Show HN: Language detection as a service

by mgaudin • 13 years ago

38 44 13 years ago

captn3m0• 13 years ago

For those who thought (like me) that this was a programming language detection service, you can take a look at github/linguist.

diasks2• 13 years ago

Looks interesting. Why not have a input on the landing page where someone can try it out without even signing up? I think then people could give it a spin before they give away their email address. Otherwise, the user just has to trust your 99% figure, which it might be helpful to give some data around, even if it is a footnote (on a corpus of x, over x period of time, etc.)

Also, I think it would be clearer if it said "A simple and scalable way to automatically classify text by language" instead of "A simple and scalable way to classify automatically text by language".

Design looks very clean though. Nice work.

EDIT: Also, your social media links at the bottom aren't hooked up yet.

himal2• 13 years ago

Hint: You can enter any email address you want.you don't have to validate it.(well, at least for now)

mdemare2• 13 years ago

Why is this better than the Google or Bing translate APIs, which also offer language detection?

donutdan4114• 13 years ago

"test it out" comes back as french...

oedj• 13 years ago

Maybe you've fallen in the 1% error rate ?

afsina• 13 years ago

Language guessing is rather hard when few letters are used especially if you use statistical methods. I think after 20 something letters you enter >%95 accuracy zone. In a simple library I wrote ( https://github.com/ahmetaa/zemberek-nlp/tree/master/lang-id Works for 60 languages but no docs yet) , for Turkish and English test results are:

For 20 letters

TR=95.90 EN=94.96

For 50 Letters

TR=99.44 EN=99.53

If 50 letters are used in a Doc, it identifies about 20000 docs per second in a decent desktop.

RBerenguel• 13 years ago

Some day I have to rewrite whatlanguageis.com (currently not working) with all the great ideas I had to improve it...

mdemare2• 13 years ago

Hmm, it takes 5+ seconds to get a response, and it chokes on the same test phrase as Google, thinking "Ik hou van vette lettertypes." is Norwegian...

ma2rten3• 13 years ago

It's probably overloaded because it's on hackernews and is based on the same features (character n-grams) as Google Translate. Your text is simply too short for character n-grams to be 100% reliable.

redox_2• 13 years ago

You should also consider full-non-ambiguous words before trying with trigrams. "marché" is only available in French, whereas "mar", "arc", ... are available in lots of languages. This should drastically improve your results.

redox_2• 13 years ago

Store only the top N common non-ambiguous words if the RAM consumption matters ;)

microtonal7• 13 years ago

Or store the lexicon in a determinisitic acyclic finite state automaton. E.g. (shameless plug):

https://github.com/danieldk/dictomaton

Though, having implemented a language guesser myself, it's only an issue with very short texts (a few words). On longer texts models based on character n-grams achieve very high accuracies.

microtonal7• 13 years ago

Also, for those who would like to know how you can implement a language guesser (sources + link to paper):

http://www.let.rug.nl/vannoord/TextCat/

Python version:

http://thomas.mangin.com/data/source/ngram.py

It's something that is fun to implement and doesn't take more than a few hours at most.

davidjgraph• 13 years ago

I'll ask plainly what others are hinting at : Is this actually your own built service, or are you a proxy for something like Google Translate API[1]?

If it's your own built service, it's critical how you explain the hows and whys of your forecast availability and scalability numbers for your chosen architecture, given who you are competing with.

[1]https://developers.google.com/translate/v2/using_rest#detect...

phpnode• 13 years ago

how does this compare in accuracy to chromium's Compact Language Detector?

https://code.google.com/p/chromium-compact-language-detector...

https://github.com/mzsanford/cld

alexott3• 13 years ago

From my experience, the CLD works pretty well in the most cases. But you need to take care for encoding detection...

dbuxton• 13 years ago

Yes, but you presumably need to get that right in order to encode as UTF-8 and send off to a third-party API...

himal2• 13 years ago

You guys might want to handle GET requests for /try URL(https://getlang.io/try) as well.currently it's returning "Server Error (500)" for GET requests.

alexott3• 13 years ago

Apache Tika (http://tika.apache.org/) also has language detector, although it maybe not so good as CLD...

razvvan• 13 years ago

If I were to implement this I'd rather use google's prediction api. At least with that you get a bit of control over what goes into the training data.

alexott3• 13 years ago

And it looks like that they are using the following library: http://code.google.com/p/language-detection/ - at least the number & list of languages is very similar :-)

ma2rten3• 13 years ago

or just the same training data...

beering• 13 years ago

Alternatively, people can just download langid.py[1] and do language detection locally. This is not a particularly hard problem - I think it's doable by undergrad ML or NLP classes.

The tricky parts are usually political - are users going to be angry if you confuse Indonesian with Malaysian, or so on?

[1] https://github.com/saffsd/langid.py

microtonal7• 13 years ago

I think it's doable by undergrad ML or NLP classes.

In fact, we had a course for high school students where they learnt how a language guesser works and where they had to change a language guesser. A simplistic method that already works very well is:

* Create an n-gram fingerprint for each language by making a list of character uni-, bi-, and trigrams ordered by their frequency in a text. Retain the (say) 300 most frequent n-grams.

* To categorize a text, create a fingerprint for that text. Then compute for each language the sum n-gram rank differences. If an n-gram does not occur, the difference is the fingerprint size. Finally, pick the language with the lowest sum.

Of course, you can do fancier things, such as training a SVM or logistic regression classifier with n-grams and words as features, etc.

An interesting variation is to be able to distinguish different languages in a text. E.g. a Dutch text with English quotes.

ma2rten3• 13 years ago

It's easy to write a language guesser, but's not easy to write a good one. Even Google Translate is not prefect (see below).

matiasb• 13 years ago

"An interesting variation is to be able to distinguish different languages in a text. E.g. a Dutch text with English quotes."

Do you know any interesting work related to the language distinction idea on the same text?

ssiddharth• 13 years ago

It might be mild OCD but it'd be great if the list of supported languages is ordered in some logical way.

efeamadasun• 13 years ago

I don't know why I can't stand this sentence "A simple and scalable way to classify automatically text by language". "Classify" and "automatically" need to switch places.

bkamapantula• 13 years ago

It's Telugu not Teligu. By Panjabi, do you mean Punjabi?

As others already mentioned, it would be good to have users try examples before signup.

chrismorgan• 13 years ago

The design is fine, but the language used on the page itself isn't quite right.

I see three spelling errors in your language list:

- Panjabi should be Punjabi;

- Teligu should be Telugu;

- Ukraininan should be Ukrainian.

There are also a few grammar problems earlier in the document, and style problems (e.g. English doesn't use a space before sentence-ending punctuation marks).

jhull• 13 years ago

I wonder how this performs on short text posts like tweets. At my last gig where we did social media text analysis we used a few different packages (chromium, guess-language, and our own ngram classifier) and still had pretty low accuracy for tweets.

AznHisoka• 13 years ago

Have you look at the metadata returned by a tweet? They also returned language, as well as location of the tweeter, which gives you some clues.

m4tthumphrey• 13 years ago

curl -XPOST -d 'hello' 'https://getlang.io/get?token=...' { "code": "fi", "name": "suomi, suomen kieli", "name_en": "Finnish" }

O_O

martingordon• 13 years ago

Matthew Kirk spoke about a neural network language predictor at RubyConf a few weeks ago. Here are his slides and code: http://modulus7.com/rubyconf/

web64• 13 years ago

I've used detectlanguage.com[1] in the past, which seems like a very similar service to getlang.io. With both of them it is hard to know what is behind the scenes...

[1] http://detectlanguage.com/

ismaelc• 13 years ago

Where's the login page? I need to get my token

Show HN: Language detection as a service

Related Stories