Make country package purely deterministic - read stochastic variables from dataset

changelog_entry.yaml

@@ -0,0 +1,6 @@
+- bump: minor
+  changes:
+    added:
+      - Add Head Start and Early Head Start takeup variables for dataset-driven stochastic modeling
+    changed:
+      - Update head_start and early_head_start to use takeup in microsimulation (defaults to True in policy calculator)

policyengine_us/variables/gov/aca/takes_up_aca_if_eligible.py

@@ -4,10 +4,11 @@
 class takes_up_aca_if_eligible(Variable):
     value_type = bool
     entity = TaxUnit
-    label = "Whether a random eligible SPM unit does not claim ACA Premium Tax Credit"
+    label = "Whether an eligible tax unit claims ACA Premium Tax Credit"
+    documentation = (
+        "Generated stochastically in the dataset using take-up rates. "
+        "No formula - purely deterministic rules engine."
+    )
     definition_period = YEAR
-
-    def formula(tax_unit, period, parameters):
-        seed = tax_unit("aca_take_up_seed", period)
-        takeup_rate = parameters(period).gov.aca.takeup_rate
-        return seed < takeup_rate
+    # For policy calculator (non-dataset), defaults to True (full take-up assumption)
+    default_value = True

policyengine_us/variables/gov/hhs/head_start/early_head_start.py

@@ -11,11 +11,18 @@ class early_head_start(Variable):
     reference = "https://headstart.gov/program-data/article/head-start-program-facts-fiscal-year-2022"
 
     def formula(person, period, parameters):
+        takes_up = person("takes_up_early_head_start_if_eligible", period)
+        is_in_microsim = hasattr(person.simulation, "dataset")
+
         p = parameters(period).gov.hhs.head_start.early_head_start
         state = person.household("state_code_str", period)
         spending = p.spending[state]
         enrollment = p.enrollment[state]
         mask = enrollment > 0
         result = np.zeros_like(p.spending[state])
         result[mask] = spending[mask] / enrollment[mask]
-        return result
+
+        if is_in_microsim:
+            return result * takes_up
+        else:
+            return result

policyengine_us/variables/gov/hhs/head_start/head_start.py

@@ -11,11 +11,18 @@ class head_start(Variable):
     reference = "https://headstart.gov/program-data/article/head-start-program-facts-fiscal-year-2022"
 
     def formula(person, period, parameters):
+        takes_up = person("takes_up_head_start_if_eligible", period)
+        is_in_microsim = hasattr(person.simulation, "dataset")
+
         p = parameters(period).gov.hhs.head_start
         state = person.household("state_code_str", period)
         spending = p.spending[state]
         enrollment = p.enrollment[state]
         mask = enrollment > 0
         result = np.zeros_like(p.spending[state])
         result[mask] = spending[mask] / enrollment[mask]
-        return result
+
+        if is_in_microsim:
+            return result * takes_up
+        else:
+            return result

policyengine_us/variables/gov/hhs/head_start/takes_up_early_head_start_if_eligible.py

@@ -0,0 +1,14 @@
+from policyengine_us.model_api import *
+
+
+class takes_up_early_head_start_if_eligible(Variable):
+    value_type = bool
+    entity = Person
+    label = "Whether an eligible person enrolls in Early Head Start"
+    documentation = (
+        "Generated stochastically in the dataset using take-up rates. "
+        "No formula - purely deterministic rules engine."
+    )
+    definition_period = YEAR
+    # For policy calculator (non-dataset), defaults to True (full take-up assumption)
+    default_value = True

policyengine_us/variables/gov/hhs/head_start/takes_up_head_start_if_eligible.py

@@ -0,0 +1,14 @@
+from policyengine_us.model_api import *
+
+
+class takes_up_head_start_if_eligible(Variable):
+    value_type = bool
+    entity = Person
+    label = "Whether an eligible person enrolls in Head Start"
+    documentation = (
+        "Generated stochastically in the dataset using take-up rates. "
+        "No formula - purely deterministic rules engine."
+    )
+    definition_period = YEAR
+    # For policy calculator (non-dataset), defaults to True (full take-up assumption)
+    default_value = True

policyengine_us/variables/gov/hhs/medicaid/takes_up_medicaid_if_eligible.py

@@ -4,10 +4,11 @@
 class takes_up_medicaid_if_eligible(Variable):
     value_type = bool
     entity = Person
-    label = "Whether a random eligible person unit does not enroll in Medicaid"
+    label = "Whether an eligible person enrolls in Medicaid"
+    documentation = (
+        "Generated stochastically in the dataset using take-up rates. "
+        "No formula - purely deterministic rules engine."
+    )
     definition_period = YEAR
-
-    def formula(person, period, parameters):
-        seed = person("medicaid_take_up_seed", period)
-        takeup_rate = parameters(period).gov.hhs.medicaid.takeup_rate
-        return seed < takeup_rate
+    # For policy calculator (non-dataset), defaults to True (full take-up assumption)
+    default_value = True

policyengine_us/variables/gov/usda/snap/takes_up_snap_if_eligible.py

@@ -4,10 +4,11 @@
 class takes_up_snap_if_eligible(Variable):
     value_type = bool
     entity = SPMUnit
-    label = "Whether a random eligible SPM unit does not claim SNAP"
+    label = "Whether an eligible SPM unit claims SNAP"
+    documentation = (
+        "Generated stochastically in the dataset using take-up rates. "
+        "No formula - purely deterministic rules engine."
+    )
     definition_period = YEAR
-
-    def formula(spm_unit, period, parameters):
-        seed = spm_unit("snap_take_up_seed", period)
-        takeup_rate = parameters(period).gov.usda.snap.takeup_rate
-        return seed < takeup_rate
+    # For policy calculator (non-dataset), defaults to True (full take-up assumption)
+    default_value = True