Join GitHub today
GitHub is home to over 50 million developers working together to host and review code, manage projects, and build software together.Sign up
GitHub is where the world builds software
Millions of developers and companies build, ship, and maintain their software on GitHub — the largest and most advanced development platform in the world.
Dick Sites (and others) at Google graciously provided a new version 2.0 of the compact language detector, here: https://code.google.com/p/cld2/ and I (email@example.com) created the Python bindings and ported the C++ test case to test.py. This has been tested on Ubuntu 14.04, with both Python 2.7.6 and 3.4.0. Updated Nov 11 2014 to the latest CLD2 release, adding new bestEffort flag (to force a guess even when confidence is low), and cutting over to the CLD2 methods that confirm incoming UTF-8 is valid. To build: * First checkout cld2, cd internal, run compile_libs.sh. This will create both libcld2.so (small tables, detects 83 languages) and libcld2_full.so (large tables, detects 163 languages). Install those libraries somewhere on your LD_LIBRARY_PATH, for example copy them into /usr/lib. * Edit both setup.py and setup_full.py: change CLD2_PATH to point to where you checked out the CLD2 sources. * python setup.py build * python setup_full.py build Note that all Python sources work with both python 2.x and 3.x so if you want to install for python3.x just repeat the above steps using python3 (or whatever python command runs python 3.x in your environment). To test both the small and full language tables: * python test.py The test produces a lot of output, due to the test cases testing the debug flags; this is normal. As long as it says OK in the end then the tests passed. To install: * python setup.py install (as root) * python setup_full.py install (as root) For documentation run: * python -c "import cld2; help(cld2.detect)" NOTE: gen_test.py and gen_enc.py were used as temporary helpers during development and are not needed for building NOTE: you must pass only valid UTF-8 bytes to the detect function, otherwise you can hit segmentation fault or get incorrect results.