Skip to main content
mySpellChecker does not include a bundled dictionary. You build your own from a text corpus. The data pipeline reads raw text, segments it into syllables and words, calculates N-gram probabilities, and packages everything into an optimized SQLite database.

​
Quick Start

Build a dictionary from the command line: 
# Build from a text corpus
myspellchecker build --input corpus.txt --output dict.db

# Build a sample database for testing
myspellchecker build --sample
Or use the Python API for programmatic control: 
from myspellchecker.data_pipeline import Pipeline

pipeline = Pipeline()
pipeline.build_database(
    input_files=["corpus.txt"],
    database_path="dict.db",
)

​
Pipeline Architecture

Dictionary building pipeline: Input Files → Ingestion → Segmentation → Frequency Counting → Packaging → SQLite Dictionary

​
What the Database Contains

TableContentPurpose
syllablesValid syllables + frequenciesSyllable validation
wordsWords + frequencies + POS tagsWord validation, suggestions
bigramsWord pair probabilitiesContext checking (2-gram)
trigramsWord triple probabilitiesContext checking (3-gram)
fourgrams4-word sequence probabilitiesContext checking (4-gram)
fivegrams5-word sequence probabilitiesContext checking (5-gram)
pos_unigrams/bigrams/trigramsPOS tag probabilitiesGrammar checking
metadataKey-value build metadataBuild info, versioning
processed_filesIngested file paths + timestampsIncremental updates
confusable_pairsConfusable word pairs + type + overlapConfusable error detection
compound_confusionsCompound vs split-word pairs + PMICompound error detection
collocationsWord pair co-occurrence + PMI/NPMICollocation error detection
register_tagsWord register labels (formal/informal)Register mixing detection

​
Next Steps

Corpus Format

Supported input formats and requirements

Building Dictionaries

CLI reference, Python API, config options

Optimization

DuckDB acceleration, Cython, parallel workers

Custom Dictionaries

Curated lexicons, domain-specific builds