Versions Compared

Key

  • This line was added.
  • This line was removed.
  • Formatting was changed.

...

Introductory Slide Deck:

View file
page
nameIFEX-quickmedium-what-and-why-v2v3.01.pdf
Data Expert Group Workshops - Interface Pillar Alignmentheight250

Compliance statement:

...

IFEX Development Meeting Minutes

2025-03-19

Participants:

Start 16:00 CET, end 16:52

Participants:
Kevin Karl – VCC
Magnus Feuer – VCC
Gunnar Andersson – MBition 


Planning for potential common technology projects in a COVESA/open-source or AUTOSAR collaboration context.

Discussing VHAL generator, IFEX-ARXML converter, and AUTOSAR PDU Serializer/Deserializer as possible COVESA FOSS projects.

Agreed that for the IFEX-ARXML converter project, if executed, it will be done in the context of AUTOSAR's structure and repositories.  

Gunnar went through the model-to-model design of the IFEX project.  If an XML-based input is to be created (e.g. ARXML, but could be other XML-based definitions), then the steps would be :

1. Create an <model>_ast.py python file, similar to ifex_ast.py and the _ast created for grpc/protobuf (pending), that defines @dataclass objects for each relevant object.

2. Parse XML with a standard python library such as lxml or etree

3. Most likely - implement custom code that loops through the XML parser output and create the AST objects accordingly.
The dacite library requires a hierarchical dictionary in which the keys are named identically to the member variables of the AST.  However, we anticipate that the AST model we want might not be a direct copy of the input XML structure but rather a simplified one (XML might include information we want to not include).  If that is the case, then a custom loop is required to interpret and build up the AST because dacite can only convert from dict to @dataclass hierarchy if the names of the fields match exactly.

4. Once that is done, implement a model-to-model translation by mirroring the protobuf_to_ifex.py implementation, which uses the generalized rule_translator.py translation engine.

5. Create output Jinja templates to enable "printout" from the new AST model (for ARXML or other).
The "new" setup (publication pending) now enables printouts from any similar AST model, whereas the old code was tied to only use the IFEX AST model as input.

2025-02-05

Start 16.00
 

Participants:

...