ConCat Schemas for Supplier & Chemical

[PeterKraus]

Hello all, this is a follow-up from #81 and a few of the questions I had on slack today. I think I got the two main classes to pretty much where I want them. There are a few outstanding issues, e.g. how to reference the Supplier from the Chemical, but I guess that will go away once I convert those to python files and deploy them on my Oasis.

Supplier

I think this will be able to get me quite far. The only pet-peeve I have is that I didn't figure out how to hide the "person selector" in the contact_person field. This is a consequence of using an Author, and I could in principle just roll-out my own Section. Other than that, it's pretty much there.

Chemical

I think the most offending issue here is how the link to the supplier is created and shown. Basically, it forces me to search for other files, then select the correct supplier (Crystal GmbH (./data)), but it's shown as the mainfile in the UI. I guess a better linking is in the works?

I wanted the delivery_date (which now that I think about it should be manufacture_date) to be just the date, but when I used DateEditQuantity (as shown here instead of the DateTimeEditQuantity the schema parsing broke.

The link to PubChem seems to work quite well. I've allowed the editing of several fields, and kept some properties visible to indicate if the name or pub_chem_cid or inchi_key get resolved properly. It would be ace if there was a way to combine searching of the CAS as well as PubChem databases, as at the moment I think we have to select only one. I'll have to add the SubstanceIdentifiers section again, so that the elemental_composition gets populated and I can search for these Chemicals using the periodic table widget.

Thanks a lot for help for now!

[aalbino2]

Hello Peter,

looks very nice, I'm glad you pushed some schemas here.

Some comments on the issues:

• how to reference the Supplier from the Chemical. Yeah, making it as an Entry archive will show it in the UI as a separate line and you have to fish it and reference it, you could automate the referencing process from a parser (check this parser code AreaA-data_modeling_and_schemas/IKZ_plugin/src/movpe_IKZ/movpe1_growth_parser/deposition_control/parser.py line 163 and on where we perform searches in the database for a specific class and then we reference it in other archives). Concerning supplier info, I think we are generally composing the supplier class in the chemical Entry (it is usually even simpler, just a string quantity called supplier), we are hence not making it a stand-alone Entry. In this last case, you would need then to parse these info in each chemical, given that you have some raw file where the chemical contains also metadata on supplier that you can parse. I know you wanted to reference it to avoid writing it again and again. Anyway, it is probably okay because, even if it is redundant info, it is not a huge array occupying space in your storage.
• the "person selector". You're right, this is a feature coming from the very specific AuthorEditQuantity. You may think to use StringEditQuantity instead and generate a simple custom section. The only thing you get from AuthorEditQuantity is the automated fishing of name in the keycloak.
• DateEditQuantity I have the feeling this is a bug, could you open a dedicated issue in this repo reporting which error comes up during processing please, we will figure it out better?
• combine searching of the CAS as well as PubChem databases. This sounds like a nice issue you may open in this repo, @hampusnasstrom will also evaluate it and see what makes more sense. Generally, I agree on merging the two things