Lemmatization is the process of grouping together the inflected forms of a word so they can be analysed as a single item, identified by the word's lemma, or dictionary form. Unlike stemming, lemmatization outputs word units that are still valid linguistic forms.
In modern natural language processing (NLP), this task is often indirectly tackled by more complex systems encompassing a whole processing pipeline. However, it appears that there is no straightforward way to address lemmatization in Python although this task can be crucial in fields such as information retrieval and NLP.
Simplemma provides a simple and multilingual approach to look for base forms or lemmata. It may not be as powerful as full-fledged solutions but it is generic, easy to install and straightforward to use. In particular, it does not need morphosyntactic information and can process a raw series of tokens or even a text with its built-in tokenizer. By design it should be reasonably fast and work in a large majority of cases, without being perfect.
With its comparatively small footprint it is especially useful when speed and simplicity matter, in low-resource contexts, for educational purposes, or as a baseline system for lemmatization and morphological analysis.
Currently, 49 languages are partly or fully supported (see table below).
The current library is written in pure Python with no dependencies:
pip install simplemma
-
pip3
where applicable -
pip install -U simplemma
for updates -
pip install git+https://github.com/adbar/simplemma
for the cutting-edge version
The last version supporting Python 3.6 and 3.7 is simplemma==1.0.0
.
Simplemma is used by selecting a language of interest and then applying the data on a list of words.
>>> import simplemma
# get a word
myword = 'masks'
# decide which language to use and apply it on a word form
>>> simplemma.lemmatize(myword, lang='en')
'mask'
# grab a list of tokens
>>> mytokens = ['Hier', 'sind', 'Vaccines']
>>> for token in mytokens:
>>> simplemma.lemmatize(token, lang='de')
'hier'
'sein'
'Vaccines'
# list comprehensions can be faster
>>> [simplemma.lemmatize(t, lang='de') for t in mytokens]
['hier', 'sein', 'Vaccines']
Chaining several languages can improve coverage, they are used in sequence:
>>> from simplemma import lemmatize
>>> lemmatize('Vaccines', lang=('de', 'en'))
'vaccine'
>>> lemmatize('spaghettis', lang='it')
'spaghettis'
>>> lemmatize('spaghettis', lang=('it', 'fr'))
'spaghetti'
>>> lemmatize('spaghetti', lang=('it', 'fr'))
'spaghetto'
For certain languages a greedier decomposition is activated by default
as it can be beneficial, mostly due to a certain capacity to address
affixes in an unsupervised way. This can be triggered manually by
setting the greedy
parameter to True
.
This option also triggers a stronger reduction through an additional iteration of the search algorithm, e.g. "angekündigten" → "angekündigt" (standard) → "ankündigen" (greedy). In some cases it may be closer to stemming than to lemmatization.
# same example as before, comes to this result in one step
>>> simplemma.lemmatize('spaghettis', lang=('it', 'fr'), greedy=True)
'spaghetto'
# German case described above
>>> simplemma.lemmatize('angekündigten', lang='de', greedy=True)
'ankündigen' # 2 steps: reduction to infinitive verb
>>> simplemma.lemmatize('angekündigten', lang='de', greedy=False)
'angekündigt' # 1 step: reduction to past participle
The additional function is_known()
checks if a given word is present
in the language data:
>>> from simplemma import is_known
>>> is_known('spaghetti', lang='it')
True
A simple tokenization function is provided for convenience:
>>> from simplemma import simple_tokenizer
>>> simple_tokenizer('Lorem ipsum dolor sit amet, consectetur adipiscing elit, sed do eiusmod tempor incididunt ut labore et dolore magna aliqua.')
['Lorem', 'ipsum', 'dolor', 'sit', 'amet', ',', 'consectetur', 'adipiscing', 'elit', ',', 'sed', 'do', 'eiusmod', 'tempor', 'incididunt', 'ut', 'labore', 'et', 'dolore', 'magna', 'aliqua', '.']
# use iterator instead
>>> simple_tokenizer('Lorem ipsum dolor sit amet', iterate=True)
The functions text_lemmatizer()
and lemma_iterator()
chain
tokenization and lemmatization. They can take greedy
(affecting
lemmatization) and silent
(affecting errors and logging) as arguments:
>>> from simplemma import text_lemmatizer
>>> sentence = 'Sou o intervalo entre o que desejo ser e os outros me fizeram.'
>>> text_lemmatizer(sentence, lang='pt')
# caveat: desejo is also a noun, should be desejar here
['ser', 'o', 'intervalo', 'entre', 'o', 'que', 'desejo', 'ser', 'e', 'o', 'outro', 'me', 'fazer', '.']
# same principle, returns a generator and not a list
>>> from simplemma import lemma_iterator
>>> lemma_iterator(sentence, lang='pt')
# don't expect too much though
# this diminutive form isn't in the model data
>>> simplemma.lemmatize('spaghettini', lang='it')
'spaghettini' # should read 'spaghettino'
# the algorithm cannot choose between valid alternatives yet
>>> simplemma.lemmatize('son', lang='es')
'son' # valid common name, but what about the verb form?
As the focus lies on overall coverage, some short frequent words (typically: pronouns and conjunctions) may need post-processing, this generally concerns a few dozens of tokens per language.
The current absence of morphosyntactic information is an advantage in
terms of simplicity. However, it is also an impassable frontier regarding
lemmatization accuracy, for example when it comes to disambiguating
between past participles and adjectives derived from verbs in Germanic
and Romance languages. In most cases, simplemma
often does not change
such input words.
The greedy algorithm seldom produces invalid forms. It is designed to work best in the low-frequency range, notably for compound words and neologisms. Aggressive decomposition is only useful as a general approach in the case of morphologically-rich languages, where it can also act as a linguistically motivated stemmer.
Bug reports over the issues page are welcome.
Language detection works by providing a text and tuple lang
consisting
of a series of languages of interest. Scores between 0 and 1 are
returned.
The lang_detector()
function returns a list of language codes along
with their corresponding scores, appending "unk" for unknown or
out-of-vocabulary words. The latter can also be calculated by using
the function in_target_language()
which returns a ratio.
# import necessary functions
>>> from simplemma import in_target_language, lang_detector
# language detection
>>> lang_detector('"Exoplaneta, též extrasolární planeta, je planeta obíhající kolem jiné hvězdy než kolem Slunce."', lang=("cs", "sk"))
[("cs", 0.75), ("sk", 0.125), ("unk", 0.25)]
# proportion of known words
>>> in_target_language("opera post physica posita (τὰ μετὰ τὰ φυσικά)", lang="la")
0.5
The greedy
argument (extensive
in past software versions) triggers
use of the greedier decomposition algorithm described above, thus
extending word coverage and recall of detection at the potential cost of
a lesser accuracy.
The functions described above are suitable for simple usage, but you
can have more control by instantiating Simplemma classes and calling
their methods instead. Lemmatization is handled by the Lemmatizer
class, while language detection is handled by the LanguageDetector
class. These in turn rely on different lemmatization strategies, which
are implementations of the LemmatizationStrategy
protocol. The
DefaultStrategy
implementation uses a combination of different
strategies, one of which is DictionaryLookupStrategy
. It looks up
tokens in a dictionary created by a DictionaryFactory
.
For example, it is possible to conserve RAM by limiting the number of
cached language dictionaries (default: 8) by creating a custom
DefaultDictionaryFactory
with a specific cache_max_size
setting,
creating a DefaultStrategy
using that factory, and then creating a
Lemmatizer
and/or a LanguageDetector
using that strategy:
# import necessary classes
>>> from simplemma import LanguageDetector, Lemmatizer
>>> from simplemma.strategies import DefaultStrategy
>>> from simplemma.strategies.dictionaries import DefaultDictionaryFactory
LANG_CACHE_SIZE = 5 # How many language dictionaries to keep in memory at once (max)
>>> dictionary_factory = DefaultDictionaryFactory(cache_max_size=LANG_CACHE_SIZE)
>>> lemmatization_strategy = DefaultStrategy(dictionary_factory=dictionary_factory)
# lemmatize using the above customized strategy
>>> lemmatizer = Lemmatizer(lemmatization_strategy=lemmatization_strategy)
>>> lemmatizer.lemmatize('doughnuts', lang='en')
'doughnut'
# detect languages using the above customized strategy
>>> language_detector = LanguageDetector('la', lemmatization_strategy=lemmatization_strategy)
>>> language_detector.proportion_in_target_languages("opera post physica posita (τὰ μετὰ τὰ φυσικά)")
0.5
For more information see the extended documentation.
Simplemma provides an alternative solution for situations where low
memory usage and fast initialization time are more important than
lemmatization and language detection performance. This solution uses a
DictionaryFactory
that employs a trie as its underlying data structure,
rather than a Python dict.
The TrieDictionaryFactory
reduces memory usage by an average of
20x and initialization time by 100x, but this comes at the cost of
potentially reducing performance by 50% or more, depending on the
specific usage.
To use the TrieDictionaryFactory
you have to install Simplemma with
the marisa-trie
extra dependency (available from version 1.1.0):
pip install simplemma[marisa-trie]
Then you have to create a custom strategy using the
TrieDictionaryFactory
and use that for Lemmatizer
and
LanguageDetector
instances:
>>> from simplemma import LanguageDetector, Lemmatizer
>>> from simplemma.strategies import DefaultStrategy
>>> from simplemma.strategies.dictionaries import TrieDictionaryFactory
>>> lemmatization_strategy = DefaultStrategy(dictionary_factory=TrieDictionaryFactory())
>>> lemmatizer = Lemmatizer(lemmatization_strategy=lemmatization_strategy)
>>> lemmatizer.lemmatize('doughnuts', lang='en')
'doughnut'
>>> language_detector = LanguageDetector('la', lemmatization_strategy=lemmatization_strategy)
>>> language_detector.proportion_in_target_languages("opera post physica posita (τὰ μετὰ τὰ φυσικά)")
0.5
While memory usage and initialization time when using the
TrieDictionaryFactory
are significantly lower compared to the
DefaultDictionaryFactory
, that's only true if the trie dictionaries
are available on disk. That's not the case when using the
TrieDictionaryFactory
for the first time, as Simplemma only ships
the dictionaries as Python dicts. The trie dictionaries have to be
generated once from the Python dicts. That happens on-the-fly when
using the TrieDictionaryFactory
for the first time for a language and
will take a few seconds and use as much memory as loading the Python
dicts for the language requires. For further invocations the trie
dictionaries get cached on disk.
If the computer supposed to run Simplemma doesn't have enough memory to generate the trie dictionaries, they can also be generated on another computer with the same CPU architecture and copied over to the cache directory.
The following languages are available, identified by their BCP 47 language tag, which typically corresponds to the ISO 639-1 code. If no such code exists, a ISO 639-3 code is used instead.
Available languages (2022-01-20):
Code | Language | Forms (10³) | Acc. | Comments |
---|---|---|---|---|
ast |
Asturian | 124 | ||
bg |
Bulgarian | 204 | ||
ca |
Catalan | 579 | ||
cs |
Czech | 187 | 0.89 | on UD CS-PDT |
cy |
Welsh | 360 | ||
da |
Danish | 554 | 0.92 | on UD DA-DDT, alternative: lemmy |
de |
German | 675 | 0.95 | on UD DE-GSD, see also German-NLP list |
el |
Greek | 181 | 0.88 | on UD EL-GDT |
en |
English | 131 | 0.94 | on UD EN-GUM, alternative: LemmInflect |
enm |
Middle English | 38 | ||
es |
Spanish | 665 | 0.95 | on UD ES-GSD |
et |
Estonian | 119 | low coverage | |
fa |
Persian | 12 | experimental | |
fi |
Finnish | 3,199 | see this benchmark | |
fr |
French | 217 | 0.94 | on UD FR-GSD |
ga |
Irish | 372 | ||
gd |
Gaelic | 48 | ||
gl |
Galician | 384 | ||
gv |
Manx | 62 | ||
hbs |
Serbo-Croatian | 656 | Croatian and Serbian lists to be added later | |
hi |
Hindi | 58 | experimental | |
hu |
Hungarian | 458 | ||
hy |
Armenian | 246 | ||
id |
Indonesian | 17 | 0.91 | on UD ID-CSUI |
is |
Icelandic | 174 | ||
it |
Italian | 333 | 0.93 | on UD IT-ISDT |
ka |
Georgian | 65 | ||
la |
Latin | 843 | ||
lb |
Luxembourgish | 305 | ||
lt |
Lithuanian | 247 | ||
lv |
Latvian | 164 | ||
mk |
Macedonian | 56 | ||
ms |
Malay | 14 | ||
nb |
Norwegian (Bokmål) | 617 | ||
nl |
Dutch | 250 | 0.92 | on UD-NL-Alpino |
nn |
Norwegian (Nynorsk) | 56 | ||
pl |
Polish | 3,211 | 0.91 | on UD-PL-PDB |
pt |
Portuguese | 924 | 0.92 | on UD-PT-GSD |
ro |
Romanian | 311 | ||
ru |
Russian | 595 | alternative: pymorphy2 | |
se |
Northern Sámi | 113 | ||
sk |
Slovak | 818 | 0.92 | on UD SK-SNK |
sl |
Slovene | 136 | ||
sq |
Albanian | 35 | ||
sv |
Swedish | 658 | alternative: lemmy | |
sw |
Swahili | 10 | experimental | |
tl |
Tagalog | 32 | experimental | |
tr |
Turkish | 1,232 | 0.89 | on UD-TR-Boun |
uk |
Ukrainian | 370 | alternative: pymorphy2 |
Languages marked as having low coverage may be better suited to language-specific libraries, but Simplemma can still provide limited functionality. Where possible, open-source Python alternatives are referenced.
Experimental mentions indicate that the language remains untested or that there could be issues with the underlying data or lemmatization process.
The scores are calculated on Universal
Dependencies treebanks on single
word tokens (including some contractions but not merged prepositions),
they describe to what extent simplemma can accurately map tokens to
their lemma form. See eval/
folder of the code repository for more
information.
This library is particularly relevant as regards the lemmatization of less frequent words. Its performance in this case is only incidentally captured by the benchmark above. In some languages, a fixed number of words such as pronouns can be further mapped by hand to enhance performance.
The following orders of magnitude are provided for reference only and were measured on an old laptop to establish a lower bound:
- Tokenization: > 1 million tokens/sec
- Lemmatization: > 250,000 words/sec
Using the most recent Python version (i.e. with pyenv
) can make the
package run faster.
- Add further lemmatization lists
- Grammatical categories as option
- Function as a meta-package?
- Integrate optional, more complex models?
The software is licensed under the MIT license. For information on the
licenses of the linguistic information databases, see the licenses
folder.
The surface lookups (non-greedy mode) rely on lemmatization lists derived from the following sources, listed in order of relative importance:
- Lemmatization lists by Michal Měchura (Open Database License)
- Wiktionary entries packaged by the Kaikki project
- FreeLing project
- spaCy lookups data
- Unimorph Project
- Wikinflection corpus by Eleni Metheniti (CC BY 4.0 License)
This package has been first created and published by Adrien Barbaresi. It has then benefited from extensive refactoring by Juanjo Diaz (especially the new classes). See the full list of contributors to the repository.
Feel free to contribute, notably by filing issues for feedback, bug reports, or links to further lemmatization lists, rules and tests.
Contributions by pull requests ought to follow the following conventions: code style with black, type hinting with mypy, included tests with pytest.
See lists: German-NLP and other awesome-NLP lists.
For another approach in Python see Spacy's edit tree lemmatizer.
To cite this software:
Barbaresi A. (year). Simplemma: a simple multilingual lemmatizer for Python [Computer software] (Version version number). Berlin, Germany: Berlin-Brandenburg Academy of Sciences. Available from https://github.com/adbar/simplemma DOI: 10.5281/zenodo.4673264
This work draws from lexical analysis algorithms used in:
- Barbaresi, A., & Hein, K. (2017). Data-driven identification of German phrasal compounds. In International Conference on Text, Speech, and Dialogue Springer, pp. 192-200.
- Barbaresi, A. (2016). An unsupervised morphological criterion for discriminating similar languages. In 3rd Workshop on NLP for Similar Languages, Varieties and Dialects (VarDial 2016), Association for Computational Linguistics, pp. 212-220.
- Barbaresi, A. (2016). Bootstrapped OCR error detection for a less-resourced language variant. In 13th Conference on Natural Language Processing (KONVENS 2016), pp. 21-26.