__init__ add for relative import when testing, and starting on writin…

…g tests

evaluation/posts/2024_06_07/index.qmd

@@ -23,6 +23,12 @@ Started on research compendium...
 
 * Moved `.csv` files (`patients.csv`, `travel_matrix.csv`, `units.csv`) from `sim/` to `data/` and modified file paths in scripts accordingly (`model.py` and `units.py`)
 * Moved `.py` files (`sim_replicate.py` and `sim_single.py`) from main folder into `sim/` and modified import in notebook accordingly (`reproduction.ipynb`)
+* Created test for successful reproduction using pytest
+    * Amended `run_replications()` to accept alternative path to save results to
+    * Copied data from my successful reproduction from `output/` to `test/data`
+    * Copied parameters from `reproduction.ipynb` into `test_model.py`
+    * Add `__init__.py` to `sim/` and modified imports in each of the `.py` scripts accordingly
+    * Started on test to run the model and compare test results against saved results in the test folder
 
 ## Suggested changes for protocol/template
 

reproduction/reproduction.ipynb

@@ -61,7 +61,7 @@
      "name": "stdout",
      "output_type": "stream",
      "text": [
-      "Running 30 reps of base_3_month => 2, 0, 1, 5, 3, 7, 11, 10, 13, 9, 8, 12, 6, 14, 17, 4, 16, 18, 15, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, "
+      "Running 30 reps of base_3_month => 0, 1, 2, 7, 4, 8, 6, 3, 5, 11, 12, 10, 15, 9, 16, 17, 14, 13, 19, 18, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, "
      ]
     },
     {

reproduction/sim/model.py

@@ -2,10 +2,10 @@
 import pandas as pd
 import numpy as np
 
-from sim.patient import Patient
-from sim.units import Dialysis_units
-from sim.allocation import AllocatePatients
-from sim.audit import Audit
+from .patient import Patient
+from .units import Dialysis_units
+from .allocation import AllocatePatients
+from .audit import Audit
 
 class Population:
     """

reproduction/sim/sim_replicate.py

@@ -1,12 +1,13 @@
-from sim.end_trial_analysis import EndTrialAnalysis
-from sim.helper_functions import expand_multi_index
-from sim.model import DialysisSim
-from sim.parameters import Scenario, Uniform, Normal
+from .end_trial_analysis import EndTrialAnalysis
+from .helper_functions import expand_multi_index
+from .model import DialysisSim
+from .parameters import Scenario, Uniform, Normal
 from joblib import Parallel, delayed
 import pandas as pd
 
 
-def run_replications(scenarios, number_of_replications=30, base_random_set=0):
+def run_replications(scenarios, number_of_replications=30, base_random_set=0,
+                     output_folder='output'):
     """
     Main simulation code. Calls multiple runs of prescribed scenarios.
 
@@ -34,6 +35,8 @@ def run_replications(scenarios, number_of_replications=30, base_random_set=0):
     base_random_set : int, optional
         To create the random number set for each replication, the replication
         number is added to this value. The default is 0.
+    output_folder : str, optional
+        Path to save .csv result files to.
 
     Returns
     -------
@@ -65,13 +68,13 @@ def run_replications(scenarios, number_of_replications=30, base_random_set=0):
 
         # Save as CSV
         p_audits.to_csv(
-            f'output/{name}_reps_{N_REPS}_patient_audit.csv', index=False)
+            f'{output_folder}/{name}_reps_{N_REPS}_patient_audit.csv', index=False)
         u_audits.to_csv(
-            f'output/{name}_reps_{N_REPS}_unit_audit.csv', index=False)
+            f'{output_folder}/{name}_reps_{N_REPS}_unit_audit.csv', index=False)
         d_audits.to_csv(
-            f'output/{name}_reps_{N_REPS}_displaced_audit.csv', index=False)
+            f'{output_folder}/{name}_reps_{N_REPS}_displaced_audit.csv', index=False)
         i_audits.to_csv(
-            f'output/{name}_reps_{N_REPS}_inpatient_audit.csv', index=False)
+            f'{output_folder}/{name}_reps_{N_REPS}_inpatient_audit.csv', index=False)
 
         # Run analysis after all replicate runs in a scenario
         analysis = EndTrialAnalysis(name, p_audits, u_audits, d_audits, i_audits)

reproduction/sim/sim_single.py

@@ -1,5 +1,5 @@
-from sim.model import DialysisSim
-from sim.parameters import Scenario, Uniform
+from .model import DialysisSim
+from .parameters import Scenario, Uniform
 
 
 if __name__ == '__main__':

reproduction/tests/test_model.py

@@ -0,0 +1,78 @@
+'''
+Model testing
+
+This module contains tests to confirm whether the simulation model is
+producing consistent results.
+'''
+
+import pytest
+import pandas as pd
+import sim.sim_replicate as sim
+from sim.parameters import Scenario
+from pathlib import Path
+
+DISPLACED = 'base_3_month_reps_30_displaced_audit.csv'
+PATH_DISPLACED_1 = Path(__file__).parent.joinpath('data', DISPLACED)
+PATH_DISPLACED_2 = Path(__file__).parent.joinpath('data', DISPLACED)
+
+@pytest.mark.parametrize(
+    'displaced_audit',
+    [
+        (PATH_DISPLACED_1),
+        (PATH_DISPLACED_2),
+    ],
+)
+
+def run_model():
+    '''
+    Run model with parameters from computational reproducibility assessment
+    '''
+    scenarios = {}
+    scenarios['base_3_month'] = Scenario(
+        run_length=150,
+        proportion_pos_requiring_inpatient=0.6)
+    number_of_replications = 30
+    base_random_set = 2700
+    output_folder='test_output'
+    sim.run_replications(
+        scenarios, number_of_replications, base_random_set, output_folder)
+
+#def test_displaced(displaced_audit):
+#    '''
+#    Test that results are consistent with those from the computational
+#    reproducibility assessment
+#    '''
+#    displaced_data = pd.read_csv(displaced_audit)
+
+'''
+# Import model
+import sim.sim_replicate as sim
+from sim.parameters import Scenario
+
+def test_reproduction():
+    # Test run model as set-up for computational reproducibility assessment.
+    # Parameters
+    scenarios = {}
+    scenarios['base_3_month'] = Scenario(
+        run_length=150,
+        proportion_pos_requiring_inpatient=0.6)
+    number_of_replications = 30
+    base_random_set = 2700
+    output_folder='test_output'
+
+    # Import expected results
+    displaced = pd.read_csv('data/base_3_month_reps_30_displaced_audit.csv')
+    inpatient = pd.read_csv('data/base_3_month_reps_30_inpatient_audit.csv')
+    patient = pd.read_csv('data/base_3_month_reps_30_patient_audit.csv')
+    unit = pd.read_csv('data/base_3_month_reps_30_unit_audit.csv')
+
+    # Run the simulation
+    sim.run_replications(
+        scenarios, number_of_replications, base_random_set, output_folder)
+
+    # Import the test results
+    test_displaced = pd.read_csv('test_output/base_3_month_reps_30_displaced_audit.csv')
+
+    # Compare against actual results
+    assert pd.testing.assert_frame_equal(displaced, test_displaced)
+'''