SeeShell

SeeShell is a project that uses machine learning to identify shell species from user-uploaded images. SeeShell consists of two parts, a webserver (back end) and an android application (front end). The server consists of an API, a MySQL Database, and a data lake containing seashell images. TODO write about client after specific technologies are decided.

Getting started

Follow the instructions below to set up the server environment, database, data, and API for SeeShell.

Environment Setup:

• Install python 3
• in the command line navigate to the source code folder, run the commands cd Server then pip install -r server_requirements.txt. Then run the commands cd ../Client then pip install -r client_requirements.txt. This will install the needed python libraries.
• Download and install MySQL Workbench and MySQL Server

Database Setup

• Ensure that the MySQL Server service is running. You can usually start the service from the MySQL Workbench or via system services.
• Open MySQL Workbench and create a new database named "seeshell".
• Open the config.json files in both the API and database folders, and update the "password" field with your MySQL root password.
• In the repository, navigate to the database folder and run the generatetables.py file with the following command: python generatetables.py
  • This will generate the required tables in the "seeshell" database.

Data Setup:

• Download the shell dataset here and unzip it to a folder: https://springernature.figshare.com/articles/dataset/all_shell_images/9122621?backTo=/collections/A_shell_dataset_for_shell_features_extraction_and_recognition/4428335
• edit config.json
  • Set rawDataPath to be the filepath to the shell data folder
  • Set genusPath and speciesPath to be locations of the output folders for preprocessed shell images. These should be empty folders, and the preprocessing script will sort the data into them.
• run PreprocessData.py

Launch the API:

• Navigate to the API folder in your repository in a terminal window
• Type flask --app SeeShellAPI.py run to launch the API (dev environment only)