Install & Compatibility
Where this runs
tested against v20260723 · pip install
no network on importno background threads
Install × environment matrix
Each cell = how many times install + import succeeded across repeated harness runs. Partial = flaky.
glibc = Debian/Ubuntu slim · musl = Alpine Linux
muslpy 3.10–3.95 runs
build_error
glibcpy 3.10–3.95 runs
installs and imports cleanly · install 25.0s · import 0.000s · 367MB
364MB installed
● package 364MB
Code
Verified usage
Verified import paths — ran on the pinned version, not inferred.
Path
✓ from sudachidict_full.dictionary import Path
This provides the file path to the installed 'full' dictionary, useful for explicit configuration with SudachiPy, though SudachiPy often finds it automatically.
This quickstart demonstrates how to use `sudachipy` with the `sudachidict-full` dictionary. After installing `sudachidict-full` alongside `sudachipy`, `sudachipy` will automatically detect and use the full dictionary by default when creating a `Dictionary` instance. The example tokenizes a simple Japanese sentence.
from sudachipy import tokenizer
from sudachipy import dictionary
# sudachidict-full must be installed for this to load the full dictionary.
# SudachiPy automatically selects the largest installed dictionary by default.
# To explicitly ensure the 'full' dictionary is used, you can pass dict_type='full'.
# tokenizer_obj = dictionary.Dictionary(dict_type='full').create()
# Create a Sudachi tokenizer instance (will use the 'full' dict if installed)
tokenizer_obj = dictionary.Dictionary().create()
mode = tokenizer.Tokenizer.SplitMode.C
text = "寿司は美味しい。"
# Tokenize the text
print(f"Original text: {text}")
morphemes = tokenizer_obj.tokenize(text, mode)
print("\nTokenization results (Surface form, Part-of-Speech, Base form):")
for m in morphemes:
print(f" {m.surface()}\t{m.part_of_speech()}\t{m.base_form()}")
# Example of getting the dictionary path (for advanced configuration)
# import sudachidict_full
# dict_path = sudachidict_full.dictionary.Path()
# print(f"\nPath to the 'full' dictionary: {dict_path}")
Debug
Known issues
gotchasudachidict-full is a data package providing dictionary files, not a standalone library for performing morphological analysis. You must install `sudachipy` separately to utilize the dictionary data.fixEnsure `sudachipy` is installed: `pip install sudachipy`.
affects: All versions of sudachidict-full
breakingStarting from version 20251022, Sudachi's internal dictionary normalization has been partly discontinued and replaced with a synonym dictionary. This change may lead to different tokenization results or altered behavior for applications relying on the previous normalization process.fixReview your application's reliance on Sudachi's normalization features and adapt post-processing or logic as needed. Test with affected inputs to understand the new behavior.
affects: >=20251022
gotchaWhen multiple Sudachi dictionaries (e.g., `sudachidict-small`, `sudachidict-core`, `sudachidict-full`) are installed, `sudachipy`'s default `dictionary.Dictionary().create()` method will automatically prioritize and load the largest available dictionary. To explicitly guarantee the 'full' dictionary is used, you can initialize with `dict_type='full'` (e.g., `dictionary.Dictionary(dict_type='full').create()`).fixEither install only `sudachidict-full` if you want the default behavior, or explicitly specify `dict_type='full'` in your `sudachipy` code for clarity and control.
affects: All versions where multiple dictionaries might be installed
Upgrade
Version history
20260723latest on PyPI · released Jul 24, 2026
Audit
Dependencies
sudachipyrequiredThis package provides dictionary data for SudachiPy. SudachiPy is required to perform morphological analysis.