From 40bf45c1a7e6567761d2adc6c875f6948693d95b Mon Sep 17 00:00:00 2001
From: Vit Novotny <witiko@mail.muni.cz>
Date: Wed, 2 Mar 2022 22:03:26 +0000
Subject: [PATCH] Add scripts.combine_tesseract_with_google_docker
---
.../combine_tesseract_with_google_docker.py | 72 +++++++++++++++++++
1 file changed, 72 insertions(+)
create mode 100644 scripts/combine_tesseract_with_google_docker.py
diff --git a/scripts/combine_tesseract_with_google_docker.py b/scripts/combine_tesseract_with_google_docker.py
new file mode 100644
index 00000000..5f4c8860
--- /dev/null
+++ b/scripts/combine_tesseract_with_google_docker.py
@@ -0,0 +1,72 @@
+import json
+from json import JSONDecodeError
+from pathlib import Path
+import shutil
+import sys
+
+from scripts.common import is_multicolumn
+
+
+TESSERACT_PATH = Path(sys.argv[1])
+GOOGLE_VISION_AI_PATH = Path(sys.argv[2])
+OUTPUT_PATH = Path(sys.argv[3])
+
+
+def read_google_json(filelike) -> str:
+ try:
+ ocr_output = json.load(filelike)
+ except JSONDecodeError:
+ return ''
+ if not ocr_output['responses'][0]:
+ return ''
+ return ocr_output['responses'][0]['fullTextAnnotation']['text']
+
+
+def main() -> None:
+ with (TESSERACT_PATH / 'list.txt').open('rt') as tf, \
+ (GOOGLE_VISION_AI_PATH / 'list.txt').open('rt') as gf:
+ tesseract_lines = list(tf)
+ google_lines = list(gf)
+
+ for tesseract_line, google_line in zip(tesseract_lines, google_lines):
+ tesseract_filename = Path(tesseract_line.rstrip('\r\n'))
+ google_filename = Path(google_line.rstrip('\r\n'))
+
+ input_tesseract_filename = (TESSERACT_PATH / tesseract_filename.stem).with_suffix('.txt')
+ input_google_filename = (GOOGLE_VISION_AI_PATH / google_filename.stem).with_suffix('.json')
+ output_filename = (OUTPUT_PATH / google_filename.stem).with_suffix('.txt')
+ decision_filename = (OUTPUT_PATH / google_filename.stem).with_suffix('.ground-truth')
+
+ for suffix in ('hocr', 'tsv', 'box', 'page'):
+ shutil.copy(
+ (TESSERACT_PATH / tesseract_filename.stem).with_suffix(f'.{suffix}'),
+ (OUTPUT_PATH / google_filename.stem).with_suffix(f'.{suffix}'),
+ )
+
+ for suffix in ('json', ):
+ shutil.copy(
+ (GOOGLE_VISION_AI_PATH / google_filename.stem).with_suffix(f'.{suffix}'),
+ (OUTPUT_PATH / google_filename.stem).with_suffix(f'.{suffix}'),
+ )
+
+ with input_tesseract_filename.with_suffix('.hocr').open('rt') as f:
+ hocr_content = f.read()
+
+ if not hocr_content.strip():
+ with output_filename.open('wt') as f:
+ print(file=f)
+ continue
+
+ if is_multicolumn(input_tesseract_filename.with_suffix('.hocr')):
+ shutil.copy(input_tesseract_filename, output_filename)
+ with decision_filename.open('wt') as f:
+ print('hocr', file=f)
+ else:
+ with input_google_filename.open('rt') as rf, output_filename.open('wt') as wf:
+ print(read_google_json(rf), file=wf)
+ with decision_filename.open('wt') as f:
+ print('json', file=f)
+
+
+if __name__ == '__main__':
+ main()
--
GitLab