[pandas] annotated tag v0.7.0rc1 created (now e77d2af)
Andreas Tille
tille at debian.org
Wed Dec 28 15:20:26 UTC 2016
This is an automated email from the git hooks/post-receive script.
tille pushed a change to annotated tag v0.7.0rc1
in repository pandas.
at e77d2af (tag)
tagging 06f608182ecd516abcd140531c65e07b30293eb0 (commit)
replaces v0.6.1
tagged by Wes McKinney
on Thu Jan 19 23:34:18 2012 -0500
- Log -----------------------------------------------------------------
Version 0.7.0 Release Candidate 1
Adam Klein (63):
updated docs with whats new 0.4.1 to 0.4.3
started whats new 0.5.0 release notes
continued documenting 0.5.0 release
more 0.5.0 documentation
added links to github
continued on 0.6.0 docs
continuing v0.6.0 docs
more working on v0.6
more working on v0.6
more working on v0.6
more working on v0.6
added test coverage, fixed DateRange join bug
added more test coverage, fixed bug in from_records with index param is integer
add note to release doc
added test coverage, fixed DateRange join bug
added more test coverage, fixed bug in from_records with index param is integer
add note to release doc
added more test coverage
rebased master
Merge branch 'testcov' of github.com:adamklein/pandas into testcov
got test_frame coverage to 100%
got pandas.io.parsers to 100% testcov
pytables at 100% coverage
100% coverage of sparse array
100% sparse coverage
whats new -> 0.7
continuing 0.7 whats new
more on boxplot docs
added docs covering new features
fixed stupid errors
cleaned version references
setup.py tweaks
working on bunch of issues
fixed alignment and printing issues
added release notes, whatsnew doc
fixed docstring copying
PR#576 changes per wesm comments
PR#576 changes per wesm comments, cleaned
ENH: made Series empty constructor work
ENH: fix to another corner case of #395, added test
ENH: set default precision back to 4
FIX: comment out float padding behavior until we nail it
FIX: comment out test too
FIX: remove test instead
FIX, ENH: #395 got float formatting working
ENH: added column justify option, refactored global config into class
FIX: common.py cleanup, test coverage
ENH: enable unstack to create Series
DOC: added stack/unstack docstrings per #531
ENH: #395 incorporate lodagro comments on disambiguating precision vs accuracy
DOC: added justify docs
Merge branch 'docupdate'
FIX: #611 namedtuple breaks to_object_array_tuples
DOC: added select method
added docs for 0.7 release
ENH: ipython_directive to warning on exceptions when building docs
ENH: ipython_directive to warning on exceptions when building docs, small fix
ENH: ipython_directive to warning on exceptions when building docs, enhancement
ENH: ipython_directive to warning on exceptions when building docs, bug fis
ENH: docs build cleanly, per #635
FIX: multiindex error #636
FIX: multiindex, comma fix
BUG: fix install_requires dateutil >= 2
Bayle Shanks (1):
(1) add a get() method like dicts have, (2) throw KeyError, not Exception, when a column isn't found
Dieter Vandenbussche (2):
Add get_attr behavior for Panel, similar to DataFrame
Add support for MaskedArray in Series constructor, with test
Graham Taylor (1):
Make _get_terminal_size_linux() work inside Emacs
Gregg Lind (4):
One approach to requirements handling
removed requirements.txt
BLD: #517 friendlier setup message with no numpy
re #520, more setup.py additions
Thomas Wiecki (1):
Fixed typo in dsintro.rst.
Wes McKinney (271):
ENH: first cut at make_dummies function
ENH: improve column reindexing performance in DataFrame by using Cython take
BUG: accept empty list to DataFrame constructor, regression from 0.6.0, GH #491
BUG: handle datetime.datetime in nanmin/nanmax, on Python 2 at least, GH #487
ENH/BUG: add length option to Series.to_string, don't print length by default, GH #489
BUG: bounds check in Cython get_value_at method, close #495
BUG: out buffer for ndarray methods needs to be same dtype, GH #486
BUG: handle length-0 objects in Series.to_string, GH #488
BUG: pickle all other index types, re: #492
BUG: can print Series/DataFrame with ndarray cell elements, GH #490
ENH: refactored Cython groupby code to not have to sort, GH #93
ENH: add AC's khash.h to start tinkering with
ENH: Cython Grouper prototype class, per #496
ENH: multi-GroupBy refactoring to be less nested, reuse more code, getting toward addressing #496
ENH: store tuples in MultiIndex, add legacy pickle testing and other speed tweaks
BUG: DateRange / subclass handling in Index.join. Handle name field in Index.union, GH #500, #501
TST: add pickle file
ENH: speed up generic multi-key groupby via SeriesGrouper/Slider Cython classes. A little black magic, GH #496
ENH: reindexing performance regression for dtype=object
ENH: hacking around with khash, appears to be working
ENH: int64 hash table testing
ENH: more helpful error message in DataFrame constructor, GH #497
ENH: handle differently-indexed results in DataFrame.apply, GH #498
ENH: more speed tweaking Series.__getitem__
TST: vbench suite
ENH: can set multiple columns at once on DataFrame in __setitem__, per #342
ENH: updated khash.h with latest version, has MSVC compatibility, re #503
BLD: don't include stdint.h
BLD: more stdint stuff
TST: don't use Py_ssize_t in the buffer for now
TST: use list_to_object_array when we can
TST: another tuple fix for NumPy 1.6.0
TST: tweak to debug 1.6.0 issue
BUG: return NumPy scalars from Series, same speed as before, GH #510
DOC: add whatsnew skeleton section
DOC: add whatsnew stuff for 0.6.2, reorg vbench suite. release notes
BUG: handle empty list passed to DataFrame.from_records
BUG: Index.format bug excluding name field with timestamps
DOC: more vbench scaffolding, rst generation
BLD: include test data pickle file, GH #508
BUG: handle rows skipped at beginning of CSV file, GH #505
DOC: release notes
DOC: fixed section headings
ENH: do some extra type-checking in engines.pyx to avoid unnecessary dict creation
ENH: refactor sql.frame_query method per #516
DOC: vb_suite improvements / code generation for docs
BUG: don't cast to int/bool when introducing NAs
DOC: sparse array / list docs
DOC: fixed CSS! so sidebar floats right
DOC: more 0.6.1 docs
ENH: rename delevel -> reset_index, add back deprecation decorator
DOC: finish what's new for v0.6.1
BUG: fix unary - on boolean DataFrame
DOC: revert docs lost in rebasing, fix warnings, rst->txt
DOC: fix ref
ENH: can pass list of dicts to DataFrame constructor, support Cython code, #526
DOC: stray character
ENH: Cythonize AxisProperty to speed up DataFrame.index/columns properties etc., GH #528
BUG: improper handling of mis-matched columns in DataFrame.combine, GH #529
BUG: Index.insert / MultiIndex.insert discarded names, GH #527
TST: some hacking around #538
ENH: perf, compute group_index only once for cython aggregation
ENH: more tinkering with khash.h for strings and pyobjects
ENH: benching / testing with khash
BUG: fix boolean inference logic in DataFrame indexing, elsewhere
TST: boolean row selection vbench test, GH #541
BUG: raise Exception when calling Series.sort on an array view, GH #316
DOC: remove redundant swaplevel section per #533
BUG: stray x broke khash.h
DOC: release notes
ENH: add reorder_levels to Series and add unit testing per #534
TST: testing + docstring for DataFrame/Panel.get function, per #521
TST: vbenchmark for Series.fillna
BUG: use centered tss in r-square with entity or time effects in panel model
ENH: int64 tinkering with khash
BUG: use int64_t in factorize
REF: starting to refactor for removing LongPanel, GH #108
REF: more refactoring for LongPanel removal
TST: vb for Index.intersection/union
ENH: more refactoring for LongPanel removal, Index._join_level method, DataFrame.reindex_axis, #108
ENH: DataFrame.{iterrows, to_panel}, MultiIndex.{copy, has_duplicates}. More
TST: core and sparse unit tests pass
TST: some panel fixes. More robust label-based slicing for MultiIndex
ENH: fix Series slicing on MultiIndex logic, get unit test suite to not error without LongPanel, still 6 test failures
TST: plm fixes, test suite passes per #108
REF: to_long -> to_frame
REF: more cleanup of LongPanel stuff
REF: a little more cleanup
BUG: properly exclude values from MultiIndex with other join methods when joining on level
ENH: skiplist implementation in C, GH #374
BLD: C portability
BUG: return NA in all-NA case in Series.corr instead of raising exception, GH #548
ENH: use nanops in Panel aggregation methods, GH #536
ENH: speed up join / align ops in cases where axes are equal
ENH: level option to Series.align/reindex, arith methods, misc related refactoring
TST: verify fix for #525
ENH: default name assignment with regular index in reset_index, GH #476
DOC: docstring re: #476
BUG: fix aggregating multiple levels in DataFrame stat ops, use cythonized groupers when possible for > 15x speedup, GH #545
BUG: fix Series agg by level also and utilize Cython methods, GH #545
BUG: use Cython grouper for Series, bugfix for NA group handling
BUG: don't need to convert numeric types already handled by converters, GH #546
BUG: fix scalar setting via ix with MultiIndex, GH #551
BUG: flip left<->right in Index._join_level, GH #547
ENH: first cuts on many-to-many joining, #249, #267
ENH: merge ops taking shape, much testing needed, GH #249
ENH: left/right merge operations working and fairly fast, #249
ENH: inner merge working, testing
TST: refactoring and consolidating join /merge unit tests
ENH: merge/join functions, compress group index with possible number of groups is insanely large, speed enhancements, refactoring
TST: modify churn script
BUG: groupby level should preserve level order, perf enhancement. docs fix. moved join op code from internals to merge.py
ENH: test suite passes with unified merge operation!! #249
ENH: test suite passes, deleted old join code, #220, #249, #267 complete
ENH: use factorizer methodology to speedup many-to-use joins, 30% speedup
DOC: release notes
ENH: perf opt, don't take when indexes are equivalent, merge unit testing / bugfix
TST: test coverage and perf enhancements for merge, always compress group labels
Merge remote branch 'adamklein/testcov' into merge-562
BUG: numpy.integer type-checking in indexing, groupby test coverage
TST: complete groupby test coverage
TST: core/index test coverage
BUG: all NA case in Series.cov
BUG: non-float rank does not handle NAs properly, test coverage
TST: test convert_dummies, pandas/core test coverage complete
TST: vbenchmark for Panel.from_dict
TST: make work on older version of pandas
TST: insert newline
ENH: logy option to Seires.plot, release notes, bump dev version to 0.7.0
BUG: fix C skiplist memory leak, use in rolling median, GH #374
DOC: new merge docs, change default merge type to inner
BUG: inner join on columns with MultiIndex, further performance enhancement
ENH: merge benchmarks, use khash to factorize
DOC: change append section header
TST: more test coverage, merge benchmark tweaks
ENH: avoid data copying in reindex_items where possible
BUG: handle list + Series combo in DataFrame constructor, still raise exception when mixing list with dict, GH #564
DOC: remove v0.6.2.txt, renamed to 0.7.0
TST: roll median leak test file
ENH: starting on general multi-append/join stuff, refactor groupby code
BUG: bug in many-to-one join in left_join cython routine
TST: vb for left series join
DOC: release note
ENH: extensively refactor BlockJoinOperation to support n > 2, Concatenator class to orchestrate concatenations, #273, #479
ENH: add multiple-join to DataFrame with new concat function, test append multiple, #115, #479, #273
TST: more test coverage for merge
BUG: 32-bit compatibility
TST: more int32/int64 nonsense
TST: augment speed test for R merges
ENH: more merge benchmarks
TST: add sqlite3 benchmark
BUG: use Cython skiplist because C one has a segfault-y issue
TST: revert plm tinker so tests pass
ENH: refactor Concatenator to work for ndim > 2, add join-multiple to Panel, #115
TST: test coverage and concat bugfix
BUG: DateOffset.__eq__ between offsets with different keywords
DOC: release notes
BUG: handle zero-padding differently per comment in#578
DOC: release notes
BUG: test concatenation of panels on other axes and fix bug
TST: test coverage
BUG: catch np.vectorize returning str/unicode dtype, fix #583
ENH: type inference internals toward #328, update R merge benchmarks
BUG: omit parsing.pyx
ENH: raise AmbiguousIndexError when location not found in iteger-indexed Series. Start of #328 API changes
ENH: tinkering around with returning base instead of view, not working yet
ENH: also address Series setitem per #328
DOC: release notes
BLD: exclude float16 typemap for earlier versions of NumPy
BLD: float128/complex256 are NumPy >= 1.6 only
ENH: refactor to add hierarchical index creation in concat. Tests needed still
ENH: concat with keys for MultiIndex works. A bit more testing / docs required
BUG: address #590 and more concat tests with hierarchical index
ENH: add justify option for column header, per #571
ENH: add bold_rows option to DataFrame.to_html, GH #586
ENH: make eng formatter a bit prettier (add leading zero to positive numbers
BUG: work around matplotlib-related buglet, Index inserting newaxis issue
TST: hacking for talk, dict benchmarks
BLD: add back fake_pyrex
ENH: fix major performance issue in multi-groupby by compressing key space, likely some win32 issues still...
BLD: win32 missing reference in util.pxd, #594, #596
DOC: add docstring re: #477
DOC: cleanup unstack docstring per #588
ENH: can pass a dict of objects, #600
ENH: can pass Series to concat function, GH #599
ENH: can pass multiple objects to Series.append, #599
ENH: factorize + groupsort to massively improve multi-column sort_index performance. Add vbench test case, GH #555
TST: add vbench for drop_duplicates, speed up panel_ctors, GH #558
TST: add vbench for nested dict DataFrame ctor, GH #540
ENH: improve performance of DataFrame/Series contructors from nested dict, GH #540
BUG: formatting of index level names in MultiIndex, per #601
BLD: 32-bit platform integer issue
BLD: 32->64 bit cast for 32-bit
DOC: release notes and minor docstring fix
TST: use lexsort to get a stable sort for NumPy 1.4
BUG: bug with concat with specified levels and unequal concatenation axes
BUG: fix mixed-type block merge bug
DOC: add reasonably complete concat docs
ENH: grouped histogram hackage
ENH: groupby refactoring to use khash, add sort option, GH #595
ENH: use bottleneck for implemented nanops if installed, GH #91
BUG: handle non-numeric aggregates in pure python Series aggregation, GH #612
TST: test both with and without bottleneck, GH #91
ENH: can pass a list of functions to DataFrame.groupby per #166
ENH: can pass list of functions to pivot_table
ENH: take a crack at #614
BUG: hack around Series mixed-type formatting issue, GH #616
BUG: wrap getting bottleneck function in try/catch, GH #91
BUG: don't cast to tuple except in failure, only test with namedtuple when it exists (py >= 2.7), GH #611
ENH: implement DataFrame.lookup for label-based vector fancy indexing, GH #338
DOC: document ignore_index in concat in docstring and sphinx docs, GH #619
TST: unit test addressing multiindex concatenation buglet described in #615
ENH: add verbose option to read_csv/read_table to print number of NA values filled in non-numeric columns per comment on #614
BUG/ENH: fix performance regression in DataFrame constructor from nested dict with integer indexes, add vbench for it, speed up _stack_dict in internals, GH #621
ENH: start tightening up API around integer slicing per #592
ENH: wrap up API changes and testing around #592
ENH: API change / refactoring in Series.__getitem__ and __setitem__ to implement #86, related tightening of integer index handling from #592
ENH: test coverage and a bugfix in concat
BUG: fix #623 and refactor Series.to_string into format.py
DOC: concat docs, add some release notes about integer-indexing API changes
ENH: add ability to pass list of dicts to DataFrame.append (GH #464)
BUG: check for non-contiguous memory in SeriesGrouper, causing segfault
ENH: hack toward #629
TST: indexing testing with minor Series.__getitem__ refactoring
TST: testing, cleanup of io.sql module
TST: add test_sql.py module
BUG: fix buglet with xs with level, GH #371
ENH: crosstab prototype function, API needs fleshing out, GH #170
ENH: add crosstab function and test
ENH: add margins to crosstab
BUG: don't lose index names when adding row margin
BUG: convert tuples in concat to MultiIndex
BUG: fix TextParser with list buglet, enable parsing of DataFrame output with index names
BUG: fix bugs related to comments in #371
ENH: xs level can take multiple levels, pass multiple levels to MultiIndex.droplevel, GH #371
DOC: release notes
BUG: DataFrame.join on keys produce wrong result, does not preserve order
TST: cast to i8 when checking margins
TST: Python 2.5 float formatting changed
TST: more Python 2.5 sadness
TST: reset printoptions in case of prior test failure on Python 2.5
TST: NumPy 1.5.1 workaround, GH #353
TST: hack more cleverly around NumPy 1.5.1 bug
TST: use flat iterator
BUG: fix multi-iter bug and pure Python aggregation given compressed ids re: GH #618
BUG: passing a string to aggfunc like 'mean' won't break _add_margins, GH #618
BUG: disable label-based slicing semantics with integer indexes in Series.__getitem__ and __setitem__, GH #641
BUG: BlockManager must have __nonzero__ -> True for pickling of empty DataFrame
ENH: add keys method to HDFStore
DOC: release notes
TST: test to verify fix of #638
DOC: release notes, what's new in 0.7.0
ENH: can pass values and custom aggregation function to crosstab
BLD: python 3 compatibility for now, GH #648
BLD: use AsEncodedString more properly
BLD: python 3 compat
ENH: add integer indexing functions iget/irow/icol, GH #628
ENH: test coverage, made pivot_table work with no rows passed and margins=True
DOC: start some docs re: integer API changes, GH #627
ENH: support for reading legacy (pre 0.4) WidePanel stored in HDFStore
TST: copy matplotlib subplots function for compat with mpl < 1.0
ENH: cleanup post cherry-pick
DOC: 0.7.0 docs, add iget_value alias and DataFrame.iget_value, GH #627
BUG: be less aggressive with type conversion in lib.map_infer
BUG: NA handling when sorting indexers
ENH: enable binary and logical Series operator to do data alignment
DOC: release notes, add io.py vbench module
ENH: don't force the conversion of Decimal objects in groupby either, skip spearmanr test in scipy < 0.8
TST: python 2.5/6 issues with Decimal
ENH: don't try to autofmt_xdate if sharex=False in DataFrame.plot
RLS: 0.7.0 Release Candidate 1
Wouter Overmeire (10):
Doc: indexing correct swaplevels --> swaplevel.
Fix keyerror in ipython session demonstrating console output formatting.
adding DataFrame.reorder_levels().
New GlobalPrintConfig should also be used when setting engineering float formatter.
Use new set_eng_float_format signature, precision is replaced by accuracy.
DOC: Clean up generated temp file foo.csv
DOC: Create temp files to demo read_csv() and read_table(), when they are needed and remove afterwards.
DOC: Since decision between summary and to_string() is based on terminal size, make sure when demo summary, that is what will be shown.
ENH: make eng formatter a bit prettier for all accuracy values
TST: Add tests for core.common.EngFormatter.
Yaroslav Halchenko (4):
DOC: added a note for behavior of append
ENH: pass figsize into _grouped_plot functions
ENH: pass **kwds into plots.boxplot from DataFrame.plot
BF: special handling for _radd_compat for any numpy under 1.6
theandygross (1):
closing tag of 'tbody' was 'body'
-----------------------------------------------------------------------
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