Skip to content
/ lingua-py Public
forked from pemistahl/lingua-py

The most accurate natural language detection library for Python, suitable for long and short text alike

License

Apache-2.0 license
0 stars 45 forks Branches Tags Activity
Star
Notifications You must be signed in to change notification settings

silvanbrummen/lingua-py

BranchesTags
 
 

Repository files navigation

lingua

build codecov supported languages docs supported Python versions pypi license


1. What does this library do?

Its task is simple: It tells you which language some text is written in. This is very useful as a preprocessing step for linguistic data in natural language processing applications such as text classification and spell checking. Other use cases, for instance, might include routing e-mails to the right geographically located customer service department, based on the e-mails' languages.

2. Why does this library exist?

Language detection is often done as part of large machine learning frameworks or natural language processing applications. In cases where you don't need the full-fledged functionality of those systems or don't want to learn the ropes of those, a small flexible library comes in handy.

Python is widely used in natural language processing, so there are a couple of comprehensive open source libraries for this task, such as Google's CLD 2 and CLD 3, langid, fastText and langdetect. Unfortunately, except for the last one they have two major drawbacks:

  1. Detection only works with quite lengthy text fragments. For very short text snippets such as Twitter messages, they do not provide adequate results.
  2. The more languages take part in the decision process, the less accurate are the detection results.

Lingua aims at eliminating these problems. She nearly does not need any configuration and yields pretty accurate results on both long and short text, even on single words and phrases. She draws on both rule-based and statistical methods but does not use any dictionaries of words. She does not need a connection to any external API or service either. Once the library has been downloaded, it can be used completely offline.

3. Which languages are supported?

Compared to other language detection libraries, Lingua's focus is on quality over quantity, that is, getting detection right for a small set of languages first before adding new ones. Currently, the following 75 languages are supported:

  • A
    • Afrikaans
    • Albanian
    • Arabic
    • Armenian
    • Azerbaijani
  • B
    • Basque
    • Belarusian
    • Bengali
    • Norwegian Bokmal
    • Bosnian
    • Bulgarian
  • C
    • Catalan
    • Chinese
    • Croatian
    • Czech
  • D
    • Danish
    • Dutch
  • E
    • English
    • Esperanto
    • Estonian
  • F
    • Finnish
    • French
  • G
    • Ganda
    • Georgian
    • German
    • Greek
    • Gujarati
  • H
    • Hebrew
    • Hindi
    • Hungarian
  • I
    • Icelandic
    • Indonesian
    • Irish
    • Italian
  • J
    • Japanese
  • K
    • Kazakh
    • Korean
  • L
    • Latin
    • Latvian
    • Lithuanian
  • M
    • Macedonian
    • Malay
    • Maori
    • Marathi
    • Mongolian
  • N
    • Norwegian Nynorsk
  • P
    • Persian
    • Polish
    • Portuguese
    • Punjabi
  • R
    • Romanian
    • Russian
  • S
    • Serbian
    • Shona
    • Slovak
    • Slovene
    • Somali
    • Sotho
    • Spanish
    • Swahili
    • Swedish
  • T
    • Tagalog
    • Tamil
    • Telugu
    • Thai
    • Tsonga
    • Tswana
    • Turkish
  • U
    • Ukrainian
    • Urdu
  • V
    • Vietnamese
  • W
    • Welsh
  • X
    • Xhosa
  • Y
    • Yoruba
  • Z
    • Zulu

4. How good is it?

Lingua is able to report accuracy statistics for some bundled test data available for each supported language. The test data for each language is split into three parts:

  1. a list of single words with a minimum length of 5 characters
  2. a list of word pairs with a minimum length of 10 characters
  3. a list of complete grammatical sentences of various lengths

Both the language models and the test data have been created from separate documents of the Wortschatz corpora offered by Leipzig University, Germany. Data crawled from various news websites have been used for training, each corpus comprising one million sentences. For testing, corpora made of arbitrarily chosen websites have been used, each comprising ten thousand sentences. From each test corpus, a random unsorted subset of 1000 single words, 1000 word pairs and 1000 sentences has been extracted, respectively.

Given the generated test data, I have compared the detection results of Lingua, fastText, langdetect, langid, CLD 2 and CLD 3 running over the data of Lingua's supported 75 languages. Languages that are not supported by the other detectors are simply ignored for them during the detection process.

Each of the following sections contains two plots. The bar plot shows the detailed accuracy results for each supported language. The box plot illustrates the distributions of the accuracy values for each classifier. The boxes themselves represent the areas which the middle 50 % of data lie within. Within the colored boxes, the horizontal lines mark the median of the distributions.

4.1 Single word detection


Single Word Detection Performance


Bar plot Single Word Detection Performance



4.2 Word pair detection


Word Pair Detection Performance


Bar plot Word Pair Detection Performance



4.3 Sentence detection


Sentence Detection Performance


Bar plot Sentence Detection Performance



4.4 Average detection


Average Detection Performance


Bar plot Average Detection Performance



4.5 Mean, median and standard deviation

The table below shows detailed statistics for each language and classifier including mean, median and standard deviation.

Open table
Language Average Single Words Word Pairs Sentences
Lingua
(high accuracy mode)		 Lingua
(low accuracy mode)		 Langdetect FastText Langid   CLD3     CLD2   Simplemma Lingua
(high accuracy mode)		 Lingua
(low accuracy mode)		 Langdetect FastText Langid   CLD3     CLD2   Simplemma Lingua
(high accuracy mode)		 Lingua
(low accuracy mode)		 Langdetect FastText Langid   CLD3     CLD2   Simplemma Lingua
(high accuracy mode)		 Lingua
(low accuracy mode)		 Langdetect FastText Langid   CLD3     CLD2   Simplemma
Afrikaans 78 64 67 36 30 55 55 - 57 38 38 11 1 22 13 - 80 62 65 23 10 46 56 - 96 93 98 74 80 98 96 -
Albanian 88 80 79 66 65 55 65 21 68 55 53 35 33 18 18 23 95 87 84 66 63 48 77 16 100 99 100 98 98 98 99 23
Arabic 99 95 97 96 91 90 67 - 97 89 93 89 84 79 19 - 99 96 98 98 90 92 82 - 100 99 100 100 98 100 99 -
Armenian 100 100 - 100 94 99 100 15 100 100 - 100 83 100 100 24 100 100 - 100 99 100 100 13 100 100 - 100 100 97 100 7
Azerbaijani 89 82 - 78 68 81 72 - 78 71 - 57 36 62 34 - 92 78 - 80 69 82 82 - 99 96 - 98 98 99 99 -
Basque 83 74 - 71 52 62 61 - 71 56 - 44 18 33 23 - 87 76 - 70 52 62 69 - 92 91 - 100 86 92 91 -
Belarusian 97 92 - 85 85 84 76 - 91 80 - 69 69 67 42 - 99 95 - 88 87 86 87 - 100 100 - 98 99 100 99 -
Bengali 100 100 100 98 92 99 63 - 100 100 100 94 92 98 19 - 100 100 100 99 88 99 69 - 100 100 100 100 97 99 99 -
Bokmal 58 49 - - 13 - - 47 38 27 - - 3 - - 15 59 47 - - 12 - - 39 75 74 - - 23 - - 86
Bosnian 33 29 - 9 5 33 19 - 28 22 - 9 2 19 4 - 32 29 - 10 4 28 15 - 39 36 - 8 8 52 36 -
Bulgarian 87 78 72 78 67 70 66 69 71 57 51 56 46 45 32 46 91 81 68 81 62 66 72 70 99 97 96 99 93 98 93 90
Catalan 70 58 54 57 38 48 38 62 50 33 25 33 5 19 4 39 74 60 51 57 29 42 30 63 86 81 86 83 81 84 79 82
Chinese 100 100 64 71 96 92 33 - 100 100 39 46 90 92 - - 100 100 55 68 97 83 2 - 100 100 97 100 100 100 98 -
Croatian 72 59 73 47 48 42 51 - 53 36 50 28 16 26 34 - 74 57 70 42 38 42 47 - 90 85 97 72 90 58 73 -
Czech 80 71 71 76 66 64 74 52 65 54 51 58 44 39 50 35 84 72 72 79 69 65 80 45 90 87 88 92 86 88 91 75
Danish 81 70 70 62 60 58 59 56 61 45 50 35 33 26 27 27 84 70 68 57 61 54 56 52 98 95 93 95 86 95 94 89
Dutch 77 64 58 78 64 58 47 57 55 36 27 55 34 29 11 34 81 61 49 81 61 47 42 46 96 94 98 100 98 97 90 90
English 81 62 60 96 85 54 56 69 55 29 23 90 84 22 12 38 89 62 59 98 71 44 55 71 99 96 99 100 99 97 100 98
Esperanto 83 66 - 76 44 57 50 - 67 44 - 51 5 22 7 - 85 61 - 79 30 51 46 - 98 92 - 100 96 98 98 -
Estonian 92 83 82 73 67 70 65 72 80 62 62 50 37 41 24 52 96 88 86 73 67 69 73 69 100 99 100 96 98 99 99 94
Finnish 96 91 93 92 83 80 77 81 90 77 84 82 62 58 44 63 98 95 95 96 88 84 89 82 100 100 100 100 100 99 98 99
French 89 77 75 83 71 55 46 70 74 52 48 62 42 22 12 46 95 83 78 86 74 49 48 69 99 97 99 99 98 94 80 94
Ganda 92 84 - - - - 61 - 80 65 - - - - 23 - 95 87 - - - - 62 - 100 100 - - - - 99 -
Georgian 100 100 - 99 99 98 100 4 100 100 - 97 97 99 100 11 100 100 - 99 100 100 100 2 100 100 - 100 100 96 100 0
German 89 80 73 89 81 66 64 83 74 57 50 76 61 40 27 64 94 84 70 93 81 62 66 86 100 99 100 100 100 98 98 99
Greek 100 100 100 99 100 100 100 75 100 100 100 98 100 100 100 71 100 100 100 100 100 100 100 61 100 100 100 100 100 100 100 93
Gujarati 100 100 100 100 100 100 100 - 100 100 100 99 100 99 100 - 100 100 100 100 100 100 100 - 100 100 100 100 100 100 100 -
Hebrew 100 100 100 100 100 - - - 100 100 100 99 100 - - - 100 100 100 100 100 - - - 100 100 100 100 100 - - -
Hindi 73 33 67 87 60 58 77 5 61 11 44 74 41 34 56 2 64 20 59 88 47 45 76 4 93 67 99 99 92 95 99 11
Hungarian 95 90 88 92 83 76 75 65 87 77 74 80 64 53 41 50 98 94 91 96 86 76 85 61 100 100 100 100 100 99 100 84
Icelandic 93 88 - 65 66 71 66 67 83 72 - 39 33 42 26 50 97 92 - 57 66 70 73 60 100 99 - 98 99 99 99 92
Indonesian 62 47 80 69 51 46 62 34 41 25 56 43 16 26 36 39 62 45 85 68 54 45 63 30 82 71 100 95 82 66 88 32
Irish 91 85 - 60 63 67 66 80 82 70 - 35 28 42 29 74 94 90 - 57 64 66 78 76 96 95 - 89 97 94 92 90
Italian 87 71 76 89 66 62 44 66 69 42 50 74 28 31 7 42 92 74 80 92 70 57 32 61 100 98 99 100 100 98 93 94
Japanese 100 100 100 87 86 98 33 - 100 100 99 72 61 97 - - 100 100 100 89 96 96 - - 100 100 100 100 100 100 100 -
Kazakh 92 90 - 88 80 82 77 - 80 78 - 72 67 62 43 - 97 93 - 90 78 83 88 - 99 99 - 100 96 99 99 -
Korean 100 100 100 99 100 99 100 - 100 100 100 98 100 100 100 - 100 100 100 100 100 100 100 - 100 100 100 100 100 98 100 -
Latin 87 73 - 50 21 62 46 70 72 49 - 24 - 44 9 50 92 76 - 41 2 58 42 66 97 93 - 85 61 83 88 93
Latvian 93 87 89 82 83 75 72 43 85 75 76 65 64 51 33 37 97 90 91 83 86 77 84 32 99 97 99 97 98 98 98 59
Lithuanian 95 88 87 81 80 72 70 72 87 76 71 61 58 42 30 65 98 89 91 83 85 75 82 64 100 98 100 99 99 99 99 86
Macedonian 83 72 86 74 51 60 60 14 64 52 71 51 15 30 27 15 86 70 88 72 44 54 70 11 99 95 100 100 94 97 84 16
Malay 30 31 - 15 11 22 18 13 25 22 - 14 2 11 9 3 36 36 - 19 9 22 22 10 28 35 - 12 22 34 23 26
Maori 92 83 - - - 52 61 - 84 64 - - - 22 12 - 92 88 - - - 43 72 - 99 98 - - - 91 98 -
Marathi 85 41 88 80 80 84 83 - 74 20 76 61 70 69 65 - 84 30 90 81 79 84 86 - 96 72 98 99 91 98 99 -
Mongolian 97 95 - 81 86 83 78 - 92 89 - 59 68 63 43 - 99 98 - 86 90 87 92 - 99 99 - 98 99 99 100 -
Nynorsk 65 52 - 29 32 - 54 30 41 25 - 8 5 - 18 9 65 49 - 18 16 - 50 24 90 81 - 61 75 - 93 55
Persian 90 80 81 90 92 76 61 12 77 62 64 79 83 57 13 12 93 80 80 92 94 70 72 5 100 98 99 100 100 99 99 18
Polish 95 90 89 92 89 77 75 90 86 77 75 80 73 51 38 84 98 93 93 97 93 80 87 88 100 99 100 100 100 99 99 99
Portuguese 81 69 61 73 54 53 54 65 58 42 30 47 19 21 20 36 85 69 55 71 44 40 48 62 98 96 98 99 98 97 94 97
Punjabi 100 100 100 100 100 100 100 - 100 100 100 99 100 99 100 - 100 100 100 100 100 100 100 - 100 100 100 100 100 100 100 -
Romanian 87 72 77 64 61 53 54 61 69 49 56 38 31 24 11 44 92 74 78 60 60 48 53 53 99 94 97 95 92 88 96 86
Russian 89 79 84 94 75 71 60 71 76 60 70 86 60 48 26 64 94 84 87 98 75 72 68 66 98 92 96 100 91 93 87 84
Serbian 88 78 - 76 64 78 69 - 74 62 - 54 39 63 29 - 90 80 - 76 63 75 78 - 99 91 - 98 89 95 99 -
Shona 90 81 - - - 76 65 - 77 56 - - - 51 24 - 95 86 - - - 79 71 - 100 100 - - - 99 99 -
Slovak 84 75 75 65 68 63 71 71 64 49 51 41 40 32 38 54 90 79 76 62 66 61 76 68 99 98 98 91 97 96 99 92
Slovene 82 67 73 59 63 63 48 78 61 39 47 32 33 29 8 62 87 68 72 54 61 60 42 76 98 93 98 90 95 99 92 96
Somali 92 85 90 24 - 69 70 - 82 64 75 4 - 38 27 - 95 91 94 15 - 70 83 - 100 100 100 52 - 100 99 -
Sotho 83 72 - - - 49 54 - 61 43 - - - 15 13 - 88 75 - - - 33 54 - 99 97 - - - 98 95 -
Spanish 70 56 57 74 65 48 43 53 44 25 26 51 37 16 12 24 69 49 47 72 59 32 34 42 97 94 97 100 98 96 85 92
Swahili 79 70 73 41 42 57 57 52 58 43 47 7 3 25 16 36 82 69 73 24 24 49 59 44 98 97 99 92 98 98 97 76
Swedish 84 72 68 76 65 61 53 66 64 46 40 51 35 30 14 43 89 76 67 78 63 56 52 65 99 95 96 98 96 96 93 89
Tagalog 78 66 76 45 42 - 50 14 52 36 50 11 2 - 9 16 82 66 77 28 26 - 44 12 98 96 99 98 98 - 95 15
Tamil 100 100 100 100 100 100 100 - 100 100 100 100 100 100 100 - 100 100 100 100 100 100 100 - 100 100 100 100 100 99 100 -
Telugu 100 100 100 100 100 99 100 - 100 100 100 100 100 99 100 - 100 100 100 100 100 100 100 - 100 100 100 100 100 99 100 -
Thai 99 99 100 100 100 99 100 - 100 100 100 100 100 100 100 - 100 100 100 100 100 100 100 - 98 98 100 100 100 98 100 -
Tsonga 83 72 - - - - 61 - 63 46 - - - - 19 - 88 73 - - - - 68 - 98 97 - - - - 97 -
Tswana 82 71 - - - - 56 - 62 44 - - - - 17 - 86 73 - - - - 57 - 98 96 - - - - 94 -
Turkish 94 87 82 86 67 69 66 82 84 71 63 70 50 41 30 71 98 91 84 88 67 70 71 80 100 99 100 100 84 97 97 94
Ukrainian 92 86 83 91 76 81 77 75 84 75 66 78 54 62 46 67 97 92 84 94 77 83 88 68 95 93 98 100 96 98 99 91
Urdu 90 80 82 63 58 61 61 - 79 65 66 40 30 39 8 - 94 78 84 50 46 53 75 - 98 96 97 99 99 92 99 -
Vietnamese 91 87 93 89 86 66 63 - 79 76 81 71 65 26 - - 94 88 98 97 93 74 90 - 99 98 100 100 100 99 100 -
Welsh 91 82 85 64 49 69 72 71 78 61 69 35 11 43 34 63 96 87 88 61 39 66 85 60 99 99 99 96 95 98 98 90
Xhosa 81 69 - - 53 66 71 - 63 45 - - 13 40 45 - 84 67 - - 49 65 71 - 98 94 - - 96 92 97 -
Yoruba 70 62 - 8 - 15 37 - 45 33 - 1 - 5 1 - 71 61 - 1 - 11 22 - 96 93 - 21 - 28 88 -
Zulu 80 71 - - 6 63 54 - 62 45 - - 0 35 18 - 83 72 - - 6 63 51 - 97 95 - - 11 92 93 -
Mean 86 77 82 74 68 69 65 55 74 61 65 58 48 48 34 41 89 78 82 74 65 67 68 51 96 93 98 92 90 93 94 73
Median 89.0 80.0 82.0 78.0 67.0 68.0 63.0 65.0 74.0 57.0 63.5 57.5 41.5 41.0 26.5 42.0 93.0 81.0 84.0 81.0 67.0 66.0 71.5 61.0 99.0 97.0 99.0 99.0 98.0 98.0 98.0 89.0
Standard Deviation 13.36 17.29 13.37 23.07 24.61 19.04 18.57 24.68 18.75 24.86 23.57 28.52 32.33 27.86 28.74 21.22 13.51 18.99 15.64 26.45 28.5 21.83 22.7 25.12 11.26 11.94 2.79 19.46 20.21 13.95 12.25 31.75

5. Why is it better than other libraries?

Every language detector uses a probabilistic n-gram model trained on the character distribution in some training corpus. Most libraries only use n-grams of size 3 (trigrams) which is satisfactory for detecting the language of longer text fragments consisting of multiple sentences. For short phrases or single words, however, trigrams are not enough. The shorter the input text is, the less n-grams are available. The probabilities estimated from such few n-grams are not reliable. This is why Lingua makes use of n-grams of sizes 1 up to 5 which results in much more accurate prediction of the correct language.

A second important difference is that Lingua does not only use such a statistical model, but also a rule-based engine. This engine first determines the alphabet of the input text and searches for characters which are unique in one or more languages. If exactly one language can be reliably chosen this way, the statistical model is not necessary anymore. In any case, the rule-based engine filters out languages that do not satisfy the conditions of the input text. Only then, in a second step, the probabilistic n-gram model is taken into consideration. This makes sense because loading less language models means less memory consumption and better runtime performance.

In general, it is always a good idea to restrict the set of languages to be considered in the classification process using the respective api methods. If you know beforehand that certain languages are never to occur in an input text, do not let those take part in the classifcation process. The filtering mechanism of the rule-based engine is quite good, however, filtering based on your own knowledge of the input text is always preferable.

6. Test report generation

If you want to reproduce the accuracy results above, you can generate the test reports yourself for all classifiers and languages by executing:

poetry install --only script
poetry run python3 scripts/accuracy_reporter.py

For each detector and language, a test report file is then written into /accuracy-reports. As an example, here is the current output of the Lingua German report:

##### German #####

>>> Accuracy on average: 89.27%

>> Detection of 1000 single words (average length: 9 chars)
Accuracy: 74.20%
Erroneously classified as Dutch: 2.30%, Danish: 2.20%, English: 2.20%, Latin: 1.80%, Bokmal: 1.60%, Italian: 1.30%, Basque: 1.20%, Esperanto: 1.20%, French: 1.20%, Swedish: 0.90%, Afrikaans: 0.70%, Finnish: 0.60%, Nynorsk: 0.60%, Portuguese: 0.60%, Yoruba: 0.60%, Sotho: 0.50%, Tsonga: 0.50%, Welsh: 0.50%, Estonian: 0.40%, Irish: 0.40%, Polish: 0.40%, Spanish: 0.40%, Tswana: 0.40%, Albanian: 0.30%, Icelandic: 0.30%, Tagalog: 0.30%, Bosnian: 0.20%, Catalan: 0.20%, Croatian: 0.20%, Indonesian: 0.20%, Lithuanian: 0.20%, Romanian: 0.20%, Swahili: 0.20%, Zulu: 0.20%, Latvian: 0.10%, Malay: 0.10%, Maori: 0.10%, Slovak: 0.10%, Slovene: 0.10%, Somali: 0.10%, Turkish: 0.10%, Xhosa: 0.10%

>> Detection of 1000 word pairs (average length: 18 chars)
Accuracy: 93.90%
Erroneously classified as Dutch: 0.90%, Latin: 0.90%, English: 0.70%, Swedish: 0.60%, Danish: 0.50%, French: 0.40%, Bokmal: 0.30%, Irish: 0.20%, Tagalog: 0.20%, Tsonga: 0.20%, Afrikaans: 0.10%, Esperanto: 0.10%, Estonian: 0.10%, Finnish: 0.10%, Italian: 0.10%, Maori: 0.10%, Nynorsk: 0.10%, Somali: 0.10%, Swahili: 0.10%, Turkish: 0.10%, Welsh: 0.10%, Zulu: 0.10%

>> Detection of 1000 sentences (average length: 111 chars)
Accuracy: 99.70%
Erroneously classified as Dutch: 0.20%, Latin: 0.10%

7. How to add it to your project?

Lingua is available in the Python Package Index and can be installed with:

pip install lingua-language-detector

8. How to build?

Lingua requires Python >= 3.8 and uses Poetry for packaging and dependency management. You need to install it first if you have not done so yet. Afterwards, clone the repository and install the project dependencies:

git clone https://github.com/pemistahl/lingua-py.git
cd lingua-py
poetry install

The library makes uses of type annotations which allow for static type checking with Mypy. Run the following command for checking the types:

poetry run mypy

The source code is accompanied by an extensive unit test suite. To run the tests, simply say:

poetry run pytest

9. How to use?

9.1 Basic usage

>>> from lingua import Language, LanguageDetectorBuilder
>>> languages = [Language.ENGLISH, Language.FRENCH, Language.GERMAN, Language.SPANISH]
>>> detector = LanguageDetectorBuilder.from_languages(*languages).build()
>>> detector.detect_language_of("languages are awesome")
Language.ENGLISH

9.2 Minimum relative distance

By default, Lingua returns the most likely language for a given input text. However, there are certain words that are spelled the same in more than one language. The word prologue, for instance, is both a valid English and French word. Lingua would output either English or French which might be wrong in the given context. For cases like that, it is possible to specify a minimum relative distance that the logarithmized and summed up probabilities for each possible language have to satisfy. It can be stated in the following way:

>>> from lingua import Language, LanguageDetectorBuilder
>>> languages = [Language.ENGLISH, Language.FRENCH, Language.GERMAN, Language.SPANISH]
>>> detector = LanguageDetectorBuilder.from_languages(*languages)\
.with_minimum_relative_distance(0.9)\
.build()
>>> print(detector.detect_language_of("languages are awesome"))
None

Be aware that the distance between the language probabilities is dependent on the length of the input text. The longer the input text, the larger the distance between the languages. So if you want to classify very short text phrases, do not set the minimum relative distance too high. Otherwise, None will be returned most of the time as in the example above. This is the return value for cases where language detection is not reliably possible.

9.3 Confidence values

Knowing about the most likely language is nice but how reliable is the computed likelihood? And how less likely are the other examined languages in comparison to the most likely one? These questions can be answered as well:

>>> from lingua import Language, LanguageDetectorBuilder
>>> languages = [Language.ENGLISH, Language.FRENCH, Language.GERMAN, Language.SPANISH]
>>> detector = LanguageDetectorBuilder.from_languages(*languages).build()
>>> confidence_values = detector.compute_language_confidence_values("languages are awesome")
>>> for language, value in confidence_values:
...     print(f"{language.name}: {value:.2f}")
ENGLISH: 0.93
FRENCH: 0.04
GERMAN: 0.02
SPANISH: 0.01

In the example above, a list is returned containing those languages which the calling instance of LanguageDetector has been built from, sorted by their confidence value in descending order. Each value is a probability between 0.0 and 1.0. The probabilities of all languages will sum to 1.0. If the language is unambiguously identified by the rule engine, the value 1.0 will always be returned for this language. The other languages will receive a value of 0.0.

There is also a method for returning the confidence value for one specific language only:

>>> from lingua import Language, LanguageDetectorBuilder
>>> languages = [Language.ENGLISH, Language.FRENCH, Language.GERMAN, Language.SPANISH]
>>> detector = LanguageDetectorBuilder.from_languages(*languages).build()
>>> confidence_value = detector.compute_language_confidence("languages are awesome", Language.FRENCH)
>>> print(f"{confidence_value:.2f}")
0.04

The value that this method computes is a number between 0.0 and 1.0. If the language is unambiguously identified by the rule engine, the value 1.0 will always be returned. If the given language is not supported by this detector instance, the value 0.0 will always be returned.

9.4 Eager loading versus lazy loading

By default, Lingua uses lazy-loading to load only those language models on demand which are considered relevant by the rule-based filter engine. For web services, for instance, it is rather beneficial to preload all language models into memory to avoid unexpected latency while waiting for the service response. If you want to enable the eager-loading mode, you can do it like this:

LanguageDetectorBuilder.from_all_languages().with_preloaded_language_models().build()

Multiple instances of LanguageDetector share the same language models in memory which are accessed asynchronously by the instances.

9.5 Low accuracy mode versus high accuracy mode

Lingua's high detection accuracy comes at the cost of being noticeably slower than other language detectors. The large language models also consume significant amounts of memory. These requirements might not be feasible for systems running low on resources. If you want to classify mostly long texts or need to save resources, you can enable a low accuracy mode that loads only a small subset of the language models into memory:

LanguageDetectorBuilder.from_all_languages().with_low_accuracy_mode().build()

The downside of this approach is that detection accuracy for short texts consisting of less than 120 characters will drop significantly. However, detection accuracy for texts which are longer than 120 characters will remain mostly unaffected.

In high accuracy mode (the default), the language detector consumes approximately 800 MB of memory if all language models are loaded. In low accuracy mode, memory consumption is reduced to approximately 60 MB.

An alternative for a smaller memory footprint and faster performance is to reduce the set of languages when building the language detector. In most cases, it is not advisable to build the detector from all supported languages. When you have knowledge about the texts you want to classify you can almost always rule out certain languages as impossible or unlikely to occur.

9.6 Detection of multiple languages in mixed-language texts

In contrast to most other language detectors, Lingua is able to detect multiple languages in mixed-language texts. This feature can yield quite reasonable results but it is still in an experimental state and therefore the detection result is highly dependent on the input text. It works best in high-accuracy mode with multiple long words for each language. The shorter the phrases and their words are, the less accurate are the results. Reducing the set of languages when building the language detector can also improve accuracy for this task if the languages occurring in the text are equal to the languages supported by the respective language detector instance.

>>> from lingua import Language, LanguageDetectorBuilder
>>> languages = [Language.ENGLISH, Language.FRENCH, Language.GERMAN]
>>> detector = LanguageDetectorBuilder.from_languages(*languages).build()
>>> sentence = "Parlez-vous français? " + \
...            "Ich spreche Französisch nur ein bisschen. " + \
...            "A little bit is better than nothing."
>>> for result in detector.detect_multiple_languages_of(sentence):
...     print(f"{result.language.name}: '{sentence[result.start_index:result.end_index]}'")
FRENCH: 'Parlez-vous français? '
GERMAN: 'Ich spreche Französisch nur ein bisschen. '
ENGLISH: 'A little bit is better than nothing.'

In the example above, a list of DetectionResult is returned. Each entry in the list describes a contiguous single-language text section, providing start and end indices of the respective substring.

9.7 Methods to build the LanguageDetector

There might be classification tasks where you know beforehand that your language data is definitely not written in Latin, for instance. The detection accuracy can become better in such cases if you exclude certain languages from the decision process or just explicitly include relevant languages:

from lingua import LanguageDetectorBuilder, Language, IsoCode639_1, IsoCode639_3

# Include all languages available in the library.
LanguageDetectorBuilder.from_all_languages()

# Include only languages that are not yet extinct (= currently excludes Latin).
LanguageDetectorBuilder.from_all_spoken_languages()

# Include only languages written with Cyrillic script.
LanguageDetectorBuilder.from_all_languages_with_cyrillic_script()

# Exclude only the Spanish language from the decision algorithm.
LanguageDetectorBuilder.from_all_languages_without(Language.SPANISH)

# Only decide between English and German.
LanguageDetectorBuilder.from_languages(Language.ENGLISH, Language.GERMAN)

# Select languages by ISO 639-1 code.
LanguageDetectorBuilder.from_iso_codes_639_1(IsoCode639_1.EN, IsoCode639_1.DE)

# Select languages by ISO 639-3 code.
LanguageDetectorBuilder.from_iso_codes_639_3(IsoCode639_3.ENG, IsoCode639_3.DEU)

10. What's next for version 1.4.0?

Take a look at the planned issues.

11. Contributions

Any contributions to Lingua are very much appreciated. Please read the instructions in CONTRIBUTING.md for how to add new languages to the library.

About

The most accurate natural language detection library for Python, suitable for long and short text alike

Resources

Readme

License

Apache-2.0 license
Activity

Stars

0 stars

Watchers

0 watching

Forks

0 forks
Report repository

Releases

No releases published

Packages

No packages published

Languages

  • Python 100.0%