Word cloud for tamil and english language

[HitheshSankararaman]

mostcommon = FreqDist(allwords).most_common(100)
wordcloud = WordCloud(width=1600, height=800, background_color='white',font_path='Nirmala.ttf',).generate(str(mostcommon))
fig = plt.figure(figsize=(30,10), facecolor='white')
plt.imshow(wordcloud) #, interpolation="bilinear")
plt.axis('off')
plt.title('Top 100 Most Common Words', fontsize=100)
#plt.tight_layout(pad=0)
plt.show()

The above code generates the word cloud.

English words are printed properly but not in correct size (high frequency words must have a bigger font size)
Eg:- "sir" must be have the biggest font size since it's been repeated more number of times (see output.csv)
Tamil words are not printed, only random letters are printed.
How to solve this ?
The words and its frequencies are present in output.csv.
output.csv

[amueller]

Does the if you want to give it a frequency distribution instead of the raw text, you need to use the generate_from_frequencies method.

[HitheshSankararaman]

Does the if you want to give it a frequency distribution instead of the raw text, you need to use the generate_from_frequencies method.

I used "generate_from_frequencies" and got the following,

If you see in the image, some tamil words are not properly printed. Eg:- சொல்லுங்க
How to rectify that ?

[amueller]

I'm sorry, I can't read tamil, so I don't know what's wrong with the processing. The most likely cause is the font is not supporting some characters.

[HitheshSankararaman]

I'm sorry, I can't read tamil, so I don't know what's wrong with the processing. The most likely cause is the font is not supporting some characters.

I am using Nirmala.ttf font . While plotting in matplotlib using Nirmala.ttf ,all the words are printed correctly with the help of mplcairo , cairo , raqm.

Only while using word cloud I am getting this error .

[amueller]

matplotlib uses pil/pillow under the hood. Can you try reproducing with pil?