- Fixes an issue with column autosizing leading to a division by zero on some systems.
- Fixes interactive support within IPython and Jupyter notebooks.
- All iterables with a known length (such as sets) can now be visualized as regular columns without an explicit conversion.
- Sets and dictionary keys are now displayed in native order by default.
A new
sort
keyword has been added to enforce ordering again. - Fixes column autosizing error with Python 3.7+
- A crash that could occur during shutdown with a running detached view and Qt5 has been fixed.
- The
gtabview
utility is now installed viaconsole_scripts
, fixing usage on Windows platforms. - Removed support for Blaze, due to broken/unmaintained upstream.
- Version information is now available as a command line via
--version
and in the module itselfgtabview.__version__
.
- Added support for PyQt5
- Sequences of bytes/strings are now correcly shown as a single column.
- More irregular/malformed data structures are now supported.
- Any missing value supported by NumPy/Pandas (such as NaT) is now displayed as an empty cell for consistency.
- Column-autosizing performance tweaks.
- Fix exception with PySide.
- Fix segmentation fault in repeated calls to view() in stand-alone programs.
- Improved column auto-sizing.
- NaNs (as None) are now also displayed as empty cells.
- Empty structures no longer cause view() to fail.
- Correctly pass user-supplied encoding to Blaze.
- Fix color palette selection on Windows.
- Added support for Blaze (http://blaze.pydata.org)
- Improved documentation.
- Headers and indexes can now be resized.
- Level names in
pd.DataFrame
objects are now shown. - Negative
start_pos
offsets are now allowed to conveniently position the cursor counting from the end of the dataset.
- Fix visualization of
np.matrix
types. - Allow data transposition.