Add @mgeier's math tagging for rst conversion[WIP] #416
Conversation
|
Is there a reason that our pandoc calls inside filters don't include |
mpacer
changed the title
|
Aha…it's because we don't use that at all because our pandoc call is structured slightly differently than nbsphinx's, which has a two step call one to go to json and then a second to convert from json to whichever format is needed. This will take a little more work than I thought, changing to WIP. |
|
Hmmm…one solution would be to perform the intermediate |
nbconvert/filters/markdown.py
Outdated
| @@ -102,4 +102,4 @@ def rawlatex2math(text): | |||
| json_data = json.loads(text, object_hook=rawlatex2math_hook) | |||
| return json.dumps(json_data) | |||
|
|
|||
| return pandoc(source, 'markdown', 'rst', extra_args=extra_args) | |||
| return pandoc(source, 'markdown', 'rst', extra_args=extra_args, filter_func=rawlatex2math) | |||
There was a problem hiding this comment.
This doesn't have def to refer back to as far as I can tell. no wonder it doesn't work.
mpacer
force-pushed
the
rst_math_latex
branch
from
c5b30ad to
d1993ab
Compare
nbconvert/filters/markdown.py
Outdated
| @@ -117,5 +117,7 @@ def rawlatex2math_hook(obj): | |||
| def rawlatex2math(text): | |||
| json_data = json.loads(text, object_hook=rawlatex2math_hook) | |||
| return json.dumps(json_data) | |||
|
|
|||
| temp_text = | |||
|
This is going the wrong direction, fix using the new syntax for intermediate actions on the json internals as a pure python process. |
|
In particular note: these results associated with #458 |
Make pandoc call include filter trying to have a secondary two processing level pandoc call
| return json.dumps(json_data) | ||
|
|
||
|
|
||
| return convert_pandoc(source, 'markdown', 'rst', extra_args=extra_args) |
There was a problem hiding this comment.
Did you mean to use the new pandoc_with_filter function you defined here? At the moment, I don't think it's using the functions functions you define above.
Should address #291.
Some note on review: I just copied that code as it was used in production with nbsphinx. It LGTM, but I've not put it through more extensive testing; I'm assuming that (if it doesn't break our tests) then it does as @mgeier describes in #291 and alludes to in #232.