Merge pull request #626 from MITx/feature/victor/answer-export

Feature/victor/answer export
2012-09-05 11:04:34 -07:00
parent 7702907662 a8cd4633c0
commit 579544ec6a
5 changed files with 109 additions and 12 deletions
--- a/lms/djangoapps/courseware/grades.py
+++ b/lms/djangoapps/courseware/grades.py
@@ -4,11 +4,14 @@ from __future__ import division
 import random
 import logging

+from collections import defaultdict
 from django.conf import settings
+from django.contrib.auth.models import User

 from models import StudentModuleCache
 from module_render import get_module, get_instance_module
 from xmodule import graders
+from xmodule.capa_module import CapaModule
 from xmodule.course_module import CourseDescriptor
 from xmodule.graders import Score
 from models import StudentModule
@@ -24,6 +27,71 @@ def yield_module_descendents(module):
        stack.extend( next_module.get_display_items() )
        yield next_module

+def yield_problems(request, course, student):
+    """
+    Return an iterator over capa_modules that this student has
+    potentially answered.  (all that student has answered will definitely be in
+    the list, but there may be others as well).
+    """
+    grading_context = course.grading_context
+    student_module_cache = StudentModuleCache(course.id, student, grading_context['all_descriptors'])
+
+    for section_format, sections in grading_context['graded_sections'].iteritems():
+        for section in sections:
+
+            section_descriptor = section['section_descriptor']
+
+            # If the student hasn't seen a single problem in the section, skip it.
+            skip = True
+            for moduledescriptor in section['xmoduledescriptors']:
+                if student_module_cache.lookup(
+                        course.id, moduledescriptor.category, moduledescriptor.location.url()):
+                    skip = False
+                    break
+
+            if skip:
+                continue
+
+            section_module = get_module(student, request,
+                                        section_descriptor.location, student_module_cache,
+                                        course.id)
+            if section_module is None:
+                # student doesn't have access to this module, or something else
+                # went wrong.
+                # log.debug("couldn't get module for student {0} for section location {1}"
+                #           .format(student.username, section_descriptor.location))
+                continue
+
+            for problem in yield_module_descendents(section_module):
+                if isinstance(problem, CapaModule):
+                    yield problem
+
+def answer_distributions(request, course):
+    """
+    Given a course_descriptor, compute frequencies of answers for each problem:
+
+    Format is:
+
+    dict: (problem url_name, problem display_name, problem_id) -> (dict : answer ->  count)
+
+    TODO (vshnayder): this is currently doing a full linear pass through all
+    students and all problems.  This will be just a little slow.
+    """
+
+    counts = defaultdict(lambda: defaultdict(int))
+
+    enrolled_students = User.objects.filter(courseenrollment__course_id=course.id)
+
+    for student in enrolled_students:
+        for capa_module in yield_problems(request, course, student):
+            for problem_id in capa_module.lcp.student_answers:
+                answer = capa_module.lcp.student_answers[problem_id]
+                key = (capa_module.url_name, capa_module.display_name, problem_id)
+                counts[key][answer] += 1
+
+    return counts
+
+
 def grade(student, request, course, student_module_cache=None, keep_raw_scores=False):
    """
    This grades a student as quickly as possible. It retuns the
--- a/lms/djangoapps/courseware/views.py
+++ b/lms/djangoapps/courseware/views.py
@@ -1,7 +1,9 @@
+import csv
 import json
 import logging
 import urllib
 import itertools
+import StringIO

 from functools import partial

@@ -219,9 +221,9 @@ def jump_to(request, course_id, location):

    # Rely on index to do all error handling and access control.
    return redirect('courseware_position',
-                    course_id=course_id, 
-                    chapter=chapter, 
-                    section=section, 
+                    course_id=course_id,
+                    chapter=chapter,
+                    section=section,
                    position=position)
@ensure_csrf_cookie
 def course_info(request, course_id):
@@ -342,7 +344,7 @@ def progress(request, course_id, student_id=None):
    # NOTE: To make sure impersonation by instructor works, use
    # student instead of request.user in the rest of the function.

-    # The pre-fetching of groups is done to make auth checks not require an 
+    # The pre-fetching of groups is done to make auth checks not require an
    # additional DB lookup (this kills the Progress page in particular).
    student = User.objects.prefetch_related("groups").get(id=student.id)

@@ -368,5 +370,3 @@ def progress(request, course_id, student_id=None):

    return render_to_response('courseware/progress.html', context)

-
-
--- a/lms/djangoapps/instructor/views.py
+++ b/lms/djangoapps/instructor/views.py
@@ -48,7 +48,7 @@ def instructor_dashboard(request, course_id):
    """Display the instructor dashboard for a course."""
    course = get_course_with_access(request.user, course_id, 'staff')

-    instructor_access = has_access(request.user, course, 'instructor')		# an instructor can manage staff lists
+    instructor_access = has_access(request.user, course, 'instructor')   # an instructor can manage staff lists

    msg = ''
    # msg += ('POST=%s' % dict(request.POST)).replace('<','&lt;')
@@ -99,7 +99,7 @@ def instructor_dashboard(request, course_id):
            msg += "git pull on %s:<p>" % data_dir
            msg += "<pre>%s</pre></p>" % escape(os.popen(cmd).read())
            track.views.server_track(request, 'git pull %s' % data_dir, {}, page='idashboard')
-            
+
    if 'Reload course' in action:
        log.debug('reloading %s (%s)' % (course_id, course))
        try:
@@ -144,6 +144,10 @@ def instructor_dashboard(request, course_id):
        return return_csv('grades_%s_raw.csv' % course_id,
                          get_student_grade_summary_data(request, course, course_id, get_raw_scores=True))

+    elif 'Download CSV of answer distributions' in action:
+        track.views.server_track(request, 'dump-answer-dist-csv', {}, page='idashboard')
+        return return_csv('answer_dist_%s.csv' % course_id, get_answers_distribution(request, course_id))
+
    elif 'List course staff' in action:
        group = get_staff_group(course)
        msg += 'Staff group = %s' % group.name
@@ -290,7 +294,7 @@ def grade_summary(request, course_id):
@ensure_csrf_cookie
@cache_control(no_cache=True, no_store=True, must_revalidate=True)
 def enroll_students(request, course_id):
-    ''' Allows a staff member to enroll students in a course.
+    """Allows a staff member to enroll students in a course.

    This is a short-term hack for Berkeley courses launching fall
    2012. In the long term, we would like functionality like this, but
@@ -300,7 +304,7 @@ def enroll_students(request, course_id):

    It is poorly written and poorly tested, but it's designed to be
    stripped out.
-    '''
+    """

    course = get_course_with_access(request.user, course_id, 'staff')
    existing_students = [ce.user.email for ce in CourseEnrollment.objects.filter(course_id=course_id)]
@@ -328,6 +332,28 @@ def enroll_students(request, course_id):
                                                       'rejected_students': rejected_students,
                                                       'debug': new_students})

+
+def get_answers_distribution(request, course_id):
+    """
+    Get the distribution of answers for all graded problems in the course.
+
+    Return a dict with two keys:
+    'header': a header row
+    'data': a list of rows
+    """
+    course = get_course_with_access(request.user, course_id, 'staff')
+
+    dist = grades.answer_distributions(request, course)
+
+    d = {}
+    d['header'] = ['url_name', 'display name', 'answer id', 'answer', 'count']
+
+    d['data'] = [[url_name, display_name, answer_id, a, answers[a]]
+                 for (url_name, display_name, answer_id), answers in dist.items()
+                 for a in answers]
+    return d
+
+
 #-----------------------------------------------------------------------------


--- a/lms/templates/courseware/instructor_dashboard.html
+++ b/lms/templates/courseware/instructor_dashboard.html
@@ -58,6 +58,9 @@ table.stat_table td {
    <input type="submit" name="action" value="Dump all RAW grades for all students in this course">
    <input type="submit" name="action" value="Download CSV of all RAW grades">

+    <p>
+    <input type="submit" name="action" value="Download CSV of answer distributions">
+
 %if instructor_access:
    <hr width="40%" style="align:left">
    <p>
--- a/lms/urls.py
+++ b/lms/urls.py
@@ -197,8 +197,8 @@ if settings.WIKI_ENABLED:
    )

 if settings.QUICKEDIT:
-	urlpatterns += (url(r'^quickedit/(?P<id>[^/]*)$', 'dogfood.views.quickedit'),)
-	urlpatterns += (url(r'^dogfood/(?P<id>[^/]*)$', 'dogfood.views.df_capa_problem'),)
+    urlpatterns += (url(r'^quickedit/(?P<id>[^/]*)$', 'dogfood.views.quickedit'),)
+    urlpatterns += (url(r'^dogfood/(?P<id>[^/]*)$', 'dogfood.views.df_capa_problem'),)

 if settings.ASKBOT_ENABLED:
    urlpatterns += (url(r'^%s' % settings.ASKBOT_URL, include('askbot.urls')), \