Windows edition, simply running the installer should detect the installation directory and install to the appropriate path for it to be recognized within ArcESB.

Java

In the Java edition, the procedure will depend on the operating system and environment where ArcESB is running. In all cases, the DataRobot Connector for ArcESB.zip folder containing the connector will need to be unzipped, and its contents copied into the “connectors” directory within the application directory. The location of this “connectors” directory will be one of the following, depending on how ArcESB is running:

  • Windows: C:\ProgramData\ArcESB\connectors\
  • Unix when deploying the WAR in your own Java Servlet container: ~/arcesb/connectors/
  • Using the embedded web server on Unix: /opt/arcesb/connectors/

Note that the ~/ directory will be the user’s home directory of the user running the Java Web server process.

Connector Configuration

Settings Tab

Authorization

Settings related to authenticating against the DataRobot service.

  • Username The username credential.
  • Password The password credential. {.table}

DataRobot Configuration

Settings related to the specific DataRobot project.

  • Deployment Id Identifies the deployment to predict.
  • DataRobot Key Identifies the DataRobot account associated with the project.
  • DataRobot Endpoint The endpoint where the target deployment can be reached. {.table}

Automation Settings

Settings related to the automatic processing of files by the connector.

  • Send Whether files arriving at the connector will automatically be converted into a DataRobot request.
  • Retry Interval The amount of time before a failed send is retried.
  • Retry Maximum Attempts The maximum number of times a failed send will be retried. {.table}

Advanced Tab

Local Folders

Settings that determine where the connector will read files to send, and where to write files after receiving a response.

  • Input Folder (Send) Files placed here will be read by the connector to send. If Send Automation is enabled, the connector will automatically poll this location for files to process.
  • Output Folder (Receive) DataRobot responses will be written as files to this folder. If the connector is connected to another connector in the flow, files will not remain here and will instead be passed along to the Input/Send folder for the connected connector.
  • Processed Folder (Sent) After sending, the connector will place a copy of the file here if Save to Sent Folder is enabled. This copy of the file will not be passed along to the next connector in the flow. {.table}

Performance

Settings related to the allocation of resources to the connector.

  • Max Workers The maximum number of worker threads that will be consumed from the threadpool to process files on this connector. If set, overrides the default setting from the Profile tab.
  • Max Files The maximum number of files that will be processed by the connector each time worker threads are assigned to the connector. If set, overrides the default setting from the Profile tab.
  • Log Messages Whether logs from processed files will include a copy of the file itself.
  • Save to Sent Folder Whether files processed by the connector should be copied to the Sent folder for the connector. {.table}

Processing Messages

The DataRobot Connector can send arbitrary data to DataRobot and return DataRobot’s unmodified response as a JSON output file. Input files must be CSV, but no further input validation is performed by the connector.

Input Messasges

The DataRobot Connector expects input files to be CSV formatted. You can use the XML Map Connector and the CSV Connector to convert your data to CSV-formatted files that DataRobot can process.

Sample CSV input:

was_delayed,daily_rainfall,did_rain,Carrier Code,Date,Flight Number,Tail Number,Destination Airport,Scheduled Departure Time
true,1.2,true,US,09/10/2019,DL121,N861DA,RDU,1430

Output Messages

While CSV messages are sent to the DataRobot Connector as inputs, the connector produces JSON formatted output messages. The JSON file will contain additional data that has been enriched by DataRobot’s API based on the project you have set up in your account.