To use this package it is necessary to install it (with
pip or by source) and also download the DAVIS 2017 Dataset. The instructions to do so are as follows.
To install the package you can run the following command on your terminal:
# Python 3 strongly recommended # Install numpy and Cython as are required to build our package pip install numpy Cython # Install the package pip install davisinteractive
In addition to installing the framework, you will need to download the
trainval DAVIS 2017 Dataset with 480p resolution from here.
Moreover, you can download the
trainval scribbles from here.
You have to unzip the scibbles zip file in the folder containing DAVIS (in /path/to supposing DAVIS is in /path/to/DAVIS).
Note: A script to download automatically the dataset and the scribbles is coming soon.
Install from Source¶
If you prefer, you can install the package downloading the source code and installing it. To do so, you need to run these commands:
# Download the code git clone https://github.com/albertomontesg/davis-interactive.git && cd davis-interactive # Install it - Python 3 recommended python setup.py install
If you want to contribute to this package you will need to have a copy of the code to work with. First, download the code from Github:
git clone https://github.com/albertomontesg/davis-interactive.git && cd davis-interactive git checkout -b my/new/branch
To have a development copy of the package installed for Python you can run the following:
pip3 install -e .
This will link the available copy of the package to your current copy so all the modifications that you made on the code will be visible by any script.
If you want to help us improve the documentation it will be necessary to have some additional packages:
pip3 install mkdocs pydoc-markdown mkdocs-material
Then you serve the documentation live on your local machine to check the changes you make on the documentation.
cd docs # Serve the documentation live pydocmd serve # Build the documentation ./build_docs.sh