Skip to content

Latest commit

 

History

History
95 lines (73 loc) · 3.13 KB

INSTALL.md

File metadata and controls

95 lines (73 loc) · 3.13 KB
Raw

Librarian Installation Guide

Download Librarian Sources:

Before getting started, please make sure Docker is installed and running. You can type systemctl status docker to check the running status of Docker daemon.

Next, run the following command to clone Librarian:

# Using SSH
git clone https://github.com/salmanee/Librarian.git

# Using HTTPS
git clone git@github.com:salmanee/Librarian.git

In the parent directory of Librarian, run the following command to start docker (which contains a working environment with all requirements installed):

docker run --rm -it -v $PWD:/home yhuai/librarian

Note that Docker only contains the environment, and the -v argument will mount your local copy of Librarian into the container.

If successful, you will see the following message at the bottom of your screen:

[root@44c29e2ea5d4 home]# ls
Librarian

Using Librarian:

All Librarian scripts are found under scripts/:

  1. To Extract binaries from the apps in sample_apps/ to UnknownLibs_bins, change the directory to scripts:
cd scripts/

Then run the following command:

python3 cluster_libs.py

Note: To extract binaries from a new set or larger set of apps, modify apps_dir and dest_folder in cluster_libs.py accordingly. Then make sure that your apps are arranged in a way similar to the structure in sample_apps/:

|-- sample_apps
|   |-- app_name_1
|   |   |-- SHA256.apk (app version1)
|   |   |-- SHA256.apk (app version2)
|   |   |-- SHA256.apk (app version3)
|   |   |   ...
|   |   |   ...
|   |   `-- SHA256.apk (app versionX)
|   |-- app_name_2
|   |-- app_name_3
|   |-- app_name_4
|   |-- app_name_5
|   ...
|   ...
|   |
|   `-- app_name_n
  1. To extract the features vector from one binary, run:
python3 scripts/Feature_Extractor/extract_feature_vector.py -i <lib.so> -o <out.json>

  1. To extract the features vectors from a set of binaries:

    3.1. Modify extracted_bins.txt to include the binaries you are interested in.

    3.2. Update the input(foldername) and output(foldername2) folders in run_extract_fv.sh.

    3.3. Then run the following command:

./scripts/Feature_Extractor/run_extract_fv.sh
  1. To compute the similarity score between two feature vectors, run the following command:
python3 scripts/Bin2Bin_Score_Calculator/binsimScore.py -f <file1.json> -f <file2.json>

Note that, the script currently shows binaries that matches with a similiarty score of 0.85 and higher. If ther similiarty sore is less than 0.85, nothing will be displayed. You an adjust this threshold in the script.

  1. To compute the similarity between a set of feature vectors:

    5.1. Modify both source_bin_FVS.txt and extracted_bin_FVS.txt to include the binaries you are interested in comparing.

    5.2. Then run:

./scripts/Bin2Bin_Score_Calculator/run_bin_sim.sh

Examples of what the output looks like when running each of the above commands are provided here.