Skip to content

Natural Language Processing Pipeline - Sentence Splitting, Tokenization, Lemmatization, Part-of-speech Tagging and Dependency Parsing

License

Notifications You must be signed in to change notification settings

KoichiYasuoka/NLP-Cube

 
 

Repository files navigation

Monthly Weekly daily Version Python 3

News

[15 April 2019] - We are releasing version 1.1 models - check all supported languages below. Both models 1.0 and 1.1 are trained on the same UD2.2 corpus; however, models 1.1 do not use vector embeddings, thus reducing the time and disk space required to download them. Some languages actually have a slightly increased accuracy, some a bit decreased. By default, NLP Cube will use the latest (at this time) 1.1 models.

To use the older 1.0 models just specify this version in the load call: cube.load("en",1.0) (en for English, or any other language code). This will download (if not already downloaded) and use this specific model version. Same goes for any language/version you want to use.

If you already have NLP Cube installed and want to use the newer 1.1 models, type either cube.load("en",1.1) or cube.load("en","latest") to auto-download them. After this, calling cube.load("en") without version number will automatically use the latest ones from your disk.


NLP-Cube

NLP-Cube is an opensource Natural Language Processing Framework with support for languages which are included in the UD Treebanks (list of all available languages below). Use NLP-Cube if you need:

  • Sentence segmentation
  • Tokenization
  • POS Tagging (both language independent (UPOSes) and language dependent (XPOSes and ATTRs))
  • Lemmatization
  • Dependency parsing

Example input: "This is a test.", output is:

1       This    this    PRON    DT      Number=Sing|PronType=Dem        4       nsubj   _
2       is      be      AUX     VBZ     Mood=Ind|Number=Sing|Person=3|Tense=Pres|VerbForm=Fin   4       cop     _
3       a       a       DET     DT      Definite=Ind|PronType=Art       4       det     _
4       test    test    NOUN    NN      Number=Sing     0       root    SpaceAfter=No
5       .       .       PUNCT   .       _       4       punct   SpaceAfter=No

If you just want to run it, here's how to set it up and use NLP-Cube in a few lines: Quick Start Tutorial.

For advanced users that want to create and train their own models, please see the Advanced Tutorials in examples/, starting with how to locally install NLP-Cube.

Simple (PIP) installation / update version

Install (or update) NLP-Cube with:

pip3 install -U nlpcube

API Usage

To use NLP-Cube *programmatically (in Python), follow this tutorial The summary would be:

from cube.api import Cube       # import the Cube object
cube=Cube(verbose=True)         # initialize it
cube.load("en")                 # select the desired language (it will auto-download the model on first run)
text="This is the text I want segmented, tokenized, lemmatized and annotated with POS and dependencies."
sentences=cube(text)            # call with your own text (string) to obtain the annotations

The sentences object now contains the annotated text, one sentence at a time. To print the third words's POS (in the first sentence), just run:

print(sentences[0][2].upos) # [0] is the first sentence and [2] is the third word

Each token object has the following attributes: index, word, lemma, upos, xpos, attrs, head, label, deps, space_after. For detailed info about each attribute please see the standard CoNLL format.

Webserver Usage

To use NLP-Cube as a web service, you need to locally install NLP-Cube and start the server:

For example, the following command will start the server and preload languages: en, fr and de.

cd cube
python3 webserver.py --port 8080 --lang=en --lang=fr --lang=de

To test, open the following link (please copy the address of the link as it is a local address and port link)

Cite

If you use NLP-Cube in your research we would be grateful if you would cite the following paper:

  • NLP-Cube: End-to-End Raw Text Processing With Neural Networks, Boroș, Tiberiu and Dumitrescu, Stefan Daniel and Burtica, Ruxandra, Proceedings of the CoNLL 2018 Shared Task: Multilingual Parsing from Raw Text to Universal Dependencies, Association for Computational Linguistics. p. 171--179. October 2018

or, in bibtex format:

@InProceedings{boro-dumitrescu-burtica:2018:K18-2,
  author    = {Boroș, Tiberiu  and  Dumitrescu, Stefan Daniel  and  Burtica, Ruxandra},
  title     = {{NLP}-Cube: End-to-End Raw Text Processing With Neural Networks},
  booktitle = {Proceedings of the {CoNLL} 2018 Shared Task: Multilingual Parsing from Raw Text to Universal Dependencies},
  month     = {October},
  year      = {2018},
  address   = {Brussels, Belgium},
  publisher = {Association for Computational Linguistics},
  pages     = {171--179},
  abstract  = {We introduce NLP-Cube: an end-to-end Natural Language Processing framework, evaluated in CoNLL's "Multilingual Parsing from Raw Text to Universal Dependencies 2018" Shared Task. It performs sentence splitting, tokenization, compound word expansion, lemmatization, tagging and parsing. Based entirely on recurrent neural networks, written in Python, this ready-to-use open source system is freely available on GitHub. For each task we describe and discuss its specific network architecture, closing with an overview on the results obtained in the competition.},
  url       = {http://www.aclweb.org/anthology/K18-2017}
}

Languages and performance

Results are reported against the test files for each language (available in the UD 2.2 corpus) using the 2018 conll eval script. Please see more info about what each metric represents here.

Note: we are in the process of training version 1.1 of the models which do not require embeddings (performance will be roughly similar, but models won't need the large vector embedding files).

Language Model Token Sentence UPOS XPOS AllTags Lemmas UAS LAS
Afrikaans
af-1.0 99.97 99.65 97.28 93.0 91.53 96.42 87.61 83.96
af-1.1 99.99 99.29 96.72 92.29 90.87 96.48 87.32 83.31
Ancient Greek
grc-1.0 100.0 18.13 94.92 95.32 84.17 86.59 72.44 67.73
grc-1.1 100.0 17.61 96.87 97.35 88.36 88.41 73.4 69.36
Arabic
ar-1.0 99.98 61.05 73.42 69.75 68.12 41.26 53.94 50.31
ar-1.1 99.99 60.53 73.27 68.98 65.95 40.87 53.06 49.45
Armenian
hy-1.0 97.34 87.52 74.13 96.76 41.51 60.58 11.41 1.7
Basque
eu-1.0 99.97 99.83 94.93 99.97 87.24 90.75 85.49 81.35
eu-1.1 99.97 99.75 95.0 99.97 88.14 90.74 85.1 80.91
Bulgarian
bg-1.0 99.94 92.8 98.51 95.6 93.99 91.59 92.38 88.84
bg-1.1 99.93 93.36 98.36 95.91 94.46 92.02 92.39 88.76
Buryat
bxr-1.0 83.26 31.52 38.08 83.26 16.74 16.05 14.44 6.5
Catalan
ca-1.0 99.98 99.27 98.17 98.23 96.63 97.83 92.33 89.95
ca-1.1 99.99 99.51 98.2 98.22 96.72 97.8 92.14 89.6
Chinese
zh-1.0 93.03 99.1 88.22 88.15 86.91 92.74 73.43 69.52
zh-1.1 92.34 99.1 86.75 86.66 85.35 92.05 71.0 67.04
Croatian
hr-1.0 99.92 95.56 97.66 99.92 89.49 93.85 90.61 85.77
hr-1.1 99.95 95.84 97.56 99.95 89.49 94.01 89.95 84.97
Czech
cs-1.0 99.99 83.79 98.75 95.54 93.61 95.79 90.67 88.46
cs-1.1 99.99 84.19 98.54 95.33 94.09 95.7 90.72 88.52
Danish
da-1.0 99.85 91.79 96.79 99.85 94.29 96.53 85.93 83.05
da-1.1 99.82 92.64 96.52 99.82 94.39 96.21 85.09 81.83
Dutch
nl-1.0 99.89 90.75 95.49 93.84 91.73 95.72 89.48 86.1
nl-1.1 99.91 90.89 95.62 93.92 92.58 95.87 89.76 86.4
English
en-1.0 99.25 72.8 95.34 94.83 92.48 95.62 84.7 81.93
en-1.1 99.2 70.94 94.4 93.93 91.04 95.18 83.3 80.32
Estonian
et-1.0 99.9 91.81 96.02 97.18 91.35 93.26 86.04 82.29
et-1.1 99.91 91.92 96.8 97.92 93.17 93.9 86.13 82.91
Finnish
fi-1.0 99.7 88.73 95.45 96.44 90.29 83.69 87.18 83.89
fi-1.1 99.65 89.23 96.22 97.07 91.8 84.02 87.83 84.96
French
fr-1.0 99.68 94.2 92.61 95.46 90.79 93.08 84.96 80.91
fr-1.1 99.67 95.31 92.51 95.45 90.8 93.0 83.88 80.16
Galician
gl-1.0 99.89 97.16 83.01 82.51 81.58 82.95 65.69 61.08
gl-1.1 99.91 97.28 82.6 82.12 80.96 82.71 62.65 58.2
German
de-1.0 99.7 81.19 91.38 94.26 80.37 75.8 79.6 74.35
de-1.1 99.77 81.99 90.47 93.82 79.79 75.46 79.3 73.87
Gothic
got-1.0 100.0 21.59 93.1 93.8 80.58 83.74 67.23 59.67
Greek
el-1.0 99.88 89.46 93.7 93.54 87.14 88.92 85.63 82.05
el-1.1 99.88 89.53 93.28 93.24 87.95 88.65 84.51 79.88
Hebrew
he-1.0 99.93 99.69 54.13 54.17 51.49 54.13 34.84 32.29
he-1.1 99.94 100.0 52.78 52.78 49.9 53.45 32.13 29.42
Hindi
hi-1.0 99.98 98.84 97.16 96.43 90.29 97.48 94.66 91.26
hi-1.1 100.0 99.11 96.81 96.28 89.74 97.4 94.56 90.96
Hungarian
hu-1.0 99.8 94.18 94.52 99.8 86.22 91.07 81.57 75.95
hu-1.1 99.88 97.77 93.11 99.88 86.79 91.18 77.89 70.94
Indonesian
id-1.0 99.95 93.59 93.13 94.15 87.65 82.19 85.01 78.18
id-1.1 100.0 94.58 92.95 92.81 86.27 81.51 84.73 77.99
Irish
ga-1.0 99.56 95.38 90.95 90.07 74.1 87.51 76.32 64.74
Italian
it-1.0 99.89 98.14 86.86 86.67 84.97 87.03 78.3 74.59
it-1.1 99.92 99.07 86.58 86.4 84.53 86.75 76.38 72.35
Japanese
ja-1.0 92.73 94.92 90.05 92.73 90.02 91.75 80.47 77.97
ja-1.1 92.42 94.92 90.28 92.42 90.28 91.66 79.94 77.79
Kazakh
kk-1.0 92.26 75.57 57.38 55.75 22.12 21.35 39.55 19.48
Korean
ko-1.0 99.87 93.9 94.66 86.92 83.81 38.7 85.52 81.39
ko-1.1 99.88 94.23 94.61 88.41 85.27 38.68 85.16 80.89
Kurmanji
kmr-1.0 89.92 88.86 53.66 52.52 25.96 53.94 12.06 5.53
Latin
la-1.0 99.97 92.5 97.95 93.75 91.76 96.9 89.2 86.29
la-1.1 99.99 92.75 98.22 94.03 92.16 97.18 89.19 86.58
Latvian
lv-1.0 99.66 96.35 93.43 82.52 79.99 89.47 83.04 77.98
North Sami
sme-1.0 99.75 98.79 86.07 87.38 71.34 80.9 66.54 56.93
Norwegian
no_bokmaal-1.0 99.92 90.93 84.24 99.92 73.68 71.68 78.24 70.83
no_bokmaal-1.1 99.92 90.32 84.69 99.92 74.84 71.47 77.71 70.63
no_nynorsk-1.0 99.96 91.08 97.33 99.96 93.87 85.82 90.33 88.02
no_nynorsk-1.1 99.96 92.18 97.47 99.96 94.75 86.07 90.23 87.98
Old Church Slavonic
cu-1.0 100.0 28.99 92.88 93.09 81.85 83.16 72.18 65.43
Persian
fa-1.0 100.0 97.91 96.34 96.17 95.51 89.4 88.35 85.08
fa-1.1 100.0 99.0 95.92 95.78 95.05 89.32 87.43 83.38
Portuguese
pt-1.0 99.69 87.88 85.02 88.39 81.35 86.23 76.38 72.99
pt-1.1 99.75 88.1 84.39 88.46 79.79 85.85 75.11 71.61
Romanian
ro-1.0 99.74 95.56 97.42 96.59 95.49 96.91 90.38 85.23
ro-1.1 99.71 95.42 96.96 96.32 94.98 96.57 90.14 85.06
Russian
ru-1.0 99.71 98.79 98.4 99.71 95.55 93.89 92.7 90.97
ru-1.1 99.73 98.5 98.48 99.73 95.37 93.8 92.88 90.99
Serbian
sr-1.0 99.97 92.61 97.61 99.97 91.54 92.93 90.89 86.92
sr-1.1 99.97 92.0 97.88 99.97 92.57 93.31 90.96 87.04
Slovak
sk-1.0 99.97 86.0 95.82 82.3 78.43 90.35 88.83 85.69
sk-1.1 99.95 86.67 95.33 81.01 76.98 89.87 87.64 83.84
Slovenian
sl-1.0 99.91 97.51 97.85 92.52 91.27 96.35 91.4 89.38
sl-1.1 99.87 97.64 97.62 93.29 90.99 96.36 91.46 89.19
Spanish
es-1.0 99.98 98.32 98.0 98.0 96.62 98.05 90.53 88.27
es-1.1 99.98 98.4 98.01 98.0 96.6 97.99 90.51 88.16
Swedish
sv-1.0 99.94 92.54 97.21 95.18 92.88 97.06 88.09 84.74
sv-1.1 99.36 91.22 92.74 0.0 0.0 89.37 78.14 71.86
Turkish
tr-1.0 99.89 97.4 90.37 89.56 81.59 87.4 65.22 58.26
tr-1.1 99.88 96.79 90.79 90.17 83.26 87.84 64.69 57.07
Ukrainian
uk-1.0 99.65 93.96 96.31 88.23 86.0 92.08 86.25 82.96
uk-1.1 99.76 93.58 96.0 88.17 85.39 92.28 84.9 81.04
Upper Sorbian
hsb-1.0 98.59 69.15 59.61 98.59 37.96 22.33 11.11 3.35
Urdu
ur-1.0 100.0 98.6 93.55 91.69 77.41 97.33 87.86 81.99
ur-1.1 100.0 98.6 92.85 91.02 77.18 97.2 87.12 80.83
Uyghur
ug-1.0 99.91 83.83 87.85 91.58 73.93 90.17 74.36 60.5
ug-1.1 99.7 84.18 88.07 90.38 75.28 92.28 75.16 62.13
Vietnamese
vi-1.0 87.2 92.88 78.35 76.43 76.18 81.47 51.59 45.49
vi-1.1 86.87 92.51 76.72 74.57 72.27 81.31 50.29 43.76

About

Natural Language Processing Pipeline - Sentence Splitting, Tokenization, Lemmatization, Part-of-speech Tagging and Dependency Parsing

Resources

License

Code of conduct

Stars

Watchers

Forks

Packages

No packages published

Languages

  • Python 100.0%