(First install packages with npm install)
This simulator can be run using the following commands:
npm run server (will start the server)
npm run sim exampleInput.json (will run the simulator)
The documentation can be viewed with the following command:
npm run docs
Simulator I/O:
- I: inputs of two types, food and exercise, each with a timestamp.
- O: output should be a “graph” of blood sugar over the course of the day, and a graph of the amount of “glycation” that occurred during the day.
Food:
- In our model, eating food will increase blood sugar linearly for two hours.
- The rate of increase depends on the glycemic index column.
Exercise:
- Exercise decreases blood sugar linearly for one hour.
Normalization:
- Blood sugar starts at 80 at the beginning of the day.
- If neither food nor exercise is affecting your blood sugar (it has been more than 1 or 2 hours), it will approach 80 linearly at a rate of 1 per minute.
Glycation:
- For every minute your blood sugar stays above 150, increment “glycation” by 1.
Decisions:
- Because normalization changes blood sugar values every minute, I used a time interval of one minute.
- Because output graphs are for a single day, I shifted and truncated timestamps so that timestamps are integers between 0 and 1440 (number of minutes in a day).
- Input will be in JSON format for simplicity.
- "Magic numbers" specified in the project details are included within functions but in a large scale project would be passed in the parameters to ensure function purity. To simplify code for this project only, magic numbers have been left in and have been tested accordingly. List of numbers include: 120 (food duration), 1/120 (food influence per minute), 60 (exercise duration), -60 (exercise influence per minute), 1440 (minutes assessed), 1441 (data points) 80 (starting blood sugar), +/-1 (normalization rate), 150 (glycation threshold).
- All input is assumed to be well-formed for the purpose of this project. This is to simplify assessment of code by omitting error checks at the beginning of every function. In production code (where function modules could be easily repurposed), functions would include error checking for all input parameters.
Solution Phases:
- Read inputs (Assume they are well-formed)
- Lookup exercise or glycemic index from DB and add to input object
- Convert inputs to influence array (which represents the influence of food or exercise on blood sugar based on time of the day). 1440 - one for each minute.
- Create blood sugar array from influence array (using the above rules of food, exercise, and normalization). 1441 points to include both midnights.
- Create a glycation array from blood sugar array (using the above rule of glycation). 1441 points to include both midnights.
- Output blood sugar and glycation arrays
- Open graphs of the output