Files
ML4T/defeat_learners/grade_best4.py

231 lines
24 KiB
Python

"""MC3-H1: Best4{LR,DT} - grading script.
Usage:
- Switch to a student feedback directory first (will write "points.txt" and "comments.txt" in pwd).
- Run this script with both ml4t/ and student solution in PYTHONPATH, e.g.:
PYTHONPATH=ml4t:MC3-P1/jdoe7 python ml4t/mc3_p1_grading/grade_learners.py
Copyright 2018, Georgia Institute of Technology (Georgia Tech)
Atlanta, Georgia 30332
All Rights Reserved
Template code for CS 4646/7646
Georgia Tech asserts copyright ownership of this template and all derivative
works, including solutions to the projects assigned in this course. Students
and other users of this template code are advised not to share it with others
or to make it available on publicly viewable websites including repositories
such as github and gitlab. This copyright statement should not be removed
or edited.
We do grant permission to share solutions privately with non-students such
as potential employers. However, sharing with other current or future
students of CS 7646 is prohibited and subject to being investigated as a
GT honor code violation.
-----do not edit anything above this line---
"""
import pytest
from grading.grading import grader, GradeResult, time_limit, run_with_timeout, IncorrectOutput
# These two lines will be commented out in the final grading script.
from LinRegLearner import LinRegLearner
from DTLearner import DTLearner
import os
import sys
import traceback as tb
import numpy as np
import pandas as pd
from collections import namedtuple
import math
import time
import functools
seconds_per_test_case = 5
max_points = 100.0
html_pre_block = True # surround comments with HTML <pre> tag (for T-Square comments field)
# Test cases
Best4TestCase = namedtuple('Best4TestCase', ['description', 'group','max_tests','needed_wins','row_limits','col_limits','seed'])
best4_test_cases = [
Best4TestCase(
description="Test Case 1: Best4LinReg",
group="best4lr",
max_tests=15,
needed_wins=10,
row_limits=(10,1000),
col_limits=(2,10),
seed=1489683274
),
Best4TestCase(
description="Test Case 2: Best4DT",
group="best4dt",
max_tests=15,
needed_wins=10,
row_limits=(10,1000),
col_limits=(2,10),
seed=1489683274
),
Best4TestCase(
description='Test for author() method',
group='author',
max_tests=None,
needed_wins=None,
row_limits=None,
col_limits=None,
seed=None,
),
]
# Test functon(s)
@pytest.mark.parametrize("description,group,max_tests,needed_wins,row_limits,col_limits,seed", best4_test_cases)
def test_learners(description, group, max_tests, needed_wins, row_limits, col_limits, seed, grader):
"""Test data generation methods beat given learner.
Requires test description, test case group, and a grader fixture.
"""
points_earned = 0.0 # initialize points for this test case
incorrect = True
msgs = []
try:
dataX, dataY = None,None
same_dataX, same_dataY = None,None
diff_dataX, diff_dataY = None,None
betterLearner, worseLearner = None, None
if group=='author':
try:
from gen_data import author
auth_string = run_with_timeout(author,seconds_per_test_case,(),{})
if auth_string == 'tb34':
incorrect = True
msgs.append(" Incorrect author name (tb34)")
points_earned = -10
elif auth_string == '':
incorrect = True
msgs.append(" Empty author name")
points_earned = -10
else:
incorrect = False
except Exception as e:
incorrect = True
msgs.append(" Exception occured when calling author() method: {}".format(e))
points_earned = -10
else:
if group=="best4dt":
from gen_data import best4DT
dataX, dataY = run_with_timeout(best4DT,seconds_per_test_case,(),{'seed':seed})
same_dataX,same_dataY = run_with_timeout(best4DT,seconds_per_test_case,(),{'seed':seed})
diff_dataX,diff_dataY = run_with_timeout(best4DT,seconds_per_test_case,(),{'seed':seed+1})
betterLearner = DTLearner
worseLearner = LinRegLearner
elif group=='best4lr':
from gen_data import best4LinReg
dataX, dataY = run_with_timeout(best4LinReg,seconds_per_test_case,(),{'seed':seed})
same_dataX, same_dataY = run_with_timeout(best4LinReg,seconds_per_test_case,(),{'seed':seed})
diff_dataX, diff_dataY = run_with_timeout(best4LinReg,seconds_per_test_case,(),{'seed':seed+1})
betterLearner = LinRegLearner
worseLearner = DTLearner
num_samples = dataX.shape[0]
cutoff = int(num_samples*0.6)
worse_better_err = []
for run in range(max_tests):
permutation = np.random.permutation(num_samples)
train_X,train_Y = dataX[permutation[:cutoff]], dataY[permutation[:cutoff]]
test_X,test_Y = dataX[permutation[cutoff:]], dataY[permutation[cutoff:]]
better = betterLearner()
worse = worseLearner()
better.addEvidence(train_X,train_Y)
worse.addEvidence(train_X,train_Y)
better_pred = better.query(test_X)
worse_pred = worse.query(test_X)
better_err = np.linalg.norm(test_Y-better_pred)
worse_err = np.linalg.norm(test_Y-worse_pred)
worse_better_err.append( (worse_err,better_err) )
worse_better_err.sort(key=functools.cmp_to_key(lambda a,b: int((b[0]-b[1])-(a[0]-a[1]))))
better_wins_count = 0
for worse_err,better_err in worse_better_err:
if better_err < 0.9*worse_err:
better_wins_count = better_wins_count+1
points_earned += 5.0
if better_wins_count >= needed_wins:
break
incorrect = False
if (dataX.shape[0] < row_limits[0]) or (dataX.shape[0]>row_limits[1]):
incorrect = True
msgs.append(" Invalid number of rows. Should be between {}, found {}".format(row_limits,dataX.shape[0]))
points_earned = max(0,points_earned-20)
if (dataX.shape[1] < col_limits[0]) or (dataX.shape[1]>col_limits[1]):
incorrect = True
msgs.append(" Invalid number of columns. Should be between {}, found {}".format(col_limits,dataX.shape[1]))
points_earned = max(0,points_earned-20)
if better_wins_count < needed_wins:
incorrect = True
msgs.append(" Better learner did not exceed worse learner. Expected {}, found {}".format(needed_wins,better_wins_count))
if not(np.array_equal(same_dataY,dataY)) or not(np.array_equal(same_dataX,dataX)):
incorrect = True
msgs.append(" Did not produce the same data with the same seed.\n"+\
" First dataX:\n{}\n".format(dataX)+\
" Second dataX:\n{}\n".format(same_dataX)+\
" First dataY:\n{}\n".format(dataY)+\
" Second dataY:\n{}\n".format(same_dataY))
points_earned = max(0,points_earned-20)
if np.array_equal(diff_dataY,dataY) and np.array_equal(diff_dataX,dataX):
incorrect = True
msgs.append(" Did not produce different data with different seeds.\n"+\
" First dataX:\n{}\n".format(dataX)+\
" Second dataX:\n{}\n".format(diff_dataX)+\
" First dataY:\n{}\n".format(dataY)+\
" Second dataY:\n{}\n".format(diff_dataY))
points_earned = max(0,points_earned-20)
if incorrect:
if group=='author':
raise IncorrectOutput("Test failed on one or more criteria.\n {}".format('\n'.join(msgs)))
else:
inputs_str = " Residuals: {}".format(worse_better_err)
raise IncorrectOutput("Test failed on one or more output criteria.\n Inputs:\n{}\n Failures:\n{}".format(inputs_str, "\n".join(msgs)))
else:
if group != 'author':
avg_ratio = 0.0
worse_better_err.sort(key=functools.cmp_to_key(lambda a,b: int(np.sign((b[0]-b[1])-(a[0]-a[1])))))
for we,be in worse_better_err[:10]:
avg_ratio += (float(we) - float(be))
avg_ratio = avg_ratio/10.0
if group=="best4dt":
grader.add_performance(np.array([avg_ratio,0]))
else:
grader.add_performance(np.array([0,avg_ratio]))
except Exception as e:
# Test result: failed
msg = "Description: {} (group: {})\n".format(description, group)
# Generate a filtered stacktrace, only showing erroneous lines in student file(s)
tb_list = tb.extract_tb(sys.exc_info()[2])
for i in range(len(tb_list)):
row = tb_list[i]
tb_list[i] = (os.path.basename(row[0]), row[1], row[2], row[3]) # show only filename instead of long absolute path
tb_list = [row for row in tb_list if (row[0] == 'gen_data.py')]
if tb_list:
msg += "Traceback:\n"
msg += ''.join(tb.format_list(tb_list)) # contains newlines
elif 'grading_traceback' in dir(e):
msg += "Traceback:\n"
msg += ''.join(tb.format_list(e.grading_traceback))
msg += "{}: {}".format(e.__class__.__name__, str(e))
# Report failure result to grader, with stacktrace
grader.add_result(GradeResult(outcome='failed', points=points_earned, msg=msg))
raise
else:
# Test result: passed (no exceptions)
grader.add_result(GradeResult(outcome='passed', points=points_earned, msg=None))
if __name__ == "__main__":
pytest.main(["-s", __file__])