Add BigQueryReader.to_recap

Recap can now convert BigQuery table schemas to Recap StructTypes.

Some notes:

1. Nested types are supported (RECORD, STRUCT)
2. Repeated/ARRAYs are supported
3. JSON types are treated as STRING
4. STRING and BYTES have a max length of 65KiB by default

For (4), I had a really hard time nailing down the exact max string size in BQ.
[This](https://cloud.google.com/bigquery/docs/reference/standard-sql/data-types)
page says they have a 2 byte header on strings/bytes, which I assume is to store
the byte length. Thus I went with 65KiB.

I'm also using [bigquery-emulator](https://github.com/goccy/bigquery-emulator)
to test. I've only included integration tests.

Closes #285

.github/workflows/ci.yaml

@@ -53,7 +53,7 @@ jobs:
     - name: Test with pytest
       run: |
         source venv/bin/activate
-        pdm run pytest tests/unit
+        pdm run pytest tests/unit -vv
 
   integration-tests:
     runs-on: ubuntu-latest
@@ -74,6 +74,15 @@ jobs:
           --health-timeout 5s
           --health-retries 5
 
+      # TODO Waiting on https://github.com/goccy/bigquery-emulator/pull/208
+      bigquery:
+        image: ghcr.io/criccomini/bigquery-emulator:0.4.3-envvar
+        env:
+          BIGQUERY_EMULATOR_PROJECT: test_project
+          BIGQUERY_EMULATOR_DATASET: test_dataset
+        ports:
+          - 9050:9050
+
       zookeeper:
         image: confluentinc/cp-zookeeper:latest
         ports:
@@ -142,4 +151,4 @@ jobs:
     - name: Test with pytest
       run: |
         source venv/bin/activate
-        pdm run pytest tests/integration
+        pdm run pytest tests/integration -vv