test: microgen - adds tests to confirm attribute extraction from classes

[chalmerlowe]

This PR introduces unit tests for the CodeAnalyzer's ability to extract class attributes.
These tests are located in test_generate_analyzer.py within the TestCodeAnalyzerAttributes
class.

Key Features of the New Tests:

• Comprehensive Coverage: Includes tests for various ways attributes can be defined:
  • Class variables with and without type hints.
  • Instance variables with and without type hints, defined within __init__.
• Parameterized Tests: Utilizes @pytest.mark.parametrize to efficiently test different
  code snippets and their expected structural output.
• Clear Assertions: Directly compares the extracted structure with the expected
  structure using built-in assertion capabilities, after normalizing the attribute
  lists for order-independent comparison.

Why this change?

These tests ensure the CodeAnalyzer accurately parses Python code to identify class and
instance attributes, which is crucial for the microgenerator's functionality.

Broader Context:

The CodeAnalyzer and its tests are essential components supporting the development of a
microgenerator for the BigQuery Python SDK. This microgenerator is being created to
automatically generate a centralized BigQueryClient.

The primary goal of this centralized client is to ease the transition for our customers from
the traditional, largely handwritten BigQuery library to the newer GAPIC-generated version. It
aims to provide a unified interface, allowing users to interact with a single BigQueryClient
rather than multiple disparate *ServiceClient classes (e.g., JobServiceClient,
TableServiceClient, ModelServiceClient).

The CodeAnalyzer plays a critical role by parsing the existing *ServiceClient classes to
extract structural information (like methods and attributes). This information is then used by
the microgenerator to build the passthrough methods on the centralized BigQueryClient,
ensuring a user experience that remains familiar and convenient for our customers.

These new tests for CodeAnalyzer's attribute extraction ensure that we are correctly
understanding the structure of the underlying service clients, which is fundamental for the
microgenerator to generate a correct and useful BigQueryClient.

[daniel-sanche]

LGTM, with suggestions

[daniel-sanche]

nit: this would probably be more readable using parentheses for the multiline string, to avoid breaking indentation:

            pytest.param(
                (
                    "class MyClass:"
                    "    CLASS_VAR = 123",
                ),
                [
                    {
                        "class_name": "MyClass",
                        "methods": [],
                        "attributes": [{"name": "CLASS_VAR", "type": None}],
                    }
                ],
                id="class_var_assign",
            ),

[daniel-sanche]

This might also be easier to read/scale if test cases are stored in a yaml file. We did something similar for bigtable here

[chalmerlowe]

This is a good idea. I will make a note of this. For now, the priority is to get a complete set of working unit tests, so not gonna jump on this refactor right away.