A curses-based DataFrame viewer inspired by TabView.
Project description
#######
dabbiew
#######
A curses-based DataFrame viewer inspired by TabView
.. image:: doc/images/dabbiew.gif
:scale: 50 %
:alt: alternate text
:align: center
*****
About
*****
This is a side project for now. I work on it because I want more green squares
on my GitHub profile.
The main difference between TabView and DabBiew is that underlying data
structure is a pandas DataFrame instead of a list of lists. This has the
advantage of potentially being able to take advantage of Dask, which supports
"Big Data" collections for distributed environments.
************
Installation
************
::
pip install dabbiew
*****
Usage
*****
Open any csv file::
dabbiew.py file.csv
************
Key Bindings
************
A move command can be repeated by typing the number of times to repeat before
issuing an action. For example, to move down 12 times, simply type ``12j`` (or
``12↓``). To perform a search, open the search bar with ``\``, enter a
substring to match, and hit return (``↵``).
================================================= ==================================
Key Action
================================================= ==================================
``v`` toggle selection mode
``esc`` cancel selection
``h`` ``j`` ``k`` ``l`` ``←`` ``↓`` ``↑`` ``→`` movement keys
``gg``, ``GG`` jump to top, bottom of DataFrame
``^``, ``$`` jump to left, right of DataFrame
``,``, ``.`` decrease, increase selection width
``<``, ``>`` decrease, increase all widths
``t``, ``y`` toggle header, index
``[``, ``]`` decrease, increase index width
``:`` toggle command mode
``/`` toggle search bar
``n``, ``p`` next, previous match
``d`` enter ipdb debug mode
``q`` quit
================================================= ==================================
************
Command Mode
************
Entering command mode (``:``) allows the user to call *any DataFrame method
which returns a Series or DataFrame* on the current selection. For instance, the
user can call ``:sum()``, ``:where(df==42)``, or even ``:where(df==41).sum()``
on a selection. The resulting Series or DataFrame is rendered on screen. To go
back to the previous view, simply quit (``q``). Note the name of the current
DataFrame is always called ``df``.
*************
Documentation
*************
To generate the source code documentation do::
cd doc
make html
and open ``_build/html/index.html``
dabbiew
#######
A curses-based DataFrame viewer inspired by TabView
.. image:: doc/images/dabbiew.gif
:scale: 50 %
:alt: alternate text
:align: center
*****
About
*****
This is a side project for now. I work on it because I want more green squares
on my GitHub profile.
The main difference between TabView and DabBiew is that underlying data
structure is a pandas DataFrame instead of a list of lists. This has the
advantage of potentially being able to take advantage of Dask, which supports
"Big Data" collections for distributed environments.
************
Installation
************
::
pip install dabbiew
*****
Usage
*****
Open any csv file::
dabbiew.py file.csv
************
Key Bindings
************
A move command can be repeated by typing the number of times to repeat before
issuing an action. For example, to move down 12 times, simply type ``12j`` (or
``12↓``). To perform a search, open the search bar with ``\``, enter a
substring to match, and hit return (``↵``).
================================================= ==================================
Key Action
================================================= ==================================
``v`` toggle selection mode
``esc`` cancel selection
``h`` ``j`` ``k`` ``l`` ``←`` ``↓`` ``↑`` ``→`` movement keys
``gg``, ``GG`` jump to top, bottom of DataFrame
``^``, ``$`` jump to left, right of DataFrame
``,``, ``.`` decrease, increase selection width
``<``, ``>`` decrease, increase all widths
``t``, ``y`` toggle header, index
``[``, ``]`` decrease, increase index width
``:`` toggle command mode
``/`` toggle search bar
``n``, ``p`` next, previous match
``d`` enter ipdb debug mode
``q`` quit
================================================= ==================================
************
Command Mode
************
Entering command mode (``:``) allows the user to call *any DataFrame method
which returns a Series or DataFrame* on the current selection. For instance, the
user can call ``:sum()``, ``:where(df==42)``, or even ``:where(df==41).sum()``
on a selection. The resulting Series or DataFrame is rendered on screen. To go
back to the previous view, simply quit (``q``). Note the name of the current
DataFrame is always called ``df``.
*************
Documentation
*************
To generate the source code documentation do::
cd doc
make html
and open ``_build/html/index.html``