def run()

in jupyter/comparison-to-datasketch/cardinality_error_experiment.py [0:0]
48 lines of code
10 McCabe index (conditional complexity)

    def run(self):
        """Runs the experiment"""
        seq_start = np.uint64(2345234)
        distinct_number = np.uint64(3462)
        previous_log_trial_index = 0
        ds_all_results = np.zeros((self.num_trials, len(self.plot_points)))
        d_all_results = np.zeros_like(ds_all_results)

        for trial in range(1, self.num_trials+1):
            #print(f"Trial = {trial}\t{self._is_power_of_two(trial)}")
            # Initialise the sketches
            hll = ds.hll_sketch(self.sketch_lgk, ds.HLL_8)
            h = d.HyperLogLogPlusPlus(p=self.sketch_lgk, hashfunc=lambda x: mmh3.hash64(x, signed=False)[0])
            plot_point_index = 0  # Return to the start of the plot points list to generate the data
            plot_point_value = self.plot_points[plot_point_index]
            total_updates = 0
            seq_start += distinct_number  # Start a new input sequence

            # Temporary result data structure
            ds_results = np.zeros((len(self.plot_points),))
            d_results = np.zeros_like(ds_results)


            for new_number in distinct_number_sequence(seq_start):
                hll.update(new_number)
                h.update(str(new_number).encode('utf8'))
                total_updates += 1
                #print(f"Trial: {trial:<5} updates: {total_updates:<5}Index: {plot_point_index:<5} Value: {plot_point_value:<5}\n")
                if total_updates == plot_point_value:
                    ds_results[plot_point_index] = (hll.get_estimate() - plot_point_value) / plot_point_value
                    d_results[plot_point_index] = (h.count() - plot_point_value) / plot_point_value
                    plot_point_index += 1
                    #print(f"PPI:{plot_point_index:<3}PPV:{plot_point_value}")

                    if plot_point_index < len(self.plot_points):
                        plot_point_value = self.plot_points[plot_point_index]
                    else:
                        #print("Already at the end")
                        break

            # After the break statement, control returns here.  Now need to decide whether to write or continue.
            ds_all_results[trial-1, :] = ds_results # subtract 1 as we use 1-based indexing for the trial count.
            d_all_results[trial - 1, :] = d_results  # subtract 1 as we use 1-based indexing for the trial count.
            if self._is_power_of_two(trial) and trial > 1:
                # write the array only a logarithmic number of times
                temporary_ds_results = ds_all_results[0:trial, : ]
                temporary_d_results = d_all_results[0:trial, :]
                print(f"#################### PARTIAL RESULTS FOR {trial} TRIALS: DATASKETCHES ####################")
                previous_log_trial_index = trial
                self.DataSketches_results_df = pd.DataFrame(temporary_ds_results.T, index=self.DataSketches_results_df.index, columns=np.arange(trial).tolist())
                self.DataSketches_results_df.to_csv(
                    self.directory_name + "/DataSketches_hll" + self.file_extension + f"trials_{trial}.csv",
                    index_label="n")
                self.datasketch_results_df = pd.DataFrame(temporary_d_results.T,
                                                            index=self.datasketch_results_df.index,
                                                            columns=np.arange(trial).tolist())
                self.datasketch_results_df.to_csv(
                    self.directory_name + "/datasketch_hll" + self.file_extension + f"trials_{trial}.csv",
                    index_label="n"
                )
                print(self.DataSketches_results_df)

                print(f"#################### PARTIAL RESULTS FOR {trial} TRIALS: datasketch ####################")
                print(self.datasketch_results_df)