Creating Namespace parser and unit test for Cecille's PR #37527 #4
Conversation
- Creating with data provided by Cecille in PR here: project-chip#37527 - Updated test.yaml to include TestSpecParsingNamespace.py to run in CI - Updated matter_testing support module to include new NamespacePathLocation function for Namespace file locations - Updated spec_parsing to include new functions build_xml_namespaces and parse_namespace, as well as XmlNamespaces and XmlTags dataclasses
- Added TestSpecParsingNamespace.py as it is a unit test and not run against an app.
j-ororke
changed the title
| class TestSpecParsingNamespace(MatterBaseTest): | ||
| def setup_class(self): | ||
| # Get the data model paths | ||
| self.dm_1_3 = os.path.join(os.path.dirname(__file__), "..", "..", "data_model", "1.3") |
There was a problem hiding this comment.
You should be able to do this just using the PrebuiltDatamodelDirectory enum - added a comment below on how to do that in the parser.
There was a problem hiding this comment.
Ah ok, thank you.
That will be much cleaner, I had not realized that was available.
| namespaces: dict[int, XmlNamespace] = {} | ||
| problems: list[ProblemNotice] = [] | ||
|
|
||
| if not os.path.exists(namespace_path): |
There was a problem hiding this comment.
please replace this with get_data_model_directory(data_model_directory, DataModelLevel.kNamespace)
(you'll have to implement the enum for the namespace, but it's one line and it should just work)
There was a problem hiding this comment.
Thank you, I have updated this section of code to now use the get_data_model_directory(data_model_directory, DataModelLevel.kNamespace) implementation that you have mentioned, as well as adding the kNamespace to DataModelLevel.
src/python_testing/matter_testing_infrastructure/chip/testing/spec_parsing.py
Outdated
Show resolved
Hide resolved
| # Validate 16-bit hex format (0xNNNN) | ||
| try: | ||
| # Remove '0x' prefix if present and try to parse | ||
| id_value = int(namespace_id.replace('0x', ''), 16) |
There was a problem hiding this comment.
If you use int( .... , 0) here, that's probably fine - I don't know that these necessarily need to be hex.
There was a problem hiding this comment.
Thank you, I have updated the code here to reflect the above change mentioned.
| )) | ||
|
|
||
| # Check format is exactly 0xNNNN where N is a hex digit | ||
| if not namespace_id.lower().startswith('0x') or len(namespace_id) != 6: |
There was a problem hiding this comment.
This is also fine - I don't think these need to be that strict.
There was a problem hiding this comment.
Ah ok, I left this section as is for now but can swap it over if needed. Was trying to be as thorough as possible, but I might have gone overboard in that regards a little bit here it appears lol.
|
|
||
| return problems | ||
|
|
||
| def test_all_namespace_files(self): |
There was a problem hiding this comment.
Thank you!
I tried to test all scenarios that I could think of with the namespace files.
It also helped a lot that I had the TestSpecParsingDeviceTypes.py code to reflect on.
- Updating to change method to using 0 instead of 16 bit enums - Updating to use PrebuiltDatamodelDirectory instead of os to set filepaths - Updated to utilizing the build_xml_namespaces function in spec_parsing.
…spec_parsing.py Co-authored-by: C Freeman <cecille@google.com>
Problem
Creates namespace parser
Summary of Changes
Creating namespace parser with data provided by Cecille:
Updates Cecille's PR #37527