BPVis

Interactive visualization tool for deep neural network.

How to run with loading from local files

• Download (clone or whatever) this repository.
• Download repository https://github.com/katka-juhasova/BP-data.
• Create directories data and modules in cloned BPVis repository. Repository tree should look now like this:

BPVis
├── app_demo.py
├── assets
├── components
├── constant.py
├── data
├── LICENSE
├── modules
├── preprocessing
├── README.md
├── requirements.txt
└── setup.py

• Content of data-part1 and data-part2 directories (from BP-data repository) move to data directory.
• Content of modules-part1 and modules-part2 directories (from BP-data repository) move to modules directory.
• Run python3 script.py <BPVis_repository_path> e.g. python3 script.py '/home/BPVis' (from BP-data repository). Directory data contains only .json representations of modules, full source code is contained in modules. This script adds path to the downloaded modules, otherwise, the modules content would be read from git url and it would take like forever to run the app.
• Install requirements. In case of some error with igraph library make sure that you downloaded correct version from https://pypi.org/project/python-igraph/. The error that might have occurred is explained here: https://stackoverflow.com/questions/36200707/error-with-igraph-library-deprecated-library.
• Run python3 app_demo.py from this repository. App is now running on http://127.0.0.1:8050/.

For visualization of other modules change lines file_left = files[0] and file_right = files[1] in app_demo.py. You can change just index or assign path to the .json file, e.g. /home/BPVis/data/30log/AST1.json or data/30log/AST1.json for Linux. In that case it might just take a little bit longer to run the app.

NOTE: interesting tree visualizations:

<BPVis_repository_path>/data/30log/AST1.json
<BPVis_repository_path>/data/30log/AST1.json