def timeseries_grade_percentiles(c, assignment_name, num_points=40):
"""
Returns a timeseries of grades with percentiles. Here is an example:
[["2015-07-17 19:00:36-0700", 0.0, 0.0, 0.0, ... 0.0, 0.0],
["2015-07-17 19:10:36-0700", 0.0, 0.0, 0.0, ... 1.0, 2.0],
["2015-07-17 19:20:36-0700", 0.0, 0.0, 0.0, ... 3.0, 4.0],
["2015-07-17 19:30:36-0700", 0.0, 0.0, 0.5, ... 5.0, 6.0],
["2015-07-17 19:40:36-0700", 0.0, 0.0, 1.0, ... 7.0, 8.0]]
"""
data_keys = range(0, 105, 5)
assignment = get_assignment_by_name(assignment_name)
if not assignment:
return
# There is a slight problem that because of DST, ordering by "started" may not always
# produce the correct result. When the timezone changes, lexicographical order does not
# match the actual order of the times. However, this only happens once a year in the middle
# of the night, so f**k it.
c.execute(
"""SELECT source, score, started FROM builds WHERE job = ? AND status = ?
ORDER BY started""",
[assignment_name, SUCCESS],
)
# XXX: There is no easy way to exclude builds started by staff ("super") groups.
# But because this graph is to show the general trend, it's usually fine if staff builds
# are included. Plus, the graph only shows up in the admin interface anyway.
builds = [(source, score, parse_time(started)) for source, score, started in c.fetchall()]
if not builds:
return []
source_set = map(lambda b: b[0], builds)
started_time_set = map(lambda b: b[2], builds)
min_started = min(started_time_set)
max_started = max(started_time_set)
assignment_min_started = parse_time(assignment.not_visible_before)
assignment_max_started = parse_time(assignment.due_date)
data_min = min(min_started, assignment_min_started)
data_max = max(max_started, assignment_max_started)
data_points = []
best_scores_so_far = {source: 0 for source in source_set}
time_delta = (data_max - data_min) / (num_points - 1)
current_time = data_min
for source, score, started_time in builds:
while current_time < started_time:
percentiles = np.percentile(best_scores_so_far.values(), data_keys)
data_points.append([format_js_compatible_time(current_time)] + list(percentiles))
current_time += time_delta
if score is not None:
best_scores_so_far[source] = max(score, best_scores_so_far[source])
percentiles = list(np.percentile(best_scores_so_far.values(), data_keys))
now_time = now()
while current_time - (time_delta / 2) < data_max:
data_points.append([format_js_compatible_time(current_time)] + percentiles)
if current_time >= now_time:
percentiles = [None] * len(percentiles)
current_time += time_delta
return data_points
def builds_one(name):
with DbCursor() as c:
c.execute('''SELECT build_name, status, score, source, `commit`, message, job, started,
log FROM builds WHERE build_name = ? LIMIT 1''', [name])
build = c.fetchone()
if not build:
abort(404)
build_info = build + (get_assignment_by_name(build[6]).full_score,)
return render_template("ta/builds_one.html",
build_info=build_info,
**_template_common())
def builds_one(name):
with DbCursor() as c:
c.execute(
'''SELECT build_name, status, score, source, `commit`, message, job, started,
log FROM builds WHERE build_name = ? LIMIT 1''', [name])
build = c.fetchone()
if not build:
abort(404)
build_info = build + (get_assignment_by_name(build[6]).full_score, )
return render_template("ta/builds_one.html",
build_info=build_info,
**_template_common())
def assignments_one(name, page):
page_size = 50
assignment = get_assignment_by_name(name)
if not assignment:
abort(404)
with DbCursor() as c:
c.execute(
'''SELECT id, name, sid, github, email, super, score, slipunits, updated
FROM grades LEFT JOIN users ON grades.user = users.id
WHERE assignment = ? ORDER BY super DESC, login''',
[name])
grades = c.fetchall()
c.execute(
'''SELECT build_name, source, status, score, `commit`, message, started
FROM builds WHERE job = ? ORDER BY started DESC LIMIT ? OFFSET ?''',
[name, page_size + 1, (page - 1) * page_size])
builds = c.fetchall()
if not builds and page > 1:
abort(404)
more_pages = len(builds) == page_size + 1
if more_pages:
builds = builds[:-1]
c.execute(
'''SELECT COUNT(*), AVG(score) FROM grades
WHERE assignment = ? AND score IS NOT NULL''', [name])
stats = c.fetchone()
if stats[0] == 0:
variance = None
stddev = None
else:
c.execute(
"SELECT AVG((score - ?) * (score - ?)) FROM grades WHERE assignment = ?",
[stats[1], stats[1], name])
variance, = c.fetchone()
stddev = sqrt(variance)
assignment_info = (
(assignment.name, assignment.full_score, assignment.min_score,
assignment.max_score, assignment.weight, assignment.due_date,
assignment.category, assignment.is_group, assignment.manual_grading,
assignment.not_visible_before, assignment.cannot_build_after,
assignment.start_auto_building, assignment.end_auto_building) +
stats + (stddev, ))
return render_template("ta/assignments_one.html",
grades=grades,
builds=builds,
assignment_info=assignment_info,
page=page,
more_pages=more_pages,
**_template_common())
def gradeslog_one(name):
with DbCursor() as c:
c.execute('''SELECT gradeslog.transaction_name, gradeslog.source, users.id, users.name,
users.github, users.super, gradeslog.assignment, gradeslog.score,
gradeslog.slipunits, gradeslog.updated, gradeslog.description
FROM gradeslog LEFT JOIN users ON gradeslog.user = users.id
WHERE gradeslog.transaction_name = ? LIMIT 1''', [name])
entry = c.fetchone()
assignment = get_assignment_by_name(entry[6])
full_score = assignment.full_score if assignment else 0.0
return render_template("ta/gradeslog_one.html",
entry=entry,
full_score=full_score,
**_template_common())
def gradeslog_one(name):
with DbCursor() as c:
c.execute(
'''SELECT gradeslog.transaction_name, gradeslog.source, users.id, users.name,
users.github, users.super, gradeslog.assignment, gradeslog.score,
gradeslog.slipunits, gradeslog.updated, gradeslog.description
FROM gradeslog LEFT JOIN users ON gradeslog.user = users.id
WHERE gradeslog.transaction_name = ? LIMIT 1''', [name])
entry = c.fetchone()
assignment = get_assignment_by_name(entry[6])
full_score = assignment.full_score if assignment else 0.0
return render_template("ta/gradeslog_one.html",
entry=entry,
full_score=full_score,
**_template_common())
def build_now():
job_name = request.form.get("f_job_name")
repo = request.form.get("f_repo")
assignment = get_assignment_by_name(job_name)
if not assignment:
abort(400)
assignment = assignment.student_view(repo)
if assignment.manual_grading:
abort(400)
with DbCursor() as c:
student = _get_student(c)
user_id, _, _, login, _, _ = student
super_value = get_super(c, user_id)
dropped = super_value is not None and super_value < 0
if assignment.is_group:
repos = get_groups(c, user_id)
else:
repos = [login]
if repo not in repos:
abort(403)
if now_compare(assignment.not_visible_before,
add_grace_period(
assignment.cannot_build_after)) != 0 or dropped:
abort(400)
branch_hash = get_branch_hash(repo, "master")
message = None
if branch_hash:
message = get_commit_message(repo, branch_hash)
# This section doesn't absolutely NEED to be consistent with the previous transaction,
# since we have not specified any actual data constraints. The only possible logical error
# would be to fulfill a request when the current user's permissions have been revoked
# between these two transactions.
with DbCursor() as c:
build_name = create_build(c, job_name, repo, branch_hash, message)
if should_limit_source(repo, job_name):
rate_limit_fail_build(build_name)
else:
job = Job(build_name, repo, "Web interface")
dockergrader_queue.enqueue(job)
return redirect(url_for("dashboard.builds_one", name=build_name))
def assignments_one(name, page):
page_size = 50
assignment = get_assignment_by_name(name)
if not assignment:
abort(404)
with DbCursor() as c:
c.execute('''SELECT id, name, sid, github, email, super, score, slipunits, updated
FROM grades LEFT JOIN users ON grades.user = users.id
WHERE assignment = ? ORDER BY super DESC, login''', [name])
grades = c.fetchall()
c.execute('''SELECT build_name, source, status, score, `commit`, message, started
FROM builds WHERE job = ? ORDER BY started DESC LIMIT ? OFFSET ?''',
[name, page_size + 1, (page - 1) * page_size])
builds = c.fetchall()
if not builds and page > 1:
abort(404)
more_pages = len(builds) == page_size + 1
if more_pages:
builds = builds[:-1]
c.execute('''SELECT COUNT(*), AVG(score) FROM grades
WHERE assignment = ? AND score IS NOT NULL''', [name])
stats = c.fetchone()
if stats[0] == 0:
variance = None
stddev = None
else:
c.execute("SELECT AVG((score - ?) * (score - ?)) FROM grades WHERE assignment = ?",
[stats[1], stats[1], name])
variance, = c.fetchone()
stddev = sqrt(variance)
assignment_info = ((assignment.name, assignment.full_score, assignment.min_score,
assignment.max_score, assignment.weight, assignment.due_date,
assignment.category, assignment.is_group, assignment.manual_grading,
assignment.not_visible_before, assignment.cannot_build_after,
assignment.start_auto_building, assignment.end_auto_building) +
stats + (stddev,))
return render_template("ta/assignments_one.html",
grades=grades,
builds=builds,
assignment_info=assignment_info,
page=page,
more_pages=more_pages,
**_template_common())
def builds_one(name):
with DbCursor() as c:
student = _get_student(c)
user_id, _, _, login, _, _ = student
group_repos = get_groups(c, user_id)
repos = [login] + group_repos
c.execute('''SELECT build_name, status, score, source, `commit`, message, job, started,
log FROM builds WHERE build_name = ? AND source in (%s)
LIMIT 1''' % (",".join(["?"] * len(repos))),
[name] + repos)
build = c.fetchone()
if not build:
abort(404)
build_info = build + (get_assignment_by_name(build[6]).full_score,)
template_common = _template_common(c)
return render_template("dashboard/builds_one.html",
build_info=build_info,
**template_common)
def builds_one(name):
with DbCursor() as c:
student = _get_student(c)
user_id, _, _, login, _, _ = student
group_repos = get_groups(c, user_id)
repos = [login] + group_repos
c.execute(
'''SELECT build_name, status, score, source, `commit`, message, job, started,
log FROM builds WHERE build_name = ? AND source in (%s)
LIMIT 1''' % (",".join(["?"] * len(repos))),
[name] + repos)
build = c.fetchone()
if not build:
abort(404)
build_info = build + (get_assignment_by_name(build[6]).full_score, )
template_common = _template_common(c)
return render_template("dashboard/builds_one.html",
build_info=build_info,
**template_common)
def build_now():
job_name = request.form.get("f_job_name")
repo = request.form.get("f_repo")
assignment = get_assignment_by_name(job_name)
if not assignment:
abort(400)
if assignment.manual_grading:
abort(400)
if now_compare(assignment.not_visible_before, assignment.cannot_build_after) != 0:
abort(400)
with DbCursor() as c:
student = _get_student(c)
user_id, _, _, login, _, _ = student
if assignment.is_group:
repos = get_groups(c, user_id)
else:
repos = [login]
if repo not in repos:
abort(403)
branch_hash = get_branch_hash(repo, "master")
message = None
if branch_hash:
message = get_commit_message(repo, branch_hash)
# This section doesn't absolutely NEED to be consistent with the previous transaction,
# since we have not specified any actual data constraints. The only possible logical error
# would be to fulfill a request when the current user's permissions have been revoked
# between these two transactions.
with DbCursor() as c:
build_name = create_build(c, job_name, repo, branch_hash, message)
job = Job(build_name, repo, "Web interface")
dockergrader_queue.enqueue(job)
return redirect(url_for("dashboard.builds_one", name=build_name))
def assignments_one(name):
with DbCursor() as c:
student = _get_student(c)
user_id, _, _, login, _, _ = student
assignment = get_assignment_by_name(name)
if not assignment:
abort(404)
slipunits_now = slip_units_now(assignment.due_date)
is_visible = now_compare(assignment.not_visible_before) >= 0
if assignment.manual_grading:
can_build = False
else:
can_build = now_compare(assignment.cannot_build_after) <= 0
if not is_visible:
abort(404)
c.execute("SELECT score, slipunits, updated FROM grades WHERE user = ? AND assignment = ?",
[user_id, name])
grade = c.fetchone()
if not grade:
grade = (None, None, None)
if assignment.is_group:
repos = get_groups(c, user_id)
else:
repos = [login]
c.execute('''SELECT build_name, source, status, score, `commit`, message, started
FROM builds WHERE job = ? AND source IN (%s)
ORDER BY started DESC''' % (",".join(["?"] * len(repos))),
[name] + repos)
builds = c.fetchall()
if builds:
most_recent_repo = builds[0][1]
else:
most_recent_repo = None
if grade[0] is not None:
c.execute('''SELECT COUNT(*) + 1 FROM grades WHERE assignment = ? AND score > ?''',
[name, grade[0]])
rank, = c.fetchone()
else:
rank = None
c.execute('''SELECT COUNT(*), AVG(score) FROM grades
WHERE assignment = ? AND score IS NOT NULL''', [name])
stats = c.fetchone()
if stats[0] == 0:
variance = None
stddev = None
else:
c.execute("SELECT AVG((score - ?) * (score - ?)) FROM grades WHERE assignment = ?",
[stats[1], stats[1], name])
variance, = c.fetchone()
stddev = sqrt(variance)
assignment_info = ((assignment.name, assignment.full_score, assignment.weight,
assignment.due_date, assignment.category, assignment.is_group,
assignment.manual_grading) + grade + (rank,) + stats + (stddev,))
template_common = _template_common(c)
return render_template("dashboard/assignments_one.html",
assignment_info=assignment_info,
builds=builds,
repos=repos,
most_recent_repo=most_recent_repo,
slipunits_now=slipunits_now,
can_build=can_build,
**template_common)
def grade_distribution(cls, c, assignment_name, max_bins=None):
"""
Returns a pre-binned histogram of grades for the assignment. Here is an example of what the
data looks like:
[{'x': 0, 'dx': 1, 'y': 3},
{'x': 1, 'dx': 2, 'y': 0},
{'x': 2, 'dx': 3, 'y': 0},
{'x': 3, 'dx': 4, 'y': 0},
{'x': 4, 'dx': 5, 'y': 1},
{'x': 5, 'dx': 6, 'y': 2},
{'x': 6, 'dx': 7, 'y': 5},
{'x': 7, 'dx': 8, 'y': 11},
{'x': 8, 'dx': 9, 'y': 7},
{'x': 9, 'dx': 10, 'y': 15},
{'x': 10, 'dx': 11, 'y': 35},
{'x': 11, 'dx': 12, 'y': 92}]
"""
assignment = get_assignment_by_name(assignment_name)
if not assignment:
return
grade_set = cls._get_grade_set(c, assignment.name)
if not grade_set:
return []
grade_set_min, grade_set_max = min(grade_set), max(grade_set)
assignment_min, assignment_max = assignment.min_score, assignment.full_score
bin_min = int(floor(min(grade_set_min, assignment_min)))
bin_max = int(ceil(max(grade_set_max, assignment_max)))
bin_range = bin_max - bin_min
if max_bins is None:
if len(str(bin_max)) < 3:
max_bins = 16
elif len(str(bin_max)) < 4:
max_bins = 13
elif len(str(bin_max)) < 5:
max_bins = 10
else:
max_bins = 7
# The minimum number of bins for factorizable ranges.
# Does not apply when range < max_bins.
min_bins = max_bins / 2
# How flexible is the max_bins? We'll allow this many more bins.
flexibility = 1.8
def factorize(n):
for candidate in range(2, n):
if n % candidate == 0:
return [candidate] + factorize(n / candidate)
return [n]
if bin_range <= max_bins:
per_bin = 1.0
num_bins = bin_range
else:
factorization = factorize(bin_range)
per_bin = 1
for factor in factorization:
if per_bin * factor * min_bins > bin_range:
break
else:
per_bin *= factor
if per_bin * max_bins > bin_range:
break
if per_bin * max_bins * flexibility > bin_range:
num_bins = bin_range / per_bin
per_bin = float(per_bin)
else:
# The best per_bin is too small to look good.
# So just fallback to naive division method.
num_bins = max_bins
per_bin = float(bin_range) / max_bins
boundaries = [int(round(per_bin * n)) for n in range(1, num_bins)]
bin_sizes = [0] * num_bins
for score in grade_set:
bin_index = int(floor((score - bin_min) / per_bin))
if bin_index >= num_bins:
bin_index = num_bins - 1
bin_sizes[bin_index] += 1
return [
{"x": x, "dx": dx, "y": y} for x, dx, y in zip([bin_min] + boundaries, boundaries + [bin_max], bin_sizes)
]
def _process_job(self, job):
build_name = job.build_name
with self.lock:
self.status = build_name
self.updated = now()
# Mark the job as In Progress
while True:
try:
with DbCursor() as c:
c.execute('''SELECT source, `commit`, message, job, started FROM builds
WHERE build_name = ? AND status = ? LIMIT 1''',
[build_name, QUEUED])
row = c.fetchone()
if row is None:
self._log("Build %s was missing from the database. Skipping." % build_name)
return
source, commit, message, job_name, started = row
owners = get_repo_owners(c, source)
owner_emails = {owner: email for owner, (_, _, _, _, _, email)
in get_users_by_ids(c, owners).items()}
c.execute("UPDATE builds SET status = ?, updated = ? WHERE build_name = ?",
[IN_PROGRESS, now_str(), build_name])
break
except apsw.Error:
self._log("Exception raised while setting status to IN_PROGRESS. Retrying...",
exc=True)
logging.exception("Failed to retrieve next dockergrader job")
self._log("Started building %s" % build_name)
try:
# if the job doesn't exist for some reason, the resulting TypeError will be caught
# and logged
assignment = get_assignment_by_name(job_name)
due_date = assignment.due_date
job_handler = get_job(job_name)
log, score = job_handler(source, commit)
# Ignore any special encoding inside the log, and just treat it as a bytes
log = buffer(log)
min_score, max_score = assignment.min_score, assignment.max_score
full_score = assignment.full_score
if score < min_score or score > max_score:
raise ValueError("A score of %s is not in the acceptable range of %f to %f" %
(str(score), min_score, max_score))
except JobFailedError as e:
self._log("Failed %s with JobFailedError" % build_name, exc=True)
with DbCursor() as c:
c.execute('''UPDATE builds SET status = ?, updated = ?, log = ?
WHERE build_name = ?''', [FAILED, now_str(), str(e), build_name])
if config.mailer_enabled:
try:
for owner in owners:
email = owner_emails.get(owner)
if not email:
continue
subject = "%s failed to complete" % build_name
send_template("build_failed", email, subject, build_name=build_name,
job_name=job_name, source=source, commit=commit,
message=message, error_message=str(e))
except Exception:
self._log("Exception raised while reporting JobFailedError", exc=True)
logging.exception("Exception raised while reporting JobFailedError")
else:
self._log("JobFailedError successfully reported via email")
return
except Exception as e:
self._log("Exception raised while building %s" % build_name, exc=True)
logging.exception("Internal error within build %s" % build_name)
with DbCursor() as c:
c.execute('''UPDATE builds SET status = ?, updated = ?, log = ?
WHERE build_name = ?''',
[FAILED, now_str(), "Build failed due to an internal error.", build_name])
return
self._log("Autograder build %s complete (score: %s)" % (build_name, str(score)))
while True:
try:
with DbCursor() as c:
c.execute('''UPDATE builds SET status = ?, score = ?, updated = ?,
log = ? WHERE build_name = ?''',
[SUCCESS, score, now_str(), log, build_name])
slipunits = slip_units(due_date, started)
affected_users = assign_grade_batch(c, owners, job_name, float(score),
slipunits, build_name, "Automatic build.",
"autograder", dont_lower=True)
break
except apsw.Error:
self._log("Exception raised while assigning grades", exc=True)
logging.exception("Failed to update build %s after build completed" % build_name)
return
if config.mailer_enabled:
try:
for owner in owners:
email = owner_emails.get(owner)
if not email:
continue
subject = "%s complete - score %s / %s" % (build_name, str(score),
str(full_score))
if owner not in affected_users:
subject += " (no effect on grade)"
else:
if slipunits == 1:
subject += " (1 %s used)" % config.slip_unit_name_singular
elif slipunits > 0:
subject += " (%s slip %s used)" % (str(slipunits),
config.slip_unit_name_plural)
send_template("build_finished", email, subject, build_name=build_name,
job_name=job_name, score=score, full_score=str(full_score),
slipunits=slipunits, log=log, source=source, commit=commit,
message=message, affected=(owner in affected_users))
except Exception:
self._log("Exception raised while reporting grade", exc=True)
logging.exception("Exception raised while reporting grade")
else:
self._log("Grade successfully reported via email")
def enter_grades_confirm():
try:
f_step = request.form.get("f_step")
if f_step not in ("1", "2"):
fail_validation("Enum out of range (probably a programming error)")
step = int(f_step)
assignment_name = request.form.get("f_assignment")
if not assignment_name:
fail_validation("Assignment name is required")
assignment = get_assignment_by_name(assignment_name)
if assignment is None:
fail_validation("Assignment not found: %s" % assignment_name)
min_score, max_score = assignment.min_score, assignment.max_score
description = request.form.get("f_description")
if not description:
fail_validation("Transaction description is required")
transaction_source = github_username()
entries = []
user_id_set = set()
with DbCursor() as c:
valid_identifiers, ambiguous_identifiers = get_valid_ambiguous_identifiers(c)
def try_add(f_student, f_score, f_slipunits):
if not any((f_student, f_score, f_slipunits)):
return
elif not f_student and (f_score or f_slipunits):
fail_validation("Expected student SID, login, or name, but none provided")
elif f_student in ambiguous_identifiers:
fail_validation("The identifier '%s' is ambiguous. Please use another." %
f_student)
else:
if step == 1:
students = get_users_by_identifier(c, f_student)
elif step == 2:
student = get_user_by_id(c, f_student)
# Let the usual error handling take care of this case
students = [student] if student else []
if not students:
fail_validation("Student or group not found: %s" % f_student)
for student in students:
user_id, student_name, _, _, _, _ = student
if user_id in user_id_set:
fail_validation("Student was listed more than once: %s" % student_name)
try:
score = float_or_none(f_score)
except ValueError:
fail_validation("Not a valid score: %s" % f_score)
try:
slipunits = int_or_none(f_slipunits)
except ValueError:
fail_validation("Slip %s amount not valid: %s" % (slip_unit_name_plural,
f_slipunits))
if slipunits is not None and slipunits < 0:
fail_validation("Slip %s cannot be negative" % slip_unit_name_plural)
if score is not None and not min_score <= score <= max_score:
fail_validation("Score is out of allowed range: %s (Range: %s to %s)" %
(f_score, str(min_score), str(max_score)))
entries.append([user_id, score, slipunits])
user_id_set.add(user_id)
if step == 1:
f_students = request.form.getlist("f_student")
f_scores = request.form.getlist("f_score")
f_slipunitss = request.form.getlist("f_slipunits")
if not same_length(f_students, f_scores, f_slipunitss):
fail_validation("Different numbers of students, scores, and slip %s " +
"reported. Browser bug?" % slip_unit_name_plural)
for f_student, f_score, f_slipunits in zip(f_students, f_scores, f_slipunitss):
try_add(f_student, f_score, f_slipunits)
f_csv = request.form.get("f_csv", "")
for row in csv.reader(StringIO.StringIO(f_csv), delimiter=",", quotechar='"'):
if len(row) != 3:
fail_validation("CSV rows must contain 3 entries")
try_add(*row)
if not entries:
fail_validation("No grade or slip %s changes entered" % slip_unit_name_plural)
if step == 1:
c.execute('''SELECT id, name, sid, login, github FROM users
WHERE id IN (%s)''' % (",".join(["?"] * len(entries))),
[user_id for user_id, _, _ in entries])
students = c.fetchall()
details_user = {}
for user_id, name, sid, login, github in students:
details_user[user_id] = [name, sid, login, github]
c.execute('''SELECT user, score, slipunits, updated FROM grades
WHERE assignment = ? AND user IN (%s)''' %
(",".join(["?"] * len(entries))),
[assignment.name] + [user_id for user_id, _, _ in entries])
grades = c.fetchall()
details_grade = {}
for user_id, score, slipunits, updated in grades:
details_grade[user_id] = [score, slipunits, updated]
entries_details = []
for entry in entries:
user_id = entry[0]
entry_details = (entry + details_user.get(user_id, [None] * 4)
+ details_grade.get(user_id, [None] * 3))
entries_details.append(entry_details)
elif step == 2:
transaction_number = get_next_autoincrementing_value(
c, "enter_grades_last_transaction_number")
transaction_name = "enter-grades-%s" % transaction_number
for user_id, score, slipunits in entries:
assign_grade_batch(c, [user_id], assignment.name, score, slipunits,
transaction_name, description, transaction_source,
manual=True, dont_lower=False)
if step == 1:
entries_csv = StringIO.StringIO()
entries_csv_writer = csv.writer(entries_csv, delimiter=",", quotechar='"')
for entry in entries:
entries_csv_writer.writerow(entry)
return render_template("ta/enter_grades_confirm.html",
entries_details=entries_details,
entries_csv=entries_csv.getvalue(),
assignment_name=assignment.name,
description=description,
full_score=assignment.full_score,
**_template_common())
elif step == 2:
if len(entries) == 1:
flash("1 grade committed", "success")
else:
flash("%d grades committed" % len(entries), "success")
return redirect(url_for("ta.enter_grades"))
except ValidationError as e:
return redirect_with_error(url_for("ta.enter_grades"), e)
def timeseries_grade_percentiles(c, assignment_name, num_points=40):
"""
Returns a timeseries of grades with percentiles. Here is an example:
[["2015-07-17 19:00:36-0700", 0.0, 0.0, 0.0, ... 0.0, 0.0],
["2015-07-17 19:10:36-0700", 0.0, 0.0, 0.0, ... 1.0, 2.0],
["2015-07-17 19:20:36-0700", 0.0, 0.0, 0.0, ... 3.0, 4.0],
["2015-07-17 19:30:36-0700", 0.0, 0.0, 0.5, ... 5.0, 6.0],
["2015-07-17 19:40:36-0700", 0.0, 0.0, 1.0, ... 7.0, 8.0]]
"""
data_keys = range(0, 105, 5)
assignment = get_assignment_by_name(assignment_name)
if not assignment:
return
# There is a slight problem that because of DST, ordering by "started" may not always
# produce the correct result. When the timezone changes, lexicographical order does not
# match the actual order of the times. However, this only happens once a year in the middle
# of the night, so f**k it.
c.execute(
'''SELECT source, score, started FROM builds WHERE job = ? AND status = ?
ORDER BY started''', [assignment_name, SUCCESS])
# XXX: There is no easy way to exclude builds started by staff ("super") groups.
# But because this graph is to show the general trend, it's usually fine if staff builds
# are included. Plus, the graph only shows up in the admin interface anyway.
builds = [(source, score, parse_time(started))
for source, score, started in c.fetchall()]
if not builds:
return []
source_set = tuple(map(lambda b: b[0], builds))
started_time_set = tuple(map(lambda b: b[2], builds))
min_started = min(started_time_set)
max_started = max(started_time_set)
assignment_min_started = parse_time(assignment.not_visible_before)
assignment_max_started = parse_time(assignment.due_date)
data_min = min(min_started, assignment_min_started)
data_max = max(max_started, assignment_max_started)
data_points = []
best_scores_so_far = {source: 0 for source in source_set}
time_delta = (data_max - data_min) / (num_points - 1)
current_time = data_min
for source, score, started_time in builds:
while current_time < started_time:
percentiles = np.percentile(tuple(best_scores_so_far.values()),
data_keys)
data_points.append([format_js_compatible_time(current_time)] +
list(percentiles))
current_time += time_delta
if score is not None:
best_scores_so_far[source] = max(score,
best_scores_so_far[source])
percentiles = list(
np.percentile(tuple(best_scores_so_far.values()), data_keys))
now_time = now()
while current_time - (time_delta / 2) < data_max:
data_points.append([format_js_compatible_time(current_time)] +
percentiles)
if current_time >= now_time:
percentiles = [None] * len(percentiles)
current_time += time_delta
return data_points
def enter_grades_confirm():
try:
f_step = request.form.get("f_step")
if f_step not in ("1", "2"):
fail_validation("Enum out of range (probably a programming error)")
step = int(f_step)
assignment_name = request.form.get("f_assignment")
if not assignment_name:
fail_validation("Assignment name is required")
assignment = get_assignment_by_name(assignment_name)
if assignment is None:
fail_validation("Assignment not found: %s" % assignment_name)
min_score, max_score = assignment.min_score, assignment.max_score
description = request.form.get("f_description")
if not description:
fail_validation("Transaction description is required")
transaction_source = github_username()
entries = []
user_id_set = set()
with DbCursor() as c:
valid_identifiers, ambiguous_identifiers = get_valid_ambiguous_identifiers(
c)
def try_add(f_student, f_score, f_slipunits):
if not any((f_student, f_score, f_slipunits)):
return
elif not f_student and (f_score or f_slipunits):
fail_validation(
"Expected student SID, login, or name, but none provided"
)
elif f_student in ambiguous_identifiers:
fail_validation(
"The identifier '%s' is ambiguous. Please use another."
% f_student)
else:
if step == 1:
students = get_users_by_identifier(c, f_student)
elif step == 2:
student = get_user_by_id(c, f_student)
# Let the usual error handling take care of this case
students = [student] if student else []
if not students:
fail_validation("Student or group not found: %s" %
f_student)
for student in students:
user_id, student_name, _, _, _, _ = student
if user_id in user_id_set:
fail_validation(
"Student was listed more than once: %s" %
student_name)
try:
score = float_or_none(f_score)
except ValueError:
fail_validation("Not a valid score: %s" % f_score)
try:
slipunits = int_or_none(f_slipunits)
except ValueError:
fail_validation(
"Slip %s amount not valid: %s" %
(slip_unit_name_plural, f_slipunits))
if slipunits is not None and slipunits < 0:
fail_validation("Slip %s cannot be negative" %
slip_unit_name_plural)
if score is not None and not min_score <= score <= max_score:
fail_validation(
"Score is out of allowed range: %s (Range: %s to %s)"
% (f_score, str(min_score), str(max_score)))
entries.append([user_id, score, slipunits])
user_id_set.add(user_id)
if step == 1:
f_students = request.form.getlist("f_student")
f_scores = request.form.getlist("f_score")
f_slipunitss = request.form.getlist("f_slipunits")
if not same_length(f_students, f_scores, f_slipunitss):
fail_validation(
"Different numbers of students, scores, and slip %s " +
"reported. Browser bug?" % slip_unit_name_plural)
for f_student, f_score, f_slipunits in zip(
f_students, f_scores, f_slipunitss):
try_add(f_student, f_score, f_slipunits)
f_csv = request.form.get("f_csv", "")
for row in csv.reader(StringIO.StringIO(f_csv),
delimiter=",",
quotechar='"'):
if len(row) != 3:
fail_validation("CSV rows must contain 3 entries")
try_add(*row)
if not entries:
fail_validation("No grade or slip %s changes entered" %
slip_unit_name_plural)
if step == 1:
c.execute(
'''SELECT id, name, sid, login, github FROM users
WHERE id IN (%s)''' %
(",".join(["?"] * len(entries))),
[user_id for user_id, _, _ in entries])
students = c.fetchall()
details_user = {}
for user_id, name, sid, login, github in students:
details_user[user_id] = [name, sid, login, github]
c.execute(
'''SELECT user, score, slipunits, updated FROM grades
WHERE assignment = ? AND user IN (%s)''' %
(",".join(["?"] * len(entries))),
[assignment.name] + [user_id for user_id, _, _ in entries])
grades = c.fetchall()
details_grade = {}
for user_id, score, slipunits, updated in grades:
details_grade[user_id] = [score, slipunits, updated]
entries_details = []
for entry in entries:
user_id = entry[0]
entry_details = (entry +
details_user.get(user_id, [None] * 4) +
details_grade.get(user_id, [None] * 3))
entries_details.append(entry_details)
elif step == 2:
transaction_number = get_next_autoincrementing_value(
c, "enter_grades_last_transaction_number")
transaction_name = "enter-grades-%s" % transaction_number
for user_id, score, slipunits in entries:
assign_grade_batch(c, [user_id],
assignment.name,
score,
slipunits,
transaction_name,
description,
transaction_source,
manual=True,
dont_lower=False)
if step == 1:
entries_csv = StringIO.StringIO()
entries_csv_writer = csv.writer(entries_csv,
delimiter=",",
quotechar='"')
for entry in entries:
entries_csv_writer.writerow(entry)
return render_template("ta/enter_grades_confirm.html",
entries_details=entries_details,
entries_csv=entries_csv.getvalue(),
assignment_name=assignment.name,
description=description,
full_score=assignment.full_score,
**_template_common())
elif step == 2:
if len(entries) == 1:
flash("1 grade committed", "success")
else:
flash("%d grades committed" % len(entries), "success")
return redirect(url_for("ta.enter_grades"))
except ValidationError as e:
return redirect_with_error(url_for("ta.enter_grades"), e)
def grade_distribution(cls, c, assignment_name, max_bins=None):
"""
Returns a pre-binned histogram of grades for the assignment. Here is an example of what the
data looks like:
[{'x': 0, 'dx': 1, 'y': 3},
{'x': 1, 'dx': 2, 'y': 0},
{'x': 2, 'dx': 3, 'y': 0},
{'x': 3, 'dx': 4, 'y': 0},
{'x': 4, 'dx': 5, 'y': 1},
{'x': 5, 'dx': 6, 'y': 2},
{'x': 6, 'dx': 7, 'y': 5},
{'x': 7, 'dx': 8, 'y': 11},
{'x': 8, 'dx': 9, 'y': 7},
{'x': 9, 'dx': 10, 'y': 15},
{'x': 10, 'dx': 11, 'y': 35},
{'x': 11, 'dx': 12, 'y': 92}]
"""
assignment = get_assignment_by_name(assignment_name)
if not assignment:
return
grade_set = cls._get_grade_set(c, assignment.name)
if not grade_set:
return []
grade_set_min, grade_set_max = min(grade_set), max(grade_set)
assignment_min, assignment_max = assignment.min_score, assignment.full_score
bin_min = int(floor(min(grade_set_min, assignment_min)))
bin_max = int(ceil(max(grade_set_max, assignment_max)))
bin_range = bin_max - bin_min
if max_bins is None:
if len(str(bin_max)) < 3:
max_bins = 16
elif len(str(bin_max)) < 4:
max_bins = 13
elif len(str(bin_max)) < 5:
max_bins = 10
else:
max_bins = 7
# The minimum number of bins for factorizable ranges.
# Does not apply when range < max_bins.
min_bins = max_bins // 2
# How flexible is the max_bins? We'll allow this many more bins.
flexibility = 1.8
def factorize(n):
for candidate in range(2, n):
if n % candidate == 0:
return [candidate] + factorize(n // candidate)
return [n]
if bin_range <= max_bins:
per_bin = 1.0
num_bins = bin_range
else:
factorization = factorize(bin_range)
per_bin = 1
for factor in factorization:
if per_bin * factor * min_bins > bin_range:
break
else:
per_bin *= factor
if per_bin * max_bins > bin_range:
break
if per_bin * max_bins * flexibility > bin_range:
num_bins = bin_range // per_bin
per_bin = float(per_bin)
else:
# The best per_bin is too small to look good.
# So just fallback to naive division method.
num_bins = max_bins
per_bin = float(bin_range) / max_bins
boundaries = [int(round(per_bin * n)) for n in range(1, num_bins)]
bin_sizes = [0] * num_bins
for score in grade_set:
bin_index = int(floor((score - bin_min) / per_bin))
if bin_index >= num_bins:
bin_index = num_bins - 1
bin_sizes[bin_index] += 1
return [{
"x": x,
"dx": dx,
"y": y
} for x, dx, y in zip([bin_min] + boundaries, boundaries +
[bin_max], bin_sizes)]
def assignments_one(name):
with DbCursor() as c:
student = _get_student(c)
user_id, _, _, login, _, _ = student
assignment = get_assignment_by_name(name)
if not assignment:
abort(404)
slipunits_now = slip_units_now(assignment.due_date)
is_visible = now_compare(assignment.not_visible_before) >= 0
if assignment.manual_grading:
can_build = False
else:
can_build = now_compare(assignment.cannot_build_after) <= 0
if not is_visible:
abort(404)
c.execute(
"SELECT score, slipunits, updated FROM grades WHERE user = ? AND assignment = ?",
[user_id, name])
grade = c.fetchone()
if not grade:
grade = (None, None, None)
if assignment.is_group:
repos = get_groups(c, user_id)
else:
repos = [login]
c.execute(
'''SELECT build_name, source, status, score, `commit`, message, started
FROM builds WHERE job = ? AND source IN (%s)
ORDER BY started DESC''' % (",".join(["?"] * len(repos))),
[name] + repos)
builds = c.fetchall()
if builds:
most_recent_repo = builds[0][1]
else:
most_recent_repo = None
if grade[0] is not None:
c.execute(
'''SELECT COUNT(*) + 1 FROM grades WHERE assignment = ? AND score > ?''',
[name, grade[0]])
rank, = c.fetchone()
else:
rank = None
c.execute(
'''SELECT COUNT(*), AVG(score) FROM grades
WHERE assignment = ? AND score IS NOT NULL''', [name])
stats = c.fetchone()
if stats[0] == 0:
variance = None
stddev = None
else:
c.execute(
"SELECT AVG((score - ?) * (score - ?)) FROM grades WHERE assignment = ?",
[stats[1], stats[1], name])
variance, = c.fetchone()
stddev = sqrt(variance)
assignment_info = (
(assignment.name, assignment.full_score, assignment.weight,
assignment.due_date, assignment.category, assignment.is_group,
assignment.manual_grading) + grade + (rank, ) + stats +
(stddev, ))
template_common = _template_common(c)
return render_template("dashboard/assignments_one.html",
assignment_info=assignment_info,
builds=builds,
repos=repos,
most_recent_repo=most_recent_repo,
slipunits_now=slipunits_now,
can_build=can_build,
**template_common)
