#!/usr/bin/python """ CS 451 Data-Intensive Distributed Computing: Assignment 5 public check script Sample usage: $ ./check_assignment5_public.py bigdatateach --env linux """ import sys import os from subprocess import call import argparse import re def check_a5(pn,env,iterations,memory): spark_cmd = ["spark-submit", "--driver-memory", memory] cp_cmd = ["cp"] mkdir_cmd = ["mkdir"] eval_cmd = "spam_eval.sh" data_prefix = "/u3/cs451/public_html/spam" cluster_options = "" if env != 'linux': spark_cmd = ["spark-submit", "--num-executors", "2", "--executor-cores", "2", "--executor-memory", "8G", "--driver-memory", memory] cp_cmd = ["hadoop","fs", "-cp"] mkdir_cmd = ["hadoop","fs","-mkdir"] eval_cmd = "spam_eval_hdfs.sh" data_prefix = "/data/cs451/" call(["mvn","clean","package"]) call(["echo", "# Train and Classify with Group-X"]) call(spark_cmd + [ "--class", "ca.uwaterloo.cs451.a5.TrainSpamClassifier".format(pn), "target/assignments-1.0.jar", "--input", "{0}/spam.train.group_x.txt".format(data_prefix), "--model", "cs451-{0}-a5-model-group_x".format(pn)]) call(spark_cmd + [ "--class", "ca.uwaterloo.cs451.a5.ApplySpamClassifier".format(pn), "target/assignments-1.0.jar", "--input", "{0}/spam.test.qrels.txt".format(data_prefix), "--model", "cs451-{0}-a5-model-group_x".format(pn), "--output", "cs451-{0}-a5-output-group_x".format(pn)]) call(["echo", "# Train and Classify with Group-Y"]) call(spark_cmd + [ "--class", "ca.uwaterloo.cs451.a5.TrainSpamClassifier".format(pn), "target/assignments-1.0.jar", "--input", "{0}/spam.train.group_y.txt".format(data_prefix), "--model", "cs451-{0}-a5-model-group_y".format(pn)]) call(spark_cmd + [ "--class", "ca.uwaterloo.cs451.a5.ApplySpamClassifier".format(pn), "target/assignments-1.0.jar", "--input", "{0}/spam.test.qrels.txt".format(data_prefix), "--model", "cs451-{0}-a5-model-group_y".format(pn), "--output", "cs451-{0}-a5-output-group_y".format(pn)]) call(["echo", "# Train and Classify with Britney"]) call(spark_cmd + [ "--class", "ca.uwaterloo.cs451.a5.TrainSpamClassifier".format(pn), "target/assignments-1.0.jar", "--input", "{0}/spam.train.britney.txt".format(data_prefix), "--model", "cs451-{0}-a5-model-britney".format(pn)]) call(spark_cmd + [ "--class", "ca.uwaterloo.cs451.a5.ApplySpamClassifier".format(pn), "target/assignments-1.0.jar", "--input", "{0}/spam.test.qrels.txt".format(data_prefix), "--model", "cs451-{0}-a5-model-britney".format(pn), "--output", "cs451-{0}-a5-output-britney".format(pn)]) call(["echo", "# Evaluation: Group-X, Group-Y, and Britney"]) call(["bash",eval_cmd,"cs451-{0}-a5-output-group_x".format(pn)]) call(["bash",eval_cmd,"cs451-{0}-a5-output-group_y".format(pn)]) call(["bash",eval_cmd,"cs451-{0}-a5-output-britney".format(pn)]) call(["echo", "# Classify with Voting Ensemble on Test"]) call(mkdir_cmd +["cs451-{0}-a5-model-fusion".format(pn)]) call(cp_cmd + ["cs451-{0}-a5-model-group_x/part-00000".format(pn), "cs451-{0}-a5-model-fusion/part-00000".format(pn)]) call(cp_cmd + ["cs451-{0}-a5-model-group_y/part-00000".format(pn), "cs451-{0}-a5-model-fusion/part-00001".format(pn)]) call(cp_cmd + ["cs451-{0}-a5-model-britney/part-00000".format(pn), "cs451-{0}-a5-model-fusion/part-00002".format(pn)]) call(spark_cmd + [ "--class", "ca.uwaterloo.cs451.a5.ApplyEnsembleSpamClassifier".format(pn), "target/assignments-1.0.jar", "--input", "{0}/spam.test.qrels.txt".format(data_prefix), "--model", "cs451-{0}-a5-model-fusion".format(pn), "--output", "cs451-{0}-a5-output-fusion-average".format(pn), "--method","average"]) call(spark_cmd + [ "--class", "ca.uwaterloo.cs451.a5.ApplyEnsembleSpamClassifier".format(pn), "target/assignments-1.0.jar", "--input", "{0}/spam.test.qrels.txt".format(data_prefix), "--model", "cs451-{0}-a5-model-fusion".format(pn), "--output", "cs451-{0}-a5-output-fusion-voting".format(pn), "--method","vote"]) call(["echo", "# Evaluation: Voting Ensemble on Test"]) call(["bash",eval_cmd,"cs451-{0}-a5-output-fusion-average".format(pn)]) call(["bash",eval_cmd,"cs451-{0}-a5-output-fusion-voting".format(pn)]) call(["echo", "# Train and Classify Single Classifier on All Data"]) call(spark_cmd + [ "--class", "ca.uwaterloo.cs451.a5.TrainSpamClassifier".format(pn), "target/assignments-1.0.jar", "--input", "{0}/spam.train.all.txt".format(data_prefix), "--model", "cs451-{0}-a5-model-all".format(pn)]) call(spark_cmd + [ "--class", "ca.uwaterloo.cs451.a5.ApplySpamClassifier".format(pn), "target/assignments-1.0.jar", "--input", "{0}/spam.test.qrels.txt".format(data_prefix), "--model", "cs451-{0}-a5-model-all".format(pn), "--output", "cs451-{0}-a5-output-all".format(pn)]) call(["echo", "# Evaluation: Single Classifier Trained on All Data"]) call(["bash",eval_cmd,"cs451-{0}-a5-output-all".format(pn)]) call(["echo", "# Data Shuffling ({0} Iterations)".format(iterations)]) for i in range(iterations): call(spark_cmd + [ "--class", "ca.uwaterloo.cs451.a5.TrainSpamClassifier".format(pn), "target/assignments-1.0.jar", "--input", "{0}/spam.train.britney.txt".format(data_prefix), "--model", "cs451-{0}-a5-model-britney-shuffle".format(pn), "--shuffle"]) call(spark_cmd + [ "--class", "ca.uwaterloo.cs451.a5.ApplySpamClassifier".format(pn), "target/assignments-1.0.jar", "--input", "{0}/spam.test.qrels.txt".format(data_prefix), "--model", "cs451-{0}-a5-model-britney-shuffle".format(pn), "--output", "cs451-{0}-a5-output-britney-shuffle".format(pn)]) call(["echo", "# Evaluation: Shuffle iteration {0}".format(i)]) call(["bash",eval_cmd,"cs451-{0}-a5-output-britney-shuffle".format(pn)]) if __name__ == "__main__": parser = argparse.ArgumentParser(description="CS 451 2021 Fall Assignment 5 Public Test Script") parser.add_argument('username',metavar='Github Username', help="Github username",type=str) parser.add_argument('-m','--memory',help="Amount of memory to give Spark jobs",type=str,default="2G") parser.add_argument('-e', '--env', help='Environment to test under.',type=str,default='linux') parser.add_argument('-i', '--iterations', help='Number of shuffle iterations',type=int,default=3) args=parser.parse_args() try: check_a5(args.username, args.env, args.iterations, args.memory) except Exception as e: print(e)