fix(ci): use dbt seeds directly, remove parquet transformation macro

- Update CI workflow to run dbt seed before build
- Remove create_test_tables macro (no longer needed)
- Add quote_character config to seed schema for CSV parsing
- Staging models now use CSV column names directly

Author: smithclay

.github/workflows/ci.yml

@@ -69,9 +69,8 @@ jobs:
       - name: Run dbt build (with test data)
         working-directory: integration_tests
         run: |
-          # Create test tables from parquet files
-          dbt run-operation create_test_tables
-          # Build all models
+          # Load seed data and build all models
+          dbt seed
           dbt build --full-refresh
 
   # Optional: SQLFluff linting (runs after compile for dbt templater)

integration_tests/macros/test_helpers.sql

@@ -7,6 +7,7 @@ seeds:
         timestamp: timestamp
         observed_timestamp: bigint
         severity_number: integer
+      quote_character: '"'
 
   - name: raw_spans
     config:
@@ -20,20 +21,25 @@ seeds:
         dropped_events_count: integer
         dropped_links_count: integer
         flags: integer
+      quote_character: '"'
 
   - name: raw_gauge
     config:
       column_types:
         timestamp: timestamp
         start_timestamp: bigint
-        value_double: double
-        value_int: bigint
+        value: double
+      quote_character: '"'
 
   - name: raw_sum
     config:
       column_types:
         timestamp: timestamp
         start_timestamp: bigint
-        value_double: double
-        value_int: bigint
+        value: double
         is_monotonic: boolean
+      quote_character: '"'
+
+  - name: claude_code_events
+    config:
+      quote_character: '"'