SymSpellPy

Warnings

• breaking SymSpellPy dropped support for Python 3.6 in v6.7.7 and Python 3.7/3.8 in v6.8.0. The library now requires Python 3.9 or higher. Ensure your environment meets this requirement.
  Fix: Upgrade your Python interpreter to version 3.9 or newer. Consider using a virtual environment to manage Python versions.
• breaking As of v6.9.0, frequency counts provided to dictionary methods (e.g., `create_dictionary_entry`, `load_dictionary`) must be 64-bit integers. Passing smaller integers or non-integer types for counts may lead to unexpected behavior or errors.
  Fix: Ensure all frequency counts in your dictionary data are parsed and stored as 64-bit integers before being passed to SymSpellPy.
• breaking In v6.9.0, internal argument names `string1` and `string2` in distance comparison methods were renamed. If you have custom distance comparers or interact with these methods directly, you may need to update your code.
  Fix: Review any custom distance comparer implementations or direct calls to distance calculation methods and adjust argument names if they use `string1` or `string2`.
• gotcha Prior to v6.7.7, `symspellpy` could inadvertently modify the root logger's configuration. Since v6.7.7, it correctly configures its own module logger. If your application relied on `symspellpy`'s old logging behavior, you may need to adjust your logging setup.
  Fix: Explicitly configure `symspellpy`'s logger using `logging.getLogger('symspellpy')` if you need to control its output, or ensure your root logger is configured independently.
• gotcha When using `ignore_term_with_digits=True` in `lookup` or `lookup_compound` methods, it is recommended to also set `ignore_non_words=True` for comprehensive filtering, as clarified in v6.7.3. Failing to do so might result in inconsistent filtering of terms with digits.
  Fix: Always use `ignore_non_words=True` in conjunction with `ignore_term_with_digits=True` to ensure proper behavior.

Install

• pip install symspellpy Install latest version

Imports

• SymSpell

  from symspellpy import SymSpell

• Verbosity

  from symspellpy import Verbosity

• DistanceAlgorithm

  from symspellpy.editdistance import DistanceAlgorithm

  DistanceAlgorithm is located in the `symspellpy.editdistance` submodule, not directly under `symspellpy`.

Quickstart

This quickstart demonstrates how to initialize SymSpell, manually create dictionary entries, and perform basic spell correction for single words and compound phrases. In a production environment, dictionaries are typically loaded from text files using `sym_spell.load_dictionary()`.

from symspellpy import SymSpell, Verbosity
import os

# Initialize SymSpell
sym_spell = SymSpell(max_dictionary_edit_distance=2, prefix_length=7)

# A simple corpus for demonstration (in a real app, load from file)
dictionary_corpus = {"apple": 10, "banana": 8, "aple": 2, "apply": 5}
for term, count in dictionary_corpus.items():
    sym_spell.create_dictionary_entry(term, count)

# Example usage: lookup a word
input_term = "aple"
suggestions = sym_spell.lookup(input_term, Verbosity.CLOSEST, max_edit_distance=2)

print(f"Input: '{input_term}'")
print("Suggestions:")
for suggestion in suggestions:
    print(f"  {suggestion.term} (distance: {suggestion.distance}, count: {suggestion.count})")

# Example usage: lookup compound words
input_phrase = "whereis th elibary"
suggestions_compound = sym_spell.lookup_compound(input_phrase, max_edit_distance=2)

print(f"\nInput phrase: '{input_phrase}'")
print("Compound suggestions:")
for suggestion in suggestions_compound:
    print(f"  {suggestion.term} (distance: {suggestion.distance}, count: {suggestion.count})")