Is there a Julia package for computing titration curves?
I’ll like to simulate a titration experiment where a weak acid with several pKa values is titrated by a strong base. Something similar to figure 17.3.5 here: 17.3: Acid-Base Titrations - Chemistry LibreTexts
I did seemingly find exactly what you want just not in Julia, rather in R language:
Note, you can use it as-is with RCall.jl, if you just need something to practically work (even without translating the code).
It’s not a whole lot of code if you want to translate (all or maybe you only need) part of it to Julia. Note, the GPL license, so it’s free to use or translate, keeping the same copyleft license. Your whole program will fall under GPL 3.0, not a problem for using, possibly for distributing. GPL 3.0 is not compatible with all licences, but that’s only a problem for distributing code, you can always legally use privately, and that also means legally within your company.
I did look up code in other languages, C, C++, MATLAB, Rust and Python. The good thing about Python code would be you can also call it, still needs its runtime, but PythonCall.jl takes care of installation of dependencies (and Python itself). I’ve seen a wrapper made for R code (Alpaca.jl), which used RCall.jl, but as with PyCall.jl you need to take care of installation of dependencies, and R.
I did also find two packages in Python, both missing a license file [EDIT: at least one of, the package chosen, is BSD-licenced] (both have all code in one file):
Maybe useful to you: Titration | kinsaurralde
The titration calculator website […] The titration calculator is written in c++ and runs on aws lambda. The code for the calculator is located at:
GitHub - kinsaurralde/titration
Again a license file is missing.
I did not see such code for Julia, you could look at this (likely incomplete) list:
There’s no R-to-Julia translator available, that I know of (nor even to C++, but info on how is still helpful):
https://adv-r.hadley.nz/rcpp.html
There’s a (partial) MATLAB-to-Julia translator available. I’ve used it, and those languages are most closely related, so might be best to find MATLAB code for this. There’s also a Python-to-Julia (py2many) but I haven’t tried it.
https://adv-r.hadley.nz/translation.html
See also: https://cran.r-project.org/web/packages/titrationCurves/vignettes/titrationCurves.pdf
Thank you very much @Palli !
That is a lot of material you found for me 
The pHcalc Python package looks quite general, yet simple, and it generates nice curves:
If nobody comes along with a better solution, I think I’ll try to port that to Julia 
Note, since there’s no license file there, you’re not on solid legal grounds. Since it’s on github you would think it’s meant to be used… I would file an issue there, ask for MIT licence file to me added first.
since there’s no license file there, you’re not on solid legal grounds.
I would expect that porting a computational method to a different programming language would be enough to give “fresh” copyright. But perhaps it depends on how closely I follow the original?
Anyway, it says it’s BSD licensed in the setup.py file. So I guess that is ok?
You would be wrong, still “derived work”. That’s also why some people avoid GPL for porting, even avoid looking at the code, or get others to look at the code, for “cleanroom implementation”.
But license = "BSD" should be good enough, it at least conveys intent (there’s no legal requirement that I know of for a main top-level license file, just makes more helpful, what I’m used to so overlooked what you found). BSD is free software, but not it comes in several forms, unclear (to me) which meant, not all GPL compatible:
https://www.gnu.org/licenses/license-list.html#OriginalBSD
So maybe you still want to get it clarified.
Great, just know you can also use PythonCall.jl (or PyCall.jl, by now I recommend the former for the features, even with the latter having faster startup, hoping that PythonCall will get optimized to start as fast), for using or helping while porting, to compare and ensure no bugs in conversion.
Is the code slow, is that an argument for porting? In general I’m a great believer in using languages together, not needing to port.
Thank you for the clarification @Palli . I learn a lot of new things today
Is the code slow, is that an argument for porting?
It does not feel very slow, and you have a point about re-using code.
I’m not super happy about the interface, but it might be faster to wrap that rather than porting the code.
On the other hand, I’ll probably learn from it 
“BSD” without context doesn’t tell me if it’s e.g. the more problematic “4-clause BSD license” or some of the other variant, down to 0-clause, but since has context, is there in metadata, maybe there’s an official meant for definition in that context. I just don’t know, but would like to know. I think people are (often) more precise in Julia-land.
“BSD” without context doesn’t tell me if it’s e.g. the more problematic “4-clause BSD license”
I missed it at first, but a bit further down it says:
'License :: OSI Approved :: BSD License',
However, OSI has 5 BSD licenses Licenses – Open Source Initiative, so you are right, not very precise.
It took a bit of time, but I finally got around to implementing a Julia version of the pHcalc python library. It can be found here: GitHub - tp2750/pHcalc.jl
Comments are very welcome.
Turns out there is now also a version in the registry: GitHub - egonik-unlp/PHcalc.jl: Implementation of the pHcalc library by Ryan Nelson at https://github.com/rnelsonchem/pHcalc