CV2-4187-Determine-what-percentage-of-an-image-is-text

app/main/controller/image_ocr_controller.py

@@ -1,13 +1,11 @@
+import json
 from flask import request, current_app as app
 from urllib3 import Retry
 from flask_restplus import Resource, Namespace, fields
 from google.cloud import vision
 import tenacity
-import json
-from google.protobuf.json_format import MessageToJson
-
 
-from app.main.lib.google_client import get_credentialed_google_client
+from app.main.lib.google_client import get_credentialed_google_client, convert_text_annotation_to_json
 
 api = Namespace('ocr', description='ocr operations')
 ocr_request = api.model('ocr_request', {
@@ -39,9 +37,8 @@ def post(self):
  if not texts:
  return
 
- res_dict = MessageToDict(response, preserving_proto_field_name=True)
  app.logger.info(
- f"[Alegre OCR] [image_uri {image.source.image_uri}] Image OCR response package looks like {json.dumps(res_dict)}")
+ f"[Alegre OCR] [image_uri {image.source.image_uri}] Image OCR response package looks like {convert_text_annotation_to_json(texts[0])}")
 
  return {
  'text': texts[0].description

app/main/lib/google_client.py

@@ -1,6 +1,7 @@
 import os
 import json
 from google.oauth2 import service_account
+from flask import current_app as app
 
 def get_credentialed_google_client(client):
  default_values = {}
@@ -26,3 +27,18 @@ def get_credentialed_google_client(client):
  except ValueError as e:
  print(f"Couldn't authenticate to google client: {str(e)}")
  return None
+def convert_text_annotation_to_json(text_annotation):
+ try:
+ text_json = {}
+ text_json['description'] = text_annotation.description
+ text_json['locale'] = text_annotation.locale
+ text_json['bounding_poly'] = []
+ for a_vertice in text_annotation.bounding_poly.vertices:
+ vertice_json = {}
+ vertice_json['x'] = a_vertice.x
+ vertice_json['y'] = a_vertice.y
+ text_json['bounding_poly'] += [vertice_json]
+ text_json = json.dumps(text_json)
+ return text_json
+ except Exception as e:
+ app.logger.exception(e)