is deleted labels (#198)

* is deleted labels

* spelling

* fix

models/bronze/labels/bronze__labels.sql

@@ -11,11 +11,14 @@ SELECT
     label_type,
     label_subtype,
     address_name,
-    project_name
+    project_name,
+    _is_deleted,
+    modified_timestamp,
+    labels_combined_id
 FROM
     {{ source(
-        'crosschain',
-        'dim_labels'
+        'crosschain_silver',
+        'labels_combined'
     ) }}
 WHERE
     blockchain = 'arbitrum'

models/doc_descriptions/labels/table_dim_labels.md

@@ -0,0 +1,5 @@
+{% docs table_dim_labels %}
+
+The labels table is a store of one-to-one address identifiers, or an address name. Labels are broken out into a "type" (such as cex, dex, dapp, games, etc.) and a "subtype" (ex: contract_deployer, hot_wallet, token_contract, etc.) in order to help classify each address name into similar groups. Our labels are sourced from many different places, but can primarily be grouped into two categories: automatic and manual. Automatic labels are continuously labeled based on certain criteria, such as a known contract deploying another contract, behavior based algorithms for finding deposit wallets, and consistent data pulls of custom protocol APIs. Manual labels are done periodically to find addresses that cannot be found programmatically such as finding new protocol addresses, centralized exchange hot wallets, or trending addresses. Labels can also be added by our community by using our add-a-label tool (https://science.flipsidecrypto.xyz/add-a-label/) or on-chain with near (https://near.social/lord1.near/widget/Form) and are reviewed by our labels team. A label can be removed by our labels team if it is found to be incorrect or no longer relevant; this generally will only happen for mislabeled deposit wallets.
+
+{% enddocs %}

models/gold/core/core__dim_labels.sql

@@ -12,19 +12,8 @@ SELECT
     label_type,
     label_subtype,
     project_name,
-    COALESCE (
-        labels_id,
-        {{ dbt_utils.generate_surrogate_key(
-            ['address']
-        ) }}
-    ) AS dim_labels_id,
-    COALESCE(
-        inserted_timestamp,
-        '2000-01-01'
-    ) AS inserted_timestamp,
-    COALESCE(
-        modified_timestamp,
-        '2000-01-01'
-    ) AS modified_timestamp
+    labels_combined_id AS dim_labels_id,
+    inserted_timestamp,
+    modified_timestamp
 FROM
     {{ ref('silver__labels') }}

models/gold/core/core__dim_labels.yml

@@ -1,7 +1,7 @@
 version: 2
 models:
   - name: core__dim_labels
-    description: '{{ doc("arb_labels_table") }}'
+    description: '{{ doc("table_dim_labels") }}'
     tests:
       - dbt_utils.unique_combination_of_columns:
           combination_of_columns:

models/silver/labels/silver__labels.sql

@@ -1,7 +1,10 @@
 {{ config(
     materialized = 'incremental',
-    unique_key = 'address',
+    unique_key = ['address', 'blockchain'],
+    incremental_strategy = 'merge',
     merge_exclude_columns = ["inserted_timestamp"],
+    cluster_by = 'modified_timestamp::DATE',
+    post_hook = "ALTER TABLE {{ this }} ADD SEARCH OPTIMIZATION ON EQUALITY(address); DELETE FROM {{ this }} WHERE _is_deleted = TRUE;",
     tags = ['non_realtime']
 ) }}
 
@@ -15,9 +18,8 @@ SELECT
     label_subtype,
     address_name,
     project_name,
-    {{ dbt_utils.generate_surrogate_key(
-        ['address']
-    ) }} AS labels_id,
+    _is_deleted,
+    labels_combined_id,
     SYSDATE() AS inserted_timestamp,
     SYSDATE() AS modified_timestamp,
     '{{ invocation_id }}' AS _invocation_id
@@ -27,10 +29,10 @@ WHERE
     1 = 1
 
 {% if is_incremental() %}
-AND insert_date >= (
+AND modified_timestamp >= (
     SELECT
         MAX(
-            insert_date
+            modified_timestamp
         )
     FROM
         {{ this }}

models/sources.yml

@@ -39,6 +39,7 @@ sources:
       - name: token_prices_all_providers_hourly
       - name: asset_metadata_all_providers
       - name: near_address_encoded
+      - name: labels_combined
   - name: crosschain_public
     database: crosschain
     schema: bronze_public