(Re)-calculate exact mass if needed

[Adafede]

Sorry messed up a bit...reopening a PR which should not conflict anymore!

[mwang87]

Maybe we should have it as a different field, like calcuatled_exactmass to distinguish it from whats in the library?

[Adafede]

Hi @mwang87!

Why would you like to keep what is in the library? What the users fill actually in is anyhow calculated...This will simply allow having consistent fields. I think so at least, I see no way this calculated mass could hurt?

[justinjjvanderhooft]

I would also be careful with changing existing masses given that other pipelines may rely on them, and add a field for the calculated masses....

[Adafede]

@justinjjvanderhooft It is actually the reason that pushed me to open that pull request.

I am part of those downstream pipelines.

There are many erroneous masses in there, so first I cleaned them on my side and then thought well...this is not the right way to do it, we should clean them upstream.

I might miss something important but what are your concerns if we clean, let's say a 162.0in 162.05282? I am of course not talking about the experimental protonated mass here, this should of course not be touched. Or did I interpret exact mass wrongly?

[justinjjvanderhooft]

Let me start with saying that I don't dispute the value of the recalculated exact masses. Regarding the terminology of GNPS, I will leave that up to the GNPS team (Ming), but I think it is best to separate user-defined (exact) masses and computed ones, also because they are likely based on user-defined elemental formulas?! In the matchms workflow, we also incorporated a cleanup of the GNPS library spectra, also adding SMILES etc., perhaps one day we can also integrate that. Again, it may be best to separate computed/computer-derived entries from manual inputed ones there....(both have their own sources of error) - anyways, my 5 cents!

[Adafede]

@justinjjvanderhooft I agree 100%!

Maybe asking for minimal but mandatory user-defined fields and computing all the rest would help harmonize everything? This goes of course beyond this PR 😊

So we are now at 10 cents!

[mwang87]

So we would want to keep whats in the library because thats what people entered. It just needs to be separate and apart in terms of naming. Whether we show it is a different story, but it needs to be clear what we have here is calculated from the structure.

[justinjjvanderhooft]

Agreed @mwang87 - also because it is good provenance! 😎

[Adafede]

I understand this means it will be included in a broader refactoring, nice 😊

Is this PR then still useful?

@mwang87 If so please advise on wanted modifications