Skip to content

Latest commit

 

History

History
53 lines (47 loc) · 2.68 KB

README.md

File metadata and controls

53 lines (47 loc) · 2.68 KB

JoernAnalyzeStackOverflowCode

From stackoverflow get Q&A data, extract code and tags, then using joern tool get partial code AST and convert AST node type to text

  • how to use stackoverflow data
  • how to use joern to do code analyse, get AST/DDG/CFG,etc
  • how to write gremlin query to visit AST tree stored in Neo4j
  • how to convert AST to Text

Data source

Joern (a platform for robust analysis of C/C++ code)

Tips for install Joern

If you follow official guide strictly, you may still encounter some problems, so make sure the following steps:

  1. java version ,better 1.8
  2. Neo4J Server 2.1.5 Community Edition, if you can't find in it's official guide: wget http://dist.neo4j.org/neo4j-community-2.1.5-unix.tar.gz
  3. choose py2neo 2.0 when install python-joern, you can use this site: https://pypi.python.org/pypi/py2neo/2.0
  4. When istalling gremlin-plugin, you need do mvn clean package, if error occurs, you should change all of the license file data to 2017

Tips for using Joern

  1. Import your code correctly, following this: http://joern.readthedocs.io/en/latest/tutorials/unixStyleCodeAnalysis.html#importing-the-code
  2. When you need to start Neo4j, using this command in case of error:

    $NEO4J/bin/neo4j start-no-wait

  3. rm .joernIndex dir before doing a new import
  4. you can't update Neo4j when it's on running, you need to close it first

File introduce

  1. code_info some file of types/keyword of different language
  2. extrac_code_from_StackOverflow java code to get question and it's answers from database
  3. code_clean_utils.py tools to do cleaning for code and text
  4. get_AST_type_joern.py gremlin query to get AST node type from Neo4j
  5. stackoverflow_data_clean.py get training and testing dataset

If you think this project may helpful, may you give me a star? :)