You can do this by setting a the color/prop cycle to use colors which have an alpha component.

The traitlet overhaul will take care of this as well

Can i set the "axes.color_cycle" values in hex natation as e.g.:
"#ff0000b4"
Because when i try to do so i receive an exception of "invalid hex color string"

take out the pound sign because our rcfile reader is really naive and
always treats that as a comment symbol.

Also note that color_cycle is deprecated in the upcoming 1.5 release,
replaced with the more versatile "prop_cycle" mechanism. However, it will
likely be awhile before we actually remove "color_cycle".

On Thu, Sep 24, 2015 at 9:30 AM, Jul3k notifications@github.com wrote:

Can i set the "axes.color_cycle" values in hex natation as e.g.:
"#ff0000b4"
Because when i try to do so i receive an exception of "invalid hex color
string"

—
Reply to this email directly or view it on GitHub
#5132 (comment)
.

Without the pound sign i get a "to_rgb: Invalid rgb arg "ff0000b4"" error.
The exact call is:

plt.rcParams['axes.color_cycle'][0] = "ff0000b4"

matplotlib version 1.4.3

Without the alpha content ("b4") it works fine. It seams to me that the parser doese not accept 4 Byte hex color codes with alpha content. I could work around this by using the to_rgba builtin method

My workaround looks as follows:

to_rgba = matplotlib.colors.ColorConverter().to_rgba
for i, c in enumerate(plt.rcParams['axes.color_cycle']):
    plt.rcParams['axes.color_cycle'][i] = to_rgba(c,0.2)

Ah, indeed, our hex color spec parser is just rgb, not rgb[a]. Sounds like
an easy feature to implement, though...

On Thu, Sep 24, 2015 at 10:21 AM, Jul3k notifications@github.com wrote:

Without the pound sign i get a "to_rgb: Invalid rgb arg "ff0000b4"" error.
The exact call is:
plt.rcParams['axes.color_cycle'][0] = "ff0000b4"

matplotlib version 1.4.3

Without the alpha content ("b4") it works fine. It seams to me that the
parser doese not accept 4 Byte hex color codes with alpha content. I could
work around this by using the to_rgba builtin method

My working code looks as follows:
to_rgba = matplotlib.colors.ColorConverter().to_rgba
for i, c in enumerate(plt.rcParams['axes.color_cycle']):
plt.rcParams['axes.color_cycle'][i] = to_rgba(c,0.2)

—
Reply to this email directly or view it on GitHub
#5132 (comment)
.

Pretty sure rcparams will take a 4-tuple an input

Yes rcparams takes a 4-tuple as an input. As @WeatherGod stated the problem was related to the fact that the color parser only checks for 3-Byte (#FFFFFF) color strings and does not understand 4-Byte (#FFFFFFAA) color strings with alpha channel. So the way i presented the color information was not suiteable to the parser

@Jul3k Can you make a new issue to upgrade out hex parser to deal with RGBA hex codes?

I opend a new issue relating to the color spec parser

I think this is closed by #6382, which updated the parser to allow 4-byte hex strings. Re-open this issue if that isn't sufficient.

I frequently need alpha independently of the color cycler. Most importantly I want different plots to have different default alphas, f.i. alpha=.5 for my scatter plots, while not using alpha for my line plots. Thus, as far as I can see, setting alpha to the color cycle is not working. Any change this is going to be implemented in the rcParams?

This is too specialized for rcParams, in particular if you want to have different values for scatter plots and line plots.