Returned JSON File From TEXT_DETECTION Is Not Properly Formatted

Question

I followed Google's documentation for their Vision API to write a program that runs TEXT_DETECTION on an image. Unlike the documentation and Google Vision the tester on their website which returns a properly formatted JSON, the JSON file I get in return is missing commas in between Objects and is missing brackets for JSONArrays.

I use the following line to get the JSON response.

BatchAnnotateImagesResponse response = vision.batchAnnotateImages(requests);

Here is my code for the TEXT_DETECTION.

    public static final Type SEARCH_TYPE = Type.TEXT_DETECTION;

    public static void main(String... args) throws Exception {
        try (ImageAnnotatorClient vision = ImageAnnotatorClient.create()) {
            String fileName = "resources/testPic.jpg";

            Path path = Paths.get(fileName);
            byte[] data = Files.readAllBytes(path);
            ByteString imgBytes = ByteString.copyFrom(data);
            List<AnnotateImageRequest> requests = new ArrayList<>();
            Image img = Image.newBuilder().setContent(imgBytes).build();
            Feature feat = Feature.newBuilder().setType(SEARCH_TYPE).build();
            AnnotateImageRequest request = AnnotateImageRequest.newBuilder().addFeatures(feat).setImage(img).build();
            requests.add(request);

            BatchAnnotateImagesResponse response = vision.batchAnnotateImages(requests);
            List<AnnotateImageResponse> responses = response.getResponsesList();

Here is a snippet of what my responses look like from BatchAnnotateImagesResponse.

 text_annotations {
    description: "almora"
    bounding_poly {
      vertices {
        x: 2307
        y: 713
      }
      vertices {
        x: 2480
        y: 711
      }
      vertices {
        x: 2480
        y: 727
      }
      vertices {
        x: 2307
        y: 729
      }
    }
  }
  text_annotations {
    description: "ryan"
    bounding_poly {
      vertices {
        x: 2458
        y: 906
      }
      vertices {
        x: 2489
        y: 933
      }
      vertices {
        x: 2476
        y: 947
      }
      vertices {
        x: 2446
        y: 920
      }
    }
  }
  text_annotations {
    description: "flanco"
    bounding_poly {
      vertices {
        x: 2441
        y: 890
      }
      vertices {
        x: 2458
        y: 905
      }
      vertices {
        x: 2443
        y: 922
      }
      vertices {
        x: 2426
        y: 907
      }
    }
  }
  text_annotations {
    description: "garbanzo beans"
    bounding_poly {
      vertices {
        x: 3780
        y: 1051
      }
      vertices {
        x: 3824
        y: 1050
      }
      vertices {
        x: 3824
        y: 1063
      }
      vertices {
        x: 3780
        y: 1064
      }
    }
  }
  text_annotations {
    description: "roberto"
    bounding_poly {
      vertices {
        x: 2111
        y: 906
      }
      vertices {
        x: 2163
        y: 905
      }
      vertices {
        x: 2163
        y: 920
      }
      vertices {
        x: 2111
        y: 921
      }
    }
  }
  text_annotations {
    description: "A10"
    bounding_poly {
      vertices {
        x: 2398
        y: 935
      }
      vertices {
        x: 2442
        y: 972
      }
      vertices {
        x: 2424
        y: 994
      }
      vertices {
        x: 2380
        y: 956
      }
    }
  }

How can I change it or fix it so that the returned file is a properly formatted JSON file?

Answer 1

You can add

com.google.protobuf.util.JsonFormat.printer().print(responses)

after

List<AnnotateImageResponse> responses = response.getResponsesList();

This will give you JSON string for your response object.

Answer 2

When using annotate, there is no json file. The response is a BatchAnnotateImagesResponse object.

If you want to generate a json file, you can use asyncBatchAnnotate instead. asyncBatchAnnotate writes json files to your gcs bucket when it finishes.