org.springframework.cloud.gcp.vision

Class DocumentOcrTemplate

java.lang.Object

org.springframework.cloud.gcp.vision.DocumentOcrTemplate

public class DocumentOcrTemplate
extends Object

Template providing convenient operations for interfacing with Google Cloud Vision's Document OCR feature, which allows you to run OCR algorithms on documents (PDF or TIFF format) stored on Google Cloud Storage.

Author:

Daniel Zou

Constructor Summary

Constructors

Constructor and Description
DocumentOcrTemplate(com.google.cloud.vision.v1.ImageAnnotatorClient imageAnnotatorClient, com.google.cloud.storage.Storage storage, Executor executor, int jsonOutputBatchSize)

Method Summary

Modifier and Type	Method and Description
DocumentOcrResultSet	readOcrOutputFile(GoogleStorageLocation jsonFile) Parses a single JSON output file and returns the list of pages stored in the file.
DocumentOcrResultSet	readOcrOutputFileSet(GoogleStorageLocation jsonOutputFilePathPrefix) Parses the OCR output files who have the specified jsonFilesetPrefix.
org.springframework.util.concurrent.ListenableFuture<DocumentOcrResultSet>	runOcrForDocument(GoogleStorageLocation document, GoogleStorageLocation outputFilePathPrefix) Runs OCR processing for a specified document and generates OCR output files under the path specified by outputFilePathPrefix.

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Detail

DocumentOcrTemplate

public DocumentOcrTemplate(com.google.cloud.vision.v1.ImageAnnotatorClient imageAnnotatorClient,
                           com.google.cloud.storage.Storage storage,
                           Executor executor,
                           int jsonOutputBatchSize)

Method Detail

runOcrForDocument

public org.springframework.util.concurrent.ListenableFuture<DocumentOcrResultSet> runOcrForDocument(GoogleStorageLocation document,
                                                                                                    GoogleStorageLocation outputFilePathPrefix)

Runs OCR processing for a specified document and generates OCR output files under the path specified by outputFilePathPrefix.

For example, if you specify an outputFilePathPrefix of "gs://bucket_name/ocr_results/myDoc_", all the output files of OCR processing will be saved under prefix, such as:

• gs://bucket_name/ocr_results/myDoc_output-1-to-5.json
• gs://bucket_name/ocr_results/myDoc_output-6-to-10.json
• gs://bucket_name/ocr_results/myDoc_output-11-to-15.json

Note: OCR processing operations may take several minutes to complete, so it may not be advisable to block on the completion of the operation. One may use the returned ListenableFuture to register callbacks or track the status of the operation.

Parameters:

document - The GoogleStorageLocation of the document to run OCR processing

outputFilePathPrefix - The GoogleStorageLocation of a file, folder, or a bucket describing the path for which all output files shall be saved under

Returns:

A ListenableFuture allowing you to register callbacks or wait for the completion of the operation.

readOcrOutputFileSet

public DocumentOcrResultSet readOcrOutputFileSet(GoogleStorageLocation jsonOutputFilePathPrefix)

Parses the OCR output files who have the specified jsonFilesetPrefix. This method assumes that all of the OCR output files with the prefix are a part of the same document.

Parameters:

jsonOutputFilePathPrefix - the folder location containing all of the JSON files of OCR output

Returns:

A DocumentOcrResultSet describing the OCR content of a document

readOcrOutputFile

public DocumentOcrResultSet readOcrOutputFile(GoogleStorageLocation jsonFile)

Parses a single JSON output file and returns the list of pages stored in the file.

Each page of the document is represented as a TextAnnotation which contains the parsed OCR data.

Parameters:

jsonFile - the location of the JSON output file

Returns:

the list of TextAnnotation containing the OCR results

Throws:

RuntimeException - if the JSON file cannot be deserialized into a TextAnnotation object