Skip to content
This repository has been archived by the owner on Feb 27, 2021. It is now read-only.

Latest commit

 

History

History
235 lines (194 loc) · 8.99 KB

CHANGELOG.md

File metadata and controls

235 lines (194 loc) · 8.99 KB

cuStrings/nvStrings 0.10.0 (TBD)

New Features

  • PR #416 Added documentation on supported timestamp specifiers

Improvement

  • PR #403 Remove driver install from ci scripts
  • PR #411 Format changes for merge with cudf repo
  • PR #415 Refactor NVText to improve compile time and logic reuse

Bug Fixes

  • PR #413 Fixed get_info character formating
  • PR #412 Fixed bug in compare when used with non-ascii character
  • PR #414 Fixed memory leaks in urldecode() and insert()

cuStrings/nvStrings 0.9.0 (21 Aug 2019)

  • PR #355 Add and use unified build script
  • PR #390 Fix local build generated file ownerships

New Features

  • PR #363 Added scatter method to nvstrings
  • PR #371 Added replace() with multiple strings parameters
  • PR #385 Added Porter Stemmer measure method
  • PR #382 Added url encode/decode method
  • PR #387 Added scalar_scatter method to nvstrings
  • PR #386 Added normalize_spaces method to nvtext
  • PR #395 Added python bindings for device memory to nvstrings

Improvements

  • PR #354 Removed rmm submodule
  • PR #359 Set default sort-type to alphabetical
  • PR #357 Improved error checking on memory allocates
  • PR #358 Added improved hash algorithm to reduce collisions
  • PR #369 Reconfigure C++ source directory structure
  • PR #370 Updated memsize() to include pointer memory too
  • PR #375 Added gtests for C++ unit testing

Bug Fixes

  • PR #353 Fixed sizer calculation for multiple replaces
  • PR #378 Fixed small memory leak in string categories
  • PR #380 Removed obsolete directory from Dockerfile
  • PR #373 Properly catch/throw exceptions raised by gather_strings
  • PR #391 Corrected overflow logic in stod

cuStrings/nvStrings 0.8.0 (27 June 2019)

New Features

  • PR #298 Added ngrams method to NVText
  • PR #299 Added gather method for booleans array
  • PR #304 Updated nvtext.ngrams to call C++ function
  • PR #325 Add test skipping functionality to build.sh
  • PR #332 Accept multiple nvstrings in cat() method
  • PR #327 Allow numeric keys in nvcategory
  • PR #396 Streamline CUDA_REL environment variable

Improvements

  • PR #316 Add .0 to integer output when using ftos/dtos
  • PR #315 Support NaN, Inf strings in stof and stod
  • PR #313 Increased regex instruction threshold for stack to 1024
  • PR #302 Add junit test output for py.test
  • PR #331 Use stack-size ranges for different regex instruction counts
  • PR #341 Support device memory in from_offsets

Bug Fixes

  • PR #314 Fixed calculation error in ip2int
  • PR #306 Set main label always for libnvstrings
  • PR #307 Update conda dependencies
  • PR #326 Update python doc version
  • PR #335 to_host no longer overwrites input pointers for null strings
  • PR #329 Fixed documentation errors in source code
  • PR #334 Regex stack-size logic moved to host code
  • PR #343 Fixed a flag which is necessary for conda install in Dockerfile
  • PR #347 Configure Sphinx to render params correctly

cuStrings/nvStrings 0.7.0 (10 May 2019)

Version jump from 0.3->0.7 is to align with other RAPIDS projects.

New Features

  • PR #295 Added nvtext.tokenize
  • PR #281 Added get_info method to show data about strings
  • PR #273 Added tokens_counts method to nvtext
  • PR #265 Added is_empty method
  • PR #260 Support for format-specified date-time to/from string conversion
  • PR #195 Added IPC python bindings
  • PR #204 Added stol/ltos and stod/dtos number converters
  • PR #194 Added method to create strings from array of floats
  • PR #193 Added IPC transfer methods to C++ interface
  • PR #188 Rename rave to nvtext and publish in nvstrings package
  • PR #180 Added edit-distance to rave module
  • PR #176 Added match_strings to nvstrings
  • PR #172 Added to/from boolean conversion methods
  • PR #171 Added conversion to/from subset of ISO8601 format
  • PR #230 Add ngrams function to nvtext module
  • PR #285 Add build script for nightly docs
  • PR #286 Add local build script to cuStrings

Improvements

  • PR #289 Improve regex perf for OR clauses
  • PR #283 Check for empty fillchar in pad methods
  • PR #278 Added insert method as alternative to slice_replace insert
  • PR #268 Support negative slicing
  • PR #264 Allow nvstrings parameter to fillna
  • PR #256 Convert device-only custring_view to header-only
  • PR #255 Removed unnecessary cudaDeviceSync calls
  • PR #237 Removed internal util.h from include dir
  • PR #216 Fixed build instructions and removed obsolete doc/build files
  • PR #178 Added compute_statistics to C++ interface
  • PR #173 Added flake8 checks to for /python files
  • PR #167 Added doxygen to NVStrings.h and NVCategory.h
  • PR #183 Align version to 0.7
  • PR #192 Rework CMakeLists and separate RMM, update conda recipe and setup.py
  • PR #181 Update python docstrings to numpydoc style
  • PR #196 Include nvtext in python module setup
  • PR #221 Create separate conda packages for libnvstrings and nvstrings
  • PR #247 Release Python GIl while calling underlying C++ API from python
  • PR #240 Initial suite of nvtext function Python unit tests
  • PR #275 Add cudatoolkit conda dependency
  • PR #291 Use latest release version in update-version CI script
  • PR #267 Add a minimal suite of nvstrings and nvcategory pytests

Bug Fixes

  • PR #276 Fixed token_count counting empty string as token
  • PR #269 Fixed exception on invalid parameter to edit_distance
  • PR #261 Fixed doxygen formatting and updated to 0.7
  • PR #248 Fixed docstring for index,rindex,find,rfind
  • PR #245 Fixed backref to continue replacing
  • PR #234 Added more type-checking to gather method
  • PR #226 Added data pre-check to create_from_index
  • PR #218 Parameter check corrected for pad methods
  • PR #217 Corrected custring.cu method signatures
  • PR #213 Fixing README link to Python API docs
  • PR #211 Re-ordering 2 methods stops invalid-dev-fn when called from cudf
  • PR #207 Fixed handling empty pattern in findall/count
  • PR #202 Allow insert on start=stop in replace
  • PR #201 Fixed some pad examples in nvstrings.py doc comments
  • PR #186 Fixed memory error in booleans to strings method
  • PR #235 Fix anaconda upload script for new conda recipe
  • PR #239 Fix definitions of upload files for conda packages
  • PR #241 Revert split package to individual recipes to avoid run_exports bug
  • PR #274 Updated RPATH/RUNPATH setting to accommodate current install location

cuStrings/nvStrings 0.3.0 (15 Mar 2019)

New Features

  • PR #148 Added fillna method to nvstrings
  • PR #105 Added ip2int/int2ip method for converting IPv4 addresses to and from ints
  • PR #99 Added nvstrings.itos (int to string) method to nvstrings
  • PR #94 Support backreferences in regexes
  • PR #81 Added copy to nvstrings
  • PR #79 Added add_strings to nvstrings bindings
  • PR #67 Adding merge_and_remap
  • PR #65 Implement len() for nvstrings
  • PR #46 Added htoi (hex to int) to nvstrings
  • PR #42 Strip multiple chars
  • PR #41 Added from_offsets method
  • PR #23 Add [] operator to nvstrings
  • PR #9 category merge python interface
  • PR #1 Added nvcategory.merge_category
  • PR #154 Default to CXX11_ABI=ON

Improvements

  • PR #163 Follow standard RAPIDS README format
  • PR #152 Add CHANGELOG.md
  • PR #145 Improve handling of whitespace in split
  • PR #144 Allow -1 for gather() position values
  • PR #141 Allow for null category values
  • PR #137 Update gpuCI scripts
  • PR #134 gather/and_remap
  • PR #130 Add bind_cpointer to nvcategory, add own parameter
  • PR #127 Splitup 6k-line source file
  • PR #125 Added nullfirst param to sort/order
  • PR #124 Split to return column of nulls if all values null
  • PR #121 Added nulls param to itos
  • PR #110 Rename split, rsplit, extract methods to match expected behavior
  • PR #108 Return None for nulls when returning Python lists
  • PR #106 Added set_null_bitmask method
  • PR #102 Added gather method
  • PR #101 Support both single instance, and lists in from_strings
  • PR #100 Support empty string '' for na_rep for nvstrings.cat
  • PR #92 Use global memory for long regexes
  • PR #89 Added remove-unused-keys
  • PR #88 Added get_value_bounds
  • PR #86 Add gather handle for ndarray
  • PR #84 Added accessor for c-pointers
  • PR #75 return null-bitmask in create_offsets
  • PR #71 Added null_count method
  • PR #68 Add byte_count to nvstrings interface
  • PR #62 add regex quantifier documentation
  • PR #61 add quantifier support for regexes
  • PR #60 add support for ndarray, buffer in getitem/gather
  • PR #49 nvcategory update keys methods: add, remove, set
  • PR #54 throw exception if enumerate is used on nvstrings instance
  • PR #39 Rename sublist1 to gather
  • PR #12, #13, #14, #136 RMM integration
  • PR #2 Refactored Reclass

Bug Fixes

  • PR #151 Update conda build script to handle build changes
  • PR #150 split-ws var init corrected
  • PR #140 Fix null placement in from_strings
  • PR #133 Corrected parameter types
  • PR #126 Fixed memory leak
  • PR #123 Fixed sort predicates for nulls
  • PR #116 Consistent split nulls
  • PR #91 Caller allocs for to_host
  • PR #87 Gather throws exception if args out of range
  • PR #85 Fix join with null-string issue
  • PR #43 cat values are ints
  • PR #38 Catbitmask
  • PR #37 use iterator instead of at() in regex eval
  • PR #31 Change NVCategory values type from unsigned to signed int32
  • PR #11 handle unescaped end-brace edge-case
  • PR #157 Fix Python not being found by CMake