My next steps are to start implementing the Fitbit commands. This means the following:
- Getting current data
- Getting active steps
- Creating a goal
- Comparing if a user has reached their goal
Later, I will make it easier to generate reports using this interface. Perhaps I can also make a command for running arbitrary Fitbit API commands.
With this being said, I will start on getting the current data and then getting the active steps today.
Getting the intraday data works. I've also added a command to get the last sync. Next I will be working to get the active steps.
To get the active steps, I will have to reference my previous implementation. I also plan to make all of the parameters set from the config file.
Once I have the active steps, I will be able to work on goal setting.
I should also write tests for goal setting and calculating active steps. It may be good to mock up the Fitbit interface and test my error handling in the requests, although it seems to work fine, as long as their API doesn't change, but my tests won't catch that; the system will just stop working.
My next step is to get the active steps. To do this, I should look at my earlier implementation. All of the parameters should be set from the config file, instead of hard-coding the various files throughout the project.
After that, I will work on goal setting.
I think for both of these, I should test them, as I am writing them. This will be a fun exercise.
Today, I have coded to get the active steps. I have also updated the command line interface to run a few commands that use this interface, namely a command to get status, check the API keys, and a command to list the devices. These work well at the moment.
The next step is to calculate goals, given the active steps data. It is likely that I should have a sort of caching system for this, so that I don't have to repeatedly pull from the Fitbit API for each day, and so that the code is faster. I will have to think of exactly how this works. Perhaps it can be, once a day is pulled, it is stored somewhere. From there, I can also compute the active steps and record the user's goal. This can be stored somewhere, too. It will probably involve adding a notion of pulling data for every day between the start of the study and the previous day. I will think more about this. Perhaps I can make a pull command, as well, to pull down data for all previous days, we'll see.
Today, I will work on goal setting. It seems that the actual goal setting is simple. It will probably be mostly a problem of caching the data.
I worked on writing the goal setting. I made and tested get weekly goal. I also tested a function to get the week number. This is the opposite of get remaining weeks, which I think is better design.
The next thing to do is tho test the weekly goal and then add the daily goal. From there, I can worry about caching.
Today, I will continue with goal setting. If all goes well, I may get to some of the caching system.
I have just considered that I should use the previous week as a baseline for the first steps goal.
I have got the goal setting working and I have tested it.
My next step is to create the data files and to use them in calculating active steps, etc. I will have to think about how these are organized.
Today I will work on daily goal setting. This should be a bit simpler than the weekly case, but we'll see.
Today, I added a command to get the daily steps goals and tested it.
- Add caching
- Fitbit data
- Active steps & goal
- Tie everything together with the CLI
- Write documentation
Today I will work on the caching system. I think that I will have to do a bit of thinking about its structure.
I wrote a simple read and write to CSV file script and tested it. This required a good bit of learning about Deno and the commands to do things on its file system. Hopefully the rest of the caching system will be easier to implement.
Today, I will hopefully finish the caching section.
I made progress adding a summary page for the general caching system. I didn't finish testing it. That is my next step.
Today I added a test for reading and writing to CSV summary files. I also added code to get a date range. This will be useful to get the caching system pulling all dates Fitbit data.
Today I hope to finish up the caching system. This may be tricky as creating the summary file involves computing many things in the correct order. For example, I have to keep track of the goals and the total active steps.
I made good progress on the caching system. I almost have all of the data that I need. I just need to add the daily steps goals. Once I have that, I can save the output to a file. I think that I will generate it every time, that seems easier than reading the data in.
One challenge with my current architecture is that it is quite hard to test the higher level components. This is because I have made a functional approach and rely on the functions. I could pass in the functions as arguments. This motivates that I could have used classes and passed in those classes, aka injecting those dependencies. I doubt that I'll have time for this, since it is a larger architectural change.
Today I hope to finish up the caching and also to write the manual, if there is time.
Today I finished exposing the required data and then I wrote a function to get the status for the current day. I tested this on real data and it seems to be working fine. I have gotten a few Fitbit errors that I believe are unrelated to my implementation: a 503 error, because their site was briefly down, and a 429 error, because I've made too many requests in a small period of time.
The next step is to update the command line interface to use my functions. After that, I should write the manual. Maybe I can finish this later today. We'll see.
I also finished the command line interface. Now I just need to check it and document how to use this function. From there, it will be ready to share.
Today is all about documentation. It would also be good to try on a Windows machine to see if it works.
I ended up testing and finding a bug where the status is not being generated correctly for day one. I tried several approaches to fix this. This will take more thinking.
Today I will work on fixing the status for the first day. If there is time, I will test on Windows.
I fixed the status for the first day, improved the status readout message, tested on windows, and began adding documentation (although not committed yet). In writing some documentation, I got to thinking about how to deploy this. I think the easiest way is to bundle it and ship it on Github. I created a Github Action to do this.
My next step is to keep writing the documentation.
Today is a documentation day.
I ended up making several small code improvements. I a way of generating the config file and also made the output of status nicer and added more information.
Today hopefully I will finish the documentation. I will also test on Windows.
Today, I think that I finished the documentation. I'll go back and read through it once and test that it all is well put together.
Today is testing on Windows day.