Move to markdown-it-py markdown parser implementation #107
Conversation
|
From a quick scan of the Circle-CI documentation, I don't see any issues. However, @choldgraf I do note that the plotly and ipywidgets no longer show, moving from:
I assume this is because this notebook is now being executed during the build, and is thus encountering jupyter/nbclient#39 |
|
In 74cf6a3 I have introduced some nice, self-documenting render tests, see |
Report line number as <cell index>*10000 + <line number>. This is a simple solution to addresses #71, that doesn't require any complex overrides of the sphinx reporting machinery.
|
In 536a1dd I have added a simple fix for cell index / line number reporting: Line number are now set as See We should also document this. |
|
I think this looks great, and agreed about the widgets not showing up. I don't think that should block us here |
Rather than having a global variable, we save the exclude paths in the sphinx env and use a seperate function `is_valid_exec_file`. Also added tests
…s#107) This commit move to markdown-it-py markdown parser implementation, concurrently with myst-parser. Additionally: - Add notebook render tests - Add simple solution for reporting correct cell index/line number: Report line number as <cell index>*10000 + <line number>. This is a simple solution to addresses executablebooks#71, that doesn't require any complex overrides of the sphinx reporting machinery. - Make tests use the actual sphinx Application - Re-write validation of which docs to execute/cache: Rather than having a global variable, we save the exclude paths in the sphinx env and use a seperate function `is_valid_exec_file`. Also added tests
Using an alpha release of executablebooks/MyST-Parser#123
All the current tests pass, with no changes to the expected outputs. I also took the opportunity to split up the code in
NotebookParser.parsein to separate functions.