Fix Decimal JSON serialization in semantic layer query results

[nustiueudinastea]

DISCLAIMER: this PR was created with the assistance of Claude Code, but I reviewed the small change set. If the project policy is to not accept these please let me know and I will close the PR.

Summary

Fix JSON serialization error when querying metrics that return Decimal values from the dbt Semantic Layer.

What Changed

• Added from decimal import Decimal import in src/dbt_mcp/semantic_layer/client.py
• Renamed DateTimeEncoder to ExtendedJSONEncoder and extended it to handle Decimal objects by converting them to float
• Updated the json.dumps() call to use the renamed encoder class

Why

When querying metrics from the dbt Semantic Layer, PyArrow tables with decimal128/decimal256 columns return Python Decimal objects via table.to_pylist(). The existing DateTimeEncoder only handled datetime and date objects, causing json.dumps() to fail with:

Object of type Decimal is not JSON serializable

MCP response showing the error:

{
  "jsonrpc": "2.0",
  "id": 3,
  "result": {
    "content": [
      {
        "type": "text",
        "text": "Object of type Decimal is not JSON serializable"
      }
    ],
    "isError": false
  }
}

Related Issues

Closes #
Related to #

Checklist

• I have performed a self-review of my code
• I have made corresponding changes to the documentation (in https://github.com/dbt-labs/docs.getdbt.com) if required -- No documentation changes required
• I have added tests that prove my fix is effective or that my feature works
• New and existing unit tests pass locally with my changes

Additional Notes

Manually verified with a query_metrics call that previously failed:

{
  "limit": 3,
  "metrics": ["sqls"],
  "group_by": [{"name": "metric_time", "grain": "MONTH", "type": "time_dimension"}],
  "order_by": [{"name": "metric_time", "descending": true}]
}

Now returns successfully:

[
  {"metric_time__month": "2025-12-01T00:00:00+00:00", "sqls": 245.0},
  {"metric_time__month": "2025-11-01T00:00:00+00:00", "sqls": 134.0},
  {"metric_time__month": "2025-10-01T00:00:00+00:00", "sqls": 9373.0}
]

[b-per]

Thanks @nustiueudinastea !

We don't have any problem with people using AI/Claude as long as there is some human analysis/review in the loop, like you did. I am ok with approving the change, but would you be able to add a changelog entry using changie like described here?

[nustiueudinastea]

Sure @b-per , let me do that and get back to you.

[nustiueudinastea]

@b-per please take a look. I added the change using changie.