Dilution

Prerequisite

Download and install LTP model: LTP which consists of two parts:ltp and ltp_data. Here we make use of ltp_data(in version 3.4.0).

install the sub module by pip: pip install pyltp pip install gensim

The LIBSVM model(version 3.23): LIBSVM has already been concluded in ./code.

Quick Start

Usage : run ./code/main.py [syndicator name] [path to ltp_data]

Examples: We put three syndications in './data/input' as examples.

Input & Data format:

Three kinds of input are required:

1. payment process : a document with sentences that describe the necessary procedures in payment process.
2. API list: a document where each line stands an API of the syndicator.
3. API parameter description: each API has a document with each line formed by:

[parameter name]***[parameter description].

Location:

All the input files are in the folder: './data/input'.
Put the payment process in './data/input/payDocument', the API list in './data/input/apiList', and the API parameter description in './data/input/apiDescription'.

File Naming:

Name the document of payment process and API list as "[Sindicator name].txt".
For API parameter description, build a directory as "[Syndicator name]" in ./data/input/apiDescription, where each API has a document named as "[API name]-[third payment name]-[post/receive].txt".

In this project we utilized 'Alipay' and 'Wechat' as the third payment.

Output:

The prediction result of logic vulnerability prediction is in './data/output/result'