New Apis

[D4ve-R]

Added

• ModelApi highlevel model api, accessible in ifcApi.modelApi
• GeomApi highlevel geometry api, accessible in ifcApi.geomApi

Changes:

• CreateModel does NOT write ifc headers anymore.
• multi-line writing,WriteLine accepts Type | Type[] now and returns number | number[]

Example

The following file was written with the new api, see ModelApi & GeomApi tests

[D4ve-R]

@beachtom i suspect the parsing error was that no IfcProject entity is in the model, can you confirm ?

[D4ve-R]

@beachtom, you are right, everything was already there it just needed some tweaking.
The new generated classes in ifc-schema are now more usable.
You don't have to pass a expressID as first argument when creating an entity with ne
e.g const app = IFC4X3.IfcApplication({...}, {...}, ...) ,
if you want to set the expressId, you can pass it as the last parameter to the constructor.
Also optional properties of ifc entity are properly translated to ts.

Couldn't find your comments, sorry, can you point me to them ?

[beachtom]

Not sure if you saw them all - the main one is for me spinning the package.json commands into shell scripts created a cross platform issue - I built on mac M1 - and I am unsure how it will respond on windows. A quick google implied to me using scripts was not a best practice approach for cross platform.

The other issue is we now have two Dockerfiles in the repo - can we make it so there is only one?

I will need to look at your latest changes later - but thanks in advance

[beachtom]

So I like the changes. Good spot on removing the IDs in the schema file to reduce size (I assume that is the reason).

My only thoughts now are: can we have something in the schema gen. Perhaps in schema_aliases (we could perhaps rename that file schema_config) that list objects used in the API

This could then generate for each of them a typescript alias i.e.

type IfcOwnerHistory = IFC2X3.IfcOwnerHistory | IFC4.IfcOwnerHistory | IFC4X3.IfcOwnerHistory

[D4ve-R]

Oops, my bad, i didn't intent on pushing the changes in package.json & the build_scripts.
I will check on the docker container, should be possible!

I think a huge benefit of removing the ID is that users don't need to care about them anymore, when creating new objects (they really shouldn't in most use-cases).
As you might have seen, i also removed a lot of special chars from the code generation, resulting in 1MB less data with still good readability.

type IfcOwnerHistory = IFC2X3.IfcOwnerHistory | IFC4.IfcOwnerHistory | IFC4X3.IfcOwnerHistory

that was on my mind too! helps a lot with writing schema generic code

[beachtom]

Great - thanks for removing that. I think this will be great once we can merge it.

A few other things I noticed:

• Commented out grid placement in geom api?
• Commented out IFCSIUNIT in model API
  -The columns test is failing I suspect ( but am not certain) the writing out there needs to be converted to use the new model API?
  -Thus the columns are also broken in the viewer
  -Examples/usage/src/modification.ts also seems to now be broken

I think most of these are caused by the fact the CreateModel no longer creates the header. I think this is quite a breaking change and may cause a few people lots of issues. Perahps we should add a boolean parameter here that is default to true to outputs the header - and the ModelAPI can set this to false.

Happy to discuss alternative approaches as well

[D4ve-R]

I will mark this as a real pull request, as soon as i'm ready to merge.

The columns test should work now, but i haven't tested the viewer examples. When i reopen the generated columns model i get parsing errors. Are you sure that the code generates valid ifc ? Because if yes, that means there might be a bug somewhere.

No all writing is still valid, there is no change in the api.

I also fixed the parsing error for empty sets and changed some test case, because they weren't working as expected.

[beachtom]

Sorry I mean to reply to this - I will keep an eye out for you to mark it - but if I don't respind just @ me

[beachtom]

@D4ve-R - I need some of the code from this to be merged to help fixing a bug - do you mind if I extract elements of this and get it merged and you can finish the rest?

[beachtom]

To grab the code I needed to merge I rebased this to main

[D4ve-R]

@beachtom no problem!