feat(ocr): German spell-check post-processing to reduce handwriting gibberish #260

Merged
marcel merged 10 commits from feat/issue-254-german-spell-check into main 2026-04-17 17:28:41 +02:00
2 changed files with 4 additions and 4 deletions
Showing only changes of commit fea24aee25 - Show all commits

View File

@@ -9,7 +9,7 @@ ILLEGIBLE_MARKER = "[unleserlich]"
CORRECTION_MARKER = "[?]"
def _collapse_adjacent_markers(tokens: list[str]) -> list[str]:
def collapse_adjacent_markers(tokens: list[str]) -> list[str]:
collapsed: list[str] = []
prev_was_marker = False
for token in tokens:
@@ -52,7 +52,7 @@ def apply_confidence_markers(words: list[dict], threshold: float | None = None)
else:
tokens.append(word["text"])
return " ".join(_collapse_adjacent_markers(tokens))
return " ".join(collapse_adjacent_markers(tokens))
def words_from_characters(prediction: str, confidences: list[float]) -> list[dict]:

View File

@@ -5,7 +5,7 @@ import os
from spellchecker import SpellChecker
from confidence import CORRECTION_MARKER, ILLEGIBLE_MARKER, _collapse_adjacent_markers
from confidence import CORRECTION_MARKER, ILLEGIBLE_MARKER, collapse_adjacent_markers
logger = logging.getLogger(__name__)
@@ -114,4 +114,4 @@ def correct_text(text: str) -> str:
else:
checked.append(ILLEGIBLE_MARKER)
return " ".join(_collapse_adjacent_markers(checked))
return " ".join(collapse_adjacent_markers(checked))