Skip to content

Latest commit

 

History

History
94 lines (69 loc) · 2.3 KB

README.md

File metadata and controls

94 lines (69 loc) · 2.3 KB

CPG Miner

A program that makes use of Joern take a CPG and some target procedure and extract the code as well as the sub-graphs of that procedure.

Getting Started

To build and launch (with a specific Log4j2 config)

sbt stage
./cpg_miner -Dlog4j.configurationFile=log4j2.xml

Usage

Generate a CPG using Joern, e.g. joern-parse <input_dir> -o <output_cpg>. Use this <output_cpg> as the input for cpg-miner.

Alternatively, you can use ./cpg_miner from-code <input_dir> and it will generate a CPG for you to ingest. This will be created at <input_dir>/cpg.bin.

Mines code snippets and sub-graphs from CPGs.
Usage: cpg_miner [from-graph|from-code] [options] <args>...

  --help
  -o, --output-dir <value>
                           The output directory to dump the mined artifacts.
  -m, --method-name <value>
                           The name of the target procedure to mine. If undefined, matches all methods.
  -c, --show-callees       Attempts to resolve and dump the methods called by the target method. Default is false.
  --combine                Combines all representations and methods into a single JSON CPG. Note this will exclude the code snippet.
Command: from-graph input-cpg

  input-cpg                The input cpg to mine from.
Command: from-code input-dir

  input-dir                The input directory to generate a CPG from

Note: The <output-dir> is not cleaned before run, but existing results will be overwritten during successive runs.

Output

The structure of the output directory is as follows:

- <method_name>
 | - ast.json       # Serialized AST
 | - cfg.json       # Serialized CFG
 | - code.<ext>     # Code snippet, where <ext> is the language file suffix
 | - ddg.json       # Serialized DDG
 | - pdg.json       # Serialized PDG (DDG + CDG)

If there are more than 1 matches for the method, then it will be suffixed by _idx where idx is a counter.

Graph Schema

The JSON graphs follow the following schema.

Graph

{
  "nodes": "<Node[]>",
  "edges": "<Edge[]>"
}

Node

{
  "id": "<number>",
  "label": "<string>",
  "name": "<string>",
  "code": "<string>"
}

Edge

{
  "src": "<number>",
  "dst": "<number>",
  "value": "<string>",
  "label": "<string>"
}