data-dwh-dbt-project/seeds/schema.yml

version: 2

seeds:
  - name: stg_seed__currencies
    description: |
      A list of valid current currencies according to ISO 4217.

      The list was obtained from https://www.six-group.com/en/products-services/financial-information/data-standards.html#scrollTo=isin
    config:
      column_types:
        iso_4217_numeric_code: varchar(3)
    columns:
      - name: iso_4217_code
        data_type: character varying
        description: The 3 character ISO 4217 code for this currency, in Uppercase.
        data_tests:
          - not_null
          - dbt_expectations.expect_column_values_to_match_regex:
              regex: "^[A-Z]{3}$"
      - name: iso_4217_numeric_code
        data_type: character varying
        description: The 3 digit ISO 4217 numeric code for this currency.
        data_tests:
          - not_null
          - dbt_expectations.expect_column_values_to_match_regex:
              regex: "^[0-9]{3}$"
      - name: decimal_positions
        data_type: int
        description: |
          The decimal positions that lead to this currency smallest unit.

          For example: since Japanese Yen (JPY) have no cents, this value is 0.

          On the other hand, since the US Dollar (USD) is composed of cents, and each dollar equals 100 cent, this value is 2.

          To convert from normal unit (Dollar) to smallest unit (Cent), multiply by `10^decimal_positions`.
          To convert from smallest unit (Cent) to normal unit (Dollar), divide by `10^decimal_positions`.
        data_tests:
          - not_null
          - dbt_expectations.expect_column_values_to_be_between:
              min_value: 0
              max_value: 8
              strictly: False

  - name: stg_seed__guest_services_vat_rates_by_country
    description: |
      A list of applicable VAT rates for guest services, by country.

      The list was provided by the Finance team. A value of 0% does not 
      necessarily mean that the country doesn't have VAT, but rather that we 
      don't need to charge it to guests from that country.

      Country names and codes _almost_ follow ISO 3166-1 (https://en.wikipedia.org/wiki/ISO_3166-1).
      The only exception sits in the Kosovo record. Kosovo does not appear as a country in ISO 3166, but is nevertheless
      a valid country in the `Country` table of the Superhog backend database. Because of this, we need to include it.
      The present codes are made up (not truly ISO 3166 codes) and match the ones present in the backend.

      Read more here: https://www.notion.so/knowyourguest-superhog/Guest-Services-Taxes-How-to-calculate-a5ab4c049d61427fafab669dbbffb3a2?pvs=4

    config:
      column_types:
        country_code: varchar(3)
    columns:
      - name: country_name
        data_type: character varying
        description: The name of the country.

        data_tests:
          - not_null
          - unique
      - name: alpha_2
        data_type: character varying
        description: |
          The two characters ISO 3166-1 Alpha-2 code for the country.

        data_tests:
          - not_null
          - unique
          - dbt_expectations.expect_column_values_to_match_regex:
              regex: "^[A-Za-z]{2}$"
      - name: alpha_3
        data_type: character varying
        description: |
          The three characters ISO 3166-1 Alpha-3 code for the country.

        data_tests:
          - not_null
          - unique
          - dbt_expectations.expect_column_values_to_match_regex:
              regex: "^[A-Za-z]{3}$"
      - name: country_code
        data_type: character varying
        description: |
          The three digit ISO 3166-1 Numeric code for the country.

        data_tests:
          - not_null
          - unique
          - dbt_expectations.expect_column_values_to_match_regex:
              regex: "^[0-9]{3}$"
      - name: vat_rate
        data_type: numeric
        description: |
          The Superhog applicable VAT rate for guests of this country. A value 
          of 0% does not necessarily mean that the country doesn't have VAT, but
          rather that we don't need to charge it to guests from that country.

        data_tests:
          - not_null
          - dbt_expectations.expect_column_values_to_be_between:
              min_value: 0
              max_value: 1
              strictly: false

  - name: stg_seed_guesty_claims_snapshot_20241010
    description: |
      A list of claims that have been paid out within the Athena/Guesty line of 
      business.

      The data was shared on 2024-10-10 by Chloe from Resolutions in a static 
      file, and was added to the DWH to support this ticket: https://guardhog.visualstudio.com/Data/_boards/board/t/Data%20Team/Stories/?workitem=22703
      This is a static snapshot and we currently have no intent of maintaining up to date.

    columns:
      - name: "Booking ID"
        data_type: character varying
        description: |
          The internal ID of this booking in Athena. Matches with the booking ID
          in the Athena verifications table.

      - name: "Claim Date"
        data_type: timestamp
        description: When was the claim received by Superhog.

      - name: "Settled Date"
        data_type: timestamp
        description: |
          When was the outcome of the claim decided by Superhog. Do not confuse
          with when was the payment executed or received.

      - name: "Paid Date"
        data_type: timestamp
        description: |
          When was the settlement amount payment executed by Superhog.

      - name: Settlement Currency
        data_type: character varying
        description: ISO4217 code of the currency in which the claim was posted.

      - name: Settlement Amount
        data_type: numeric
        description: |
          How much Superhog decided to pay out to the partner as part of this 
          claim, defined in the settlement currency.

  - name: stg_seed__athena_price_history
    description: |
      A price history for the Athena fee per night.

      Yes, I know. It's terrible that we keep this here. Oh boy, how I wish it 
      wasn't like this!

    columns:
      - name: start_at_utc
        data_type: timestamp
        description: |
          The start of the time range where this record is applicable.

      - name: end_at_utc
        data_type: timestamp
        description: The end of the time range where this record is applicable.

      - name: fee_per_night_gbp
        data_type: numeric
        description: |
          How much we charge per night in this time range.