Add benchmark test suite for model operations

[abrookins]

Summary

Adds a comprehensive benchmark test suite using pytest-benchmark to document performance characteristics of redis-om-python model operations.

Benchmarks Included

The suite covers 25 benchmarks across these categories:

Instantiation

• Pydantic BaseModel (baseline)
• HashModel (simple and complex)
• JsonModel (simple and complex)

Save Operations

• HashModel save
• JsonModel save
• JsonModel with embedded documents

Get Operations

• HashModel get by primary key
• JsonModel get by primary key
• JsonModel with embedded documents

Query Operations

• Find all records
• Find by indexed field
• Find by embedded field
• Find with sorting
• Find with pagination

Update Operations

• HashModel field updates
• JsonModel field updates
• JsonModel embedded document updates

Batch Operations

• Batch add for HashModel
• Batch add for JsonModel
• Batch add for JsonModel with embedded documents

Running Benchmarks Locally

# Run benchmarks with timing output
pytest tests/test_benchmarks.py -v --benchmark-only

# Run benchmarks as part of normal test suite
pytest tests/test_benchmarks.py -v

CI Integration

Benchmarks run in a dedicated CI job that:

• Executes all 25 benchmarks with full timing measurements
• Generates a summary table in the GitHub Actions run summary
• Uploads benchmark results as a JSON artifact for historical comparison

To view benchmark results:

1. Go to the Actions tab for this PR
2. Click on the "CI" workflow
3. Click Summary in the left-hand sidebar
4. View the summary table (scroll down)

Interpreting Results

The benchmark output shows:

• Min/Max/Mean: Timing statistics
• OPS: Operations per second (higher is better)

Changes

• Added tests/test_benchmarks.py with 25 benchmark tests
• Added pytest-benchmark as a dev dependency
• Added dedicated "Benchmarks" job to CI workflow with summary reporting
• Updated make_sync.py to exclude benchmark tests from sync generation (async benchmarks are sufficient)

Related to #640