From 09da789bc818930eeea4c3cacaabf560a1cfb0fe Mon Sep 17 00:00:00 2001 From: Darshan Gangurde <53848504+darsh169@users.noreply.github.com> Date: Wed, 23 Jun 2021 16:53:40 +0530 Subject: [PATCH] Update generate_tesseract_results.py Saving candidates ( fix #26 ) Please review @CS-savvy --- generate_tesseract_results.py | 6 ++++++ 1 file changed, 6 insertions(+) diff --git a/generate_tesseract_results.py b/generate_tesseract_results.py index 7c90063..5093128 100644 --- a/generate_tesseract_results.py +++ b/generate_tesseract_results.py @@ -5,7 +5,10 @@ from tqdm import tqdm import os import json +from utils import config +import extract_candidates +candidates_dir=str(config.CANDIDATE_DIR)+'/' def get_tesseract_results(image_path): image = cv2.imread(image_path) @@ -29,5 +32,8 @@ def get_tesseract_results(image_path): for image in tqdm(images[:1], desc='Generating Tesseract Results'): image_name = os.path.splitext(os.path.split(image)[-1])[0] result = get_tesseract_results(image) + candidates=extract_candidates.get_candidates(results)#extract_candidates from ocr results with open(os.path.join(tesseract_results, image_name + '.json'), 'w') as f: json.dump(result, f) + with open(candidates_dir+image_name+'.json','w') as c: + json.dump(candidates,c)