API Reference¶

High-performance text correction for OCR output.

class search_and_replace.OCRCorrector[source]¶

Bases: object

Corrects common OCR character confusions (0/O, 1/l, rn/m, etc.).

correct(text)[source]¶

Apply OCR confusion corrections.

Return type:: str
Parameters:: text (str)

class search_and_replace.PatternCorrector(patterns)[source]¶

Bases: object

Hyperscan-based pattern matching for fuzzy single-character errors.

Parameters:: patterns (Sequence[tuple[str, int]]) – List of (word, max_errors) tuples.

correct(text)[source]¶

Apply pattern corrections.

Return type:: str
Parameters:: text (str)

class search_and_replace.Replacer(replacements)[source]¶

Bases: object

Direct string replacement.

Parameters:: replacements (Sequence[tuple[str, str]])

correct(text)[source]¶

Apply all replacements.

Return type:: str
Parameters:: text (str)

class search_and_replace.SpellCorrector(words=None, dictionary=None, max_distance=2)[source]¶

Bases: object

Levenshtein-based spelling correction using SymSpell.

Parameters:

words (list[str] | None) – Custom word list. If None and dictionary is None, uses bundled dictionary.
dictionary (Path | str | None) – Path to dictionary file (one word per line).
max_distance (int) – Maximum edit distance (default: 2).

correct(word)[source]¶

Correct a single word.

Return type:: str
Parameters:: word (str)

correct_text(text)[source]¶

Correct all words in text.

Return type:: str
Parameters:: text (str)

search_and_replace.load_patterns(path)[source]¶

Load pattern list: word,max_errors per line.

Return type:: list[tuple[str, int]]
Parameters:: path (Path)

search_and_replace.load_replacements(path)[source]¶

Load replacement list: search,replace per line.

Return type:: list[tuple[str, str]]
Parameters:: path (Path)

search_and_replace.process_directory(input_dir, output_dir, pattern_data=None, replacement_data=None, *, pattern='*.txt', resume=False, jobs=None)[source]¶

Process all matching files in parallel.

Parameters:

input_dir (Path) – Source directory.
output_dir (Path) – Destination directory.
pattern_data (Sequence[tuple[str, int]] | None) – List of (word, max_errors) for PatternCorrector.
replacement_data (Sequence[tuple[str, str]] | None) – List of (search, replace) for Replacer.
pattern (str) – Glob pattern (default: *.txt).
resume (bool) – Skip existing output files.
jobs (int | None) – Worker count (default: CPU count).

Return type:

tuple[int, int]

Returns:

(processed_count, skipped_count)

Text correction classes.

class search_and_replace.correctors.SpellCorrector(words=None, dictionary=None, max_distance=2)[source]¶

Bases: object

Levenshtein-based spelling correction using SymSpell.

Parameters:

words (list[str] | None) – Custom word list. If None and dictionary is None, uses bundled dictionary.
dictionary (Path | str | None) – Path to dictionary file (one word per line).
max_distance (int) – Maximum edit distance (default: 2).

correct(word)[source]¶

Correct a single word.

Return type:: str
Parameters:: word (str)

correct_text(text)[source]¶

Correct all words in text.

Return type:: str
Parameters:: text (str)

class search_and_replace.correctors.OCRCorrector[source]¶

Bases: object

Corrects common OCR character confusions (0/O, 1/l, rn/m, etc.).

correct(text)[source]¶

Apply OCR confusion corrections.

Return type:: str
Parameters:: text (str)

class search_and_replace.correctors.PatternCorrector(patterns)[source]¶

Bases: object

Hyperscan-based pattern matching for fuzzy single-character errors.

Parameters:: patterns (Sequence[tuple[str, int]]) – List of (word, max_errors) tuples.

correct(text)[source]¶

Apply pattern corrections.

Return type:: str
Parameters:: text (str)

class search_and_replace.correctors.Replacer(replacements)[source]¶

Bases: object

Direct string replacement.

Parameters:: replacements (Sequence[tuple[str, str]])

correct(text)[source]¶

Apply all replacements.

Return type:: str
Parameters:: text (str)

Batch file processing and I/O.

search_and_replace.batch.load_patterns(path)[source]¶

Load pattern list: word,max_errors per line.

Return type:: list[tuple[str, int]]
Parameters:: path (Path)

search_and_replace.batch.load_replacements(path)[source]¶

Load replacement list: search,replace per line.

Return type:: list[tuple[str, str]]
Parameters:: path (Path)

search_and_replace.batch.process_directory(input_dir, output_dir, pattern_data=None, replacement_data=None, *, pattern='*.txt', resume=False, jobs=None)[source]¶

Process all matching files in parallel.

Parameters:

input_dir (Path) – Source directory.
output_dir (Path) – Destination directory.
pattern_data (Sequence[tuple[str, int]] | None) – List of (word, max_errors) for PatternCorrector.
replacement_data (Sequence[tuple[str, str]] | None) – List of (search, replace) for Replacer.
pattern (str) – Glob pattern (default: *.txt).
resume (bool) – Skip existing output files.
jobs (int | None) – Worker count (default: CPU count).

Return type:

tuple[int, int]

Returns:

(processed_count, skipped_count)

Command-line interface.

search_and_replace.cli.main(argv=None)[source]¶

Main entry point.

Return type:: int
Parameters:: argv (list[str] | None)