name: "f2dq_integration_tests"
version: "1.0"
config-version: 2

# Changed when running tests
profile: "f2dq_postgres"

target-path: "target"
clean-targets: ["target", "dbt_modules", "dbt_packages"]

models:
  f2dq:
    enabled: true

sources:
  f2dq_integration_tests:
    +f2dq_monitored: true
    +f2dq_time_filter: null
  
vars:
  f2dq:store_table_samples: true
  # The expected_*.csv seeds encode re_data's original bucket anchoring
  # ([run_date - 1d, run_date)) and only the rows that bucket produces. The
  # package default now anchors the bucket on the run date and also emits
  # `global__*` metrics; both change the result set, and the seed values cannot
  # be regenerated without the upstream multi-database harness. Pin the legacy
  # behaviour here so this suite keeps exercising the metric engine and the
  # `global__`-vs-windowed naming contract. The corrected behaviour is covered
  # by the contract tests in the control_plane test suite.
  f2dq:emit_global_metrics: false
  f2dq:time_window_end: '{{ run_started_at.strftime("%Y-%m-%d 00:00:00") }}'
  f2dq:time_window_start: '{{ (run_started_at - modules.datetime.timedelta(1)).strftime("%Y-%m-%d 00:00:00") }}'
  f2dq:anomaly_detector:
    name: modified_z_score
    threshold: 0.6
  f2dq:max_columns_in_query: 1

  f2dq:select:
    - tag:testtag
    - sample_table
    - sample_without_time_filter
    - sample_with_anomaly
    - f2dq_source_test_table

  f2dq:metrics_groups:
    integration_test_group:
      table:
        - row_count
        - freshness
        - my_distinct_table_rows

      column:
        numeric:
          - min
          - max
          - avg
          - stddev
          - variance
          - nulls_count
          - nulls_percent
          - diff # my own custom metric

        text:
          - min_length
          - max_length
          - avg_length
          - nulls_count
          - nulls_percent
          - missing_percent
          - missing_count
  
  f2dq:default_metrics:
    - integration_test_group

seeds:
  +schema: seeds
  +quote_columns: false

  f2dq_integration_tests:
    monitoring:
      sample_with_anomaly:
        +f2dq_monitored: true
        +f2dq_time_filter: creation_time

      sample_without_time_filter:
        +f2dq_monitored: true
        +f2dq_time_filter: null

      sample_table:
        +f2dq_monitored: true
        +f2dq_time_filter: creation_time
        
        +f2dq_columns: 
          - event_type
          - value1
          - value2
          - null_value
        
        +f2dq_metrics:
          table:
            - my_custom_table_metric # my own custom metric
            - distinct_table_rows
          column:
            event_type:
              - regex_test:
                  regex: ([A-Za-z0-9]+)
              - match_regex:
                  regex: ^sell
              - match_regex_percent:
                  regex: ^sell
              - not_match_regex:
                  regex: ^buy
              - not_match_regex_percent:
                  regex: ^buy
              - distinct_values
              - duplicate_values
              - duplicate_rows
              - unique_rows

      expected_z_score:
        +column_types:
          time_window_end: "TIMESTAMP"

      expected_metrics:
        +column_types:
          time_window_start: "TIMESTAMP"
          time_window_end: "TIMESTAMP"