update layout

build/lib/wiggum/__init__.py

@@ -0,0 +1,43 @@
+import pandas as pd
+import numpy as np
+import scipy.stats as stats
+import itertools as itert
+
+#from .detectors import upper_triangle_df, upper_triangle_element, isReverse,
+# from .detectors import detect_simpsons_paradox
+
+
+from .ranking_processing import DEFAULT_SP_DEF, trend_quality_sp
+
+# from .ranking_processing import (mark_designed_rows, compute_angle,
+#     compute_slope_all, add_slope_sp, add_angle_col, get_SP_views,
+#     count_sp_views, add_slope_cols, add_view_count,
+#     add_weighted, add_view_score,get_trend_row,label_SP_rows,get_views,
+#     get_SP_rows)
+
+
+# from .data_augmentation import add_quantile, add_all_dpgmm
+
+
+from .labeled_dataframe import (LabeledDataFrame, possible_roles, var_types,
+                            simple_type_mapper)
+
+# trend types
+from .trends import (All_Pearson, Continuous_Pearson, Spearman_Correlation,
+                    Kendall_Correlation, Mean_Rank_Trend,Linear_Trend,
+                    All_Linear_Trend,Binary_Pearson_Trend,Binary_Mean_Rank_Trend,
+                    Median_Rank_Trend,Binary_Median_Rank_Trend,
+                    Binary_Accuracy_Trend, Binary_TPR_Trend,Binary_PPV_Trend,
+                    Binary_TNR_Trend, Binary_FDR_Trend, Binary_F1_Trend,
+                    Binary_Error_Trend,Binary_FNR_Trend)
+
+#tren lists
+from .trends import (all_trend_types,default_binary_trends,default_distance_trends)
+
+# __all__ = ['detect_simpsons_paradox','mark_designed_rows', 'compute_angle',
+#             'add_slope_sp','compute_slope_all','add_angle_col','count_sp_views',
+#             'add_slope_cols','add_view_count',
+#             'add_weighted','get_subgroup_trends_1lev','add_all_dpgmm',
+#             'get_subgroup_trends_2lev','add_quantile','get_correlations',
+#             'add_view_score','get_trend_row','label_SP_rows','get_views',
+#             'get_SP_rows']

build/lib/wiggum/auditing_reporting.py

@@ -0,0 +1,194 @@
+import pandas as pd
+import numpy as np
+import scipy.stats as stats
+import itertools as itert
+
+
+class _AuditReporting():
+    """
+    This mixin class contians methods for reproducible auditing and report generation
+    """
+
+    def annotate(self,row_idx, annotate_col, comment):
+        '''
+        add text of comment to the annotate_col column of result_df in the
+        row_idx row
+
+        Parameters
+        -----------
+        row_idx : integer
+            row index to annotate
+        annotate_col : string
+            name of annotation column (may be existing or new)
+        comment : string or number
+            content to add as annotation
+
+        Returns
+        -------
+        changed row
+        '''
+
+        # if new annotation column, initialize it empty
+        if not(annotate_col in self.result_df.columns):
+            self.result_df[annotate_col] = ''
+
+        # add specific annotation
+        self.result_df.loc[row_idx,annotate_col] = comment
+
+        return self.result_df.loc[row_idx]
+
+
+    def filter_annotate(self,feat1 = None,feat2 = None,group_feat= None,
+                            subgroup= None,subgroup2= None,trend_type=None,
+                            annotate_col = None, comment = 'x'):
+        '''
+        add text of comment to the annotate_col column of result_df in the
+        rows specified by a filter using the trend related variables
+
+        Parameters
+        -----------
+        feat1 : str, list, or  None
+            trend variable name or None to include all
+        feat2 : str, list, or  None
+            trend variable name or None to include all
+        group_feat : str, list, or  None
+            groupoby variable name or None to include all
+        subgroup : str, list, or  None
+            value of groupby_feat or or None to include all
+        annotate_col : string
+            name of annotation column (may be existing or new) or none to be
+        auto generated from filter
+        comment : string or number
+            content to add as annotation
+
+        Returns
+        -------
+        changed rows
+        '''
+        if annotate_col is None:
+            filter_vars = [feat1, feat2, group_feat, subgroup, subgroup2,
+                            trend_type]
+            filter_params = [var for var in filter_vars if not(var is None)]
+            annotate_col = '_'.join(filter_params)
+
+        filt_df = self.get_trend_rows(feat1, feat2, group_feat, subgroup,
+                        subgroup2, trend_type)
+
+        for row in filt_df.index:
+            self.annotate(row, annotate_col,comment)
+
+        filt_df[annotate_col] = comment
+
+        return filt_df
+
+    def delete_annotate(self,row_idx,annotate_col):
+        """
+        delete a comment in an annotate_col column of result_df in the
+        row and annotate column specified
+
+        Parameters
+        -----------
+        row_inx : integer
+            row index to delete
+        annotate_col : string
+            name of annotation column (should be existing already) to delete from
+        Returns
+        -------
+        row deleted from
+
+        """
+        self.result_df.loc[row_idx,annotate_col] = ''
+
+        return self.result_df.loc[row_idx]
+
+    def get_report_table(self,report_cols,report_rows):
+        """
+        generate a df that can be used for tables
+
+        Parameters
+        -----------
+        report_cols : list
+            list of integers representing column indexes for report
+        report_rows : list
+            list of integers representing row indexes for report
+        Returns
+        -------
+        rows and columns of the report
+        """
+        report_df = self.result_df.iloc[report_cols,report_rows]
+        return report_df
+
+    def detect_annotate():
+
+        pass
+
+    def save_report_table(self,report_cols,report_rows,filename):
+        """
+        generate a csv file to save the report
+
+        Parameters
+        -----------
+        report_cols : list
+            list of integers representing column indexes for report
+        report_rows : list
+            list of integers representing row indexes for report
+        filename : string
+            .csv filename to save the report to
+        Returns
+        -------
+        rows and columns written to the report
+        """
+        # call above and then save to csv
+        report_df = self.get_report_table(report_cols,report_rows)
+        report_df.to_csv(filename)
+        return 
+
+
+
+    def count_values_above_thres(self,column,threshold):
+        """
+        count all the values in a column above a certain threshold
+
+        Parameters
+        -----------
+        column : string
+            name of column to peform thresholding on
+        threshold : float
+            threshold that should be exceeded
+        Returns
+        -----------
+        number of values above the threshold
+        """
+        valueCount = len(self.result_df[self.result_df[column] > threshold])
+        return valueCount
+    def result_df_stat_summary(self,filename):
+        """
+        generate and save to file(s) summary statistics on the results
+
+        Parameters
+        -----------
+        filename : string
+            the filename of the saved summary files
+
+        Returns
+        -------
+
+
+        """
+
+        # generate state tables and save each individually
+        state_table = self.result_df
+        state_table.to_csv(filename + ".csv")
+
+        # generate narrative for a text file summary
+        trendNum = state_table['trend_type'].value_counts()
+        dataGroups = state_table['subgroup'].value_counts()
+        new_df = self.result_df[((self.result_df["agg_trend"] < 0) == (self.result_df["subgroup_trend"] < 0)) == False]
+        total_rows = len(new_df.index)
+        # thresholds that you would like to add to the report
+        valueCount = self.count_values_above_thres('distance',.5)
+        valueCount2 = self.count_values_above_thres('distance',.2)
+        valueCount3 = self.count_values_above_thres('distance',.3)
+        outputFile = open(filename + ".txt",'w')
+        outputFile.write("Summary Statistics for " + filename + "\n"  + "\n" +"Subgroups Present: \n" + dataGroups.to_string() +"\n\nTrends Detected: \n"+ trendNum.to_string() + "\n" + "\nNumber of Reversals found: " + str(total_rows) + "\n" +"\nDistance above .2 threshold: " + str(valueCount2)+"\nDistance above .3 threshold: " + str(valueCount)+ "\nDistance above .5 threshold: " + str(valueCount))
+        return