Help wanted: Example sentences for testing languages

[ines]

When testing models and spaCy's performance in general, it's always nice to have a few text examples handy – especially for the languages you don't speak. We noticed this while testing the new Spanish and multi-language NER models. It's not always easy to find good, representative sentences, and copy-pasting whatever is on the Wikipedia front page that day isn't ideal either.

So here's the idea: For each language, we want to add a simple list of text examples that can be imported and processed – for example, to test if a model is working as expected. Ideally, those examples should include different types of sentences, grammatical constructions and named entities. The examples could also be used in spaCy's test suite.

Examples could live with the language's data in spacy.lang.

examples = [
    'Apple is looking at buying U.K. startup for $1 billion',
    'Autonomous cars shift insurance liability toward manufacturers',
    'San Francisco considers banning sidewalk delivery robots',
    'London is a big city in the United Kingdom.'
]

from spacy.lang.en import examples

doc = nlp(examples[0]) # use one example
print([(t.text, t.pos_, t.tag_) for t in doc])

from spacy import displacy
docs = [nlp(doc) for doc in examples]
displacy.serve(*docs) # visualise all example sentences on one page

✅ Languages with models (8/8)

• English en (can be covered by us)
• German de (can be covered by us)
• French fr (thanks @Gregory-Howard & @Tpt)
• Spanish es (thanks @lucianosilvi)
• Portuguese pt (thanks @oximer)
• Italian it (thanks @MartinoMensio)
• Dutch nl (thanks @redevries)
• Greek el (thanks @Eleni170 & @giannisdaras)

Languages with only tokenizers (29/40)

• Albanian sq (thanks @AlbionaHoti)
• Arabic ar (thanks @tzano)
• Bengali bn (thanks @roshni-b)
• Bulgarian bg (thanks @KristiyanVachev)
• Catalan ca (thanks @mpuig)
• Chinese zh (thanks @renewang)
• Croatian hr (thanks @isaric)
• Danish da (thanks @mollerhoj)
• Finnish fi (thanks @tjkemp)
• Hebrew he (thanks @beneyal)
• Hindi hi (thanks @weavermonkey)
• Hungarian hu (thanks @oroszgy)
• Indonesian id (thanks @geovedi)
• Japanese ja (thanks @tokestermw)
• Lithuanian lt (thanks @rokasramas)
• Norwegian nb (thanks @luvogels)
• Persian fa (thanks @azarezade)
• Polish pl (thanks @adam-ra)
• Romanian ro (thanks @janimo)
• Russian ru (thanks @gsoul)
• Sinhala si (thanks @keshan)
• Swedish sv (thanks @RickardSjogren)
• Tamil ta (thanks @loghijiaha)
• Tatar tt (thanks @aliiae)
• Telugu te (thanks @sainathadapa)
• Thai th (thanks @korakot & @wannaphongcom)
• Turkish tr (thanks @DuyguA)
• Ukrainian uk (thanks @juliamakogon)
• Urdu ur (thanks @mirfan899)
• Afrikaans af
• Czech cs
• Icelandic is
• Irish ga
• Kannada kn
• Latvian lv
• Luxembourgish lb
• Serbian sr
• Slovak sk
• Slovenian sl
• Vietnamese vi

"This is a sentence" example sentence (28/49)

If you happen to speak one of the tokenization-only languages above, a translation of "This is a sentence." would also be very helpful! The new model docs have a "Quickstart" widget that shows the install command, plus an example sentence. If we ever add a model for one of the alpha languages, it'd be nice to have the example sentence available (and mistakes in the example would be pretty embarrassing, haha).

• Albanian sq: "Kjo është një fjali." (thanks @AlbionaHoti)
• Arabic ar: هذه جملة (thanks @Ahmed0Sultan)
• Bulgarian bg: "Това е изречение" (thanks @KristiyanVachev)
• Danish da: "Dette er en sætning." (thanks @philiphoyos)
• Dutch nl: "Dit is een zin." (thanks @redevries)
• English en: "This is a sentence."
• French fr: "C'est une phrase."
• German de: "Dies ist ein Satz."
• Greek el: "Αυτή είναι μια πρόταση." (thanks @Eleni170)
• Hebrew he: "זהו משפט." (thanks @beneyal)
• Hindi hi: "यह एक वाक्य है।" (thanks @weavermonkey & @psmehta21)
• Hungarian hu: "Ez egy mondat." (thanks @oroszgy)
• Indonesian id: "Ini adalah sebuah kalimat." (thanks @prdx)
• Italian it: "Questa è una frase." (thanks @MartinoMensio)
• Japanese ja: "これは文章です。" (thanks @tokestermw)
• Norwegian nb: "Dette er en setning." (thanks @luvogels)
• Polish pt: "To jest zdanie." (thanks @adam-ra)
• Portuguese pt: "Esta é uma frase."
• Romanian ro: "Aceasta este o propoziție" (thanks @janimo)
• Russian ru : Это предложение. (thanks @buriy)
• Sinhala si: "මෙය වාක්‍යයකි." (thanks @keshan)
• Spanish es: "Esto es una frase."
• Tamil ta: "இது ஒரு வாக்கியம்." (thanks @arvindcheenu)
• Telugu te: "ఇది ఒక వాక్యం." (thanks @sainathadapa)
• Thai th: "นี่คือประโยค" (thanks @cvorasucha)
• Turkish tr: "Bu bir cümledir." (thanks @DuyguA)
• Ukrainian uk: "Це речення." (thanks @juliamakogon)
• Urdu ur: "یہ ایک جملہ ہے۔" (thanks @mirfan899)
• Afrikaans af
• Bengali bn
• Catalan ca
• Chinese zh
• Croatian hr
• Czech cs
• Finnish fi
• Icelandic is
• Irish ga
• Kannada kn
• Latvian lv
• Lithuanian lt
• Luxembourgish lb
• Persian fa
• Serbian sr
• Slovak sk
• Slovenian sl
• Swedish sv
• Tatar tt
• Vietnamese vi

---

We haven't fully decided how the text examples should be implemented The examples will live in a file examples.py in the language data – for example lang/en/examples.py. Currently, there's only a sentences list, but more examples can be added in the future.

If you want to help out and contribute a few examples, you can submit a PR, or simply post them here! 🙂

[Gregory-Howard]

In french :
Translations of examples given before.

examples = [
"Apple cherche a acheter une startup anglaise pour 1 milliard de dollard",
"Les voitures autonomes voient leur assurances décalées vers les constructeurs",
"San Francisco envisage d'interdire les robots coursiers",
"Londres est une grande ville du Royaume-Uni"
]

Title of news :

examples = [
"L’Italie choisit ArcelorMittal pour reprendre la plus grande aciérie d’Europe",
"Apple lance HomePod parce qu'il se sent menacé par l'Echo d'Amazon",
"La France ne devrait pas manquer d'électricité cet été, même en cas de canicule",
"Nouvelles attaques de Trump contre le maire de Londres"
]

[Tpt]

Some questions in English and French:

en_examples = [
"Where are you?",
"Who is the president of France?",
"Where is the capital of the United States?",
"When is Barack Obama born?"
]

fr_examples = [
"Où es-tu ?",
"Qui est le président de la France ?",
"Où est la capitale des Etats-Unis ?",
"Quand est né Barack Obama ?"
]

[oroszgy]

A related idea: you might want to test these strings as well to check that none of them causes a problem: https://github.com/minimaxir/big-list-of-naughty-strings

[ines]

Thanks for the examples so far! 👍

@oroszgy Thanks, that's a good idea for the tokenizer tests, especially the crazy unicode stuff! I'm actually pretty curious to see if spaCy can handle all of them – I'll start adding a test with a selection of them on develop. (In v2.0 alpha, there are currently still some issues in the tokenizer serialization across Python versions, which we suspect might be unicode-related. So this might be pretty helpful for debugging this issue as well.)

[unography]

@ines Hi, how could one contribute for the Indian languages, like Hindi and Bengali?

[ines]

@eldor4do There's no Hindi tokenization support yet (although it looks like some people have started working on it, see #641).

Bengali tokenization is already supported – so a few sentences in Bengali would be great 👍 Here's what we have in the tokenizer tests so far. Just had a look at some of the test examples in Google Translate and they seem pretty good – so more examples like that would be nice.

[lucianosilvi]

For Spanish:

es_examples = [
    'Apple está buscando comprar una startup del Reino Unido por mil millones de dólares',
    'Los coches autónomos delegan la responsabilidad del seguro en sus fabricantes',
    'San Francisco analiza prohibir los robots delivery',
    'Londres es una gran ciudad del Reino Unido',
    'El gato come pescado',
    'Veo al hombre con el telescopio',
    'La araña come moscas',
    'El pingüino incuba en su nido'
]

[adam-ra]

For Polish:

pl_examples = [
    'Poczuł przyjemną woń mocnej kawy.',
    'Istnieje wiele dróg oddziaływania substancji psychoaktywnej na układ nerwowy.',
    'Powitał mnie biało-czarny kot, płosząc siedzące na płocie trzy dorodne dudki.',
    'Nowy abonament pod lupą Komisji Europejskiej',
    'Czy w ciągu ostatnich 48 godzin spożyłeś leki zawierające paracetamol?',
    'Kto ma ochotę zapoznać się z innymi niż w książkach przygodami Muminków i ich przyjaciół, temu polecam komiks Tove Jansson „Muminki i morze”.'
]

[beneyal]

Hebrew examples:

he_this_is_a_sentence = 'זהו משפט.'

he_examples_news = [
    'סין מקימה קרן של 440 מיליון דולר להשקעה בהייטק בישראל',
    'רה"מ הודיע כי יחרים טקס בחסותו',
    'הכנסת צפויה לאשר איכון אוטומטי של שיחות למוקד 100',
    'תוכנית לאומית תהפוך את ישראל למעצמה דיגיטלית'
]

he_examples_misc = [
    'סע לשלום, המפתחות בפנים.',
    'מלצר, פעמיים טורקי!',
    'ואהבת לרעך כמוך.',
    'היום נעשה משהו בלתי נשכח.'
]

he_examples_questions = [
    'איפה הילד?',
    'מיהו נשיא צרפת?',
    'מהי בירת ארצות הברית?',
    "איך קוראים בעברית לצ'ופצ'יק של הקומקום?",
    'מה הייתה הדקה?',
    'מי אומר שלום ראשון, זה שעולה או זה שיורד?',
]

[RickardSjogren]

For Swedish:

examples = [
    'Apple överväger att köpa brittisk startup för 1 miljard dollar.',
    'Självkörande bilar förskjuter försäkringsansvar mot tillverkare.',
    'San Fransisco överväger förbud mot leveransrobotar på trottoarer.'.
    'London är en storstad i Storbritannien.'
]

[adam-ra]

“This is a sentence.” -> „To jest zdanie.” (Polish)

[oximer]

For Portuguese BR

pt_examples = [
'Apple está querendo comprar uma startup do Reino Unido por 100 milhões de dólares',
'Carros autônomos empurram a responsabilidade do seguro para os fabricantes.'
'São Francisco considera banir os robôs de entrega que andam pelas calçadas',
'Londres é a maior cidade do Reino Unido'
]

[mollerhoj]

For Danish:

examples = [
    'Apple overvejer at købe et britisk statup for 1 milliard dollar',
    'Selvkørende biler flytter forsikringsansvaret over på producenterne',
    'San Francisco overvejer at forbyde leverandørrobotter på fortov',
    'London er en stor by i Storbritannien'
]

[luvogels]

For Norwegian NB:

examples = [
    'Apple vurderer å kjøpe britisk oppstartfirma for en milliard dollar',
    'Selvkjørende biler flytter forsikringsansvaret over på produsentene ',
    'San Francisco vurderer å forby robotbud på fortauene',
    'London er en stor by i Storbritannia.'
]

“This is a sentence.” -> "Dette er en setning."

[MartinoMensio]

For Italian:

examples = [
    'Apple vuole comprare una startup del Regno Unito per un miliardo di dollari',
    'Le automobili a guida autonoma spostano la responsabilità assicurativa verso i produttori',
    'San Francisco prevede di bandire i robot di consegna porta a porta',
    'Londra è una grande città del Regno Unito.'
]

Edit: "This is a sentence" --> "Questa è una frase"

[khudhairJassim]

Arabic
examples = [
     "تتطلع شركة أبل إلى شراء شركة ناشئة في المملكة المتحدة مقابل مليار دولار" ،
     "السيارات الخصوصي تحول مسؤولية التأمين نحو المصنعين" ،
     " سان فرانسيسكو تاخذ بعين الاعتبار حظر روبوتات التوصيل على رصيف المشاة" ،
     ".لندن هي مدينة كبيرة في المملكة المتحدة"
]

[loghijiaha]

I would like to support to spaCy by adding Tamil language .@ines

[mpuig]

For Catalan:

This is a sentence -> Això és una frase

[juliamakogon]

Ukrainian: Це речення. нд, 17 лют. 2019 о 17:41 Marc Puig <[email protected]> пише:

…

For Catalan: This is a sentence -> Això és una frase — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <https://github.com/explosion/spaCy/issues/1107#issuecomment-464471841>, or mute the thread <https://github.com/notifications/unsubscribe-auth/ABcqpVXlFSuMVSFo35q_hzCur7h484bvks5vOXgigaJpZM4NxRts> .

[VariableVasasMT]

Hindi:

sentences = [
    "एप्पल एक यूक स्टार्टअप को १ अरब डॉलर के लिए खरीदने का सोच रहा है|",

"एप्पल एक यूके की स्टार्टअप को १ अरब डॉलर मे खरीदने का सोच रहा है|"

"गाड़ियां अब बीमा दायित्व को अब उत्पादक के ऊपर दाल रहे हैं|",

should be स्वचालित गाड़ियां अब बीमा के दायित्व को अब उत्पादक के ऊपर डाल रही हैं|

"सं फ्रांसिस्को फुटपाथ वितरण रोबोट पर प्रतिबंध लगाने का विचार कर रहा है|",
"लन्दन यूके में एक बड़ा शहर है|",
"कहाँ हो?",
"फ्रांस का राष्ट्रपति कौन है?"
"अमेरिका का राजधानी क्या है?",

"अमेरिका की राजधानी क्या है?"

"बराक ओबामा का जनम कब हुआ?"
]

`hindi_this_is_sentence = "य हक वाकया है|"`

यह एक वाक्य है|

[prdx]

I saw Indonesian model is in progress, here is the Indonesian translation:

This is a sentence -> Ini adalah sebuah kalimat

[AlbionaHoti]

Albanian:

examples = [
   'Apple po shqyrton blerjen e nje shoqërie të U.K. për 1 miliard dollarë',
   'Makinat autonome ndryshojnë përgjegjësinë e sigurimit ndaj prodhuesve',
   'San Francisko konsideron ndalimin e robotëve të shpërndarjes',
   'Londra është një qytet i madh në Mbretërinë e Bashkuar.' 
]

This is a sentence -> Kjo është një fjali 

[KristiyanVachev]

For Bulgarian:

This is a sentence -> Това е изречение

examples = [
    'Епъл иска да купи английски стартъп за 1 милиард долара.' 
    'Автономните коли прехвърлят застрахователната отговорност към производителите.' 
    'Сан Франциско обмисля забрана на роботи доставящи по тротоари. ' 
    'Лондон е голям град в Обединеното Кралство.'
]

There are some specifics about the use of Bulgarian.

• In 'Apple is looking at buying U.K. startup for $1 billion', we don't have a word for UK and we use the equivalent of "English startup". It may be more accurate to say "великобритански стартъп" - "great-britaneese startup", but I have never seen it being used in practice.
• In 'San Francisco considers banning sidewalk delivery robots', we don't structure phrases like "sidewalk delivery robots" but more like "robots that deliver on sidewalks".

Spacy is great and I would like to contribute more for the Bulgarian pack in the future.

[ship-it-mind]

"This is a sentence" example sentence in Arabic

This is a sentence  -> هذه جملة

[mzeidhassan]

Do you need further help with Arabic, @ines ? More examples, translation review, etc.?

[buriy]

"This is a sentence." example sentence

Russian:
Это предложение.

[isarvindone]

Translated the examples below into Tamil:

'Apple is looking at buying U.K. startup for $1 billion',
'Autonomous cars shift insurance liability toward manufacturers',
'San Francisco considers banning sidewalk delivery robots',
'London is a big city in the United Kingdom.'

translates to

"இது ஒரு வாக்கியம்.",
"ஆப்பிள் நிறுவனம் யு.கே. தொடக்க நிறுவனத்தை ஒரு லட்சம் கோடிக்கு வாங்கப் பார்க்கிறது",
"தன்னாட்சி கார்கள் காப்பீட்டு பொறுப்பை உற்பத்தியாளரிடம் மாற்றுகின்றன",
"நடைபாதை விநியோக ரோபோக்களை தடை செய்வதை சான் பிரான்சிஸ்கோ கருதுகிறது",
"லண்டன் ஐக்கிய இராச்சியத்தில் ஒரு பெரிய நகரம்."

[rasyidf]

There is something wrong with the Indonesian Example sentences, it contains sensitive phrases as terorism and bombing, while there are tons of examples available.

sending a pull request for it.

[jairwen]

Simplified Chinese / zh_cn

examples = [
Apple is looking at buying U.K. startup for $1 billion',
'Autonomous cars shift insurance liability toward manufacturers',
'San Francisco considers banning sidewalk delivery robots',
'London is a big city in the United Kingdom.'
]

->

examples = [
苹果公司正考虑用1亿美元收购英国初创公司',
'自动驾驶汽车使得保险责任转移至生产厂家',
'旧金山考虑禁止人行道式送餐机器人',
'伦敦是英国的一个大城市'
]

This is a sentence -> 这是一个句子。

[elisno]

Icelandic examples:

examples = [
    'Apple íhugar að kaupa breskt sprotafyrirtæki fyrir 1 milljarð dollara',
    'Sjálfstýrðir bílar færa tryggingarábyrgð yfir á framleiðendur',
    'San Francisco íhugar að banna afhendingarvélmenni á gangstéttum',
    'London er stór borg í Bretlandi.'
]

Note: Similarly to the Bulgarian translations (as with other translations that I see here), U.K. startup would turn into British startup when translating English -> Icelandic -> English.

afhendingarvélmenni is just some neologism for delivery robots, which I suspect that current automatic translations would opt for.

As mentioned for other languages, the translation for sidewalk delivery robots would be structured as robots on sidewalks that do deliveries (if it were back-translated).
Perhaps one could go overboard coin an even longer word by translating sidewalk delivery robot as gangstéttarafgreiðsluvélmenni, but I'm afraid it won't claim the record as the longest Icelandic word. 🤪

---

Icelandic "This is a sentence" example sentence:

Þetta er setning