Log inGet started

Custom OCR Java

Suggest Edits

The JAVA OCR SDK supports custom-built API from the API Builder.
If your document isn't covered by one of Mindee's Off-the-Shelf APIs,
you can create your own API using the API Builder.

For the following examples, we are using our own W9s custom API
created with the API Builder.

Quick Start

String path = "/path/to/the/file.ext";
LocalInputSource inputSource = new LocalInputSource(path);
CustomEndpoint myEndpoint = new CustomEndpoint(
    "wnine",
    "john",
    // "1.1" // optional
);

MindeeClient mindeeClient = new MindeeClient(apiKey);
  
PredictResponse<CustomV1> response = mindeeClient
    .parse(inputSource, myEndpoint);

// Print a summary of the response
System.out.println(response.toString());

If the version argument is set, you'll be required to update it every time a new model is trained.
This is probably not needed for development but essential for production use.

The CustomV1 Object

The CustomV1 object contains the results of document-level and page-level predictions.

All the fields which are present in the API builder are available.

The fields are defined when creating your custom API.

Document-Level Predictions

The document-level predictions are ultimately contained in the CustomV1Document object.

The predictions are stored in two properties: classificationFields and fields.

Both are a Map where the key is a String of the name of the field.

For the values:

  • classificationFields have a ClassificationField object as value, each containing a single String value.
  • fields have a ListField object as value, each containing a list of all String values extracted for this field.

Here are some example usages:

String path = "/path/to/the/file.ext";
LocalInputSource inputSource = new LocalInputSource(path);

PredictResponse<CustomV1> response = mindeeClient
    .parse(CustomV1.class, inputSource);

CustomV1 inference = response.getDocument().getInference();
  
// === Getting a single field === \\

ListField employerName = inference.getPrediction().getFields().get("employer_name");

// get the field as a string
System.out.println(employerName.toString());

// get the field as a string with a custom separator
System.out.println(employerName.getContentsString("_"));

// get the list of string values in the field
System.out.println(employerName.getContentsList());

// == Getting all fields === \\

for (Map.Entry<String, ListField> entry : inference.getPrediction().getFields().entrySet()) {
  ListField field = entry.getValue();

  // Not really needed, just showing that the method exists ;-)
  if (field.isEmpty()) {
      continue;
  }

  // We can print directly as in the single field example above ...
  System.out.println(field.toString());

  // ... or go through each value
  for (ListFieldValue value : field.getValues()) {

    // The actual value (word)
    System.out.println(value.getContent());

    // The page on which the value was found
    System.out.println(value.getPageId());
  }
}

Page-Level Predictions

The page-level predictions are ultimately contained in the CustomV1Page object.

In the response, there is a list of these objects, each one representing a single page.

The prediction results are stored as a key-value HashMap<String, ListField>.

Here are some example usages:

String path = "/path/to/the/file.ext";
LocalInputSource inputSource = new LocalInputSource(path);

PredictResponse<CustomV1> response = mindeeClient
    .parse(CustomV1.class, inputSource);

CustomV1 inference = response.getDocument().getInference();
  
for (Page<CustomV1Page> page : inference.getPages()) {
  CustomV1Page pagePrediction = page.getPrediction();

  for (Map.Entry<String, ListField> entry : pagePrediction.entrySet()) {
    ListField field = entry.getValue();
    
    // get the field as a string
    System.out.println(field.toString());

    // get the field as a string with a custom separator
    System.out.println(field.getContentsString("_"));

    // get the list of strings in the field
    System.out.println(field.getContentsList());
  }
}

Questions?

Join our Slack

Updated about 2 months ago