This is the server backend for Project Felina, the wildlife image analysis service.

It manages user authentication and image and metadata upload and storage. It also provides a facility for executing native computer vision executables that can access the database. The API responses are JSON and the database is MySQL.

Get Node and npm, then

git clone https://github.com/felina/server.git
cd server
npm install
cp db_settings.json.example db_settings.json
cp config.json.example config.json

You will need an instance of MySQL with the schema already prepared. To do this:

tools/db_install.sh

This will connect to a server on localhost as root by default. Edit the script if you need to change the connection settings.

You should then update config.json with your AWS settings, and db_settings with your MySQL connection settings.

You will need to enable various settings in your AWS S3 buckets. The public image bucket in particular will require you to enable the static webserver and add the following redirect rule:

  <RoutingRules>
    <RoutingRule>
    <Condition>
      <HttpErrorCodeReturnedEquals>404</HttpErrorCodeReturnedEquals>
      <KeyPrefixEquals>thm_</KeyPrefixEquals>
    </Condition>
    <Redirect>
      <ReplaceKeyPrefixWith></ReplaceKeyPrefixWith>
    </Redirect>
    </RoutingRule>
  </RoutingRules>

node src/main.js

Create a new file test/db_settingsTest.json that is the same as config/db_settings.json but replace the line "database" : "felina" with "database" : "felinaTest", otherwise the server has issues.

Edit the file test/testConfig.json and change the paths of the two tests images to ones on your machine.

The test scripts require some dependencies to be installed on your system. If you're on Debian, or a derivative, you can use the setup script provided (tested on Debian 7):

./test/setup.sh

To perform the tests run

./test/testing.py

Example testcase:

def test_name():
    print_test('Test name')
    r = requests.get(url=path + test_path, data=data_for_request, cookies=cookie) # .get or .post endpoint with parameters
    response_handle(r, 'Explanation of error if res is not what is expected', True) # Bool is expected res value
    response_object_check(r, jsr.test_name(test_params)) # Compare the response object to the expected response object
    ppass()

When adding a testcase you must also add an expected result JSON object to test/jsonResponses.py. Try to make it dynamic eg. the example response for an uploaded image computes the hash of the image to compare with the returned id.

MIT