Fetching contributors…
Cannot retrieve contributors at this time
263 lines (179 sloc) 9.69 KB

CLD - Compact Language Detector

This package contains the CLD (Compact Language Detection) library, extracted from the source code for Google's Chromium library at, specifically revision 105735. The original extraction was done by Mike McCandless and was altered to improve the build and packaging as well as add new bindings for various languages.

The LICENSE is the same as Chromium's LICENSE and is included in the LICENSE file for reference.

Installing the C++ library


Building the C++ library requires a C++ compiler (duh) as well as pkg-config so future programs can locate the needed dependancies.


The shared C++ library can be built from source using the following commands:

$ git clone git://
$ cd cld
$ ./configure
$ make
$ make check # optional. Runs the tests


Once built from source:

$ make install

Or, Mac OSX users of Homebrew can build and install via:

$ brew install


None of the ports are 100% completed yet but the preliminary APIs are introduced below



The libcld C++ library must be installed (see above)


# 'gem install ...' in the near future
$ git clone
$ cd cld/ports/ruby
$ rake gem
$ gem install pkg/*gem


require 'CLD'
require 'pp' # just for illustration

cld =
res = cld.detect_language('I am the very measure of a modern major general')

pp res
    @language=#<CLD::Language:0x007fb7728b66a0 @code="en", @name="ENGLISH">,
  #<CLD::Language:0x007fb7728b6740 @code="en", @name="ENGLISH">,


Full RDoc documentation is at and includes even more examples.



The libcld C++ library must be installed (see above). Tested with node version 10.0.26.


$ git clone
$ cd cld/ports/node
$ make -f Makefile.example test
$ npm install /PATH/FOR/CLD/cld/ports/node


Assuming you're in /PATH/FOR/YOUR/PROJECT above:

var LanguageDetector = require("cld/cld.node").LanguageDetector;
var detector = new LanguageDetector();

// Sync - Returns two letter language code of the most likely candidate language
var simpleResult = detector.detectSync("This is my sample text");
// Returns 'en' in this case

// Async - Returns detailed result structure
detector.detect("This is my sample text", function(result) {
  // 'result' contains:
  // { languageCode: 'en',
  //  reliable: false,
  //  details:
  //   [ { languageCode: 'en', normalizedScore: 20.25931928687196, percentScore: 64 },
  //     { languageCode: 'fr', normalizedScore: 8.221993833504625, percentScore: 36 },
  //     { languageCode: 'un', normalizedScore: 0, percentScore: 0 } ] }


Both the detectSync and detect methods take an option second parameter that is a hash of options. The available options (and defaults) are:

  • tld (none): The TLD of the domain that the data came from. This is used as a hint for some ambiguous languages, for example the following snippet from there tests:

    var detector = new LanguageDetector();
    var ambig =  "  常用漢字  ";
    detector.detectSync(ambig)                  #=> "zh-TW"
    detector.detectSync(ambig, { tld: "cn" })   #=> "zh"
    detector.detectSync(ambig, { tld: "jp" })   #=> "ja"
  • html (false): The string to be processed HTML. If this is the case then markup will be ignored in the calculations.

  • allowExtendedLanguages (true): Return language from outside of the core set of supported languages (where wuality is not as good)

  • pickSummaryLanguage (false): If true then the summary language may not be the top match, based on some other factors.

  • skipWeakMatches (false): Include weak matches in the results


As a JNI library the Java port may require some environment settings to work correctly. For example on a fresh Ubuntu install I needed to install the Oracle JDK, set JAVA_HOME and LD_LIBRARY_PATH for linking and compilation to work correctly.


The libcld C++ library must be installed (see above). See the notes above about compilation and runtime linker environments.


# Maven repository information in the future
$ git clone
$ cd cld/ports/java
# mvn test (optional)
$ mvn install


CompactLanguageDetector compactLanguageDetector = new CompactLanguageDetector();
LanguageDetectionResult result = compactLanguageDetector.detect("This is my sample text");
if (result.isReliable()) {
  // getProbableLocale returns a java.util.Locale
  System.out.println("Pretty sure that's " + result.getProbableLocale().getDisplayName());
} else {
  for (LanguageDetectionCandidate candidate : result.getCandidates()) {
    System.out.println("Maybe it's " + candidate.getLocale().getDisplayName());


Full Javadocs are at



The libcld C++ library must be installed (see above)


$ git clone
$ cd cld/ports/python
$ make install # This will prompt for your password


import cld

detectedLangName, detectedLangCode, isReliable, textBytesFound, details = cld.detect("This is my sample text", pickSummaryLanguage=True, removeWeakMatches=False)
print '  detected: %s' % detectedLangName
print '  reliable: %s' % (isReliable != 0)
print '  textBytes: %s' % textBytesFound
print '  details: %s' % str(details)

# The output look lie so:
#  detected: ENGLISH
#  reliable: True
#  textBytes: 25
#  details: [('ENGLISH', 'en', 64, 20.25931928687196), ('FRENCH', 'fr', 36, 8.221993833504625)]


Once you've compiled & installed the Python bindings detection is easy.

First, you must get your content (plain text or HTML) encoded into UTF8 bytes. Then, detect like this:

topLanguageName, topLanguageCode, isReliable, textBytesFound, details = cld.detect(bytes)

The code and name of the top language is returned. isReliable is True if the top language is much better than 2nd best language. textBytesFound tells you how many actual bytes CLD analyzed (after removing HTML tags, collapsing areas of too-many-spaces, etc.). details has an entry per top 3 languages that matched, that includes the percent confidence of the match as well as a separate normalized score.

The detect method takes optional params:

  • isPlainText (default is False): set to True if you know your bytes don't have any XML/HTML markup

  • includeExtendedLanguages (default is True): set to False to exclude "extended" languages added by Google

  • hintTopLevelDomain (default is None): set to the last part of the domain name that the content came from (for example if the URL was, pass the string 'cz'). This gives a hint that can bias the detector somewhat.

  • hintLanguageCode (default is None): set to the possible language. For example, if the web-server declared the language, or the content itself embedded an http-equiv meta tag declaring the language, pass this (for example, "it" for Italian). This gives a hint that can bias the detector somewhat.

  • hintEncoding (default is None): set to the original encoding of the content (note you still must pass UTF-8 encoded bytes). This gives a hint that can bias the detector somewhat. NOTE: this is currently not working.

  • pickSummaryLanguage (default is False): if False, CLD will always return the top matching language as the answer. If True, it will sometimes pick 2nd or 3rd match (for example, if English and X match, where X (not UNK) is big enough, assume the English is boilerplate and return X). In simple testing accuracy seems to suffer a bit (XX to YY %) when this is True so I've defaulted to False.

  • removeWeakMatches (default is True): if a match isn't strong enough, delete it. This ensures some amount of confidence when a language is returned.

The module exports these global constants:

  • cld.ENCODINGS: list of the encoding names CLD recognizes (if you provide hintEncoding, it must be one of these names).

  • cld.LANGUAGES: list of languages and their codes (if you provide hintLanguageCode, it must be one of the codes from these codes).

  • cld.EXTERNAL_LANGUAGES: list of external languages and their codes. Note that external languages cannot be hinted, but may be matched if you pass includeExtendedLanguages=True (the default).

  • cld.DETECTED_LANGUAGES: list of all detectable languages, as best I can determine (this was reverse engineered from a unit test, ie it contains a language X if that language was tested and passes for at least one example text).