Add ACAlgorithm python example

examples/CMakeLists.txt

@@ -7,12 +7,14 @@ add_custom_command(TARGET copy-python-examples
             copy ${CMAKE_SOURCE_DIR}/examples/dedupe.py
                  ${CMAKE_SOURCE_DIR}/examples/mine_typos.py
                  ${CMAKE_SOURCE_DIR}/examples/anomaly_detection.py
+                 ${CMAKE_SOURCE_DIR}/examples/algebraic_constraints.py
                  ${CMAKE_BINARY_DIR}/examples
         COMMAND ${CMAKE_COMMAND} -E
             copy ${CMAKE_SOURCE_DIR}/examples/datasets/duplicates.csv
                  ${CMAKE_SOURCE_DIR}/examples/datasets/Workshop.csv
                  ${CMAKE_SOURCE_DIR}/examples/datasets/cargo_data_1.csv
                  ${CMAKE_SOURCE_DIR}/examples/datasets/cargo_data_2.csv
                  ${CMAKE_SOURCE_DIR}/examples/datasets/cargo_data_3.csv
+                 ${CMAKE_SOURCE_DIR}/examples/datasets/cargo_march.csv
                  ${CMAKE_BINARY_DIR}/examples/datasets
 )

examples/algebraic_constraints.py

@@ -0,0 +1,50 @@
+import desbordante
+import pandas
+import operator
+
+TABLE = 'datasets/cargo_march.csv'
+HEADER = 0
+SEPARATOR = ','
+P_FUZZ = 0.85
+FUZZINESS = 0.2
+BUMPS_LIMIT = 0
+WEIGHT = 0.1
+BIN_OPERATION = '-'
+AC_SEED = 11
+ITERATIONS_LIMIT = 4
+OPERATIONS = {
+    '+': (operator.add, 'Sum'),
+    '-': (operator.sub, 'Difference'),
+    '*': (operator.mul, 'Product'),
+    '/': (operator.truediv, 'Ratio'),
+}
+operation, operation_name = OPERATIONS[BIN_OPERATION]
+
+algo = desbordante.ACAlgorithm()
+
+df = pandas.read_csv(TABLE, sep=SEPARATOR, header=HEADER)
+df_without_id = df[['Delivery date', 'Dispatch date']]
+
+algo.load_data(df=df_without_id)
+
+algo.execute(p_fuzz=P_FUZZ, fuzziness=FUZZINESS, bumps_limit=BUMPS_LIMIT, weight=WEIGHT,
+             bin_operation=BIN_OPERATION, ac_seed=AC_SEED, iterations_limit=ITERATIONS_LIMIT)
+
+ac_ranges = algo.get_ac_ranges()
+for ac_range in ac_ranges:
+    l_col = df_without_id.columns[ac_range.column_indices[0]]
+    r_col = df_without_id.columns[ac_range.column_indices[1]] 
+    print(f'Discovered ranges for ({l_col} {BIN_OPERATION} {r_col}) are:')
+    print(ac_range.ranges)
+
+ac_exceptions = algo.get_ac_exceptions()
+print()
+print(f'Rows in which the result of the chosen operation ({BIN_OPERATION}) is outside of discovered ranges:')
+for ac_exception in ac_exceptions:
+    id, delivery_date, dispatch_date = df.iloc[ac_exception.row_index]
+    print(f'id: {id}')
+    print(f'Dispatch date: {dispatch_date}')
+    print(f'Delivery date: {delivery_date}')
+    print(f'{operation_name}: {operation(delivery_date, dispatch_date)}')
+    print()
+

examples/datasets/cargo_march.csv

@@ -0,0 +1,45 @@
+id,Delivery date,Dispatch date
+0,3,1
+1,8,1
+2,8,2
+3,10,4
+4,7,4
+5,14,10
+6,21,17
+7,30,1
+8,30,27
+9,26,22
+10,25,20
+11,25,20
+12,19,14
+13,16,11
+14,16,1
+15,26,4
+16,23,7
+17,26,10
+18,30,9
+19,26,6
+20,23,3
+21,20,1
+22,23,4
+23,27,8
+24,29,11
+25,28,10
+26,18,7
+27,28,10
+28,19,1
+29,20,3
+30,22,11
+31,9,2
+32,29,12
+33,27,5
+34,22,5
+35,30,14
+36,25,9
+37,25,9
+38,30,14
+39,16,11
+40,30,15
+41,20,1
+42,30,9
+43,29,12