Mailman 3 April 2017 - Matplotlib-users

plotnine v0.1.0, a grammar of graphics for Python
by Hassan Kibirige April 25, 2017

April 25, 2017

Matplotlib users, I am happy to announce plotnine -- a grammar of graphics package for python. plotnine is a fork and rewrite of "ggplot for python" (now known as ggpy). The grammar implementation is a detailed adaptation of ggplot2 and the plotting backend is all Matplotlib. Although this is the first release, it should be pleasant to work with and where it falls short, I apologise. You can checkout the code at https://github.com/has2k1/plotnine, the documentation at http://plotnine.readthedocs.io/en/stable/ and if you encounter any issues please report them at https://github.com/has2k1/plotnine/issues. Hassan

1 0

Line mode of RectangleSelector broken or no longer intuitive
by Lars Bittrich April 13, 2017

April 13, 2017

Dear all, I recently upgraded to matplotlib 2.0 and found a regression with RectangleSelector with drawtype='line'. A line is no longer draw between start point and end point, but from lower left to upper right of an imaginary rectangle. So you can no longer select a downward slope. This behavior may have been around for some time and most users will just use RectangleSelector with a rectangle and not a line. There has been a mail on this list regarding the return values only: http://matplotlib.1069221.n5.nabble.com/widgets-RectangleSelector-return-va… But also the interactive drawing is affected. I looked into the corresponding code and found quite a lot of sorting of x and y values separately. To restore the old behavior one would need many changes to the worse from my point of view. Please note, that the docstring for RectangleSelector still advertises the ‘line’-mode even if the ‘box’ mode is default now. This example clearly illustrates the problem. However, I suggest and request for people like me, who want so select a line in a plot to add a LineSelector instead of changing RectangleSelector to the old behavior and maybe break some code for the ‘box’-mode again. This would be much more intuitive for new users. I added such a LineSelector and attached a patch for widgets.py based on a much reduced copy of RectangleSelector. There still is some code duplicated in the _release method, which could be resolved with a common base class, but would require more changes. Additionally, there are some changes in the docstring of RectangleSelector to use the box-mode and change the print statement to clearly state that start point and end point are no longer returned. In future one could add a deprecation warning and point users to LineSelector if the RectangleSelector is used in line-mode. Best regards, Lars

2 1

Testing of interactive functions
by Juan Nunez-Iglesias April 13, 2017

April 13, 2017

Hello, We just got an exciting PR in scikit-image to do manual image segmentation with matplotlib: https://github.com/scikit-image/scikit-image/pull/2584 The manual_lasso_segmentation function in particular is astoundingly simple — a testament to the ease of use of the mpl interactive bindings. Thanks! Because it relies on interactive clicking and click-and-dragging after a blocking call to plt.show(block=True), though, it’s tricky if not impossible to test. I’m wondering whether mpl has a testing framework for the interactive components? Juan. PS: Also, the layout of the undo button is a bit tricky… If you have any tips for that, we are all ears! =)

3 2

Why different figure size for screen/savefig even dpi is the same?
by Shiyao Ma April 8, 2017

April 8, 2017

Hi, This, http://imgur.com/a/q9UvR contains two pictures. One is drawn by the plt.show(), the other is by plt.savefig and then opened by the pdf reader. All the settings are the same, including the dpi. ( I am using matplotlib2) Why does the figure size differ ? The one shown in the pdf reader seems normal, however, the one in the default plt.show() is a little bit small. I am on macos with retina, if that makes a difference. Best.

1 0

width of bar plot when using pandas DatetimeIndex
by Pierre Haessig April 8, 2017

April 8, 2017

Hi, I wonder if there is a documented behavior when making a barplot with datetimes (more precisely pandas DatetimeIndex) as the x values. I've pasted an example below with matplotlib 2.0 and pandas 0.19.2. In this experiment, it seems that the width of the bars is interpreted as a number of days, so if I want to plot monthly data without blanks, I have to put width = 31. So my question is: /is it a reliable behavior that the bar width is a number of days when working with a DatetimeIndex? / Notice that it is not coherent with pandas behavior: when addiding +1 on a monthly datetime index it adds on month. Also, with the default width value (0.8), some of the bars are not visible and I don't understand why, but maybe it's a separate issue. best, Pierre The example: a = pd.date_range('2015', '2016',freq='M') ts = pd.Series([3.,2]*6, a) plt.bar(a, ts); plt.bar(a, ts, width=31);

1 0

Subplots problem: Last tick-formatting not working
by lorny April 6, 2017

April 6, 2017

Hi, I'm new here and also quite a python/matplotlib-noob (as you will be able to recognize in my code-example). I have a problem with *plt.subplots*: When I plot any amount of bars (or different graphs/diags) in a for-loop through the subplots, and I try to format the yaxis-ticklabels (using the ticker module), the ticklabels are adapting all the subplots *except the last subplot yaxis* Anyone any ideas or hints? Thanks a lot for any support! Best regards Here is the code: import matplotlib.pyplot as plt import matplotlib.ticker as ticker import numpy as np from sympy import * from random import randint init_printing(use_unicode=True) np.seterr(all='ignore') def graph(numsubplots): fig1, axarr = plt.subplots(numsubplots, ncols=1, num="Subplot Problem", figsize=(10, 15), dpi=100) fontsize = 16 for i in range(0, len(axarr)): ax = axarr[i] n = 5 indx = np.arange(1, n+1) width = randint(1, 5)/10 ax.barh(indx - width/2, [4, 7, 2, 1, 9], width, color='y', label="Me") ax.barh(indx + width/2, [3, 3, 0, 3, 3], width, color='r', label="The other") ax.set_xlabel('Hours', fontsize=fontsize) ax.set_ylabel('Day', fontsize=fontsize) ticklabelformat(ax, which='y', step=1, formating='day %d') ax.set_title('Hours working a day', fontsize=fontsize) ax.legend(fontsize=fontsize) ax.grid(True) for i in fig1.axes: plt.setp(i.get_xticklabels(), visible=True, rotation=30) plt.yscale('linear') plt.tight_layout() fig1.subplots_adjust(hspace=0.3) plt.show() def ticklabelformat(axe, which='xy', step=1, formating='%d'): majorlocator = ticker.MultipleLocator(step) majorformatter = ticker.FormatStrFormatter(formating) if which == 'y': axe.yaxis.set_major_locator(majorlocator) axe.yaxis.set_major_formatter(majorformatter) elif which == 'x': axe.xaxis.set_major_locator(majorlocator) axe.xaxis.set_major_formatter(majorformatter) else: axe.yaxis.set_major_locator(majorlocator) axe.yaxis.set_major_formatter(majorformatter) axe.xaxis.set_major_locator(majorlocator) axe.xaxis.set_major_formatter(majorformatter) if __name__ == "__main__": graph(3) -- View this message in context: http://matplotlib.1069221.n5.nabble.com/Subplots-problem-Last-tick-formatti… Sent from the matplotlib - users mailing list archive at Nabble.com.

2 2

Arbitrary artist data on SVG elements
by Joshua Klein April 3, 2017

April 3, 2017

Hello, I often embed figures as SVG graphics in web pages. As part of this process, I usually do the following 1. Set gids on artists and link that gid to a set of data describing that part of a graphic in an external dictionary. This includes things like setting the element’s class, extra contextual information, information that would be good to show in a tooltip, ids of related elements, and so forth. 2. Serialize the figure into a file-like object, use an element tree implementation’s XMLID to get an element id map and Element objects 3. Iterate over my data dictionary from (1) and set keys in the mapped Element’s attrib dictionary, using the id map from (2) 4. Use the element tree implementation’s tostring function to serialize the updated Element objects back into a string and then send the string out as a response to a web request. 5. After receiving the SVG string from the server on the client, add the SVG to the page’s DOM and then hang event handlers on it (or pre-specify delegated handlers) that use the added attributes to configure interactive behavior. I looked at the Artist type and saw no good place to store “arbitrary data”. Before I start working on this I wanted to know if anyone else had a better solution. I would also like to know if the devs would be opposed to a PR that adds an extra dictionary/attribute to every Artist instance created. Another alternative solution would be to find a way to push my dictionary mapping gids to extra attributes into the SVGRenderer and have it pass them as **extras to XMLWriter.element when it processes individual artists. Here’s a generic example of what I do currently: def plot_with_extras_for_svg(*data, **kwargs): # Do the plotting, generating the id-linked data in `id_mapper` ax, id_mapper = plot_my_data(*data, **kwargs) xlim = ax.get_xlim() ylim = ax.get_ylim() # compute the total space used in both dimensions when dealing with # negative axis bounds x_size = sum(map(abs, xlim)) y_size = sum(map(abs, ylim)) # Map the used axis space to the drawable region dimensions aspect_ratio = x_size / y_size canvas_x = 8. canvas_y = canvas_x / aspect_ratio # Configure the artist to draw within the new drawable region bounds fig = ax.get_figure() fig.tight_layout(pad=0.2) fig.patch.set_visible(False) fig.set_figwidth(canvas_x) fig.set_figheight(canvas_y) ax.patch.set_visible(False) # Perform the first serialization buff = StringIO() fig.savefig(buff, format='svg') # Parse XML buffer from `buff` and configure tag attributes root, ids = ET.XMLID(buff.getvalue()) root.attrib['class'] = 'plot-class-svg' for id, attributes in id_mapper.items(): element = ids[id] element.attrib.update({("data-" + k): str(v) for k, v in attributes.items()}) element.attrib['class'] = id.rsplit('-')[0] # More drawable space shenanigans min_x, min_y, max_x, max_y = map(int, root.attrib["viewBox"].split(" ")) min_x += 100 max_x += 200 view_box = ' '.join(map(str, (min_x, min_y, max_x, max_y))) root.attrib["viewBox"] = view_box width = float(root.attrib["width"][:-2]) * 1.75 root.attrib["width"] = "100%" height = width / (aspect_ratio) root.attrib["height"] = "%dpt" % (height * 1.2) root.attrib["preserveAspectRatio"] = "xMinYMin meet" # Second serialization svg = ET.tostring(root) plt.close(fig) return svg Thank you

2 5

bug for savefig to eps in the latest version of matplotlib
by Xu, Bin April 3, 2017

April 3, 2017

Dear all, I recently update my matplotlib to the latest version, and found pyplot.savefig to eps file didn't work well like previous version. Test script: ================================== #!/usr/bin/python -tt from matplotlib.colors import LogNorm import matplotlib.pyplot as plt import numpy as np # normal distribution center at x=0 and y=5 x = np.random.randn(100000) y = np.random.randn(100000) + 5 plt.hist2d(x, y, bins=40, norm=LogNorm()) plt.colorbar() #plt.show() plt.savefig('test.eps') #plt.savefig('test.png') ================================== The result of plt.show() is just like the result of test.png, while test.eps has a weired large file size (more than 20 MB in other case). In the meantime, the facecolor for 'NaN' value in test.eps is black, different from test.png and previous version. I consider it is a bug here. How to reach the desired style (regular size, and white color for 'NaN' value in image)? Thanks a lot. -- ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Xu,Bin

3 4

How is GUI startup implemented?
by Jean CRUYPENYNCK April 1, 2017

April 1, 2017

Hi all! I’m currently writing some Python/C++ library using Boost.Python on macOS. When I import it, I get the “famous rocket icon“ bouncing in the macOS Dock and I can’t understand why. I know that this icon appears on macOS when Python detects that a program will interact with the GUI. It’s something that happens with matplotlib as well, when running from matplotlib import pyplot with the default backend. I’m wondering how this behavior is implemented, i.e. where in the source code does matplotlib call the GUI process when I execute the from matplotlib import pyplot statement? I believe that could help me solving my problem with Boost.Python. Thank you in advance for your answer, Jean.

2 1