PDF Embedded fonts with python3 mpl reported as 'Unknown' by pdffonts and pdf readers

[jclevesque]

Say I want to save a PDF figure with a simple script :

import matplotlib.pyplot as plt
f = plt.figure()
ax = f.add_subplot(111)
ax.plot([1,2,3], [4,5,6])
ax.set_title('fonts')
f.savefig('test.pdf')

If I run this script with python2, then the output of pdffonts test.pdf is the following :

name                                 type              emb sub uni object ID
------------------------------------ ----------------- --- --- --- ---------
NimbusSanL-Regu                      Type 1            yes no  no      17  0
CMMI12                               Type 1            yes no  no      13  0
CMR12                                Type 1            yes no  no      21  0

However, if I run the script with python3, then the output of pdffonts test.pdf becomes :

name                                 type              emb sub uni object ID
------------------------------------ ----------------- --- --- --- ---------
Unknown                              Type 1            yes no  no      21  0
Unknown                              Type 1            yes no  no      17  0
Unknown                              Type 1            yes no  no      13  0

Both figures are visually identical. This is a problem for me because I'm trying to make sure my PDFs contain only the allowed font types for a conference paper submission.

This is with matplotlib 1.3 for both python 2.7.3 and python 3.2.3.

[audurand]

This is also an issue for me.

[pelson]

@jkseppan - any ideas on this?

[jkseppan]

@jclevesque - can you make the offending pdf file available at some web address? I'd like to peek inside. My guess is that because of Python 3's unicode-vs-bytes differentiation the font name gets encoded in a different encoding than before (e.g. UCS-2 vs ASCII) but I don't have a Python 3 installation handy to test on.

[audurand]

I attached an example of offending pdf along with the Python3 script used
to produce it.

2014-06-18 7:20 GMT-04:00 Jouni K. Seppänen notifications@github.com:

@jclevesque https://github.com/jclevesque - can you make the offending
pdf file available at some web address? I'd like to peek inside. My guess
is that because of Python 3's unicode-vs-bytes differentiation the font
name gets encoded in a different encoding than before (e.g. UCS-2 vs ASCII)
but I don't have a Python 3 installation handy to test on.

—
Reply to this email directly or view it on GitHub
#3049 (comment)
.

[jkseppan]

@audurand - Sorry, I can't see the pdf file linked from this github issue. Could you either put it up on Dropbox or similar, or email it directly to me at jks@iki.fi?

[audurand]

Here is an example of offending pdf along with the Python3 script used to
produce it. If you cannot see the files I will upload them on Dropbox,

2014-07-09 2:15 GMT-04:00 Jouni K. Seppänen notifications@github.com:

@audurand https://github.com/audurand - Sorry, I can't see the pdf file
linked from this github issue. Could you either put it up on Dropbox or
similar, or email it directly to me at jks@iki.fi?

—
Reply to this email directly or view it on GitHub
#3049 (comment)
.

[jenshnielsen]

I can reproduce this issue. But only with usetex=True

[jkseppan]

Yes, it seems to be related to the AFM parsing in type1font.py.