making SymptomsByBatchcodesTableFactoryTest pass

src/SymptomsByBatchcodesTableFactory.py

@@ -4,5 +4,33 @@ import pandas as pd
 class SymptomsByBatchcodesTableFactory:
 
     @staticmethod
-    def createSymptomsByBatchcodesTable(vaxTable, symptomsTable):
+    def createSymptomsByBatchcodesTable(VAERSVAX, VAERSSYMPTOMS):
-        return pd.DataFrame({'A': []})
+        return pd.merge(
+            SymptomsByBatchcodesTableFactory._get_VAERSVAX_WITH_VAX_LOTS(VAERSVAX),
+            SymptomsByBatchcodesTableFactory._getSymptomsTable(VAERSSYMPTOMS),
+            on = 'VAERS_ID').set_index(['VAX_LOT1', 'VAX_LOT2'])[['SYMPTOMS']]
+    
+    @staticmethod
+    def _get_VAERSVAX_WITH_VAX_LOTS(VAERSVAX):
+        return pd.concat(
+            [VAERSVAX, SymptomsByBatchcodesTableFactory._getVaxLotsTable(VAERSVAX)],
+            axis=1).drop_duplicates(subset=['VAX_LOT1', 'VAX_LOT2']).reset_index()
+
+    @staticmethod
+    def _getVaxLotsTable(VAERSVAX):
+        VAX_LOT_LIST_Table = VAERSVAX.groupby("VAERS_ID").agg(VAX_LOT_LIST = pd.NamedAgg(column = 'VAX_LOT', aggfunc = list))
+        return pd.DataFrame(
+            VAX_LOT_LIST_Table['VAX_LOT_LIST'].tolist(),
+            columns = ['VAX_LOT1', 'VAX_LOT2'],
+            index = VAX_LOT_LIST_Table.index)
+
+    @staticmethod
+    def _getSymptomsTable(VAERSSYMPTOMS):
+        return pd.concat(
+            [
+                VAERSSYMPTOMS['SYMPTOM1'],
+                VAERSSYMPTOMS['SYMPTOM2'],
+                VAERSSYMPTOMS['SYMPTOM3'],
+                VAERSSYMPTOMS['SYMPTOM4'],
+                VAERSSYMPTOMS['SYMPTOM5']
+            ]).dropna().drop_duplicates().to_frame(name = "SYMPTOMS").reset_index()

src/SymptomsByBatchcodesTableFactoryTest.py

@@ -40,8 +40,20 @@ class SymptomsByBatchcodesTableFactoryTest(unittest.TestCase):
         assert_frame_equal(
             symptomsByBatchcodesTable,
             TestHelper.createDataFrame(
-                columns = ['Blood pressure orthostatic abnormal', 'COVID-19', 'Coma', 'Computerised tomogram', 'Exposure to SARS-CoV-2', 'Head injury', 'Headache', 'Laboratory test', 'Magnetic resonance imaging', 'SARS-CoV-2 antibody test negative', 'SARS-CoV-2 test positive', 'Unresponsive to stimuli', 'X-ray'],
+                columns = ['SYMPTOMS'],
-                data = [  [1,                                     1,          1,      1,                        1,                        1,             1,         1,                 1,                            1,                                   1,                          1,                         1]],
+                data = [  ['Blood pressure orthostatic abnormal'],
+                          ['Head injury'],
+                          ['SARS-CoV-2 test positive'],
+                          ['COVID-19'],
+                          ['Headache'],
+                          ['Unresponsive to stimuli'],
+                          ['Coma'],
+                          ['Laboratory test'],
+                          ['X-ray'],
+                          ['Computerised tomogram'],
+                          ['Magnetic resonance imaging'],
+                          ['Exposure to SARS-CoV-2'],
+                          ['SARS-CoV-2 antibody test negative']],
                 index = pd.MultiIndex.from_tuples(
                     names =   ['VAX_LOT1', 'VAX_LOT2'],
-                    tuples = [['1808982',  'EW0175']])))
+                    tuples = [['1808982',  'EW0175']] * 13)))