Download

The following 3 tools are currently implemented

Block level backup tool

• Requires elevation. Please run as admin.
  
• Creates volume level images using either VSS snapshots or from live volume.
  
• Creates volume level images of already existing persistent VSS snapshots on the system.
  
• Ability to exclude files (only for VSS backups) by deleting them from the source snapshot.
  
• Support for MBR and GPT partitions. Volumes > 2TB are by default saved as VHDX with GPT partitioning.
  
• Supported virtual disk formats :
  
  • Dynamic vhd (d-vhd)
    
  • Fixed vhd (f-vhd)
    
  • Dynamic vhdx (d-vhdx)
    
  • Raw volume image (raw)
    
• Volume images can either be generated locally or streamed to an FTP server.
  
• URI format for streaming to FTP:
  
  • ftps://username:password@hostname:port/a/b/c
    
    

FTP(S) client

• Supports plain and secure FTPS (using openssl async bio).
  
• Ability to browse while transfers are in progress.
  
• Directory listing support for Linux, Windows and MLSD.
  
• FTPS supports TLS1.3
  
  

Camera tool

• Face detection, Motion detection, Face recognition.
  
• Uses default opencv built-in models and DNN based detectors.
  
• Detection configurability on a per camera basis.
  
• Ability to save detection results on a per camera basis.
  
• The release has default(non-CUDA) opencv build for CPUs.
  
• For NVIDIA GPU it is recommended to do a custom opencv CUDA/CUDNN variant build via vcpkg.
  
• The code either ways supports both CUDA/CPU backend and target for DNN inference; in order.
  
• Set CVL_MODELS_ROOT env variable to the local MODELS folder from repo for detections to work.
  
• Define a new camera with face detection and specify a results folder. Run face detection for 5-10 secs and then stop. Then use the facerec "train" option to register the above face detection results with facerec. (via camera settings)
  
• The system can be tested by defining a "Window capture" source in OBS studio against youtube running inside the browser. Then expose the webbrowser feed over RTSP(using OBS RSTP server plugin). This can then be captured via a cam configuration in offset using the obs rtsp url for running detections.
  

vcpkg dependencies

git clone https://github.com/microsoft/vcpkg.git
cd vcpkg && bootstrap-vcpkg.bat
vcpkg.exe install openssl crc32c rapidjson zlib opencv4[contrib,core,dnn,ffmpeg,highgui,jpeg] --recurse
sudo apt-get install build-essential zip curl vim bison nasm meson pkg-config
./vcpkg install tinyxml2 libzip openssl crc32c rapidjson zlib opencv4[contrib,core,dnn,ffmpeg,highgui,jpeg] --recurse
use the resulting toolchain file in cmake configure step as highlited under the build section

QT-6.9.0 source build

make sure ninja and python3.9 are under PATH
set PATH=D:\Python39;%PATH%
where python
D:\Python39\python.exe
C:\Users\nmam\AppData\Local\Microsoft\WindowsApps\python.exe
where ninja
C:\Windows\System32\ninja.exe
C:\Program Files\Microsoft Visual Studio\2022\Community\Common7\IDE\CommonExtensions\Microsoft\CMake\Ninja\ninja.exe

\qt-everywhere-src-6.9.0\configure.bat -prefix D:\Qt-6.9.0\install -skip qtconnectivity -nomake examples -nomake tests -skip speech -skip scxml -skip qtsensors -skip qtserialbus -skip qtspeech -skip qtdoc -skip qtandroidextras -release

Qt is now configured for building. Just run 'cmake --build . --parallel'
Once everything is built, you must run 'cmake --install .'
Qt will be installed into 'D:/QT-6.8.2/install'

Build

git clone https://github.com/n-mam/offset.git
cd offset && mkdir build && cd build
SET Qt6_DIR=D:\QT-6.9.0\install\lib\cmake\Qt6
cmake -DCMAKE_TOOLCHAIN_FILE=D:/vcpkg/scripts/buildsystems/vcpkg.cmake ..
cmake -DCMAKE_TOOLCHAIN_FILE=~/vcpkg/scripts/buildsystems/vcpkg.cmake ..
cmake --build . --config Release

Run as admin (elevation needed only for the block level backup tool):
SET PATH=D:\QT-6.9.0\install\bin;%PATH%
SET CVL_MODELS_ROOT=E:\offset\cvl\MODELS\ (only for camera app)
qml\Release\offset.exe

Deploy:

windeployqt --qmldir D:\offset\app\qml D:\offset\build\app\Release\offset.exe

vc redist is bundled with package zip; in case your system does not have that installed already

Contact:

Telegram: https://t.me/neelabhm