genanki is a python lib to generate Anki apkg files directly. No need to run Anki in the background anymore.

Originally, the choices option field use <div>B something.</div> to make a new line.
In the old versions, Anki will treat this kind of syntax as new lines and recognize it as an option.

    options = options[0] + \
        list2str(options[1:], "<div>", "</div>", keepsuffix=True)

However, in Anki 2.1.41 and later versions:

<br> tags will now be used by default instead of the previous <div> tags, which solves some issues with multiple lines in cloze deletions and MathJax (thanks to Henrik).

Now both <br> and <div> are supported thus I will still stick to the old method for backward compatibility.
However, after the new version is universally adapted, I'd change it to <br> for better consistency.

Describe the bug

$\ce{K_3[Fe(CN)_6]_3}$ is formatted as \(\ce{K<em>3[Fe(CN)</em>6]_3}\) .

To Reproduce

from helper.formatHelper import formatText

text = "$\ce{K_3[Fe(CN)_6]_3}$"

print(formatText(text))

Expected behavior

It should be rendered as \(\ce{K_3[Fe(CN)_6]_3}\) to adapt to LaTeX syntax.

In the past, I set the alignment to left for proper list display style. However, I've found a method to set center-left alignment, so I'll switch back to center.

It is notable that Choices will also be center-aligned.

To avoid a collision, the names of Note Types would be changed.

Consider using logging.

When importing a file for the second time, try to overwrite the duplicates rather than add another note.
This requires a method to recognize the original note.
Maybe an ID should be added to the markdown file, but it would pollute the text.
So I'd rather create an ID for a file and insert it into the ending of the file. Then add a property for cards generated by the file.
Each time a file is imported, do an O(n^2) fuzzy match to see which note belongs to which block, and updating notes would be rather easy.

I'd like to know your opinions about this function.

Use pure string parsing instead of regex to avoid errors.

Lines with only spaces should be recognized as blank lines.

Separate config system from system args parsing.

directly render Markdown into Html for Anki support.

Customize SiYuan Attribute Tag. Default is ankilink.

Use any syntax you like.

**test** [test] (test)

$ dollars

For now, it would probably crash AnkiLink as it would assume that the text includes a math formula but fail to find the right bracket.

This function should be configurable.

Embed config in markdown files so that anyone with the file can import without extra params.
A useful function when collaborating with others.
There should be file-level config, subtitle-level config, and block-level config.

This function may require parsing markdown syntax into a tree.

A test case to clarify what I want.

!!!Deck=Test!!!  

!!!Tag=TestTag!!! #set tag to TestTag

Example Q  
Example A

## SubTitle

!!!Deck=Test2!!! #overwrite deck

!!!Tag=tag2!!! #overwrite tag

Something like that...
hmm..
!!!Tag:+tag2!!! #add tag for block

This is a block.  
No tag2 for this.

### SubSubTitle

Inherit parent config.

## Another SubTitle

Now Deck is Test and Tag is TestTag.  

call the function once the process starts.
Create anki note type in the script automatically when started for the first time rather than import a test deck manually.

This has been deprecated as it requires that Anki is opened during initializing.
Now the model is created when a related exception occurred.

For now, it is hardcoded.

The tags are missing when exporting to an APKG file.
This may have something to do with Genanki I guess.

Support Multiple Line Questions in a more elegant way.

For now, you have to use <br> inline HTML for multiple line questions.

Question1 <br> Question2  
Answer.

I'd like to add a built-in note type for multiple-line questions.

!Question1  
!Question2  
Answer  
Answer2

Maybe a ! prefix would be used.

If the target deck doesn't exist, create it via anki connect.

Since Deck is assigned in AnkiLink, the configuration in AnkiIn doesn't work.
Add a direct export method to support this config.

Sometimes a block shouldn't be recognized as a note.
Add syntax for block skipping.

support numbering clozes, an example:

Cloze1 is **[1]something** and Cloze2 is **[1]something** like that.

Cloze1 is {{c1::something}} and Cloze2 is {{c1::something}} like that.

Save the configuration in SiYuan block Inline Attribute List for config.

That's a pretty big shift.

Now we expose the whole config module to the user, indicating that you can change any config as you like.

Configs are in TOML format.

This would solve #21 #11

Similar to import notes via Anki connect, but don't require its running.
It may be complicated to implement and is not a promised feature. I'd look further on it in the future.

Test Case:

判断：向 $\ce{ZnS}$ 和 $\ce{NaS}$ 的悬浊液中滴加 $\ce{CuSO4}$ 溶液，有黑色沉淀生成，说明   $\ce{K_{sp}(CuS) < K_{sp}(ZnS)}$.
错误，考虑 $\ce{NaS}$ 使 $\ce{S^{2-}}$ 过量。