rootpy: 0.7.0

Le résumé fourni par la source

What's new in this release: bugfixes Fix #49. Don't require python>=2.6 for install, otherwise we get python3 Check for sys.version < (2, 6) in setup.py Eliminate ipython printing (Bool_t)1 when ROOT starts up Emit an error wherever the diaper pattern is used Sanitize ROOT's message output to ascii, using repr() if it isn't Register HistStack with class registry correctly general improvements Documentation Defer configuring defaults until finalSetup (#74) Switch to pyparsing for C++ template parsing Use XDG base directory specificiation Add defaults.extra_initializations for further ROOT setup Add userdata.BINARY_PATH, unique per (arch, rootversion) Use .cache/rootpy/{arch}-{rootversion}/{modules,dicts} to store generated binaries defaults: Disable AutoDict generation Introduce rootpy/util/quickroot which can get to vital root symbols without finalSetup (10x speed improvement in import time) Implement TPyDispatcherProcessedEvent using rootpy.compiled Falling back to batch mode should be a WARNING not ERROR Numpy code is now moved to the rootpy/root_numpy repository Use hash of unique name rather than generating a UUID for dictionary shared object name Automatic testing of each commit / pull request on travis-ci PEP8 Python code style compliance Add TitleView class to plotting.views functionality Unit tests Rewrite registry code so magical imports are no longer required to make rootpy aware of what ROOT classes are subclassed in rootpy Alias ROOT CamelCase methods with snake_case methods Improvements to the setup.py script New examples for using Trees, Hists, etc Make name and title properties for Objects new externs Add extern/pyparsing.py Add extern/byteplay Add module_facade Add inject_closure_values to close over global variables docs Add logger documentation Sphinx: Introduce ipython highlighting Document a bit more about python's vanilla logging Update STL docs Add note in CONTRIBUTING.rst about debug python builds scripts Use argparse (argparse is now in rootpy.extern) Unify scripts into one rootpy script Remove scripts that are still WIP Use the new rootpy logger Your roosh command history is saved and can be searched Multi-canvas and multi-file support in roosh Improvements to the root2hdf5 script, that now uses root_numpy for faster conversion to HDF5 rootpy.interactive (NEW!) Add interactive.rootwait, provides wait_for_zero_canvases which blocks program execution until all canvases are closed Add canvas_events for closing canvases on middle click rootpy.context (NEW!) Add preserve_{current_canvas,batch_state} and invisible_canvas context managers Add preserve_current_directory context manager rootpy.plotting Add xaxis/yaxis properties to plotting.Efficiency Implement new fill_array function in root_numpy to fill histograms with NumPy arrays. Histograms in rootpy now have a fill_array method TStyle is now subclasses in rootpy as Style. Styles can be used at context managers using the "with" statement Add the official ATLAS style and example Hist2Ds can be "ravel()"ed, like NumPy's ravel() method that converts a 2D array into a 1D array by repeating the second axis along the first Add SetColor to all Plottables to set all colors simultaneously Improvements to the root2matplotlib module Make style attributes properties Add xaxis, yaxis, zaxis properties for Hist classes Plottable now gracefully handles deprecation of properties rootpy.logger (NEW!) New logging module for internal and (optionally) external use Automatically coloured status level if we're attached to a terminal Default rootpy logging level to INFO unless os.environ['DEBUG'] is present Redirecting ROOT's errors as python exceptions Add a log symbol to each subpackage Add @log.trace() decorator which can be used to log function entry/exit/duration Add log.showstack() to log current stack whenever a message is emitted by log If there is no logging handler when the first message is emitted, automatically add one Automatically log python stack trace if there is a segfault rootpy.compiled (NEW!) Adds an interface for compiling C++ code on demand with CompileMacro Add support for inline C++ code definitions rootpy.compiled.register_file("mycode.cxx", ["mysymbol"]) then rootpy.compiled.mysymbol will be generated when requested. mycode.cxx is located relative to the module where register_file is called rootpy.compiled.register_code(".. C++ source code ..", ["mysymbol"]) .. same as above but without needing a cxx file on disk rootpy.memory (NEW!) Add log.showdeletion() to show TObject cleanup keepalive: function to ensure objects are kept alive exactly as long as needed rootpy.util.hook (NEW!) Add ability to hook ROOT methods and append properties/methods to existing classes rootpy.util.cpp (NEW!) New module for parsing C++ syntax. rootpy.stl (NEW!) Compiling arbitrary templated types. rootpy.tree Make root_numpy an optional dependency and remove copy from rootpy Reading branches on demand: Only call GetEntry on branches that are accessed Add support for branches of array types Improvements to the documentation The categories module is rewritten and now properly parses the category syntax (unit tests are included) Improvements to Tree.Draw Dictionaries for branches of STL types are now compiled automatically Reverse order of axes expressions in Tree.Draw to match hist argument order so X:Y:Z maps onto filling along X:Y:Z instead of Z:Y:X

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.