SpeciesFileGroup/taxonworks

Workbench for biodiversity informatics.

Chandra Earl

@sunray1

Thanks for answering all my questions today and the walk through - appreciate all your help!

Matt

@mjy

Thanks for taking the time to engage and describe your needs, and very glad (for those who weren't there) that you got things running with Docker locally!

Hernán Lucas Pereira

@LocoDelAssembly

What is the correct procedure to make pairing_is_allowed to work when doing origin relationships?

sorry..,

What is the correct procedure to make pairing_is_allowed to work for all subclasses when doing origin relationships?

Matt

@mjy

See concerns/shared/origin_relationship.rb

!! You must redundantly provide STI subclasses and parent classes if you want to allow both.  Providing
   a superclass does *not* provide the subclasses.

I fought that for a while and decided to go explicit, and by subclass. Since there are limited possibilities so far it wasn't too bad. Are you hitting issues in Datasets?

Hernán Lucas Pereira

@LocoDelAssembly

Yes, it is now crashing, so was trying to add is_origin_for/originates_from on both ends (originally only is_origin_for Person.to_s at ImportDataset).

Matt

@mjy

do you also have the inverse in both?

is_origin_for

originates_from

you likley need

Person::Vetted?

and Person::Unvetted?

Hernán Lucas Pereira

@LocoDelAssembly

Yes, that is the problem now and added originates_from ImportDataset.to_s at Person::Unvetted

Matt

@mjy

(good idea with ImportDataset.to_s approach btw

We should update everything to be that way...

)

Hernán Lucas Pereira

@LocoDelAssembly

Guess I'll have to list all descendants? Would also require adding is_origin_for to all subclasses of ImportDataset?

Matt

@mjy

I believe so, yes.

How many subclasses do you have?

Hernán Lucas Pereira

@LocoDelAssembly

["ImportDataset::DarwinCore", "ImportDataset::DarwinCore::Occurrences", "ImportDataset::DarwinCore::Checklist", "ImportDataset::DarwinCore::Unknown"]

(In reality only the ones in the middle would be associated with people)

Matt

@mjy

OK. Not too bad so far.

I think in this case it's good we have explict

To further define people can only come from some of those

and not all

Hernán Lucas Pereira

@LocoDelAssembly

Should I set is_origin_for for Person::Vetted? (I create always unvetted, but eventually them transition to vetted automatically, even when importing more rows without the user doing anything)

Matt

@mjy

No, don't.

The actuall class stored in the DB is the base class.

It's just a validation step on create IIRC

I.e. add it only if we make it possible in a UI.

Hernán Lucas Pereira

@LocoDelAssembly

 OriginRelationship.group(:old_object_type,:new_object_type).count
   (0.4ms)  SELECT COUNT(*) AS count_all, "origin_relationships"."old_object_type" AS origin_relationships_old_object_type, "origin_relationships"."new_object_type" AS origin_relationships_new_object_type FROM "origin_relationships" GROUP BY "origin_relationships"."old_object_type", "origin_relationships"."new_object_type"
 => {["ImportDataset", "Person"]=>399}

Alright, so Unvetted it is. Thanks @mjy

Hernán Lucas Pereira

@LocoDelAssembly

  originates_from(*
    [
      ImportDataset::DarwinCore::Checklist,
      ImportDataset::DarwinCore::Occurrences
    ].map(&:to_s)
  )

But perhaps I just change originates_from/is_origin_for method in development branch to call to_s on all elements, right? (Which should be fully compatible with existing code). Any downside?

Matt

@mjy

It does make a requirement that the model is loaded already.

So maybe we shouldn't.

Plus side- model must exist.

Downside- creates load constraints.

Matt

@mjy

@LocoDelAssembly @jlpereira do we need to update https://github.com/SpeciesFileGroup/install_taxonworks/blob/master/development/native/macos.md to reference mime-magic updates?

Hernán Lucas Pereira

@LocoDelAssembly

Guess will have to be, because migrating to ActiveStorage causes those untrackable problems of "has_one_attached method unknown" when it shouldn't at all...

BTW, is Xcode actually required? It is a VERY heavy download even after installing it I noticed that brew downloaded it again anyway (although command line tools only)

José Luis Pereira

@jlpereira

what you mean about the mime-magic?

Matt

@mjy

brew install shared-mime-info

José Luis Pereira

@jlpereira

I didnt have to do anything different to get it work

Matt

@mjy

@LocoDelAssembly I didn't know brew loaded xcode. If it does then we don't need it again, though I can't imagine a dev env without it to start.

Hernán Lucas Pereira

@LocoDelAssembly

Try bundle update && git checkout Gemfile.lock

Matt

@mjy

@jrflood that ^

Ah, but bundle update fails for him.

I don' thtink that will work.

Hernán Lucas Pereira

@LocoDelAssembly

Try bundle update && git checkout Gemfile.lock

If it fails then you need really need to install shared-mime-info otherwise will work and Gemfile.lock will be restored so updates are not pushed.

Where communities thrive

People

Repo info

Activity