Hey I#m the author of the Typsium package

[Ants-Aare]

Would you be interested in merging the two? It's kind of filling the same use case, would be better to have all of the functionality in one place. For example I saw that you don't have support for parsing content, but your support for bonds and precipitation arrows etc is looking really nice. Idk let me know what you think, I've sent you an invite to the Typsium organisation.

[pacaunt]

Hi @Ants-Aare,

I agree that the packages serve the same use case and should me merged. However, from my perspective, I'm not a fan of content parsing due to two reasons:

1. Content can be very arbitrary. It may contain many unexpected stuff so that inspecting them could be broken in the future. E.g. context, layout stuff.
2. Content makes use a lot of special symbol for their special syntax. So, there will be less symbol left to parse. Moreover, some symbol can join each other and become another symbol which becomes very hard to handle in my opinion.

Btw, I'm not a programmer, so you can see from my code that sometimes they are quite crappy and I'm not so good at reviewing the other people's code. (and that's why I wrote it myself, not made a pr in your package. Not because it was bad, i'm just not that good to know how to adapt from it) It is good to merge the package, but I am not sure how to cleanly maintain it.

That's all of my concern. I would like to know your opinion too about maintaing the package if they will be joined.

Thanks!

[tryptophawa]

Hi @Ants-Aare,

I agree that the packages serve the same use case and should be merged. However, from my perspective, I'm not a fan of content...

Hi, I'm also an author of the Typsium package. This is just an invitation to share some functions between the two programs, and there is also an issue (Typsium/typsium#41) suggesting this.

For your first concern, we have used elambic to extract text from the content, parse it, and then restore the content. It is not a problem so far, as it is still powerful enough. I don't think people will want to draw a table or try to measure things within the function, so we're trying to recover the text's weight, colour, etc.

For the second concern, it does affect us, so we have left a fallback to normal mode, which is just to cope with text.

The merging is not to say you have to merge your package completely with typsium; it is just an invitation to join our typsium organisation and maybe help us add more features to typsium if you are willing.

[pacaunt]

I would like to join the organization, and I will try adding more features as far as I could. I will try look into the source code of Typsium/or merge the features of both packages when I have time.

Thanks for the clarification and the invitation!

[Ants-Aare]

I took some time to look at your package a bit more closely, I think your parsing implementation is a lot better (and mostly a lot cleaner!) than ours. But I think that drawing using elembic elements is the superior way of displaying the formulas, you already have a little bit of this issue with how you need to bubble up all of the parameters from the bond drawing code as parameters to the #ch method. If you were to make more parts of the package configurable you would have a long list of parameters in ch, I think being able to do something like this is a lot more native to typst: `
#ce("H2C==CH2")

#show: set-bond(
stroke: stroke(paint:red,
thickness:1pt, cap:"round",
dash:"dotted")
)
#ce("H2C==CH2")
`
I added the precipitation arrows and bonds to typsium, so in that regard we have almost feature parity with chemformula. One thing that I have recently tried to add is math equations but for chemical reactions, so our ce function can be drawn, labeled and referenced like an equation, and it can show up in a table of contents, etc. But I’m not quite happy with the implementation. I have also been thinking about adding something similar to this kind of behaviour https://typst.app/universe/package/chem-par to typsium, but couldn’t really justify it because the existing package is already pretty good, maybe if you have some new ideas in this direction it would could be useful to add to the package. The Alchemist package has been using its own formula parser until now and I’ve been thinking about adding typsium as a dependency or adding a way for users to set their preferred package. Because there are pretty regularly coming in PRs in that package that have to do more with formula parsing and less with actual molecule drawing. A more powerful system in the background could be useful.

Just some ideas if you are interested in contributing to our organisation! There are lots of other ideas we’ve been discussing and some entirely new packages in the works, but things move slow for hobby projects :)

[Ants-Aare]

We also have a slowly growing test suite, so if you want to go crazy, feel free to do so. We can see whenever things break or things don’t parse correctly. I’m actually thinking about using some of your regex rules to update typsium, so it’s useful to know what each change is doing