vars-localize

Tool for creating localizations within the VARS database.

Author: Kevin Barnard (kbarnard@mbari.org)

This project is made with PyQt5.

Dependencies

Dependencies are managed primarily through Anaconda. Instructions for direct pip installation are listed below as well.

Note: These instructions are designed for macOS or Linux users.

Using Anaconda

To setup the app within an Anaconda environment, follow the proceeding steps:

1. Create environment

From the root directory, run:

conda env create -f environment.yml

This will create an Anaconda environment named vars-localize with all required dependencies.

2. Activate Anaconda environment

To activate the environment, run:

$ conda activate vars-localize

Using pip directly

If setup with Anaconda does not work or isn't preferred, the dependencies can be installed through pip3 with Python 3.

Run the following command to install the required packages in your system.

$ pip install -r requirements.txt

Configuration (required)

1. Set the Annosaurus API key

Setting the API key is necessary for the application to save localization data back to VARS. Change this in .env:

API_KEY=foo

Production Annosaurus API keys may be acquired from Brian Schlining (brian@mbari.org).

2. Configure the proper M3 endpoints

The application settings can be configured within config/config.ini. The settings below should be modified to switch from test to production mode:

user_site = %(test_user_site)s/accounts/v1
anno_site = %(test_anno_site)s/anno/v1
kb_site = %(test_kb_site)s/kb/v1
vam_site = %(test_vam_site)s/vam/v1

To do this, simply replace all instances of test with prod in this block only.

Usage

To start the application, run the following from the project directory:

python vars-localize.py

Once the application launches, log in with your VARS username.

Search for a concept in the bar at the top left of the application, then select a concept from the list of results to populate a tree of imaged moments in the pane below. Select an observation from the children in the subtree of the imaged moment, and draw a bounding box around the observed concept by clicking and dragging.

You can double-click on any created localizations to pull up a dialog of properties. Here, you can change the observation's concept, modify the localization bounds, or delete the localization. Additionally, localizations can be resized by dragging the corners of the box when this dialog is closed.