WIP: Data frame method for epochs

[dengemann]

Related to #133 here is a first proposal of how to interface with pandas data frames.
One important questions certainly is whether to extend support to raw and source.

What do you think?

[agramfort]

woooottt ! that's pretty nice.

I've sent you a PR with a little clean up

if you want to pimp it up a bit more before we merge it would be great.

[dengemann]

:-) great you like it. actually i did it directly after the biomag but i wasn't sure whether to go for it..
also consider my last post on #133

On 07.10.2012, at 22:12, Alexandre Gramfort notifications@github.com wrote:

woooottt ! that's pretty nice.

I've sent you a PR with a little clean up

if you want to pimp it up a bit more before we merge it would be great.

—
Reply to this email directly or view it on GitHub.

[agramfort]

:-) great you like it. actually i did it directly after the biomag but i wasn't sure whether to go for it..
also consider my last post on #133

I am not scared about merging such code as it's not intrusive and most
of the code is in an example.

It could make a few figures with pandas kung fu before we merge it
would be great. I am thinking also about an exported as
nitime.TimeSeries

[dengemann]

On 07.10.2012, at 22:18, Alexandre Gramfort notifications@github.com wrote:

:-) great you like it. actually i did it directly after the biomag but i wasn't sure whether to go for it..
also consider my last post on #133

I am not scared about merging such code as it's not intrusive and most
of the code is in an example.

:-)

It could make a few figures with pandas kung fu before we merge it
would be great.

the example already does some figures. its just that they aren't that nicely formatted (all the stuff done in plot_evoked, for example).
i will look into making them nicer with a few commands.
do you think we should confine ourselves to a few but nicer examples? if so, do you have preferences for certain examples?

I am thinking also about an exported as
nitime.TimeSeries
yes, we were discussing that.
i could look into this in the course of the next days. should be quite straight-forward, as i already experimented with that.
—
Reply to this email directly or view it on GitHub.

[agramfort]

the example already does some figures. its just that they aren't that
nicely formatted (all the stuff done in plot_evoked, for example).
i will look into making them nicer with a few commands.
do you think we should confine ourselves to a few but nicer examples? if
so, do you have preferences for certain examples?

you're the one to judge.

an example should produce one or a few nice figures, illustrate a common
usecase, be not too long (otherwise it's more a full tutorial) and
well documenter
in the top docstring and in the code.

i could look into this in the course of the next days. should be quite
straight-forward, as i already experimented with that.

cool !

[dengemann]

On 08.10.2012, at 14:26, Alexandre Gramfort notifications@github.com wrote:

the example already does some figures. its just that they aren't that
nicely formatted (all the stuff done in plot_evoked, for example).
i will look into making them nicer with a few commands.
do you think we should confine ourselves to a few but nicer examples? if
so, do you have preferences for certain examples?

you're the one to judge.

an example should produce one or a few nice figures, illustrate a common
usecase, be not too long (otherwise it's more a full tutorial) and
well documenter
in the top docstring and in the code.

have a look at my last push, wdyt?
for now i would be happy with that.
more meaningful examples could be generated as #133 proceeds, e.g. creating dfs with epoch x condition x tsl indices.

i could look into this in the course of the next days. should be quite
straight-forward, as i already experimented with that.

cool !

:-)

Reply to this email directly or view it on GitHub.

[agramfort]

example is broken due to renaming :-/

please add a description in the top docstring of what's happening
and show some basics pandas commands (show columns, lines names, shapes
etc...)

[agramfort]

merged by rebase after one more fix :)

how do you manage to have broken examples? do you still work with python setup.py install?

if so consider using : "pip install -e --user" from mne-python folder

[dengemann]

nope, just excessive ipython sessions...
a name remained somewhere bound to some value and then the example still is working. i should more often reset / use run. bad habits from R times :-)

On 08.10.2012, at 23:35, Alexandre Gramfort notifications@github.com wrote:

merged by rebase after one more fix :)

how do you manage to have broken examples? do you still work with python setup.py install?

if so consider using : "pip install -e --user" from mne-python folder

—
Reply to this email directly or view it on GitHub.