Environment Setup (macOS)

python3 -m venv venv  
source venv/bin/activate
pip3 install .

Malicious Contracts Detector

Important note: You must use exactly the same pre-processing steps that were used in training the model in production/testing. This includes the same vectorization/tokenization and any processing to the opcodes. See at example opcode in src/opcode_example.txt.

Set up

In the main directory run pip3 install -e . .

Training the model

Use src/model_training/train.ipynb to train the model and follow the instructions in the notebook. Make sure you save the new model and new vectorizer as model.pkl and vectorizer.pkl to the working_model directory before running it in test/production. Models are saved in the model_training directory.

Using the model

Call predict([opcode]) from src/working_model/main.py with a string of opcode, opcodes and operands separated by spaces like: PUSH1 0x80 PUSH1 0x40 MSTORE PUSH1 0x04.

To test the model save the opcodes in .txt files in working_model/test_opcodes, and run src/working_model/test.py

If you're reading from file, the format of the opcodes should be a string as above, or:

[1] PUSH1 0x80

[3] PUSH1 0x40

[4] MSTORE

[6] PUSH1 0x04

If your opcode has any other formatting, make sure you pre-process it to fit the expected format (a string of opcodes and operands)