1. Requirements¶
This section defines the technical requirements for the implementation.
1.1. Language and library structure¶
Library is to be developed in Python 3 and using multiple modules (files). Each file should collect related functionality.
1.2. Documentation¶
All classes, methods (a function that belongs to a class) and functions must be documented with a ‘docstring’. The docstring shall explain what the class or function does. For a function, the docstring shall make clear the purpose, what arguments are expected and what is returned.
Simple examples of Python docstrings can be viewed here.
1.3. Development practices¶
The quality of a flood warning software library is paramount; implementation errors could put lives at risk and lead to substantial financial losses. You are therefore required to adopt software engineering best practices. Your team is required to:
Provide automated tests for your implementations using pytest to demonstrate the quality of the system (see Test framework)
Use automated continuous integration testing (see Automated testing)
1.4. Data source¶
The system is to be built on the (near) real-time river level data at the nearly 2000 monitoring stations that is made available by the Department for Environment Food and Rural Affairs (DEFRA) at https://environment.data.gov.uk/. For most stations river level data is updated every 15 minutes. The data service is summarised at https://data.gov.uk/dataset/real-time-and-near-real-time-river-level-data1.
- REST interface for data retrieval
Data is fetched from https://environment.data.gov.uk/ using a REST interface. With a suitably formed URL (a string), as defined in the service documentation, the server returns the requested data as a JSON object. JSON objects are represented in Python as data structures made up of dictionaries, lists and strings. JSON objects are straightforward to manipulate from Python. The interface to the DEFRA service is documented at https://environment.data.gov.uk/flood-monitoring/doc/reference.