This section describes the settings for using Postman for prediction via the REST API of the Neuton Platform.
Postman is a widespread software for API development, testing, and debugging. You can download and install a version for your operating system from the official website https://www.postman.com. Before using Postman for prediction, you must train your model and enable Web Predictions for your solution. Web prediction is available only for solutions with the turned-off Digital Signal Preprocessing (DSP) option.
You have two options to send prediction requests in Postman:

Step 1: Modify your new data according to these rules in a text editor:

You can see an example below for a sample test dataset. Data from the original test dataset (CSV file):

Humid, Temp
26.2573,22.248
36.548,22.1493
Modified data for using in Postman requests:
{"data":"Humid,Temp\n26.2573,22.248\n36.548,22.1493"}

Copy all files from this archive to your project and include the header file:

Step 2: Copy /v2/predict URL from Neuton Access Instructions, you will use this URL in a Postman request URL specification.

Step 3: Open the Postman application and on a new tab select “POST” request type, then select “Body”, click the “raw” radio button and select “JSON” from the dropdown list as highlighted on the screen below.

Paste /v2/predict URL (from Step 2) in the request URL and modified data (from Step 1) and click “Send”.

Step 4: Copy returned UUID

Step 5: Open a new Postman tab. To view the data processing status returned by the platform, paste the URL from Step 2, add “/status/UUID” to the end of the request URL and click “Send”. Neuton should return a “SUCCESS” status.

Step 6: To get prediction results, open a new tab in Postman, paste the URL from Step 2, and add “/result/UUID” (copy UUID from Step 4) to the end of the request URL, and click “Send”.

In the prediction results you can find the predicted target variable value, probabilities for each class (for classification task type) or confidence interval (for regression task type) and model-to-data relevance indicator value for each row.

Step 1: Copy the /v2/predict/file URL from Access Instructions, you will use this URL in the Postman request URL specification.

Step 2: Open the Postman application and on a new tab select “POST” request type, then select “Body”, click the “form-data” radio button, type “file” in the KEY field and mark the check box, and then navigate to the file with the new data using the “Select files” button in the VALUE field, as highlighted on the screen below. Paste /v2/predict/file URL (from Step 1) in the request URL and click the “Send” button.

Step 3: Copy returned UUID.

Step 4: Open a new Postman tab. To view the data processing status returned by the Neuton platform, paste the URL from Step 1 in the request URL, replace “/file” with “/status/UUID” at the end of the URL, and click “Send”. Neuton should return a “SUCCESS” status.

Step 5: To get prediction results, open a new tab in Postman, paste the URL from Step 1, replace “/file” with “/result/UUID” at the end of the request URL (copy UUID from Step 3) and click “Send”.