Follow {tidyverse} principles

[wendtke]

In addition to reading

• tidy eval
• tidyverse style guide
• tidyverse design principles
• tidy tools manifesto

consider (@geanders suggestions):

• 1. input/output data in same format (allows functions to retain order)
• 2. common prefix to function name (psyphr_read_wb())
• 3. check tidy eval book on how to manage column-naming conventions within mutate to allow users to bring in non-MindWare data
• 4. create unique geom
• 5. check lubridate or tidyr for examples of maintaining consistency across functions and packages (e.g., verbose = TRUE option within function)
• 6. create umbrella package with modular packages within it to wrangle raw and output, visualize, analyze (see Evaluate BIDS Schema #52)

[iqis]

1. I wonder what does "allow functions to retain order" mean?
2. This can be a good option if we have a lot of functions and some may have namespace collision with other packages. I doubt if that's necessary for now.
3. Let's bring this up in the next talk.
4. Good idea.
5. ditto
6. ditto; do we want to test the water with a creating package for BIDS? I've taken a look at their Python package, it's pretty simple compared to WFDB.

[wendtke]

1. I wonder what does "allow functions to retain order" mean?

I think it means we don't have to reorganize the order of functions within .R scripts.

2. This can be a good option if we have a lot of functions and some may have namespace collision with other packages. I doubt if that's necessary for now.

It could also help with user-friendliness; newer users who aren't as familiar with function names and purposes might like the tab-completion option of psyphr_* or psyphr_tidy_*.

3. Let's bring this up in the next talk.

Okay, I made a new doc.

4. Good idea.

She said creating geoms is difficult and would help when the time came.

5. ditto

Ditto what? Good idea or paste on Monday's agenda?

6. ditto; do we want to test the water with a creating package for BIDS? I've taken a look at their Python package, it's pretty simple compared to WFDB.

I think we should get this package out the door first then explore the modular options. We should be very deliberate in how we name things, though. Maybe "psyphr" should be the umbrella package.