Merge pull request #1427 from sbenthall/i1373

allow use of string expressions instead of lambdas when definition new format models.

Documentation/CHANGELOG.md

@@ -8,13 +8,14 @@ For more information on HARK, see [our Github organization](https://github.com/e
 
 ## Changes
 
-### 0.15.2 (in development)
+### 0.16.0 (in development)
 
 Release Date: TBD
 
 #### Major Changes
 
-none
+- Allows structural equations in model files to be provided in string form [#1427](https://github.com/econ-ark/HARK/pull/1427)
+- Introduces `HARK.parser' module for parsing configuration files into models [#1427](https://github.com/econ-ark/HARK/pull/1427)
 
 #### Minor Changes
 
@@ -50,7 +51,7 @@ This release drops support for Python 3.8 and 3.9, consistent with SPEC 0, and a
 - Object-oriented solver code has been moved to /HARK/ConsumptionSaving/LegacyOOsolvers.py, for legacy support of downstream projects.
 - AgentTypeMonteCarloSimulator now requires model shock, parameter, and dynamics information to be organized into 'blocks'. The DBlock object is introduced. [#1411](https://github.com/econ-ark/HARK/pull/1411)
 - RBlock object allows for recursive composition of DBlocks in models, as demonstrated by the AgentTypeMonteCarloSimulator [#1417](https://github.com/econ-ark/HARK/pull/1417/)
-- Transtion, reward, state-rulle value function, decision value function, and arrival value function added to DBlock [#1417](https://github.com/econ-ark/HARK/pull/1417/)
+- Transtion, reward, state-rule value function, decision value function, and arrival value function added to DBlock [#1417](https://github.com/econ-ark/HARK/pull/1417/)
 - All methods that construct inputs for solvers are now functions that are specified in the dictionary attribute `constructors`. [#1410](https://github.com/econ-ark/HARK/pull/1410)
 - Such constructed inputs can use alternate parameterizations / formats by changing the `constructor` function and providing its arguments in `parameters`.
 - Move `HARK.datasets` to `HARK.Calibration` for better organization of data and calibration tools. [#1430](https://github.com/econ-ark/HARK/pull/1430)

HARK/model.py

@@ -11,6 +11,7 @@
 )
 from inspect import signature
 import numpy as np
+from HARK.parser import math_text_to_lambda
 from typing import Any, Callable, Mapping, List, Union
 
 
@@ -143,7 +144,16 @@ class DBlock:
 
     Parameters
     ----------
-    ...
+    shocks: Mapping(str, Distribution)
+        A mapping from variable names to Distribution objects,
+        representing exogenous shocks.
+
+    dynamics: Mapping(str, str or callable)
+        A dictionary mapping variable names to mathematical expressions.
+        These expressions can be simple functions, in which case the
+        argument names should match the variable inputs.
+        Or these can be strings, which are parsed into functions.
+
     """
 
     name: str = ""
@@ -152,6 +162,11 @@ class DBlock:
     dynamics: dict = field(default_factory=dict)
     reward: dict = field(default_factory=dict)
 
+    def __post_init__(self):
+        for v in self.dynamics:
+            if isinstance(self.dynamics[v], str):
+                self.dynamics[v] = math_text_to_lambda(self.dynamics[v])
+
     def get_shocks(self):
         return self.shocks
 

HARK/models/consumer.py

@@ -53,7 +53,7 @@
             "b": lambda k, R, PermGroFac: k * R / PermGroFac,
             "m": lambda b, theta: b + theta,
             "c": Control(["m"]),
-            "a": lambda m, c: m - c,
+            "a": "m - c",
         },
         "reward": {"u": lambda c, CRRA: c ** (1 - CRRA) / (1 - CRRA)},
     }

HARK/parser.py

@@ -0,0 +1,26 @@
+from sympy.utilities.lambdify import lambdify
+from sympy.parsing.sympy_parser import parse_expr
+
+
+class Expression:
+    def __init__(self, text):
+        self.txt
+        self.expr = parse_expr(text)
+        self.npf = self.func()
+
+        # first derivatives.
+        self.grad = {
+            sym.__str__(): self.expr.diff(sym) for sym in list(self.expr.free_symbols)
+        }
+
+    def func(self):
+        return lambdify(list(self.expr.free_symbols), self.expr, "numpy")
+
+
+def math_text_to_lambda(text):
+    """
+    Returns a function represented by the given mathematical text.
+    """
+    expr = parse_expr(text)
+    func = lambdify(list(expr.free_symbols), expr, "numpy")
+    return func