Python CLI

The Python wheel installs suzume, a command for analysis and tag extraction. It uses the same bundled native library and dictionaries as the Python API.

Installation

pip install suzume

The command is available on the same platforms as the Python package: Linux x86_64 (manylinux2014 / manylinux_2_17) and macOS arm64 on macOS 11 or later. See Python Bindings for the full compatibility list.

Input

Pass text as arguments or pipe UTF-8 text to standard input:

suzume "東京へ行く"
printf 'りんごを食べる\n' | suzume

analyze is an optional alias. These commands are equivalent:

suzume --mode search "東京の公園"
suzume analyze --mode search "東京の公園"

Use -- before input that begins with a hyphen.

Output formats

Select the format with -f or --format:

Format	Output
morpheme	Default. surface TAB POS TAB lemma TAB start TAB end
tsv	Same fields as morpheme
tags	tag TAB POS, one tag per line
json	An object containing input, normalized_text, and morphemes
chasen	ChaSen-like fields followed by EOS

suzume --format json "ＡＢＣを検索"
suzume --format chasen "食べている"
suzume --format tags "東京の公園に行く"

Backslashes, tabs, and line breaks inside one-line TAB output are escaped.

Analysis options

Option	Description
-m, --mode MODE	Use normal, search, or split segmentation
--normalize-vu	Normalize ヴ variants instead of preserving them
--lowercase	Normalize ASCII letters to lowercase
--preserve-symbols	Keep punctuation-like SYMBOL tokens; content symbols and emoji remain OTHER either way
--no-lemmatize	Disable post-analysis lemma correction
--merge-compounds	Merge consecutive noun compounds
--skip-env-config	Ignore scorer configuration environment variables

suzume --mode split --lowercase "ABCと東京都"
suzume --normalize-vu --preserve-symbols "ヴァイオリン、ビオラ"

Currency and unit signs, arrows, technical marks, and emoji remain OTHER by default. --preserve-symbols adds punctuation-like tokens such as 、 and 。.

Run suzume --help for the current option summary and suzume --version for the installed native library version.

Dictionaries

Use -d or --dict to load a UTF-8 text dictionary or compiled .dic file. The option is repeatable:

suzume --dict terms.tsv --dict names.dic "東京スカイツリーへ行く"

Option	Description
-d, --dict PATH	Load a text or compiled .dic dictionary; repeat to load more
--no-core-dict	Do not load the bundled core dictionary
--no-user-dict	Do not load the bundled user dictionary

Warnings produced while the analyzer is being created are written to standard error. Warnings produced by a dictionary passed with --dict are not currently printed by the CLI.

Tag output

Choose --format tags to generate keyword tags:

suzume --format tags --tag-max-tags 5 "東京都の天気予報を確認する"

Option	Default	Description
--tag-pos POS	all	Keep one POS; repeat for more
--tag-exclude-basic	off	Exclude tags whose lemma contains only hiragana
--tag-use-surface	off	Use surface forms instead of lemmas
--tag-min-length N	2	Set the minimum tag length
--tag-max-tags N	0	Limit the number of tags (0 means unlimited)
--include-particles	off	Include particles
--include-auxiliaries	off	Include auxiliaries
--include-formal-nouns	off	Include formal nouns
--include-low-info	off	Include low-information words
--tag-keep-duplicates	off	Keep duplicate tags

--tag-pos accepts only noun, verb, adjective, and adverb. To include particles or auxiliaries, omit --tag-pos and use --include-particles or --include-auxiliaries.

suzume --format tags \
  --tag-pos noun \
  --tag-pos verb \
  --tag-min-length 1 \
  "本を読む"

JSON and offsets

Morpheme start and end values are character offsets into the normalized text, not necessarily the original input. JSON output includes that exact string as normalized_text:

suzume --lowercase --format json "ABCを検索"

Use JSON when normalization can change the input and downstream code needs to slice text by the reported offsets.

Scope

The Python suzume command covers analysis and tag extraction. It does not compile or validate dictionaries and does not run the native test or benchmark commands.

Use the separately built suzume-cli for those developer workflows. See Native CLI Reference and Native Build.

See also

• Python Bindings for the Python API and distribution details.
• Installation for all available packages and native builds.