intègre le vieillissement dans la pipeline

leximpact_prepare_data/pipeline_survey_scenario.py

@@ -9,7 +9,7 @@
 from openfisca_core.taxbenefitsystems import TaxBenefitSystem
 
 from openfisca_survey_manager.input_dataframe_generator import set_table_in_survey
-from openfisca_survey_manager import default_config_files_directory
+from openfisca_survey_manager.paths import default_config_files_directory
 
 from leximpact_prepare_data.scenario_tools.helpers_survey_scenario import get_copules
 

leximpact_prepare_data/run_pipeline.py

@@ -5,6 +5,8 @@
 from leximpact_prepare_data.pipeline_tax_and_benefit_system import pipeline_tbs
 import os
 import click
+import pandas as pd
+from openfisca_survey_manager.input_dataframe_generator import set_table_in_survey
 
 config = Configuration(project_folder="leximpact-prepare-data")
 aggregates_path = config.get("AGREGATS_PATH")
@@ -18,7 +20,7 @@
 @click.option(
     "-erfs",
     "--annee_erfs",
-    default=2019,
+    default=2021,
     help="ERFS-FPR year",
     show_default=True,
     type=int,
@@ -29,7 +31,7 @@
 @click.option(
     "-calcul",
     "--annee_de_calcul",
-    default=2022,
+    default=2025,
     help="POTE year",
     show_default=True,
     type=int,
@@ -37,22 +39,56 @@
 def run_pipeline(annee_erfs, annee_pote, annee_de_calcul):
     log.debug("Create FranceTaxBenefitSystem")
 
-    annee_pote = str(annee_pote)
     # Instanciation du scenario
     pipeline_survey_scenario = PipelineErfsSurveyScenario(
         config_files_directory=config_files_directory,
         annee_donnees=annee_erfs,
-        period=annee_de_calcul,
+        period=annee_pote,
         baseline_tax_benefit_system=pipeline_tbs,
         collection="openfisca_erfs_fpr",
         survey_name=f"openfisca_erfs_fpr_{annee_erfs}",
     )
 
-    pipeline_survey_scenario.build_imputation(year=annee_pote)
+    pipeline_survey_scenario.build_imputation(year=str(annee_pote))
 
     pipeline_survey_scenario.save_current_survey(
         variables=pipeline_survey_scenario.used_as_input_variables,
         collection="leximpact",
-        survey_name=f"leximpact_{annee_de_calcul}",
+        survey_name=f"leximpact_{annee_pote}",
+        period=annee_de_calcul,
+    )
+
+    del pipeline_survey_scenario
+
+    survey_scenario = PipelineErfsSurveyScenario(
         period=annee_de_calcul,
+        annee_donnees=annee_pote,
+        collection="leximpact",
+        survey_name=f"leximpact_{annee_pote}",
+        config_files_directory=config_files_directory,
+        baseline_tax_benefit_system=pipeline_tbs,
+    )
+    for year in [annee_de_calcul, annee_de_calcul - 1, annee_de_calcul - 2]:
+        variables = survey_scenario.used_as_input_variables
+        if year in [annee_de_calcul - 1, annee_de_calcul - 2]:
+            variables = variables + ["rfr"]
+        data_frame_by_entity = survey_scenario.simulations[
+            "baseline"
+        ].create_data_frame_by_entity(variables=variables, period=year, index=True)
+
+        collection = "leximpact"
+        survey_name = f"leximpact_{annee_de_calcul}"
+        for entity, input_dataframe in data_frame_by_entity.items():
+            assert isinstance(input_dataframe, pd.DataFrame)
+            print(
+                f"set_table_in_survey of {entity} for {year} in {collection}.{survey_name}"
+            )
+            set_table_in_survey(
+                input_dataframe,
+                entity,
+                period=year,
+                collection=collection,
+                survey_name=survey_name,
+                config_files_directory=survey_scenario.data["config_files_directory"],
+                source_format="parquet",
             )