Reworked examples

Author: jmafoster1

causal_testing/testing/causal_test_case.py

@@ -39,9 +39,9 @@ def __init__(
         :param expected_causal_effect: The expected causal effect (Positive, Negative, No Effect).
         :param control_value: The control value for the treatment variable (before intervention).
         :param treatment_value: The treatment value for the treatment variable (after intervention).
-        :param estimate_type: A string which denotes the type of estimate to return
-        :param effect_modifier_configuration:
-        :param estimator: An Estimator class instance
+        :param estimate_type: A string which denotes the type of estimate to return.
+        :param effect_modifier_configuration: The assignment of the effect modifiers to use for estimates.
+        :param estimator: An Estimator class object
         """
         self.base_test_case = base_test_case
         self.control_value = control_value

causal_testing/testing/causal_test_suite.py

@@ -31,18 +31,20 @@ def add_test_object(
         self,
         base_test_case: BaseTestCase,
         causal_test_case_list: Iterable[CausalTestCase],
-        estimators_classes: Iterable[Type[Estimator]],
+        estimators: Iterable[Type[Estimator]],
         estimate_type: str = "ate",
     ):
         """
         A setter object to allow for the easy construction of the dictionary test suite structure
 
         :param base_test_case: A BaseTestCase object consisting of a treatment variable, outcome variable and effect
         :param causal_test_case_list: A list of causal test cases to be executed
-        :param estimators_classes: A list of estimators, one for each causal test case
+        :param estimators: A list of estimator classes (NOT instances) to be used to execute each of the test cases.
+        Each estimator will be applied to each test case, so this will typically just be a single element list.
+        However, if you want to compare the outputs of different estimators, you may include more than one class here.
         :param estimate_type: A string which denotes the type of estimate to return
         """
-        test_object = {"tests": causal_test_case_list, "estimators": estimators_classes, "estimate_type": estimate_type}
+        test_object = {"tests": causal_test_case_list, "estimators": estimators, "estimate_type": estimate_type}
         self.data[base_test_case] = test_object
 
     def execute_test_suite(

examples/covasim_/vaccinating_elderly/example_vaccine.py

@@ -50,7 +50,6 @@ def setup_test_case(verbose: bool = False):
             cum_vaccinations,
             max_doses,
         },
-        constraints={pop_size.z3 == 50000, pop_infected.z3 == 1000, n_days.z3 == 50},
     )
 
     # 4. Construct a causal specification from the scenario and causal DAG

examples/poisson-line-process/README.md

@@ -6,6 +6,7 @@ To run this case study:
 1. Ensure all project dependencies are installed by running `pip install .` in the top level directory
    (instructions are provided in the project README).
 2. Change directory to `causal_testing/examples/poisson-line-process`.
-3. Run the command `python example_poisson_process.py`
+3. Run the command `python example_pure_python.py` to demonstrate causal testing using pure python.
+3. Run the command `python example_json_frontend.py` to demonstrate the same causal tests using JSON.
 
 This should print a series of causal test results and produce two CSV files. `intensity_num_shapes_results_random_1000.csv` corresponds to table 1, and `width_num_shapes_results_random_1000.csv` relates to our findings regarding the relationship of width and `P_u`.

examples/poisson/causal_tests.json renamed to examples/poisson-line-process/causal_tests.json

@@ -3,7 +3,8 @@
     {
       "name": "width__num_lines_abs",
       "mutations": {"width": "Increase"},
-      "estimator": "WidthHeightEstimator",
+      "estimator": "LinearRegressionEstimator",
+      "formula": "num_lines_abs ~ I(intensity * (width + height))",
       "estimate_type": "ate",
       "effect_modifiers": ["intensity", "height"],
       "expected_effect": {"num_lines_abs": "PoissonWidthHeight"},
@@ -138,7 +139,8 @@
     {
       "name": "height__num_lines_abs",
       "mutations": {"height": "Increase"},
-      "estimator": "WidthHeightEstimator",
+      "estimator": "LinearRegressionEstimator",
+      "formula": "num_lines_abs ~ I(intensity * (width + height))",
       "estimate_type": "ate",
       "effect_modifiers": ["intensity", "width"],
       "expected_effect": {"num_lines_abs": "PoissonWidthHeight"},
@@ -192,7 +194,8 @@
     {
       "name": "intensity__num_lines_abs",
       "mutations": {"intensity": "Increase"},
-      "estimator": "WidthHeightEstimator",
+      "estimator": "LinearRegressionEstimator",
+      "formula": "num_lines_abs ~ I(intensity * (width + height))",
       "effect_modifiers": ["height", "width"],
       "estimate_type": "ate",
       "expected_effect": {"num_lines_abs": "PoissonIntensity"},

examples/poisson-line-process/dag.dot

@@ -1,5 +1,4 @@
 digraph poisson_line_process {
-  rankdir=LR;
   width -> num_lines_abs;
   width -> num_shapes_abs;
   width -> num_lines_unit;