How does date plotting work? about bokeh

[pzwang]

I think the fundamental problem is that pandas.date_range gives you back a DatetimeIndex object. I'm trying to see now how best to convert it to something that we can serialize and plot...

[pzwang]

An update: the following code works better and displays a plot, but the grid lines and axes are still off and not rendering right.

from bokeh import mpl
import numpy as np
import pandas as pd
p = mpl.PlotClient()
p.notebooksources()
pd_dates = pd.date_range('2012-01-02', '2012-01-05', freq='H')
dates = pd_dates.astype(int) / 1000000
values = np.arange(len(dates))
plot = p.plot(x=dates, y=values, color='orange', width=700, is_x_date=True, )
plot.notebook()

(Bokeh currently doesn't know how to serialize a Pandas DatetimeIndex object, so we have to convert to a Numpy integer array.)

[hhuuggoo]

I think the grid lines issues are fixed so I'm closing this out we should do a better job of handling datetimes on the python side, maybe auto convert to millis?