plotgrades: fixes for new blackboard settings, added --group-by

1 files changed, 43 insertions, 14 deletions

diff --git a/plotgrades.py b/plotgrades.py
index 165d0bc..f3c8baa 100755
--- a/plotgrades.py
+++ b/plotgrades.py
@@ -2,14 +2,16 @@
 import argparse
 import csv
 from functools import partial
+import itertools
 import matplotlib.pyplot as plt
 import numpy
+from operator import itemgetter
 import re
 
 def readcsv(filename):
     """Read a (BlackBoard) CSV file into a header and data"""
     with open(filename, 'r') as csvfile:
-        reader = csv.reader(csvfile, delimiter=',', quotechar='"')
+        reader = csv.reader(csvfile, delimiter='\t', quotechar='"')
         data = []
         header = next(reader)
         for row in reader:
@@ -37,9 +39,9 @@ def remove_zeros(matrix):
     want to distort the boxplots. This function removes zeros on a per-column
     basis."""
     m_new = []
-    for col in numpy.transpose(matrix):
+    for col in matrix:
         m_new.append([v for v in col if v != 0.0])
-    return numpy.transpose(m_new)
+    return m_new
 
 def normalise(matrix):
     """Normalise grades to a 0-100 range
@@ -53,15 +55,16 @@ def normalise(matrix):
 
 def strip_header(h):
     """Strip common BlackBoard additions in the CSV header"""
-    return h.split('instructor.download.column.total')[0]
+    return h.split('[Total Pts:')[0]
 
 def remove_empty_lists(headers, data):
     """Remove empty columns and the corresponding headers from a matrix"""
     new_headers, new_data = [], []
-    for h, d in zip(headers, data):
-        if d != []:
+    for h, col in zip(headers, numpy.transpose(data)):
+        if not all([x == 0.0 for x in col]):
             new_headers.append(h)
-            new_data.append(d)
+            new_data.append(col)
+
     return new_headers, new_data
 
 def header_regex_callback(header, regex='', invert=False):
@@ -97,23 +100,47 @@ def check_participant_callback(headers, data, participant_callback):
             new_data.append(d)
     return new_data
 
+def make_groups(headers, data, column):
+    i = 0
+    for h in headers:
+        if column in h:
+            break
+        i += 1
+
+    data = list(numpy.transpose(data))
+    data.sort(key=itemgetter(i))
+    data = [list(x) for _, x in itertools.groupby(data, itemgetter(i))]
+    new_headers = []
+    new_data = []
+    j = 0
+    for h in headers:
+        for group in data:
+            if not all([x[j] == 0.0 for x in group]):
+                new_headers.append(h + ' (group: ' + str(group[0][i]) + ')')
+                new_data.append([x[j] for x in group])
+        j += 1
+
+    return new_headers, new_data
+
 def plotgrades(headers, data, skip=0,
-        participant_callback=lambda x:True, header_callback=lambda x:True):
+        participant_callback=lambda x:True, header_callback=lambda x:True,
+        group_participants_column=None):
     """Plot grades corresponding to headers in a boxplot"""
     data = check_participant_callback(headers, data, participant_callback)
-    headers, data = headers[skip:], data[skip:]
+    headers, data = map(strip_header, headers[skip:]), data[skip:]
     data = parse_floats(data)
-    data = remove_zeros(data)
     headers, data = remove_empty_lists(headers, data)
+    if group_participants_column is not None:
+        headers, data = make_groups(headers, data, group_participants_column)
+    data = remove_zeros(data)
     headers, data = check_header_callback(headers, data, header_callback)
     data = normalise(data)
-    headers = [strip_header(h) for h in headers]
 
     if len(data) > 0:
         plt.boxplot(data)
     
         ax = plt.gca()
-        plt.xticks(range(0, len(data) + 1), [''] + headers, rotation=90)
+        plt.xticks(range(1, len(data) + 1), headers, rotation=90)
         ax.set_ylim([0,100])
     
         plt.show()
@@ -128,6 +155,8 @@ def parse_args():
             help='Restrict what grades are shown with a regex')
     pars.add_argument('-wh', '--where-has', metavar='regex', default='',
             help='Only count participants with a grade for a matching item')
+    pars.add_argument('-g', '--group-by', metavar='column', default=None,
+            help='Group participants by this column')
     pars.add_argument('-iw', '--invert-where', action='store_true',
             help='Invert --where regex')
     pars.add_argument('-iwh', '--invert-where-has', action='store_true',
@@ -146,8 +175,8 @@ def main():
             header_callback=partial(header_regex_callback,
                 regex=args.where, invert=args.invert_where),
             participant_callback=partial(participant_wherehas_callback,
-                regex=args.where_has, invert=args.invert_where_has))
+                regex=args.where_has, invert=args.invert_where_has),
+            group_participants_column=args.group_by)
 
 if __name__ == '__main__':
     main()
-