summaryrefslogtreecommitdiff
Commit message (Collapse)AuthorAgeFilesLines
* Missed backticks in README.developSeth M Morton2016-05-081-2/+2
|
* Grammar in documentation.Seth M Morton2016-05-082-2/+2
|
* Updated some documentation misses.Seth M Morton2016-05-083-6/+16
|
* natsort version 5.0.0 release.v5.0.05.0.0Seth M Morton2016-05-083-9/+46
| | | | | | | | | | | | | | | | | | | | | - ``ns.LOCALE``/``humansorted`` now accounts for thousands separators. - Refactored entire codebase to be more functional (as in use functions as units). Previously, the code was rather monolithic and difficult to follow. The goal is that with the code existing in smaller units contributing will be easier. - Deprecated ``ns.TYPESAFE`` option as it is now always on (due to a new iterator-based algorithm, the typesafe function is now cheap). - Increased speed of execution (came for free with the new functional approach because the new factory function paradigm eliminates most ``if`` branches during execution). - For the most cases, the code is 30-40% faster than version 4.0.4. - If using ``ns.LOCALE`` or ``humansorted``, the code is 1100% faster than version 4.0.4. - Improved clarity of documentaion with regards to locale-aware sorting. - Added a new ``chain_functions`` function for convenience in creating a complex user-given ``key`` from several existing functions.
* Fixed locale error if no locale is loaded.Seth M Morton2016-05-081-1/+4
|
* Fixed flakes error in profile_natsort.Seth M Morton2016-05-081-1/+0
|
* Profile code now tests what matters.Seth M Morton2016-05-081-101/+40
|
* Natsort now imports from "icu".Seth M Morton2016-05-081-10/+10
| | | | "PyICU" is deprecated.
* Made documentation great again.Seth M Morton2016-05-088-284/+298
| | | | | Added more clarification, made quick start more navigable, and added more locale information.
* New coverage guards.Seth M Morton2016-05-051-4/+4
|
* Added unit tests for LOCALENUM functionality.Seth M Morton2016-05-054-17/+167
| | | | | | | | | | LOCALENUM is also now dependent on FLOAT. Additionally, it was found that the broken locale problem extends to the thousands separator character. A lookup table has been implemented for locales with known problems. This closes issue #36.
* Separated LOCALE into LOCALEALPHA and LOCALENUM.Seth M Morton2016-05-055-40/+57
| | | | This is so users can control if they want numeric modifications or not.
* Removed locale_help.py.Seth M Morton2016-05-058-103/+17
| | | | | The remaining function, locale_convert_function, was replaced by get_strxfrm from natsort.compat.locale.
* Pulled groupletters out of locale_help.Seth M Morton2016-05-0516-102/+79
| | | | | | This function has been moved to utils. Additionally, the python version checking has been unified.
* Added thousands and decimal separator replacement.Seth M Morton2016-05-045-25/+63
| | | | | | This is turned on with LOCALE. Unfortunately, the thousands separators are not reliable with the OSX implementation of locale (of course), by PyICU is still good. Need to add more tests and documentation.
* Hopefully made Travis-CI startup shorter.Seth M Morton2016-05-042-3/+1
| | | | Also removed a stray print.
* Fixed a Python 2.6 string formatting error.Seth M Morton2016-05-041-1/+1
|
* Fixed PyICU TypeError.Seth M Morton2016-05-044-19/+11
| | | | | | With the LOCALE | CAPITALFIRST options there was a type mismatch between numbers and strings. This has been resolved by replacing "null_sep" with ''.
* Improved chain_functions speed.Seth M Morton2016-05-032-6/+16
| | | | | This new method essentially hard-codes the chained function using eval.
* Ensured numbers won't break with LOCALE and CAPITALFIRST or PATH.Seth M Morton2016-05-035-10/+66
| | | | | | | PATH or LOCALE and CAPITALFIRST modify the output. The output will be modified appropriately for numbers. Improved ns enum documentation.
* Clarified post_string_parse_function tests.Seth M Morton2016-05-031-0/+3
|
* Added tests for natsort_keygen.Seth M Morton2016-05-033-8/+53
| | | | | These tests are not entirely thorough, but they demonstrate the function well enough. The rest of the testing is good.
* Fixed rare false failure in test_fake_fastnumbers.Seth M Morton2016-05-031-1/+6
|
* Consolidated test PEP8 exceptions.Seth M Morton2016-05-031-14/+1
|
* Split tests into many more files.Seth M Morton2016-05-0313-687/+860
| | | | | | The testing files were getting too long, so they have been split into more targeted files. This will make further development less daunting.
* Reduced false failures due to unicode whitespace.Seth M Morton2016-05-032-1/+18
|
* 3rd attempt to fix Python 2.6 testing bug.Seth M Morton2016-05-022-16/+16
|
* 2nd attempt to fix 2.6 testing bug.Seth M Morton2016-05-021-1/+2
|
* Fixed testing bug with Python 2.6.Seth M Morton2016-05-021-3/+6
|
* Removed excess packages from tox.ini.Seth M Morton2016-05-021-16/+1
|
* Fixed Python2/3 incompatibility bugs.Seth M Morton2016-05-021-15/+18
|
* Redefined _natsort_key in a functional style.Seth M Morton2016-05-025-408/+194
| | | | | | | | | | | | | | This required changing the input to be the functions created by all the function factories that have been recently committed. This had a cascading effect that required a re-write of several unit tests, as well as the top level natsort_keygen function, which now has the job of creating the appropriate functions on the fly to pass to _natsort_key. A future commit will add more testing to natsort_keygen. For the moment, handling of locale-specific numbers is disabled, but will be added back in a future commit.
* locale_convert is now locale_convert_function.Seth M Morton2016-05-024-15/+32
| | | | | It now returns the appropriate function for the given configuration to do the locale conversion. No unnecessary if statements.
* Removed "if" in _post_string_parse_function.Seth M Morton2016-05-021-2/+3
| | | | | The function returned from _post_string_parse_function had an unneeded if statement within it. It has been removed.
* Removed the now unused _ALL_BUT_PATH.Seth M Morton2016-05-021-1/+0
|
* Clarified a statement in the README.Seth M Morton2016-05-021-2/+2
|
* Removed a line from coverage.Seth M Morton2016-05-021-1/+1
|
* Cleaned some of the top-level functions.Seth M Morton2016-05-021-29/+4
| | | | | | | The main sorting functions had some code to handle TYPECAFE addition. This is no longer needed since TYPESAFE is default. Further, calls to _kwargs_to_alg are pushed to natsort_keygen.
* Redefined _parse_string_function.Seth M Morton2016-05-021-9/+17
| | | | | The new definition returns a function with no if statement in an attempt to micro-optimize this function.
* Added a docstring to chain_functions.Seth M Morton2016-05-024-1/+44
|
* Added new string and path splitting function generators.Seth M Morton2016-05-013-124/+98
| | | | These complete the required components to make natsort more functional.
* Removed _fix_nan function.Seth M Morton2016-05-012-25/+0
| | | | The work of this function is now being done elsewhere.
* Solved some testing errors.Seth M Morton2016-05-015-19/+18
|
* Fixed setup.py classifiers.Seth M Morton2016-05-011-13/+19
|
* Fixed old accidental bug in bytes parsing.Seth M Morton2016-04-181-1/+1
|
* Changed _ungroupletters into a factory function.Seth M Morton2016-04-182-30/+48
| | | | It is now called _post_string_parse_function.
* Created _parse_bytes_function.Seth M Morton2016-04-183-1/+56
| | | | | This is a factory function that will return a function to parse a bytes string in the appropriate manner according user options.
* Created _parse_number_function.Seth M Morton2016-04-182-6/+42
| | | | | This is a factory function that will return a function to parse a number in the appropriate manner according user options.
* Switched to try: except in fastnumbers.Seth M Morton2016-04-151-12/+12
| | | | This will only be used in Python2.
* Fixed typo in README.Seth M Morton2016-04-151-1/+1
|