[pandas] annotated tag v0.4.0 created (now 1fbb3c7)

Andreas Tille tille at debian.org
Wed Dec 28 15:20:22 UTC 2016


This is an automated email from the git hooks/post-receive script.

tille pushed a change to annotated tag v0.4.0
in repository pandas.

        at  1fbb3c7   (tag)
   tagging  9e406c50b77949433ed8f7fd094fd30b5421a35e (commit)
 tagged by  Wes McKinney
        on  Mon Sep 12 16:32:15 2011 -0400

- Log -----------------------------------------------------------------
Version 0.4.0

Dieter Vandenbussche (1):
      Minor change to CleanCommand so build works with stdeb

Joon Ro (2):
      added describe and scoreatpercentile in frame.py
      changed _pfixed in common.py to add a leading space to positive numbers

Justin Berka (1):
      BUILD: Fixed issue #80, xls/csv file copying

Shane Conway (4):
      Added functions to pull data off Yahoo! and FRED
      Added function to pull data from Fama/French data
      Updated data reader functions
      Updated data reader functions

Skipper Seabold (1):
      ENH: Add join to Panel

Tim McNamara (1):
      README updated for better Github output

Wes McKinney (853):
      Initial directory structure.
      adding trunk
      oops
      added svn:ignore
      first commit with cleaned up code
      minor edit
      fixed isinf reference
      latest edits, miscellaneous cleanup and bug fixes from development
      added stats empty directory
      fixed inconsistency with dateCol parameter
      moved operator factory methods out of class level
      fixed error in groupby, added yahoo finance example
      added pandas.stats.linmodel
      added pandas.stats.linmodel
      adding test directory
      added pandas.stats to package list
      added regression examples
      cleaned up linmodel class and made to work with scikits.statsmodels
      improved regression example
      improved regression example
      removed reference to regress function
      couple unfortunate hacks with numpy ascii and unicode string types in groupby
      misc fixes
      some Series constructor cleanup, made DataFrame.toDataMatrix better
      misc cleanup
      revamped DataFrame.reindex to better handle mixed-type data, and also be a lot faster
      made sum, count, mean functions more robust to mixed-type data
      got rid of usages of np.unique
      made index attribute a property
      added prototype WidePanel, LongPanel
      added slice function to LongPanel
      fleshing out docstrings, API. added filterItems
      using ndarray.take
      added sandbox directory and restricted least-squares class
      merging latest developments, in particular LongPanel, WidePanel
      diff cruft
      reorganized cython code. rolling moment implementations
      added OLS and PanelOLS functionality
      changes for cython 0.12
      misc fixed to Cython
      import fixes, removed old unit tests
      removed linmodel module
      set to ignore pyc files
      made build script for lib
      minor change
      modified build script
      modified build script
      working on min_periods for regular OLS, leaving things in a semi-broken state for now
      still mid-refactor on min_periods
      finished min_periods implementation, cleanup of shared code between MovingOLS and MovingPanelOLS
      fixed ewma problem
      added time rule mappings in datetools.py and connected with shift, asfreq methods'
      fixed flatiter issue
      shift method has timeRule argument
      removed old f_stat_raw
      typo
      fixed numpy.flatiter issue
      made Panel objects Picklable
      changed Factor class a bit, and made LongPanel.fromRecords mostly functional
      DataMatrix will classify booleans as objects, and Series.reindex with boolean arrays will produce object array with NaNs if things are missing. Need some unit tests for these
      test suite in broken state for now
      DataMatrix.apply will preserve column order
      Series.interpolate, LongPanel.getAxisDummies improvements
      reverted map_indices from buffer version, for now
      fixed rolling_rank corner case
      PanelOLS will accept LongPanel arguments
      added new test files
      fix to DataMatrix constructor
      made test case stubs for all Index, Series, DataFrame, DataMatrix methods
      deleted miscellaneous unneeded methods, bugfix to DataMatrix.__init__
      panel test method stubs
      sphinx quickstart
      minor fix to DataMatrix constructor
      unit tests
      unit test coverage for Series and Index, various reorganization in release prep
      more unit tests for DataFrame and DataMatrix
      moving toward release
      fixed problem with Index.__array_finalize__ when calling tolist
      working on sphinx docs
      hard coding version number for now
      misc minor changes
      few more unit tests for Series
      a little more Series test coverage
      fixed up docstrings. changed Series.__new__ back to NOT copy the input data (causes problems in DataMatrix.__getitem__ when you want to modify a column in-place)
      few more unit tests
      more DataFrame unit test coverage
      added lagged_y_predict to PanelOLS
      miscellaneous changes, fix to Series.reindex when using fillMethod 'pad' on empty Series
      removeD
      added source directory back
      adding back
      latest Series docs, installation, front page material
      latest Makefile
      not weighting untransformed data in PanelOLS. changed DataFrame and DataMatrix.toString to take a buffer instead of a to_stdout flag
      DataFrame documentation and other extensive edits
      simpler column reindexing in DataFrame and DataMatrix, some cleanup for release
      release 0.1
      bar chart option in Series.plot
      optimized combineFrame, added DataFrame.filter function
      increased test coverage of DataFrame / DataMatrix, miscellaneous fixes
      more unit test coverage for DataFrame etc.
      got rid of basically pointless asarray flag in DataFrame/Matrix statistical methods
      fixed inconsistencies with weighted panel regression, some code cleanup
      removed usages of asarray flag in stats package
      * Cleaned up and sped up DataMatrix and DataFrame pickling
      fixed DataMatrix.copy to make a copy of the objects matrix as well
      removed unneeded import
      consolidated filter methods, removed DataMatrix.getTS
      fleshing out WidePanel functionality. miscellaneous minor changes
      fixed unit tests
      added note about NumPy 1.3
      added more statistical methods to WidePanel and finally grinding out unit tests for Panel classes
      added rolling moment interface methods
      many more unit tests, expansion of pandas.* namespace. expanded WidePanel functionality
      Python 2.4 compatibility and enabling to be used without having SciPy installed
      restored dateutil import
      accept record arrays
      added VAR implementation
      miscellaneous fixes and cleanup
      import fix
      edits to docs. unit tests
      adjustments to platform independent make.py
      miscellaneous changes to codebase, docs in flux
      fixed r-squared calculation when response variable is assumed to be zero-mean (no intercept used)
      fixed truncate bug
      more unit test coverage
      fix for MSVC2008 compatibility
      added IndexableSkiplist and first version of a faster rolling median algo
      simplified rolling moments
      rolling_median using skiplist
      loads of cleanup, tests, and DataFrame / DataMatrix API consistency. removed DataFrame.fromDict and fromMatrix methods, incorporated into default constructor
      further test coverage (technically 100%) for frame.py and matrix.py
      misc edits
      fixed ewma function
      removed Series.fromDict and integrated with Series.__new__
      merged in some minor changes
      ported tdates C extension to Cython, removed unneeded include files, etc.
      cython 12.1
      more unit tests, miscellaneous fixes for tdates.c removal
      added testing module
      test coverage
      fixed typo
      more test coverage
      renamed dayOfWeek parameter
      bugfix + unit tests
      test coverage
      working a bit on docs
      removed ext line from ts conf
      moving ols speed enhancements and import optimization
      unit test
      more unit testing, starting merge function cleanup
      bit of getMergeFunc cleanup
      grinding unit tests. groupby cleanup. cython code cleanup and testing. reindex function cleanup
      test coverage
      merge / join function cleanup
      removed duplicated code
      code rearrangement
      array interface to DataFrame classes
      test coverage. deleted deprecated filter* methods. working on docs for 0.2 release
      rename methods
      cap, floor, and unit tests
      test coverage
      minor edit
      unit tests
      fixing up setup.py scripts so test directories get included and extension module too
      removed unnecessary Python.pxi
      cannot safely call ndarray.any on an object array i guess
      fixed issue 9; miscellaneous refactorings and unit tests
      renaming
      DateRange enhancements. fixed failing fama-macbeth unit test
      latest
      unit tests
      unit tests
      reindex _backfill bugfix (switched to buffer interface), min / max performance enhancements
      wrote wirth.c in Cython
      no mutation
      whoops
      unit tests
      added io.pyx file, first version of HDFStore
      fix
      display bugfix
      bugfix for float-Series + int-Series
      benchmark suite
      DataMatrix._resultFrame optimization (removed fancy indexing)
      working on docs + unit tests
      code reorg in Series, miscellaneous groupby refactoring. Changed Series.values() to be a property so the interface is consistent with DataMatrix and WidePanel
      expanding docs toward 0.3 release
      shortened moments.py
      docstring fixed, setup.py versioning help
      unit tests
      lowered version, spot of code reformatting
      renamed parameters in DateRange, promoted isnull and notnull out of Cython code
      bugfix for unit test breakage in isnull
      fixed reported issue with adding / subtracting Tick objects
      renamed cap methods to clip
      committing tests and changes in broken state, to be fixed ASAP
      fixed plm brokenness. changed boolean reindexing to cast to float
      Series.order was failing on object dtype. misc other cleanup
      rolling_min and max impl
      formatting issues and copy issues
      fixed DateRange state pickling issue, turn off rpy unit tests for now
      misc documentation, some work on rpy2 interface. near git migration
      Added .gitignore file
      test
      Updated NumPy sphinxext, Cython 0.13 upgrades, and misc
      Bugfix in IndexableSkiplist that was breaking rolling_max
      Fixed DateRange.union logic, docs, changed XDateRange class to a generator function
      Fixed test cases
      Misc cleanup, docs, pylint directives, etc.
      API changes, cleanup, unit testing
      statsmodels dataset API change
      added matplotlib plot_directive extension to config, few moment plots etc.
      corr_with stub, replaced remove_na usage with valid()
      notes nearing 0.3 release
      temporary version.py post git migration
      fixed brokenness with version.py
      fixed brokenness with version.py
      fixed version.py template
      fixed version.py template
      misc cleanup in rolling moments, work on var.py
      added sql module, still in prototype stage
      fixed IPython debugger usage
      adding reindex_like functions
      DataFrame.corrwith and unit test coverage
      changed fillMethod argument name and added DeprecationWarning
      unit test to raise DeprecationWarning so I don't forget
      removed cap/floor functions from DataMatrix, replaced by clip*. Unit test coverage
      misc code reorg
      added release notes
      refactoring EW moments to have uniform interface, unit test coverage
      Rolling moment plots in docs, API consistency, etc.
      fixed docs for rolling_{cov,corr]
      fixed GH issue 7
      fixed GH issue 4
      'fixed' GH issue 15
      working toward consistent column ordering in DataFrame vs. DataMatrix
      consistent column ordering fixes
      added python 2.5 build file
      python 2.7 builder
      beta for now
      catch dateutil import error and print message
      updated docstring
      python 2.7 test script
      Ran nose with Python 2.7 and fixed warnings issues, patched up missing import errors in unit tests (e.g. SciPy, statsmodels) so users without them can run the tests without having a bunch of stuff print to the screen
      recompiled Cython with 0.14, and added reindex_like to WidePanel
      reverted to Cython 0.13 due to bug
      bumped to beta2
      updated NumPy version in PyPI info
      Fixed median issue GH #26
      Merge branch 'master' of https://github.com/wesm/pandas
      added WeekOfMonth date offset
      unpickling compatibility with old pickle files
      cols() returns list type
      maintain sortedness of columns on insertion again (when possible)
      fixed object insertion into float matrix of same column name
      changed arg name
      fix DataMatrix.combineFirst issues + unit tests
      cast to float in Series.map when have missing values in int array
      fixed NaN casting issue in int arrays for _join_on
      fix bug in Index.__add__
      fixed issue GH #19
      refactoring, module needs more work
      refactoring DataFrame arith methods, working on comparison methods
      unit tests for WeekOfMonth
      random edits
      add axis argument to Series.plot
      unit tests
      unit tests, coverage
      trying to fixed mixed-type ambiguity in DataFrame.sum
      made ols interface function easier to use, fixed up docs a touch
      fixed integer indexing bug
      fixed int64 indexing problem
      misc enhancements to rpy, DataFrame.toRecords
      fixed DateRange buglet with DateOffset caching-- added CacheableOffset class to make it easier to differentiate
      fixed statsmodels api in ols/unit tests. misc test coverage
      fillna function and warnings in fill(), misc unit tests and things
      whoohoo! 100% line coverage for pandas.core
      unit tests, getXS deprecation warning
      added ability to boolean index DataFrame objects
      0.3.0 release notes
      more 0.3.0 release notes
      fixed NumPy 1.4.0 NaN-handling issue in DataFrame.combine
      changed repo loc in docs
      deprecated Series.fromValue, moved logic into constructor
      can build latex. misc edits
      fixed up Series.fromValue deprecation changes
      fixed GH #27
      fixed paths for include files
      version 0.3.0
      added hist functions to DataFrame and Series, added use_index option to {DataFrame, Series}.plot
      fixed truediv issue in DataFrame with importing division from __future__
      speed optimization in DataMatrix.__getitem__
      deprecation warnings in panel.py
      refactoring groupby a bit, some speed optimizations, unit tests. transform can propagate things like np.mean now
      updated groupby docs
      Merge branch 'master' of github.com:wesm/pandas
      fixed NumPy 1.6 broadcasting bug, groupby bugfix
      fixed DataFrame.toCSV bug, test coverage
      Merge branch 'master' of https://github.com/joonpy/pandas into joonpy-master
      tweaked describe and renamed scoreatpercentile to quantile
      unit tests
      added quantile and describe to Series
      tweaked toString output for DataFrame and DataMatrix
      starting on sparse data structures
      starting on sparse data structures
      work on arith operators. need to test
      operators appear to work, and relatively fast
      sparse progress
      more sparse work
      about to move test_sparse
      moved test_sparse
      broke something. sigh
      fixed issue
      shell of SparseSeries subclass working
      refactored to remove SparseVector class
      SparseSeries and SparseDataFrame work, piecing things together
      further sparse work, testing, etc.
      block union routine finally working
      operators with fill values appear to be working
      fixed bug in arith ops for SparseSeries
      misc refactorings and work related to sparse data structures
      mini fix to plot function
      further sparse unit testing
      skeleton of sparse WidePanel, unit tests and stubs, more necessary functionality
      lookup bugfixes and tests
      moved test_sparse to test_libsparse
      SparseSeries getitem working correctly now
      unit tests
      unit tests, identified weird NumPy issue with r* operators and sort of worked around
      SparseDataFrame reindex unit tests
      more sparse unit tests. barebones panel stacking routine written and decently optimized
      little refactoring
      sparse_reindex, homogenize and tests
      working on SparseWidePanel. reduced some code dup in core classes
      more SparseWidePanel work
      correct pickling of sparse frame/panel. setitem/delitem for sparse panel
      working on sparse panel reindexing and other things
      sparse panel reindex unit test
      addressed panel constructor case
      reindex test
      most efficient SparseWidePanel.to_long that I can think of
      more to_long unit tests, minor bugfix
      major/minor xs
      some plm optimizations. sparse wide panel *may* be working in it now
      added sparse extension to setup
      ExcelFile class, enhanced excel date handling, per GH#9
      removed trycall function
      unit test coverage right about 100% for sparse data structures, yes \!
      modest pytz handling in DateRange. starting on AxisIndexer in frame.py
      custom NA values, allow skiprows in ExcelFile
      changed / fixed inconsistency / minor optimization in Index.union when one contained in the other, fixed resulting failing unit tests. forced SparseSeries fill_value to always be float64 type
      added sparse.c for now
      added fancy indexing 'ix' property to DataFrame, unit tests. TODO: update docs for ix property
      spelling fail
      csv parser improvements, universal newlines
      consistent column ordering in simpleParser, select function in DataFrame, boolean indexing fix in Series
      fixed breaking unit test due to slicing int labels
      fixed DataFrameIndexer bug, wrote unit tests, select unit tests
      refactoring
      get view when calling xs in DataMatrix
      finished select implementation, refactored _get_axis* methods, unit tests
      fixed inference of window_type='rolling' in ols when passing window
      code coverage, corner case when modifying sparse fill_value
      boolean indexing with Series must match index labels
      some refactoring, starting work on fast time series reductions
      broken, but want to commit new branch
      working on pytz support in DateRange
      moving around unit tests, working on fancy get/setitem for objects. fixed buglet in parseCSV
      Series fancy setitem implementation, unit tests, etc.
      more DateRange tzinfo work
      DateRange unit test coverage, tz_validate impl
      Merge branch 'master' into datamatrix-refactor
      boolean indexing work, needs to be part of larger refactor
      added to release notes, removed some deprecated methods, continued work on DataMatrix refactor
      i like where this is going
      this is becoming epic. deep breaths
      refactoring, refactoring. trying not to make a huge mess
      trim trim trim. DataFrame unit tests still run. basic block structure prototype looking good, much implementation work remains...
      renamed proto to internals
      added refactor notes file
      time to start playing whack-a-mole with the test suite
      starting toward reindex methods in new prototype
      continuing with the refactor, eliminating unnecessary code
      DataFrame unit tests pass. lots of work to do to deal with the type hierarchy
      whack-a-mole with the test suite. we're getting there
      brought back dtype casting. making progress
      picklability, column reordering issues, more refactoring
      more bug fixes. more work needed on the mixed-type front
      test suite passes, yay
      entire unit test suite passes
      WidePanel.to_sparse, bugfix, unit tests
      putting SparseWidePanel through more unit tests
      more unit test generalization for sparse structure
      fixed merge conflicts
      switched statsmodels t() usage to tvalues
      a few stats unit tests
      Merge branch 'master' into datamatrix-refactor
      faster count method for SparseDataframe
      Merge branch 'master' into datamatrix-refactor
      unit test for filter with objects, fix
      SparseSeries.shift efficient impl
      handle non-NA fill_value case in SparseSeries.shift. complete test line coverage
      fixed merge conflict
      rename buglet
      rolling moments will return SparseDataFrame now
      fix merge conflict'
      bit the bullet and did the big merge. now time to investigate lots of potential problems
      DataMatrix is no more. test suite passes. lots of missing test coverage / things to look into now
      more cleanup, moved some things to internals. removing references to DataMatrix
      need some changes to block internals, getting started there. misc other impl
      intermediate state atm
      hmm, this probably isn't going to work. need to think on it a bit
      unit tests pass...winning!
      hammering out block internals issues
      more unit tests
      removed VAR impl
      as_matrix will return right type in exceptional cases. more tests. calling it a day
      finished test coverage for internals. more testing still needed
      line coverage for frame.py complete
      consolidated DataFrame unit tests
      removed test_matrix.py
      DataFrame.join optimization. unit tests for consol
      sparse unit tests post refactor, getting closer
      100% test coverage, but perhaps deceptive
      removed pandas/core/matrix.py
      renaming DataMatrix to DataFrame, misc cleanup
      more renaming
      some consolidation, deprecations, cleanup
      once over of DataFrame. looks good to me. more review of SparseDataFrame needed
      append works with SparseDataFrame
      more sparse testing: rename, fillna, apply
      more tests. basically satisfied with SparseDataFrame at the moment
      CRLF -> LF
      CRLF -> LF
      merged datamatrix-refactor branch, fixed conflicts
      CRLF fix, optimization
      fix CRLF
      hasty in removing matrix.py, need for legacy pickles
      read_csv and read_table renamings in parsers.py. updated release notes
      release notes
      added var impl back for now
      WeekOfMonth enhancement w/ unit tests
      fixed API change in examples/finance.py
      minor doc edits for DataFrame/DataMatrix consolidation, bumped version in version.py
      mixed-type fixes in DataFrame statistical methods
      test coverage
      hacking on HDFStore, incorporating submitted patches and optimizing
      unit tests pass
      sped up write_table
      optimized WidePanel read / write method + tests
      remove unit tests, enable to remove full node. docs
      100% test for HDFStore as it now stands
      provide legacy HDFStore file support
      cleanup
      mixed type storage support for DataFrame in HDFStore
      lazy import of PyTables
      updated docs
      updated docs
      error message disabling storing mixed-type DataFrame in table format
      more unit tests, handle all types of Series
      DataFrame fancy indexing gives view
      added missing Series.ix case for completeness, optimized DataFrame.ix[...] to reduce unnecessary computation in some cases
      removed usage of common.get_indexer
      eager evaluation of GroupBy groups
      updated release notes
      starting flexible arith methods on Series
      BUG: Index.__ne__ was not implemented
      BUG: Index.__hash__ should raise an exception
      BUG: DataFrame.ix[0] should return view
      unit testing for flexible Series addition operators
      avoiding code dup in fancy_index
      Pre-emptive unit tests for fancy setitem on DataFrame
      updated readme about github
      little bit of fancy indexing refactoring
      created indexing.py to organize fancy indexing code
      fixed import error
      unit tests for fancy setitem pass
      mostly feature complete for homogeneous fancy DataFrame setitem
      fancy indexing test coverage
      more deprecations, release notes, parsers docs and added new functions to pandas namespace
      pivot multiple columns at once, need tests etc. still
      generalized pivot function to be able to pivot multiple columns into a WidePanel
      updated bsd license
      Cython code reorg and co-opted pyzmq build framework so do not have to check in cythonized .c code into git
      removed stats/setup.py
      updated release notes
      BUG: fillna bug in Series GH#54
      groupby now excludes NA values just like R does. avoids NA-handling issues
      updated release notes about groupby
      release notes
      added set_printoptions function to common module, similar to NumPy print options
      release notes and added set_printoptions to pandas namespace
      changed float_format to precision
      printoptions docstring
      implemented rolling_quantile and unit tests
      updated release notes
      added generic rolling_apply function
      release notes'
      BUG: summing boolean DataFrame
      release notes
      tinkering with more general block structure
      BUG: boolean DataFrame comparison had been broken after internals refactoring
      starting unit tests for parsing functions, way overdue
      added test csv files
      oops, added test_parsers file, unit test coverage
      Merge pull request #55 from dieterv77/CleanCommand
      parsers won't NA out inf values by default. unit tests
      Merge branch 'master' of github.com:wesm/pandas
      test coverage, more still needed
      unit test for specifying custom date parser
      starting in on the mixed-type panel redesign
      removed usages of indexMap, misc panel stuff
      biting the bullet, working to make generic n-dimensional structure to back DataFrame and WidePanel and friends
      added ascending argument to Series.order
      testing IO roundtripping with HDFStore
      Merge branch 'master' into widepanel-refactor
      more nd generalization
      BUG: fixed join_on regression from refactor
      Merge branch 'master' into widepanel-refactor
      BUG: support legacy legacy DataFrame pickle
      Merge branch 'master' into widepanel-refactor
      continuing changing blockmanager to be n-dimensional
      amazingly all the unit tests pass post redesign. much cleanup still needed
      panel tests pass. but others fail still
      internals unit tests pass now
      fixed failing sparse panel and ols unit tests
      unit tests all pass!!
      made BlockManager.xs more generic to support nd
      generic slice with BlockManager for panel
      removed core/setup.py and test coverage for internals
      more panel test coverage'
      test coverage
      always copy data when inserting into DataFrame. failing reverse op dtype=object unit test
      BUG: in as_matrix with subset of columns with single-block
      BUG: fixed GH #56
      Merge branch 'master' into widepanel-refactor
      enhanced xs functionality returns lower dim BlockManager
      test coverage
      SparseDataFrame + DataFrame -> SparseDataFrame, simplified arith methods
      removed collection.py
      enable boolean indexing with object-dtyped array
      some generification
      added fill_value option to DataFrame.{add, mul, sub, div}
      FIX: broke sparse unit tests in the last couple days
      ENH: console formatting enhancements
      contributed patches
      casting to uint8 for numpy < 1.5.1 compat
      add check about numpy version...
      raise skiptest for < python 2.7
      work in python 2.6
      pickle with highest protocol, doesn't work in python 2.6 without
      highest protocol pickle
      ENH: read_table can infer integers. changed DataFrame.fromcsv to use it
      setuptools resulting in unhappiness
      int32/int64 issues
      FIX: numpy.distutils thinks we're using Pyrex
      ENH: rkern patch to debug
      FIX: critical bug fix in block formation
      FIX: unit test for bug
      more unit tests
      ENH: block integrity check (ensure consistent ordering)
      REF: removed pytools and test_pytools
      ENH: make DataFrame a bit more amenable to subclassing
      ENH: cumsum/cumprod refactor, consistency
      ENH: made things more generic, reduced code dup
      fix merge conflict
      FIX: DataFrame.__ne__ was not implemented
      ENH: added BlockManager.insert function
      ENH: use _constructor properties to get factory function instead of type(self), for easier subclassing
      TST: modify DataFrame.values in place
      ENH: DataFrame.insert method implementation
      DOC: noted DataFrame.insert in release notes
      ENH: DataFrame.__init__ will accept structured arrays. general performance tweak in constructor too
      ENH: significant performance optimization in DataFrame constructor with ndarrays
      BUG: reindexing length-0 DataFrame contributed fix
      ENH: contrib perf optimization in panel homogenization routine
      BUG: ndim not handled correctly in BlockManager pickling
      FIX: legacy pickle format
      starting to make a mess of GroupBy but should be interesting
      starting on fast cythonized multi-groupby
      some progress
      ENH: got multi-groupby working in Cython. Lot more work needed
      TST: tests pass except for multi-column. renamed self.klass to DataFrame in test_frame.py
      ENH: refactoring of groupby. NA handling in new Cython methods
      ENH: got DataFrameGroupBy column selection working, clean up constructors
      ENH: some refactoring plus mystery bugfix in LongPanel.sort. need to investigate
      ENH: working on generic python multi-groupby generator
      ENH: Index can store tuples now
      groupby test module
      Merge branch 'multi-groupby' of github.com:wesm/pandas into multi-groupby
      ENH: groupby delegation wrapper working
      ENH: pure python multi-grouping working, UI needs work. as does cython integration
      ENH: perf opt in Index. lazy enforcement of uniqueness, prevent unnecessary calls to map_indices
      ENH: generic drop method for pandas objects
      ENH: multi-groupby prototype working with DataFrame. nice
      TST: type inference bugfix and test coverage
      DOC: release notes
      DOC: more thanks
      BUG: explicit dtype in otherwise all-NA DataFrame
      TST: raise exception when try to create empty int columns
      BUG: typo caused regression
      TST: for groupby regression bugfix
      ENH: integrated cython aggregators
      TST: multi-column groupby test passes now
      TST: test on single column too
      ENH: Series can be created without passing an explicit Index
      ENH: multi-group flat iteration works
      TST: do roundtrip in try: finally:
      TST: unit tests for groupby
      BUG: unnecessary line
      TST: unit testing for groupby
      FIX: numpy 1.4.0 bug in unit test
      BUF: missing parens in catching multiple exception types
      ENH: proper masking using counts in groupby. unit test and release notes
      ENH: multi-grouping using functions works too
      STY: small simplification
      BUG: apparently True == 1 and False == 0
      BUILD: changing to distutils setup instead of numpy.distutils setup
      ENH: DataFrame.dropna method and deprecated dropIncompleteRows and dropEmptyRows
      DOC: release notes about dropna
      BUG: calling select on empty DataFrame generates Exception
      DOC: illustrative example for DataFrame.pivot
      STY: switched everyone to use generic select function
      TST: unit test on indexing single column
      TST: failing unit test for mixed-type LongPanel
      ENH: working on multi-level index object
      ENH: test_panel passes, but a lot of other stuff is failing
      TST: unit test suite passes with refactored LongPanel
      ENH: added Index.format functions and improved datetime formatting when no time information
      ENH: removed LongPanelIndex in favor of the generic multilevel, patched up code and unit tests
      ENH: renamed MultiLevelIndex to MultiIndex. implemented slicing, cleanup
      ENH: sortlevel function, refactoring for hierarchical indexing
      ENH: sortlevel refactoring
      ENH: multilevel index in groupby. tests do not pass yet
      ENH: stack and unstack methods
      ENH: stack/unstack working. need to fix to_long_mixed test still
      ENH: mixed-type WidePanel can make roundtrip to LongPanel without losing proper mixed-typeness
      ENH: added delevel method, not certain about name
      BUG: need to mask out aggregated levels
      TST: failing unit test for groupby
      BUG: fixed groupby bug with missing labels
      ENH: multi-level index from_arrays method, reindexing-related methods, and Series changes so alignment will work. need to write a lot of tests
      ENH: start enabling proper fancy indexing with multi-level index
      ENH: indexing defer to _constructor
      ENH: example of rolling_apply_frame
      BUG: prevent NumPy upcasting float to string, regression from prior version
      TST: Index/MultiIndex testing and minor refactoring
      ENH: partial indexing with multi-level index, and various related tests
      TST: MultiIndex unit tests
      BUG: fix NumPy type upcasting to string, regression from 0.3.0
      BUG: index needs to be casted to ndarray in Series.plot
      TST: MultiIndex test coverage
      TST: unit tests coverage for MultiIndex
      BUG: fix GH #98
      ENH: sparse formatting of multi-level indexes, and will print out columns, too
      BUG: rewrote sparse formatting to fix buglet. created test_multilevel.py
      ENH: working on hierarchical __getitem__ in DataFrame
      ENH: hierarch xs working
      BUG: unstacking with more than 2 levels was not correct
      ENH: unstacking refactor, DataFrame can have a level expanded now, but needs work
      ENH: missing unstacked levels will be omitted
      ENH: handle FactorVector in rpy2 DataFrame conversion
      MSC: don't assert Index.ndim == 1 because matplotlib sometimes adds axes
      BUG: need to select unique groups from level when unstacking into regular index
      ENH: implemented DataFrame._count_level to replace LongPanel.count
      BUG: int block for intercept causing slow ols perf
      ENH: some NA handling when converting from rpy2 DataFrame
      ENH: drop levels when selecting higher level from Series or DataFrame. refactoring needed
      BUG: multi-level column formatting
      BUG: unicode handling in DataFrame.toString/info
      STY: removed random code
      ENH: AmbiguousIndexError not raised anymore, prefer label-based indexing, other multi-lev compat
      ENH: added a cache of DataFrame columns to improve perf with many calls to __getitem__
      ENH: some integration of multi-level indexing with .ix fancy indexing
      ENH: support picklability of Index subclasses
      TST: unit tests, and custom delete for Index objects
      ENH: Index.reindex refactoring and getting drop to work with MultiIndex. address GH #101
      ENH: enable group by level of MultiIndex
      TST: test coverage in DataFrame and _count_level
      BUG: error message in Series.__getitem__ with KeyError was made uninformative
      ENH: accept arbitrary iterables in fancy indexing
      ENH: switch to Cython agg for sum/mean for 1 group.
      TST: unit test for Cython group mean and added similar NA fix
      BUG: agg methods should return NaN when all NaN
      BUG: NAs were not being properly excluded in multi-key groupby
      BUG: tuples could not be inserted as new column keys in DataFrame
      ENH: perf enhancement in Index.insert
      ENH: multi-grouping with axis > 0
      ENH: string names of functions can be passed to groupby
      BUG: fixed mysterious major memory leak with method wrappers in groupby
      BUG: fix major memory leak in Series caused by caching view as array
      STY: refactored to prior state
      ENH: perf enhancement in Series.groupby
      ENH: groupby refactoring, testing, and perf opt. Series, DataFrame take functions
      ENH: refactoring, generic GroupBy.apply with inference
      BUG: tests pass now
      TST: added memory leak script
      ENH: added groups attribute back and implemented GroupBy.__len__, addressing GH #99 and GH #95
      ENH: application resulting in hierarchical index working in some cases
      ENH: axis=1 working mostly with hierarch groupby output
      BUG: cast to ndarray
      ENH: revert to old GroupBy.get_group for DataFrame due to perf
      ENH: more groupby tests, tweaking apply result output inference
      TST: completed groupby test (line) coverage
      TST: Index/MultiIndex tests, other misc coverage
      BUG: fix groupby regression with differently-indexed Series. test coverage
      ENH: proper item slicing in internals and tests
      ENH: copy option in reindex methods. cython fast_unique
      ENH: huge perf enhancement in homogenization speed
      ENH: more perf tweaking with homogenization
      ENH: don't copy in panel homogenization step
      ENH: use new fast_unique in Index.union
      ENH: no boundscheck / wraparound in map_indices
      ENH: WidePanel constructor can handle dict
      BUG: groupby aggregate regression
      ENH: Panel.pivot testing and performance enhancement
      TST: Series test coverage
      TST: sparse test coverage
      TST: misc test coverage. moved some code to sandbox
      TST: round out test coverage in pandas.core
      TST: HDFStore with compression test coverage
      BUG: address GH #106, and various ols unit tests
      TST: more ols test coverage
      TST: finished line coverage for ols.py
      TST: BAM! plm test coverage. comparisons with R still needed
      STY: ran pep8
      ENH: return tuple in multi-iteration
      DOC: Series docstring cleanup, reorg, deprecations
      DOC: DataFrame docs, sort unit tests, other cleanup. deprecated tgroupby
      DOC: DataFrame docstring edits and misc tweaks
      DOC: some Index docstring cleanup
      ENH: asof return nan so play nice with groupby
      TST: fixed index unit test
      ENH: Factor refactoring (heh) and docs
      DOC: index.py docstrings etc.
      DOC: panel docs
      DOC: documentation pass on groupby and rest of pandas.core
      BUG: removed __new__ tinker from last night
      DOC: docstring edits and misc test coverage
      ENH: implemented MultiIndex.insert function + tests
      DOC: removed var.rst, fewer warnings now
      TST: don't have time to figure out how to make pandas.test() work so removing it
      DOC: starting work on overhauling the sphinx docs. switched to agogo theme
      DOC: ipython directive from skipper's repo
      DOC: hack so pure_python only executed once for now
      ENH: convert multi-line pure python to continuation lines using the ast
      DOC: small amount of cleanup in ipython_directive
      BUG: Series from dict buglet
      DOC: skeleton of reorg'd docs and starting new tutorial
      DOC: reformatted release notes, deprecated DataFrame.tapply
      DOC: release notes reformat. note about hierarchical indexing
      DOC: release notes
      DOC: adding API file, filling things out
      DOC: more docs work on tutorial
      DOC: more docs, tutorial
      DOC: a bit more docs, Series valid->dropna
      DOC: more docs, org
      ENH: isnull/notnull return Series now. doc edits
      DOC: missing data docs, Series API
      DOC: more missing data docs
      DOC: more missing data docs
      DOC: chopping up old tutorial
      DOC: readme edits to be more like the docs, mostly finished missing data section
      DOC: more on basics
      ENH: more flexible renaming. corrwith bugfix
      DOC: more hacking on the basics
      DOC: even more work on tut. reindexing how-to getting there
      ENH: implemented sort for WidePanel
      DOC: yet more basics docs and reorg
      ENH: sort_index impl and consistency across classes
      DOC: finished basic sorting docs.
      BUG: rolling_cov bias correction did not take into account changing number of observations
      DOC: release notes
      DOC: release notes reformat
      DOC: release notes reformat
      DOC: minor readme reorg
      DOC: more reorg, deleted old series and dataframe rst files. starting on indexing docs
      DOC: boolean indexing with DataFrame
      DOC: fancy indexing -> advanced indexing to disambiguate
      DOCS: more indexing docs
      DOC: finished fancy indexing. starting on hierarch indexing
      DOC: more multiindex
      DOC: a bit more, and noticed a bug :(
      ENH: implemented lexsort_depth and is_lexsorted
      TST: more unit tests for lexsort_depth check
      DOC: mostly finished indexing chapter
      DOC: starting to rewrite the groupby docs. slowly
      DOC: making some progress on groupby docs
      DOC: first cut on new groupby docs done
      ENH: can do as_index=False in groupby. removed dummy 'index' column from DataFrame.to_csv output and added index_label option
      DOC: groupby doc about as_index=False. some joining/merging docs
      ENH: DataFrame.pivot now returns MultiIndex columns with multiple fields. Return view when indexing homogeneous hierarhical columns. various related unit tests
      ENH: indexing object with MultiIndex works if indexing the top level. address GH #120
      DOC: starting on pivot docs. some indexing additions given new functionality
      DOC: release notes. and rewrote _make_long_index
      REF: removed make_long_index method and switched things to directly create MultiIndex
      ENH: stacking starting with MultiIndex implemented, more unit testing needed
      ENH: swaplevel methods
      ENH: stack/unstack bug fixes and unit tests
      ENH: stack/unstack seem to be working. unit tests pass
      BUG: buglet in to_csv I introduced earlier today
      REF: moved stack/unstack/pivot code into new reshape module
      TST: test coverage
      DOC: reshape docs. man that is cool
      BUG: groupby bug with flex aggregate functions. and hierarchical columns weren't being passed on
      TST: test coverage
      DOC: note about swaplevels
      DOC: time series docs, timeRule->time_rule renaming
      REF: removed timeRule argument altogether
      DOC: more time series docs and release notes
      DOC: time series docs. bit of LongPanel refactoring
      REF: LongPanel is now no longer a 'Panel'. added level option to Series.count. fixed up OLS classes now that LongPanel works a bit differently. more testing / spot checking of results needed
      BUG: let downstream code convert list to ndarray, asarray not tuplesafe
      REF: renamed WidePanel -> Panel. legacy pickle support remains and a reference to WidePanel in the pandas namespace
      DOC: plotting function enhancements and plotting / stats docs
      ENH: more flexible boolean setting with DataFrame
      BUG: r2 needs to use transformed data
      DOC: stats docs
      BUG: change to _trans to fix unit test
      DOC: start IO docs
      ENH: enable storage of hierarchical index objects in HDFStore. addresses GH #128
      TST: ensure that MultIIndex names are stored correctly
      ENH: added add_prefix / add_suffix methods to DataFrame
      ENH: enable joining with overlap and lsuffix/rsuffix options
      BUG: self->this, need unit tests
      ENH: extensive indexing.py refactoring and implemented fancy/advanced indexing
      BUG: multi-key groupby now works with multiple functions. starting to add *args, **kwargs. address GH #133
      BUG: address groupby bug described in GH #131
      TST: unit testing for *args/**kwargs in groupby. address GH #132
      ENH: Panel.join. some cleanup, testing needed
      ENH: add omitna option to Series stat methods and fix up unit tests
      ENH: skipna in DataFrame stat functions
      ENH: skipna in Panel methods. need to unit test
      BUG: PanelOLS couldn't cope with object DataFrame input
      ENH: skipna and unit testing for Panel
      DOC: docs about skipna
      DOC: sparse docs. getting closer to release
      TST: DataFrame test coverage
      TST: Panel.join unit tests and misc panel test coverage
      TST: add_prefix/suffix to sparse. test coverage
      BUG: win64 buglet in pytables IO
      BUG: git code churn typo
      DOC: release notes, setup.py updating
      BUG: to_wide was returning incorrect results with inconsistent long panel index. added HDFStore handling for duplicates in table format
      DOC: doc warning fixes
      BUG: Python 2.5 compat in fillna
      BUG: commented out block of cython code causing compiler warning
      BLD: windows test batch files
      Merge branch 'master' of github.com:wesm/pandas
      BLD: include sandbox subpackage. set released to True
      Version 0.4.0
      BLD: edit MANIFEST.in to include docs
      RLS: Version 0.4.0, fixed version.py

-----------------------------------------------------------------------

No new revisions were added by this update.

-- 
Alioth's /usr/local/bin/git-commit-notice on /srv/git.debian.org/git/debian-science/packages/pandas.git



More information about the debian-science-commits mailing list