FOSSLight Scanner

FOSSLight Scanner is released under the Apache-2.0. Current python package version.

FOSSLight Scanner can perform an analysis for open source compliance at once. It can perform open source analysis of source code, binary and dependency.

It works using the following scanners:

  1. FOSSLight Source Scanner : Analyze the source code and generate open source analysis result.
  2. FOSSLight Dependency Scanner : Analyze the dependencies used through package manager or build system and generate open source analysis result. 
  3. FOSSLight Binary Scanner : Analyze the binaries and generate open source analysis result.

Github Repository : https://github.com/fosslight/fosslight_scanner
License : Apache-2.0

📋 Prerequisite

  1. FOSSLight Scanner needs a Python 3.8+.
  2. To use the function to extract OSS information (OSS Name, OSS Version, License) from Binary DB, see the database setting guide.
  3. (Only for windows) Install Microsoft Build Tools (Microsoft Visual C++ 14.0+) from https://visualstudio.microsoft.com/ko/visual-cpp-build-tools/

🎉 How to install

It can be installed using pip3. It is recommended to install it in the python 3.8 + virtualenv environment.

$ pip3 install fosslight_scanner

🚀 How to run

How to run by mode & Parameters

$ fosslight [Mode] [option1] <arg1> [option2] <arg2>...
     Parameters:
        Mode: Multiple modes can be entered by separating them with , (ex. source,binary)
            all                     Run all scanners(Default)
            source                  Run FOSSLight Source Scanner
            dependency              Run FOSSLight Dependency Scanner
            binary                  Run FOSSLight Binary Scanner
            compare                 Compare two FOSSLight reports

        Options:
            -h                      Print help message
            -p <path>               Path to analyze (ex, -p {input_path})
                                     * Compare mode input file: Two FOSSLight reports (supports excel, yaml)
                                       (ex, -p {before_name}.xlsx {after_name}.xlsx)
            -w <link>               Link to be analyzed can be downloaded by wget or git clone            
            -f <formats> [<format> ...]     FOSSLight Report file format (excel, csv, opossum, yaml, spdx-tag, spdx-yaml, spdx-json, spdx-xml)
                                     * Compare mode result file: supports excel, json, yaml, html
                                     * Multiple formats can be specified separated by space.
            -e <path>               Path to exclude from analysis (ex, -e {dir} {file})
            -o <output>             Output directory or file
            -c <number>             Number of processes to analyze source
            -r                      Keep raw data
            -t                      Hide the progress bar
            -v                      Print FOSSLight Scanner version
            -s <path>               Path to apply setting from file (check format with 'setting.json' in this repository)
                                     * Direct cli flags have higher priority than setting file
                                       (ex, '-f yaml -s setting.json' - result file extension is .yaml)
            --no_correction         Enter if you don't want to correct OSS information with sbom-info.yaml
                                     * Correction mode only supported xlsx format.
            --correct_fpath <path>  Path to the sbom-info.yaml file
            --ui                          Generate UI mode result file

        Options for only 'all' or 'bin' mode
            -u <db_url>             DB Connection(format :'postgresql://username:password@host:port/database_name')

        Options for only 'all' or 'dependency' mode
            -d <dependency_argument>        Additional arguments for running dependency analysis

Ex.1 Local Source Analysis

fosslight all -p /home/source_path
fosslight all -o test_result_wget -w "https://github.com/LGE-OSS/example.git"

Ex.3 Compare the BOM of two FOSSLight reports with yaml or excel format and check the oss status (change/add/delete).

fosslight compare -p FOSSLight_before_proj.yaml FOSSLight_after_proj.yaml -o test_result

How to call execution parameters as json

  1. Write and save the values ​​for each execution parameter as a json file in setting.json format.
  2. When running, call setting.json with -s.
    fosslight -s setting.json
    

    🛈 The value called during execution takes precedence over the parameters written in the json file. ex. When called with ‘-f yaml -s setting.json', the output file is in yaml format.

📁 Result

Result for the mode that analyze the open source (all, source, dependency, binary)

test_result/
├── fosslight_log
│   └── fosslight_log_220214_1824.txt
├── fosslight_report_all_220214_1824.xlsx
└── fosslight_raw_data (with -r option)
    ├── fosslight_src_220214_1824.xlsx
    ├── fosslight_bin_220214_1824.xlsx
    └── fosslight_dep_220214_1824.xlsx
  • fosslight_report_all_(datetime).xlsx : FOSSLight Report format file in which source code analysis, binary analysis, and dependency analysis results are written.
  • fosslight_raw_data directory: The folder where the analysis result raw data file is created (with -r option)
    • fosslight_src_(datetime).xlsx : FOSSLight Report of FOSSLight Source Scanner
    • fosslight_dep_(datetime).xlsx : FOSSLight Report of FOSSLight Dependency Scanner
    • fosslight_bin_(datetime).xlsx : FOSSLight Report of FOSSLight Binary Scanner

fosslight_report_(datetime).xlsx

  1. Exclude: Checked Row test(s), doc(s), hidden files or folders are checked as Exclude.
  2. When sbom-info.yaml is loaded, the loaded data is appended and the analysis results for duplicate files checked as Exclude.
  3. Comment :
    Add/Loaded by ** : Row loaded from **
    Excluded by ** : Row excluded due to **

Result for compare mode

test_result/
├── fosslight_log
│   └── fosslight_log_20220817_114259.txt
└── fosslight_compare_20220817_114259.xlsx
  • fosslight_compare_(datetime).xlsx : Two BOM comparison results in the form of (add/delete/change) table.

🐳 How to install and run using Docker

  1. Build image using Dockerfile.
    $docker build -t fosslight .
    
  2. Run with the image you built.
    ex. Output: /Users/fosslight_scanner/test_output, Path to be analyzed: tests/test_files
    $docker run -it -v /Users/fosslight_scanner/test_output:/app/output fosslight -p tests/test_files -o output