sudacode/owocr
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Make furigana filter independent from --language option

Browse Source
This commit is contained in:
AuroraWright
2025-10-20 14:13:28 +02:00
parent d5ebb292d5
commit 665837a764
1 changed files with 7 additions and 5 deletions
Download Patch File Download Diff File Expand all files Collapse all files

12
owocr/run.py
View File

@@ -307,7 +307,7 @@ class TextFiltering:
self.json_output = config.get_general('output_format') == 'json' self.json_output = config.get_general('output_format') == 'json'
self.frame_stabilization = 0 if config.get_general('screen_capture_delay_secs') == -1 else config.get_general('screen_capture_frame_stabilization') self.frame_stabilization = 0 if config.get_general('screen_capture_delay_secs') == -1 else config.get_general('screen_capture_frame_stabilization')
self.line_recovery = not self.json_output and config.get_general('screen_capture_line_recovery') self.line_recovery = not self.json_output and config.get_general('screen_capture_line_recovery')
self.furigana_filter = self.language == 'ja' and config.get_general('furigana_filter') self.furigana_filter = config.get_general('furigana_filter')
self.last_frame_data = (None, None) self.last_frame_data = (None, None)
self.last_last_frame_data = (None, None) self.last_last_frame_data = (None, None)
self.stable_frame_data = None self.stable_frame_data = None
@@ -562,7 +562,7 @@ class TextFiltering:
if all(not current_text_line for current_text_line in current_lines): if all(not current_text_line for current_text_line in current_lines):
return None return None
if self.furigana_filter and isinstance(current_result_ocr, OcrResult): if self.furigana_filter and self.language == 'ja' and isinstance(current_result_ocr, OcrResult):
for p in current_result_ocr.paragraphs: for p in current_result_ocr.paragraphs:
current_lines_ocr.extend(p.lines) current_lines_ocr.extend(p.lines)
@@ -700,7 +700,9 @@ class TextFiltering:
lines_ocr = [] lines_ocr = []
for line in result: for line in result:
text_line = self._normalize_line_for_comparison(line) if not line:
lines.append(line)
text_line = ''.join(self.cj_regex.findall(line))
lines.append(text_line) lines.append(text_line)
if all(not text_line for text_line in lines): if all(not text_line for text_line in lines):
return result return result
@@ -711,12 +713,12 @@ class TextFiltering:
for i, text in enumerate(lines): for i, text in enumerate(lines):
filtered_line = result[i] filtered_line = result[i]
logger.opt(colors=True).debug(f"<magenta>Line: '{text}'</magenta>")
if not text: if not text:
filtered_lines.append(filtered_line) filtered_lines.append(filtered_line)
continue continue
logger.opt(colors=True).debug(f"<magenta>Line: '{text}'</magenta>")
is_furigana = self._furigana_filter(lines, lines_ocr, text, i) is_furigana = self._furigana_filter(lines, lines_ocr, text, i)
if is_furigana: if is_furigana:
continue continue