<== blog page - 2/3 ==> Blog automation (4) documentation (8) example (2) latex (2) notebook (4) sphinx (10)
blog page - 2/3¶
Issue with module babel¶
2015-08-17
I went through the following issue with the latest version of babel (2.0). It happens when I try to generate the documentation for this module. It gave me the following exception:
File "c:\python34_x64\lib\site-packages\sphinx\__init__.py", line 51, in main
sys.exit(build_main(argv))
File "c:\python34_x64\lib\site-packages\sphinx\__init__.py", line 61, in build_main
from sphinx import cmdline
File "c:\python34_x64\lib\site-packages\sphinx\cmdline.py", line 23, in <module>
from sphinx.application import Sphinx
File "c:\python34_x64\lib\site-packages\sphinx\application.py", line 37, in <module>
from sphinx.builders import BUILTIN_BUILDERS
File "c:\python34_x64\lib\site-packages\sphinx\builders\__init__.py", line 23, in <module>
from sphinx.util import i18n, path_stabilize
File "c:\python34_x64\lib\site-packages\sphinx\util\i18n.py", line 15, in <module>
from babel.messages.pofile import read_po
File "c:\python34_x64\lib\site-packages\babel\messages\__init__.py", line 12, in <module>
from babel.messages.catalog import *
File "c:\python34_x64\lib\site-packages\babel\messages\catalog.py", line 23, in <module>
from babel.dates import format_datetime
File "c:\python34_x64\lib\site-packages\babel\dates.py", line 28, in <module>
from babel.util import UTC, LOCALTZ
File "c:\python34_x64\lib\site-packages\babel\util.py", line 278, in <module>
from babel import localtime
File "c:\python34_x64\lib\site-packages\babel\localtime\__init__.py", line 21, in <module>
from babel.localtime._win32 import _get_localzone
File "c:\python34_x64\lib\site-packages\babel\localtime\_win32.py", line 18, in <module>
tz_names = get_global('windows_zone_mapping')
File "c:\python34_x64\lib\site-packages\babel\core.py", line 58, in get_global
_global_data = pickle.load(fileobj)
TypeError: an integer is required (got type str)
The exception disappeared after I reverted to babel 1.3.
Convert a notebook into slides¶
2015-05-16
I thought it would be easy to convert a notebook into slides. I would just have to execute nbconvert. I went through two issues. The first one came from reveal.js. My first tries did not work. I decided to take the version included in the module sphinxjp.themes.revealjs and I also updated the output of nbconvert to remove external links as much as possible.
The second issue was that all my notebooks did not include any metadata
indicated to indicate whether or not a new slide or subslide should start.
So I create a simple function which does that on a notebook
add_tag_slide.
It does not overwrite existing metadata but start new slides for every section
and new subslide if the current one becomes too long:
from pyquickhelper.ipythonhelper import read_nb
nb = read_nb("your notebook.ipynb")
nb.add_tag_slide()
nb.to_json("the modified notebook.ipynb")
It is too simple to be perfect, it is difficult to guess the size
of the rendering of some objects (images, javascript...).
You can check the results for this notebook:
example pyquickhelper.
That what the function nb2slides
is doing first and then converts it into slides:
from pyquickhelper import nb2slides
nb2slides(("your notebook.ipynb", "convert.slides.html")
Create a script to read this blog¶
2015-05-10
The module now includes a function
write_module_scripts
creates a script auto_rss_server.py which grabs the latest blog post
from this stream, runs a server and opens the default browser to read them.
It uses the module
pyrsslocal.
Here is the code to read this blog:
from pyquickhelper import write_module_scripts, __blog__
write_module_scripts("blog", blog_list=__blog__)
The blog list can be replaced by any other one. Here is its content:
<?xml version="1.0" encoding="UTF-8"?>
<opml version="1.0">
<head>
<title>blog</title>
</head>
<body>
<outline text="pyquickhelper"
title="pyquickhelper"
type="rss"
xmlUrl="http://www.xavierdupre.fr/app/pyquickhelper/helpsphinx/_downloads/rss.xml"
htmlUrl="http://www.xavierdupre.fr/app/pyquickhelper/helpsphinx/blog/main_0000.html" />
</body>
</opml>
Frequent commands and automation¶
2015-05-06
The script setup.py accepts several options
such as install or build. It also accepts
unittest to run the unit tests or build_sphinx
to build the documentation.
It usually requires to have a command line windows opened
as well as an editor to write programs.
On Windows, the module now produces a series of scripts
to automate tasks such as running the unit tests,
building the documentation. They are not included in the sources
anymore but the can be obtained by typing:
python setup.py build_script
The scripts can now be produced for every module using pyquickhelper to automate setup, unit tests and documentation.
Visualize differences between two files in a notebook¶
2015-04-23
It is now possible to visualize the differences between two files directly from a notebook:
See function create_visual_diff_through_html_files.
Producing a version for Python 2.7 from Python 3¶
2015-04-17
I tried to make most of the unit tests run under Python 2.7. Most of the function deals with strings for the documentation and it is a real pain to think again about str, unicode, bytes. Some of the functions only works in Python 3 but the goal was more to see what needed to be done.
The first issue came from exception:
try:
something
except Exception as e:
raise Exception("other message") from e
This syntax is not available and to avoid losing it, I decided to have two separate versions of the same module. I created a function py3to2_convert_tree <pyquickhelper.pycode.py3to2.py3to2_convert_tree> which copies the source and deals with this case (it removes everything it can).
I also had an issue with code like isinstance(v, long) as the long type
does not exists. So I added the string int #long#
to be replaced by long by a function. The one
I use the most is str #unicode# replaced by unicode.
The second issue is the function open. I usually use the following trick:
if sys.version_info[0]==2:
from codecs import open
It does not solve everything (strings become unicode) and will melt
with strings because I do not use u"..." which I could implement in
the function
py3to2_convert_tree <pyquickhelper.pycode.py3to2.py3to2_convert_tree>.
Running out of courage, I disabled some unit tests because they were not passing due to the encoding issues. I had to add in some files but maybe I should have added that everywhere:
from __future__ import print_function
The following is added everywhere:
from __future__ import unicode_literals
Migration to IPython 3.1¶
2015-04-16
It took me some time to do the migration to IPython 3.1. The code which automatically generates the documentation had to be updated to follow the new format of the notebooks. I had to redo the configuration of ipython to have the graph inline... I hope ipython does not rename itself into jupyter. The design is better but the notebooks crash from time to time. I guess the code becomes better each time a migration happens. But it takes quite some time. However, the changes are not reversible. Ipython needs to be updated otherwise the automated generation of the documentation will not work. For Jenkins, just remind that the server needs your credentials othewise it does not easily find pandoc.
Conversion to latex is taking for ever¶
2015-04-16
The unit tests are now scheduled using Jenkins. When the documentation is generated for the first time after a fresh installation of the machine (latex, sphinx...), the compilation can take for ever. This is due to extra packages needed by latex. When the process is run from a command line, a windows shows up asking for approval before going on installing the missing latex packages. As this command line is showing up in the output, it is just needed to executed from a command line window to import the missing package. After this short break, the latex compilation runs fine on Jenkins.
Anaconda does not have the latest version of Sphinx¶
2015-04-16
As of today, when I type conda update sphinx,
The Anaconda distribution does not upgrade to 1.3
and I need it to get the extension
sphinx.ext.napoleon. So I need to type
pip install sphinx --upgrade instead.
<== blog page - 2/3 ==> 2015-04 (8) 2015-05 (4) 2015-08 (2) 2015-10 (1) 2015-12 (3) 2016-01 (1) 2016-02 (3) 2016-04 (1) 2016-06 (1)