Data Description

We provide test data (test_data) and training data can be constructed by ourselves. The strategy adopted in our paper is to use test data from nine of the projects combined into training data to predict one of the remaining projects. The training set should be used to build your machine learning model. For the test set we provided, it was constructed several times. oldName and newName were used to construct label_class, edge was used to calculate changeNum, and oldStmt_body and newStmt_body were used to filter the data for the input model. After the filtering is completed, oldStmt and newStmt are used as model inputs and label_class is the expected output.

Column	Variable	Definition	Type	Content
2	label_class	Whether the method name needs to be renamed	Integer	0=No,1=Yes
4	oldname	the historical method name	Text	outDegrees
5	newname	the current method name	Text	getDegrees
6	oldStmt	the historical method(masked name)	Text	public DataSet<Tuple2<K, Long>> _(){return vertices.join(edges).where(0).equalTo(0).map(new VertexKeyWithOne<K, EV, VV>()).groupBy(0).sum(1);}
7	newStmt	the current method(masked name)	Text	public DataSet<Tuple2<K, Long>> _(){return outDegrees().union(inDegrees()).groupBy(0).sum(1);}
8	edge	the set of the current entity and its related entity pairs	Text	<outDegrees,Graph>,<outDegrees,flink.graphs>
9	changeNum	number of renaming activities performed in the related entity	Text	1
10	oldStmt_body	the historical method body	Text	{return vertices.join(edges).where(0).equalTo(0).map(new VertexKeyWithOne<K, EV, VV>()).groupBy(0).sum(1);}
11	newStmt_body	the current method body	Text	{return outDegrees().union(inDegrees()).groupBy(0).sum(1);}

Quick Start

(1) Prepare data Note that:Preparing data will take a long time, the output data is the test data we provided(test_data).

1. Collect the renaming data of the project https://github.com/konL/MethodRenamePrediction/blob/dfaada6d4065241de25f587c36239d2b81ebdf46/RenamePrediction_preprocess/src/main/java/Extractor/HistoryAnalysis.java
2. Filter method renaming data https://github.com/konL/MethodRenamePrediction/blob/07fff20c2f3933273a23bf4ea4d10cec61029c3f/RenamePrediction_preprocess/src/main/java/create_proj_method.java
3. Generate the old and new versions of the database based on the renaming data in step 2 https://github.com/konL/MethodRenamePrediction/blob/07fff20c2f3933273a23bf4ea4d10cec61029c3f/RenamePrediction_preprocess/src/main/java/createVerDB.java
4. generate test data

• Collect old info and new info from the old and new versions of the database https://github.com/konL/MethodRenamePrediction/blob/dfaada6d4065241de25f587c36239d2b81ebdf46/RenamePrediction_preprocess/src/main/java/createEmbedding/creatResultFile.java
• Collect related entities and calculate the number of changeNum https://github.com/konL/MethodRenamePrediction/blob/main/RenamePrediction/DataUtils/createChangeFile.py
• Filter method https://github.com/konL/MethodRenamePrediction/blob/dfaada6d4065241de25f587c36239d2b81ebdf46/RenamePrediction/DataUtils/delStmt_processing.py
• Mask method name https://github.com/konL/MethodRenamePrediction/blob/dfaada6d4065241de25f587c36239d2b81ebdf46/RenamePrediction/DataUtils/delDeclare.py

5. Generate training data https://github.com/konL/MethodRenamePrediction/blob/dfaada6d4065241de25f587c36239d2b81ebdf46/RenamePrediction/DataUtils/createTraindata.py

(2) Training and prediction

6. Training and prediction The test data of a project and its corresponding training data are selected for prediction, and the Precision, Recall and F-measure of the project are finally output. https://github.com/SerVal-DTF/debug-method-RenamePrediction/prediction/keras4bert_loaddata.py