Support Pydantic models for report format

[garrett-zetier]

For the report format it appears that attrs is used to define the format of the dumped JSON report. For purposes of ingesting the JSON report generated by unblob it would be useful to have a pydantic model defined that can support serialization and deserialization of the report. This would be useful for any tools wanting to ingest from the unblob JSON report file.

I have implemented a pydantic model and can submit a pull request for this, but wanted to gauge interest before doing so to see if this is something the team wanted to maintain moving forward.

[qkaiser]

We rely on Pydantic for many things internally so it does make sense. We see the value of providing a model.

What do you think @vlaci @e3krisztian ?

[vlaci]

It would make sense, as it'd also provide a JSON schema.

Also, we could use dataclasses instead of attrs in other places, dropping that dependency.

[garrett-zetier]

My initial approach was to just provide a Pydantic model for the unblob report, along with a unit test for validating serialization and deserialization using it (and generating a JSON schema).

Would the preference be to replace attrs with Pydantic in report.py?

I can provide a pull request, but I initially only have a less obtrusive implementation that doesn't modify report.py.

[vlaci]

My initial approach was to just provide a Pydantic model for the unblob report, along with a unit test for validating serialization and deserialization using it (and generating a JSON schema).

Would the preference be to replace attrs with Pydantic in report.py?

I can provide a pull request, but I initially only have a less obtrusive implementation that doesn't modify report.py.

My thinking was that replacing the Report classes would be the simplest approach, their asdict method could be repurposed to call the relevant pydantic function.

[qkaiser]

@wilypence looks like we're in agreement as to how we want to go forward. Feel free to submit a PR and we'll do our best to assist you with getting it merged. The team won't get in the way.

[garrett-zetier]

I have a commit that seems to be working well with all the tests. I have replaced all the models involved in the JSON report with Pydantic models and have verified that the report content, format, and layout are the same. To open a PR, should I create a fork of unblob, or can I get contributor access?

[qkaiser]

You must create a fork and submit a PR from there.

[garrett-zetier]

Thanks! PR has been submitted and linked to this issue.