fix(memory): Reduce memory and cpu usage for standardizeColNames

[tonywu1999]

Motivation and Context

Using the dataset from the Fragpipe-MSstats Nature Protocols paper, I discovered that there are sections of the MSstatsConvert code that are not efficient in space and time.

In one case, when running .standardizeColnames here, I saw that the memory available decreased by 70GB (despite our dataset being 20GB). This code also took 6.5 minutes to run.

Changes

• Refactor .standardizeColnames to be more efficient w.r.t. space and time.
• The idea is that stringi is memory and time intensive when processing large vectors. But if we have millions of rows, there are usually many duplicate run names, so we only need to run stringi operations on unique run names. Then, we can do a table lookup for all runs, which is O(1) for each lookup.

Testing

• Added and modified unit tests that verified that number of rows in converter datasets didn't change from before and that standardizeColnames does not change the order or length of the input.
• Using the Nature Protocols dataset
  • Memory used within the function decreased from 70GB to 20GB.
  • For the original implementation, memory available had decreased by 70GB when comparing before and after the function call. On the other hand, memory available remain unchanged for this new implementation.
  • Runtime decreased from 6.5 minutes to 12 seconds.
  • Although RStudio memory used isn't necessarily reliable, it reported an interesting metric after the converter completed & after running garbage collection:
    • Original implementation RStudio memory used: 68GB
    • New implementation RStudio memory used: 21GB
    • My explanation here is that our old implementation caused so much overhead that even garbage collection couldn't properly clean up unused memory.

Checklist Before Requesting a Review

• I have read the MSstats contributing guidelines
• My changes generate no new warnings
• Any dependent changes have been merged and published in downstream modules

[mstaniak]

Well done, thank you

[devonjkohler]

Looks great!