policyengine-core 3.2.0__tar.gz → 3.4.0__tar.gz

{policyengine_core-3.2.0/policyengine_core.egg-info → policyengine_core-3.4.0}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: policyengine-core
-Version: 3.2.0
+Version: 3.4.0
 Summary: Core microsimulation engine enabling country-specific policy models.
 Home-page: https://github.com/policyengine/policyengine-core
 Author: PolicyEngine
@@ -32,8 +32,6 @@ Requires-Dist: nptyping<2
 Requires-Dist: psutil<6
 Requires-Dist: wheel<1
 Requires-Dist: h5py<4,>=3
-Requires-Dist: microdf_python<1,>=0.3.0
-Requires-Dist: tqdm<5,>=4.46.0
 Requires-Dist: requests<3,>=2.27.1
 Requires-Dist: pandas>=1
 Requires-Dist: plotly<6,>=5.6.0

{policyengine_core-3.2.0 → policyengine_core-3.4.0}/policyengine_core/charts/bar.py

@@ -1,7 +1,7 @@
 import pandas as pd
 from .formatting import *
 import plotly.express as px
-from microdf import MicroSeries
+from policyengine_core.weighting import MicroSeries
 from typing import Callable
 import numpy as np
 

{policyengine_core-3.2.0 → policyengine_core-3.4.0}/policyengine_core/data/dataset.py

@@ -186,6 +186,8 @@ class Dataset:
                         raise ValueError(
                             f"Could not save {variable} to {file}. The value is {value}."
                         )
+        elif self.data_format == Dataset.FLAT_FILE:
+            values.to_csv(file, index=False)
 
     def load_dataset(
         self,
@@ -371,17 +373,16 @@ class Dataset:
         Returns:
             Dataset: The dataset.
         """
-        file_path = Path(file_path)
         dataset = type(
             "Dataset",
             (Dataset,),
             {
-                "name": file_path.stem,
-                "label": file_path.stem,
+                "name": "dataframe",
+                "label": "DataFrame",
                 "data_format": Dataset.FLAT_FILE,
                 "file_path": "dataframe",
                 "time_period": time_period,
-                "load": lambda: dataframe,
+                "load": lambda self: dataframe,
             },
         )()
 

{policyengine_core-3.2.0 → policyengine_core-3.4.0}/policyengine_core/enums/enum.py

@@ -53,10 +53,11 @@ class Enum(enum.Enum):
         # Confusingly, Numpy uses "S" to refer to byte-string arrays
         # and "U" to refer to Unicode-string arrays, which are also
         # referred to as the "str" type
-        if array.dtype.kind == "S":
+        if isinstance(array[0], Enum):
+            array = np.array([item.name for item in array])
+        if array.dtype.kind == "S" or array.dtype == object:
             # Convert boolean array to string array
             array = array.astype(str)
-
         if isinstance(array, np.ndarray) and array.dtype.kind in {"U", "S"}:
             # String array
             indices = np.select(

{policyengine_core-3.2.0 → policyengine_core-3.4.0}/policyengine_core/parameters/parameter.py

@@ -224,3 +224,12 @@ class Parameter(AtInstantLike):
             if value_at_instant.instant_str <= instant:
                 return value_at_instant.value
         return None
+
+    def relative_change(self, start_instant, end_instant):
+        start_instant = str(start_instant)
+        end_instant = str(end_instant)
+        end_value = self._get_at_instant(end_instant)
+        start_value = self._get_at_instant(start_instant)
+        if end_value is None or start_value is None:
+            return None
+        return end_value / start_value - 1

{policyengine_core-3.2.0 → policyengine_core-3.4.0}/policyengine_core/populations/group_population.py

@@ -7,6 +7,8 @@ from policyengine_core import projectors
 from policyengine_core.entities import Entity, Role
 from policyengine_core.enums import EnumArray
 from policyengine_core.populations.population import Population
+from policyengine_core.periods.period_ import Period
+from typing import Optional, Container
 
 if TYPE_CHECKING:
     from policyengine_core.simulations import Simulation
@@ -21,6 +23,20 @@ class GroupPopulation(Population):
         self._members_position: ArrayLike = None
         self._ordered_members_map = None
 
+    def __call__(
+        self,
+        variable_name: str,
+        period: Period = None,
+        options: Optional[Container[str]] = None,
+    ):
+        variable = self.simulation.tax_benefit_system.variables.get(
+            variable_name
+        )
+        if variable.entity.is_person:
+            return self.sum(self.members(variable_name, period, options))
+        else:
+            return super().__call__(variable_name, period, options)
+
     def clone(
         self, simulation: "Simulation", members: Population
     ) -> "GroupPopulation":

{policyengine_core-3.2.0 → policyengine_core-3.4.0}/policyengine_core/populations/population.py

@@ -34,6 +34,9 @@ class Population:
         result.ids = self.ids
         return result
 
+    def has_any_input(self, variable_name: str) -> bool:
+        return len(self.get_holder(variable_name).get_known_periods()) > 0
+
     def empty_array(self) -> numpy.ndarray:
         return numpy.zeros(self.count)
 

{policyengine_core-3.2.0 → policyengine_core-3.4.0}/policyengine_core/reforms/reform.py

@@ -148,6 +148,8 @@ class Reform(TaxBenefitSystem):
                         for period, value in period_values.items():
                             if "." in period:
                                 start, stop = period.split(".")
+                                start = instant_(start)
+                                stop = instant_(stop)
                                 parameter.update(
                                     start=start, stop=stop, value=value
                                 )

{policyengine_core-3.2.0 → policyengine_core-3.4.0}/policyengine_core/simulations/microsimulation.py

@@ -1,6 +1,6 @@
 from typing import Dict, Type
 
-from microdf import MicroDataFrame, MicroSeries
+from policyengine_core.weighting import MicroDataFrame, MicroSeries
 import numpy as np
 from policyengine_core.data.dataset import Dataset
 from policyengine_core.periods import Period

{policyengine_core-3.2.0 → policyengine_core-3.4.0}/policyengine_core/simulations/simulation.py

@@ -76,6 +76,9 @@ class Simulation:
     macro_cache_write: bool = True
     """Whether to write to the macro cache."""
 
+    start_instant: str = None
+    """The earliest data input instant of the simulation."""
+
     def __init__(
         self,
         tax_benefit_system: "TaxBenefitSystem" = None,
@@ -155,6 +158,10 @@ class Simulation:
                     )
             if isinstance(dataset, type):
                 self.dataset: Dataset = dataset(require=True)
+            elif isinstance(dataset, pd.DataFrame):
+                self.dataset = Dataset.from_dataframe(
+                    dataset, self.default_input_period
+                )
             else:
                 self.dataset = dataset
             self.build_from_dataset()
@@ -242,23 +249,33 @@ class Simulation:
                 + "Make sure you have downloaded or built it using the `policyengine-core data` command."
             ) from e
 
+        if self.dataset.data_format == Dataset.FLAT_FILE:
+            data_copy = {col: data[col].values for col in data.copy().columns}
+            data = {col: data[col].values for col in data.columns}
+
         person_entity = self.tax_benefit_system.person_entity
         entity_id_field = f"{person_entity.key}_id"
+
+        def get_eternity_array(name):
+            if self.dataset.data_format == Dataset.FLAT_FILE:
+                # Look for any column with variablename__timeperiod
+                for col in data:
+                    if col.split("__")[0] == name:
+                        return data[col]
+            elif self.dataset.data_format == Dataset.TIME_PERIOD_ARRAYS:
+                return data[name][list(data[name].keys())[0]]
+            return data[name]
+
         if self.dataset.data_format != Dataset.FLAT_FILE:
             assert (
                 entity_id_field in data
             ), f"Missing {entity_id_field} column in the dataset. Each person entity must have an ID array defined for ETERNITY."
         elif entity_id_field not in data:
-            data[entity_id_field] = np.arange(len(data))
-        if self.dataset.data_format != Dataset.FLAT_FILE:
-            get_eternity_array = lambda ds: (
-                ds[list(ds.keys())[0]]
-                if self.dataset.data_format == Dataset.TIME_PERIOD_ARRAYS
-                else ds
+            data[entity_id_field] = np.arange(
+                len(get_eternity_array("person_id"))
             )
-        else:
-            get_eternity_array = lambda ds: ds
-        entity_ids = get_eternity_array(data[entity_id_field])
+
+        entity_ids = get_eternity_array(entity_id_field)
         builder.declare_person_entity(person_entity.key, entity_ids)
 
         for group_entity in self.tax_benefit_system.group_entities:
@@ -267,10 +284,18 @@ class Simulation:
                 assert (
                     entity_id_field in data
                 ), f"Missing {entity_id_field} column in the dataset. Each group entity must have an ID array defined for ETERNITY."
+                entity_ids = get_eternity_array(entity_id_field)
             elif entity_id_field not in data:
-                data[entity_id_field] = np.arange(len(data))
+                entity_id_field_values = get_eternity_array(
+                    f"person_{group_entity.key}_id"
+                )
+                if entity_id_field_values is not None:
+                    entity_ids = np.arange(
+                        len(np.unique(entity_id_field_values))
+                    )
+                else:
+                    entity_ids = np.arange(len(data[list(data.keys())[0]]))
 
-            entity_ids = get_eternity_array(data[entity_id_field])
             builder.declare_entity(group_entity.key, entity_ids)
 
             person_membership_id_field = (
@@ -283,14 +308,14 @@ class Simulation:
             elif person_membership_id_field not in data:
                 data[person_membership_id_field] = np.arange(len(data))
             person_membership_ids = get_eternity_array(
-                data[person_membership_id_field]
+                person_membership_id_field
             )
 
             person_role_field = f"{person_entity.key}_{group_entity.key}_role"
             if person_role_field in data:
-                person_roles = get_eternity_array(data[person_role_field])
+                person_roles = get_eternity_array(person_role_field)
             elif "role" in data:
-                person_roles = get_eternity_array(data["role"])
+                person_roles = get_eternity_array("role")
             elif self.default_role is not None:
                 person_roles = np.full(len(entity_ids), self.default_role)
             else:
@@ -305,6 +330,10 @@ class Simulation:
 
         self.build_from_populations(builder.populations)
 
+        if self.dataset.data_format == Dataset.FLAT_FILE:
+            # Ensure we're back to all person-level data.
+            data = data_copy
+
         if self.dataset.data_format != Dataset.FLAT_FILE:
             for variable in data:
                 if variable in self.tax_benefit_system.variables:
@@ -333,9 +362,6 @@ class Simulation:
                     )
 
                 if variable_name not in self.tax_benefit_system.variables:
-                    logging.warn(
-                        f"Variable {variable_name} not found. Skipping."
-                    )
                     continue
 
                 variable_meta = self.tax_benefit_system.get_variable(
@@ -353,9 +379,11 @@ class Simulation:
                 else:
                     entity_level_data = data[variable]
 
-                self.set_input(variable, time_period, entity_level_data)
+                self.set_input(variable_name, time_period, entity_level_data)
 
-        self.default_calculation_period = self.dataset.time_period
+        self.default_calculation_period = (
+            self.dataset.time_period or self.default_calculation_period
+        )
 
         self.tax_benefit_system.data_modified = False
 
@@ -684,6 +712,8 @@ class Simulation:
                 ):
                     # Variables with a calculate-output property specify
                     last_known_period = sorted(known_periods)[-1]
+                    if last_known_period.start > period.start:
+                        return holder.default_array()
                     array = holder.get_array(last_known_period)
                 else:
                     array = holder.default_array()
@@ -1139,10 +1169,12 @@ class Simulation:
 
         If a ``set_input`` property has been set for the variable, this method may accept inputs for periods not matching the ``definition_period`` of the variable. To read more about this, check the `documentation <https://openfisca.org/doc/coding-the-legislation/35_periods.html#automatically-process-variable-inputs-defined-for-periods-not-matching-the-definitionperiod>`_.
         """
+        period = periods.period(period)
+        if self.start_instant is None or self.start_instant > period.start:
+            self.start_instant = period.start
         variable = self.tax_benefit_system.get_variable(
             variable_name, check_existence=True
         )
-        period = periods.period(period)
         if (variable.end is not None) and (period.start.date > variable.end):
             return
         self.get_holder(variable_name).set_input(
@@ -1438,6 +1470,25 @@ class Simulation:
         with h5py.File(cache_file_path, "w") as f:
             f.create_dataset("values", data=value)
 
+    def to_input_dataframe(
+        self,
+    ) -> pd.DataFrame:
+        """Exports a DataFrame which can be loaded back to a new Simulation to reproduce the same results.
+
+        Returns:
+            pd.DataFrame: The DataFrame containing the input values.
+        """
+
+        df = pd.DataFrame()
+
+        for variable in self.tax_benefit_system.variables:
+            for period in self.get_holder(variable).get_known_periods():
+                values = self.calculate(variable, period, map_to="person")
+                if values is not None:
+                    df[f"{variable}__{period}"] = values
+
+        return df
+
 
 class NpEncoder(json.JSONEncoder):
     def default(self, obj):

{policyengine_core-3.2.0 → policyengine_core-3.4.0/policyengine_core.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: policyengine-core
-Version: 3.2.0
+Version: 3.4.0
 Summary: Core microsimulation engine enabling country-specific policy models.
 Home-page: https://github.com/policyengine/policyengine-core
 Author: PolicyEngine
@@ -32,8 +32,6 @@ Requires-Dist: nptyping<2
 Requires-Dist: psutil<6
 Requires-Dist: wheel<1
 Requires-Dist: h5py<4,>=3
-Requires-Dist: microdf_python<1,>=0.3.0
-Requires-Dist: tqdm<5,>=4.46.0
 Requires-Dist: requests<3,>=2.27.1
 Requires-Dist: pandas>=1
 Requires-Dist: plotly<6,>=5.6.0

{policyengine_core-3.2.0 → policyengine_core-3.4.0}/policyengine_core.egg-info/requires.txt

@@ -11,8 +11,6 @@ nptyping<2
 psutil<6
 wheel<1
 h5py<4,>=3
-microdf_python<1,>=0.3.0
-tqdm<5,>=4.46.0
 requests<3,>=2.27.1
 pandas>=1
 plotly<6,>=5.6.0

{policyengine_core-3.2.0 → policyengine_core-3.4.0}/setup.py

@@ -23,8 +23,6 @@ general_requirements = [
     "psutil<6",
     "wheel<1",
     "h5py>=3,<4",
-    "microdf_python>=0.3.0,<1",
-    "tqdm>=4.46.0,<5",
     "requests>=2.27.1,<3",
     "pandas>=1",
     "plotly>=5.6.0,<6",
@@ -50,7 +48,7 @@ dev_requirements = [
 
 setup(
     name="policyengine-core",
-    version="3.2.0",
+    version="3.4.0",
     author="PolicyEngine",
     author_email="hello@policyengine.org",
     classifiers=[