renommage/déplacement row to data

Author: maxcorbeau

dags/enrich/tasks/business_logic/enrich_dbt_model_row_to_suggest_data.py

@@ -0,0 +1,22 @@
+from enrich.config import SUGGEST_PREFIX
+
+
+def dbt_model_row_to_suggest_data(row: dict) -> dict:
+    """Construct the pydantic model_params.data dict from a
+    dbt model's row based on fields prefixed with SUGGEST_PREFIX"""
+    pre = SUGGEST_PREFIX
+    keys_ok = [k for k in row.keys() if k.startswith(f"{pre}_")]
+    keys_ok.remove(f"{pre}_cohort")
+
+    # Validation
+    keys_fail = [
+        k
+        for k in row.keys()
+        if pre in k and k not in keys_ok and not k.startswith(f"{pre}_")
+    ]
+    if keys_fail:
+        msg = f"Colonnes invalides avec {pre} mais sans {pre}_: {keys_fail}"
+        raise KeyError(msg)
+
+    # Construct the data dict
+    return {k.replace(pre + "_", ""): row[k] for k in keys_ok}

dags/enrich/tasks/business_logic/enrich_dbt_model_to_suggestions.py

@@ -5,32 +5,15 @@
 from cluster.tasks.business_logic.cluster_acteurs_parents_choose_new import (
     parent_id_generate,
 )
-from enrich.config import COHORTS, COLS, SUGGEST_PREFIX
+from enrich.config import COHORTS, COLS
+from enrich.tasks.business_logic.enrich_dbt_model_row_to_suggest_data import (
+    dbt_model_row_to_suggest_data,
+)
 from utils import logging_utils as log
 
 logger = logging.getLogger(__name__)
 
 
-def row_to_suggest_data(row: dict) -> dict:
-    """Construct the data dict from all row props starting with SUGGEST_PREFIX"""
-    pre = SUGGEST_PREFIX
-    keys_ok = [k for k in row.keys() if k.startswith(f"{pre}_")]
-    keys_ok.remove(f"{pre}_cohort")
-
-    # Validation
-    keys_fail = [
-        k
-        for k in row.keys()
-        if pre in k and k not in keys_ok and not k.startswith(f"{pre}_")
-    ]
-    if keys_fail:
-        msg = f"Colonnes invalides avec {pre} mais sans {pre}_: {keys_fail}"
-        raise KeyError(msg)
-
-    # Construct the data dict
-    return {k.replace(pre + "_", ""): row[k] for k in keys_ok}
-
-
 def changes_prepare(
     model,
     model_params: dict,
@@ -100,7 +83,7 @@ def changes_prepare_closed_replaced(
     today = datetime.now(timezone.utc).strftime("%Y-%m-%d")
     # Parent
     parent_id = parent_id_generate([str(row[COLS.SUGGEST_SIRET])])
-    parent_data = row_to_suggest_data(row)
+    parent_data = dbt_model_row_to_suggest_data(row)
     parent_data["identifiant_unique"] = parent_id
     parent_data["source"] = None
     parent_data["statut"] = ActeurStatus.ACTIF

dags_unit_tests/enrich/tasks/test_enrich_suggestions.py renamed to dags_unit_tests/enrich/tasks/test_enrich_dbt_model_row_to_data.py

@@ -1,18 +1,18 @@
 import pytest
-from enrich.tasks.business_logic.enrich_dbt_model_to_suggestions import (
-    row_to_suggest_data,
+from enrich.tasks.business_logic.enrich_dbt_model_row_to_suggest_data import (
+    dbt_model_row_to_suggest_data,
 )
 
 
-class TestEnrichSuggestionsRowToSuggestData:
+class TestEnrichDbtModelRowToData:
 
     def test_row_to_suggest_data(self):
         row = {
             "suggest_cohort": "cohort",
             "suggest_siret": "12345678901234",
             "foo": "bar",
         }
-        data = row_to_suggest_data(row)
+        data = dbt_model_row_to_suggest_data(row)
         assert data == {"siret": "12345678901234"}
 
     @pytest.mark.parametrize(
@@ -23,9 +23,9 @@ def test_raise_if_inconsistent_suggest_keys(self, key):
         row = {"suggest_cohort": "cohort"}  # must always be present
         row[key] = "12345678901234"
         with pytest.raises(KeyError, match="Colonnes invalides"):
-            row_to_suggest_data(row)
+            dbt_model_row_to_suggest_data(row)
 
     def test_raise_if_missing_cohort(self):
         row = {"suggest_siret": "12345678901234"}
         with pytest.raises(ValueError, match="not in list"):
-            row_to_suggest_data(row)
+            dbt_model_row_to_suggest_data(row)