I am using matplotlib to plot some data, however labels for plots are send via argparse. The problem is - I want some of the labels to contain greek letters. Which "greek letter code" type should I use to handle it?
I've used argparse to read in command line arguments for my legends for matplotlib. It translates Greek characters just fine when I specify something like '$\Delta G$' as an input argument (output as ΔG).
Related
Each word in SpaCy is represented by a vector of length 300. How can I plot these words on a scatter plot to get a visual perspective on how close any 2 words are?
There's a new package called whatlies that does exactly this: https://rasahq.github.io/whatlies/
See a short spacy example: https://spacy.io/universe/project/whatlies
When working with small-to-medium-sized texts, ScatterText is a tool which can be used to discover words that have distinguishing features. It also enables users to create interactive scatter plots that contain non-overlapping term labels.
Intall via -https://pypi.org/project/scattertext/
import spacy
import scattertext as st
nlp = spacy.load('en')
corpus = st.CorpusFromPandas(convention_df,
category_col='party',
text_col='text',
nlp=nlp).build()
I want to use tf.data.TextLineDataset() to read Chinese sentences, then use the map() function to divide into the single word, but tf.split doesn't work for Chinese.
I also hope someone can help us kindly with the issue.
It is my current solution:
read Chinese sentence from the file with Utf-8 coding format.
tokenize the sentences with some tool like jieba.
construct the vocab table.
convert source/target sentence according to vocab table.
convert to the dataset using from_tensor_slices.
get iterator from the dataset.
do other things.
if using TextLineDataset to load chinese sentences directlly, the content of dataset is something strange , displayed with byte flow.
maybe we can consider every byte as one character in english kind of language.
can anyone confirm with this or has any other suggestion, plz?
The above answer is one common option when handling non-English style language like Chinese, Korean, Japanese, etc.
You can also use the code below.
BTW, as you know, TextLineDataSet will read text content as a byte string.
So if we want to handle Chinese, we need to first decode it to unicode.
Unfortunately, there is no such option in tensorflow.
We need to choose other method like py_funct to do this.
#!/usr/bin/env python
# -*- coding: utf-8 -*-
import tensorflow as tf
def preprocess_func(x):
ret= "*".join(x.decode('utf-8'))
return ret
str = tf.py_func(
preprocess_func,
[tf.constant(u"我爱，南京")],
tf.string)
with tf.Session() as sess:
value = sess.run(str)
print(value.decode('utf-8'))
output: 我*爱*，*南*京
I am having problems saving a figure created by matplotlib as a .eps or .ps when I enable usetex=True. This works when this is not enabled. Here's an example:
plt.plot([1,2,3], [1,2,3], 'b.')
plt.text(2,2,r'\textbf{(a)} \lambda_{1} value', usetex=True, fontsize=16, fontname='Times New Roman')
plt.savefig('check.eps')
I receive this error:
File "C:\Python27\lib\site-packages\matplotlib\backends\backend_ps.py", line 671, in draw_tex
thetext = 'psmarker%d' % self.textcnt
AttributeError: 'RendererPS' object has no attribute 'textcnt'
I also cannot set the font to Time New Roman using the text command when I enable usetex=True.
Even if it is a bit late, I would like to give an answer, if someone else searches for this error.
Saving eps files starts the RendererPS backend, which checks whether plt.rcParams['text.usetex'] is set to True to initialise its textcnt attribute. Since usetex is set only for the text object, the backend does not expect to have to handle LaTeX and throws an error when it tries to. These kind of inconsistencies concerning LaTeX rendering are sadly present in a number of places in matplotlib.
If you cannot achieve the text formatting with the standard matplotib functionality, one solution is to set usetex globally: plt.rcParams['text.usetex'] = True. This render all the text of the figure with LaTeX (e.g. also tick labels or axis labels). I would recommend to do this in any case, to have consistent visuals.
Concerning the font, the fontname argument only affects standard matplotlib formatting. For LaTeX rendering, you have to specify the font like you would do in LaTeX. To get a font like Times New Roman, you need to load the respective package e.g. mathptmx (the times package is deprecated) by plt.rcParams['text.latex.preamble'] = [r'\usepackage{mathptmx}']. Of course, the package has to be installed in your local LaTeX installation. The standard font family in matplotlib is sans-serif, so this has to be changed to serif with plt.rcParams['font.family'] = 'serif'.
The final code is:
import matplotlib.pyplot as plt
# LaTeX setup
plt.rcParams['text.latex.preamble'] = [r'\usepackage{mathptmx}'] # load times roman font
plt.rcParams['font.family'] = 'serif' # use serif font as default
plt.rcParams['text.usetex'] = True # enable LaTeX rendering globally
plt.plot([1,2,3], [1,2,3], 'b.')
plt.text(2,2,r'\textbf{(a)} $\lambda_1$ value', fontsize=16)
plt.savefig('check.eps')
Note, that you have to enclose the math variable in dollar signs $\lambda_1$, otherwise you get an error or warning in LaTeX. (Also, single character subscripts do not need to be enclosed in curly brackets).
As a side note: I encountered the same error, when trying to save a figure as eps after turning usetex off again.
Working in IJulia. Desperately trying to make a custom colormap.
Tried the line:
matplotlib.colors.ListedColormap([(1,0,0),(0,1,0),(0,0,1)],"A")
which resulted in the following error
type PyObject has no field colors while loading In[16], in expression starting on line 1
which apparently means that I cannot use matplotlib directly, but only the functions which are in PyPlot.
I cannot involve matplotlib with an import (as this is invalid in IJulia).
I have noted that others have had help on similar problems, but that doesn't solve mine.
By using the PyCall package which PyPlot is using to wrap matplotlib you can obtain a colormap like this:
using PyCall
#pyimport matplotlib.colors as matcolors
cmap = matcolors.ListedColormap([(1,0,0),(0,1,0),(0,0,1)],"A")
In order to access fields in a PyObject you need to index the object with a symbol like:
cmap[:set_over]((0,0,0))
This is equivalent to: cmap.set_over((0,0,0)) in python. For other good examples of how to plot different kinds of plots using PyPlot, see these examples: https://gist.github.com/gizmaa/7214002
You don't need to use PyCall to call Python directly (although this is, of course, an option). You can also just use the PyPlot constructors for ColorMap to construct a colormap from (r,g,b) arrays or an array of colors as defined in the Julia Color package. See the PyPlot ColorMap documentation. For example:
using PyPlot, Color
ColorMap("A", [RGB(1,0,0),RGB(0,1,0),RGB(0,0,1)])
Is there any possibility to get lines and points into a legend text in matplotlib?
I had something in mind like the following
x=np.linspace(0,10,100)
ys=np.sin(x)
yc=np.cos(x)
pl.plot(x,ys,'--',label='sin')
pl.plot(x,yc,':',label='derivative of --')
pl.legend()
pl.show()
except that instead of the -- there should be the same symbol with the corresponding color just as in front of the legend label sin.
After reading around in the matplotlib source code I finally found a solution that works perfect for me and that does not need any position tweaking etc. as it used matplotlibs internal V- and HPackers.
import numpy as np
import pylab as pl
x=np.linspace(0,10,100)
ys=np.sin(x)
yc=np.cos(x)
pl.plot(x,ys,'--',label='sin')
pl.plot(x,yc,':',label='derivative of')
leg=pl.legend()
# let the hacking begin
legrows = leg.get_children()[0].get_children()[1]\
.get_children()[0].get_children()
symbol = legrows[0].get_children()[0]
childs = legrows[1].get_children().append(symbol)
pl.show()
The result looks as follows:
This is a little bit of a hack, but it accomplishes your goal and places all of the pieces (i.e. the legend and the text) on the plot in the appropriate order.
import pylab
pl.plot(x,ys,'--',label='sin', color='green')
pl.plot(x,yc,':',label='derivative of --',color='blue')
line1= pylab.Line2D(range(10), range(10), marker='None', linestyle='--',linewidth=2.0, color="green")
line2= pylab.Line2D(range(10), range(10), marker='None', linestyle=':',linewidth=2.0, color="blue")
leg = pl.legend((line1,line2),('sin','derivative of '),numpoints=1, loc=1)
pylab.text(9.4, 0.73, '- -', color='green')
leg.set_zorder(2)
pl.show()
Instead of relying on the default colors for the lines, I set them such that they can be referenced specifically in the legend. There are extra spaces left in the text for 'the derivative' for the second line in the legend, so we can place text (aka corresponding symbol/color of the sin line) on top of it. Then you specify the symbol/color of the text and place it such that it lines up with the text in the legend. Finally you specify the order, via zorder, to set the text on top.