CathyZhang0822 commented on a change in pull request #11935: [MXNET-691][WIP] Add LabelBot prediction module URL: https://github.com/apache/incubator-mxnet/pull/11935#discussion_r207051236
########## File path: mxnet-bot/LabelBotPredict/elastic_beanstalk_server/predict_labels.py ########## @@ -0,0 +1,106 @@ +# Licensed to the Apache Software Foundation (ASF) under one +# or more contributor license agreements. See the NOTICE file +# distributed with this work for additional information +# regarding copyright ownership. The ASF licenses this file +# to you under the Apache License, Version 2.0 (the +# "License"); you may not use this file except in compliance +# with the License. You may obtain a copy of the License at +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, +# software distributed under the License is distributed on an +# "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY +# KIND, either express or implied. See the License for the +# specific language governing permissions and limitations +# under the License. + +from sklearn.preprocessing import LabelEncoder +from SentenceParser import SentenceParser +from DataFetche import DataFetcher +import numpy as np +import pickle +import re +import logging + +logging.basicConfig(level=logging.INFO) +LOGGER = logging.getLogger(__name__) + +def tokenize(row): + row = re.sub('[^a-zA-Z0-9]', ' ', row).lower() + words = set(row.split()) + return words + +def rule_based(row): + # return a list of rule_based labels + l = [] + # 'feature request' in the issue's title + if "feature request" in row.lower(): + l.append("Feature") + # 'c++' in the issue's title + if "c++" in row.lower(): + l.append("C++") + tokens = tokenize(row) + ci_keywords = ["ci", "ccache", "jenkins"] + flaky_keywords = ["flaky"] + gluon_keywords = ["gluon"] + cuda_keywords = ["cuda", "cudnn"] + scala_keywords = ["scala"] + # one of ci keywords in the issue's title + for key in ci_keywords: + if key in tokens: + l.append("CI") + break + # one of flaky keywords in the issue's title + for key in flaky_keywords: + if key in tokens: + l.append("Flaky") + # one of gluon keywords in the issue's title + for key in gluon_keywords: + if key in tokens: + l.append("Gluon") + # one of scala keywords in the issue's title + for key in scala_keywords: + if key in tokens: + l.append("Scala") + # one of cudo keywords in the issue's title + for key in cuda_keywords: + if key in tokens: + l.append("CUDA") + return l + +def predict(issues): + # get Machine Learning models' predictions + # return Rule-based predictions and Machine Learning models' predictions together + # step1: fetch data + DF = DataFetcher() + df_test = DF.fetch_issues(issues) + # step2: data cleaning + SP = SentenceParser() + SP.data = df_test + SP.clean_body('body', True, True) + SP.merge_column(['title', 'title', 'title', 'body'], 'train') + test_text = SP.process_text('train', True, False, True) + # step3: word embedding + tv = pickle.load(open("Vectorizer.p", "rb")) + test_data_tfidf = tv.transform(test_text).toarray() + tv_op = pickle.load(open("Vectorizer_Operator.p", "rb")) + test_data_tfidf_operator = tv_op.transform(test_text).toarray() + labels = pickle.load(open("Labels.p", "rb")) + le = LabelEncoder() + le.fit_transform(labels) + # step4: classification + clf = pickle.load(open("Classifier.p", "rb")) + probs = clf.predict_proba(test_data_tfidf) + # pickup top 2 predictions which exceeds threshold 0.3 + best_n = np.argsort(probs, axis=1)[:, -2:] + clf_g = pickle.load(open("Classifier_Gaussian.p", "rb")) + ops_pre = clf_g.predict(test_data_tfidf_operator) + recommendations = [] + for i in range(len(best_n)): + l = rule_based(df_test.loc[i, 'title']) + l += [le.classes_[best_n[i][j]] for j in range(-1, -3, -1) if probs[i][best_n[i][j]] > 0.3] Review comment: Will set the threshold as the input ---------------------------------------------------------------- This is an automated message from the Apache Git Service. To respond to the message, please log on GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org With regards, Apache Git Services
