Integrate comments from review

.pre-commit-config.yaml

@@ -74,6 +74,13 @@ repos:
           - --wrap
           - '88'
         files: (README\.md)
+  - repo: https://github.com/kynan/nbstripout
+    rev: 0.7.1
+    hooks:
+      - id: nbstripout
+        args:
+          - --drop-empty-cells
+          - --keep-output
   - repo: https://github.com/pre-commit/mirrors-mypy
     rev: v1.11.2
     hooks:

src/lcm/create_params_template.py

@@ -14,7 +14,7 @@ def create_params_template(
     user_model: Model,
     variable_info: pd.DataFrame,
     grids: dict[str, Array],
-    default_params: dict[str, int | float] | None = None,
+    default_params: dict[str, float] | None = None,
 ) -> ParamsDict:
     """Create parameter template from a model specification.
 
@@ -28,7 +28,7 @@ def create_params_template(
             np.nan} for beta-delta discounting.
 
     Returns:
-        ParamsDict: A nested dictionary of model parameters.
+        A nested dictionary of model parameters.
 
     """
     if default_params is None:
@@ -63,8 +63,8 @@ def _create_function_params(user_model: Model) -> dict[str, dict[str, float]]:
         user_model: The model as provided by the user.
 
     Returns:
-        dict: A dictionary for each model function, containing a parameters required in
-            the model functions, initialized with jnp.nan.
+        A dictionary for each model function, containing a parameters required in the
+            model functions, initialized with jnp.nan.
 
     """
     # Collect all model variables, that includes choices, states, the period, and
@@ -103,8 +103,8 @@ def _create_stochastic_transition_params(
         grids: A dictionary of grids consistent with user_model.
 
     Returns:
-        dict: A dictionary of parameters required for stochastic transitions,
-            initialized with jnp.nan matrices of the correct dimensions.
+        A dictionary of parameters required for stochastic transitions, initialized with
+            jnp.nan matrices of the correct dimensions.
 
     """
     stochastic_variables = variable_info.query("is_stochastic").index.tolist()

src/lcm/discrete_problem.py

@@ -26,12 +26,12 @@
 from jax import Array
 from jax.ops import segment_max
 
-from lcm.typing import ParamsDict, SegmentInfo, Shock
+from lcm.typing import ParamsDict, SegmentInfo, ShockType
 
 
 def get_solve_discrete_problem(
     *,
-    random_utility_shock_type: Shock,
+    random_utility_shock_type: ShockType,
     variable_info: pd.DataFrame,
     is_last_period: bool,
     choice_segments: SegmentInfo | None,
@@ -61,9 +61,9 @@ def get_solve_discrete_problem(
 
     choice_axes = _determine_dense_discrete_choice_axes(variable_info)
 
-    if random_utility_shock_type == Shock.NONE:
+    if random_utility_shock_type == ShockType.NONE:
         func = _solve_discrete_problem_no_shocks
-    elif random_utility_shock_type == Shock.EXTREME_VALUE:
+    elif random_utility_shock_type == ShockType.EXTREME_VALUE:
         raise NotImplementedError("Extreme value shocks are not yet implemented.")
     else:
         raise ValueError(f"Invalid shock_type: {random_utility_shock_type}.")

src/lcm/entry_point.py

@@ -1,6 +1,7 @@
 import functools
+from collections.abc import Callable
 from functools import partial
-from typing import Literal
+from typing import Literal, cast
 
 import jax
 import jax.numpy as jnp
@@ -19,6 +20,7 @@
 from lcm.simulate import simulate
 from lcm.solve_brute import solve
 from lcm.state_space import create_state_choice_space
+from lcm.typing import ParamsDict
 
 
 def get_lcm_function(
@@ -28,37 +30,34 @@ def get_lcm_function(
     debug_mode: bool = True,
     jit: bool = True,
     interpolation_options: dict | None = None,
-):
+) -> tuple[Callable, ParamsDict]:
     """Entry point for users to get high level functions generated by lcm.
 
+    Return the function to solve and/or simulate a model along with a template for the
+    parameters.
+
     Advanced users might want to use lower level functions instead, but can read the
     source code of this function to see how the lower level components are meant to be
     used.
 
     Notes:
     -----
-    - Further targets could be "likelihood" or "simulate"
-    - We might need additional arguments such as solver_options that we want to take
-      separate from a model specification.
-    - create_params needs to work with a processed_model instead of a user model.
-    - currently all the preparations are hardcoded to generate the arguments needed
-      by solve_brute. In the long run, this needs to inspect the signature of the
-      solver, or generate only what is needed using dags.
-    - there is a hack to make the state_indexers empty in the last period which needs
+    - There is a hack to make the state_indexers empty in the last period which needs
       to be replaced by a better solution, when we want to allow for bequest motives.
 
     Args:
-        model (dict): User model specification.
-        targets (str or iterable): The requested function types. Currently only
-            "solve", "simulate" and "solve_and_simulate" are supported.
-        debug_mode (bool): Whether to log debug messages.
-        jit (bool): Whether to jit the returned function.
-        interpolation_options (dict): Dictionary of keyword arguments for interpolation
+        model: User model specification.
+        targets: The requested function types. Currently only "solve", "simulate" and
+            "solve_and_simulate" are supported.
+        debug_mode: Whether to log debug messages.
+        jit: Whether to jit the returned function.
+        interpolation_options: Dictionary of keyword arguments for interpolation
             via map_coordinates. If None, the default options are used.
 
     Returns:
-        callable: A function that takes params and returns the requested targets.
-        dict: A parameter dict where all parameter values are initialized to NaN.
+        - A function that takes params (and possibly other arguments, such as initial
+          states in the simulate case) and returns the requested targets.
+        - A parameter dictionary where all parameter values are initialized to NaN.
 
     """
     # ==================================================================================
@@ -154,7 +153,7 @@ def get_lcm_function(
         # create list of emax_calculators
         # ==============================================================================
         calculator = get_solve_discrete_problem(
-            random_utility_shock_type=_mod.shocks,
+            random_utility_shock_type=_mod.random_utility_shocks,
             variable_info=_mod.variable_info,
             is_last_period=is_last_period,
             choice_segments=choice_segments[period],
@@ -195,7 +194,7 @@ def get_lcm_function(
     elif targets == "solve_and_simulate":
         _target = partial(simulate_model, solve_model=solve_model)
 
-    return _target, _mod.params
+    return cast(Callable, _target), _mod.params
 
 
 def create_compute_conditional_continuation_value(

src/lcm/interfaces.py

@@ -5,7 +5,7 @@
 from jax import Array
 
 from lcm.grids import ContinuousGrid, DiscreteGrid, Grid
-from lcm.typing import ParamsDict, Shock
+from lcm.typing import ParamsDict, ShockType
 
 
 @dataclass(frozen=True)
@@ -73,28 +73,28 @@ class InternalModel:
     """Internal representation of a user model.
 
     Attributes:
-        grids (dict): Dictionary that maps names of model variables to grids of feasible
-            values for that variable.
-        gridspecs (dict): Dictionary that maps names of model variables to
-            specifications from which grids of feasible values can be built.
-        variable_info (pd.DataFrame): A table with information about all variables in
-            the model. The index contains the name of a model variable. The columns are
-            booleans that are True if the variable has the corresponding property. The
-            columns are: is_state, is_choice, is_continuous, is_discrete, is_sparse,
-            is_dense.
-        functions (dict): Dictionary that maps names of functions to functions. The
-            functions differ from the user functions in that they all except the
-            filter functions take ``params`` as keyword argument. If the original
-            function depended on model parameters, those are automatically extracted
-            from ``params`` and passed to the original function. Otherwise, the
-            ``params`` argument is simply ignored.
-        function_info (pd.DataFrame): A table with information about all functions in
-            the model. The index contains the name of a function. The columns are
-            booleans that are True if the function has the corresponding property. The
-            columns are: is_filter, is_constraint, is_next.
-        params (dict): Dict of model parameters.
-        n_periods (int): Number of periods.
-        shocks (Shock): Type of shocks.
+        grids: Dictionary that maps names of model variables to grids of feasible values
+            for that variable.
+        gridspecs: Dictionary that maps names of model variables to specifications from
+            which grids of feasible values can be built.
+        variable_info: A table with information about all variables in the model. The
+            index contains the name of a model variable. The columns are booleans that
+            are True if the variable has the corresponding property. The columns are:
+            is_state, is_choice, is_continuous, is_discrete, is_sparse, columns are:
+            is_state, is_choice, is_continuous, is_discrete, is_sparse, is_dense.
+        functions: Dictionary that maps names of functions to functions. The functions
+            differ from the user functions in that they all except the filter functions
+            take ``params`` as keyword argument. If the original function depended on
+            model parameters, those are automatically extracted from ``params`` and
+            passed to the original function. Otherwise, the ``params`` argument is
+            simply ignored.
+        function_info: A table with information about all functions in the model. The
+            index contains the name of a function. The columns are booleans that are
+            True if the function has the corresponding property. The columns are:
+            is_filter, is_constraint, is_next.
+        params: Dict of model parameters.
+        n_periods: Number of periods.
+        random_utility_shocks: Type of random utility shocks.
 
     """
 
@@ -106,4 +106,4 @@ class InternalModel:
     params: ParamsDict
     n_periods: int
     # Not properly processed yet
-    shocks: Shock
+    random_utility_shocks: ShockType

src/lcm/process_model.py

@@ -16,7 +16,7 @@
 )
 from lcm.interfaces import InternalModel
 from lcm.model import Model
-from lcm.typing import ParamsDict, Shock
+from lcm.typing import ParamsDict, ShockType
 
 
 def process_model(user_model: Model) -> InternalModel:
@@ -66,7 +66,8 @@ def process_model(user_model: Model) -> InternalModel:
         functions=functions,
         function_info=function_info,
         params=params,
-        shocks=Shock.NONE,  # currently no additive utility shocks are supported
+        # currently no additive utility shocks are supported
+        random_utility_shocks=ShockType.NONE,
         n_periods=user_model.n_periods,
     )
 

src/lcm/simulate.py

@@ -14,12 +14,12 @@
 
 def simulate(
     params,
+    initial_states,
     state_indexers,
     continuous_choice_grids,
     compute_ccv_policy_functions,
     model,
     next_state,
-    initial_states,
     logger,
     solve_model=None,
     vf_arr_list=None,
@@ -30,6 +30,8 @@ def simulate(
 
     Args:
         params (dict): Dict of model parameters.
+        initial_states (list): List of initial states to start from. Typically from the
+            observed dataset.
         state_indexers (list): List of dicts of length n_periods. Each dict contains one
             or several state indexers.
         continuous_choice_grids (list): List of dicts of length n_periods. Each dict
@@ -41,8 +43,6 @@ def simulate(
             state and choice variables. For stochastic variables, it returns a random
             draw from the distribution of the next state.
         model (Model): Model instance.
-        initial_states (list): List of initial states to start from. Typically from the
-            observed dataset.
         logger (logging.Logger): Logger that logs to stdout.
         solve_model (callable): Function that solves the model. Is only required if
             vf_arr_list is not provided.

src/lcm/typing.py

@@ -10,7 +10,7 @@
 ParamsDict = dict[str, Any]
 
 
-class Shock(Enum):
+class ShockType(Enum):
     """Type of shocks."""
 
     EXTREME_VALUE = "extreme_value"

tests/test_discrete_problem.py

@@ -15,7 +15,7 @@
     _solve_discrete_problem_no_shocks,
     get_solve_discrete_problem,
 )
-from lcm.typing import Shock
+from lcm.typing import ShockType
 
 
 @pytest.fixture
@@ -50,7 +50,7 @@ def test_aggregation_without_shocks(cc_values, segment_info, collapse, n_extra_a
     )
 
     solve_discrete_problem = get_solve_discrete_problem(
-        random_utility_shock_type=Shock.NONE,
+        random_utility_shock_type=ShockType.NONE,
         variable_info=var_info,
         is_last_period=False,
         choice_segments=segment_info,
@@ -155,7 +155,7 @@ def test_get_solve_discrete_problem_illustrative():
     )  # leads to choice_axes = [1]
 
     solve_discrete_problem = get_solve_discrete_problem(
-        random_utility_shock_type=Shock.NONE,
+        random_utility_shock_type=ShockType.NONE,
         variable_info=variable_info,
         is_last_period=False,
         choice_segments=None,

tests/test_model_functions.py

@@ -34,7 +34,7 @@ def h():
         functions={"f": f, "g": g, "h": h},
         function_info=function_info,
         params=None,
-        shocks=None,
+        random_utility_shocks=None,
         n_periods=None,
     )
     combined_constraint = get_combined_constraint(model)

tests/test_next_state.py

@@ -68,7 +68,7 @@ def f_weight_b(state):  # noqa: ARG001
         gridspecs=None,
         variable_info=None,
         params=None,
-        shocks=None,
+        random_utility_shocks=None,
         n_periods=1,
     )
 

tests/test_state_space.py

@@ -68,7 +68,7 @@ def absorbing_retirement_filter(retirement, lagged_retirement):
         variable_info=None,
         functions=functions,
         function_info=function_info,
-        shocks=None,
+        random_utility_shocks=None,
         n_periods=100,
         params={},
     )