Issue

Mirroring internal CR 208912038

Description

This change contains refactors to improve code reuse and clean up the logical flow. Specifically, it:

• updates the data! and processed_data! macro to define the function to instantiate all the data structs for record/report and add them to PerformanceData/VisualizationData, which allows us to
  • remove the ctor static functions defined in each data module that have same implementation
  • have a centralized place to collect all data names and register all the data structs (which pave the way for the skip-data feature)
  • have a simpler flow where only the needed data are initialized, instead of initializing all of them (which has side effects) and then removing unwanted ones based on user input
  • remove the lock wrapped around the PerformanceData and VisualizationData objects, which incurs extra costs, since now everything is deterministically sequential
• Consolidate the naming mechanism of all data. Now we have data name = rust module name = rust file name = js file name (through get_data_name_from_type<T>). This pattern benefits from Rust's design where the module definition resides in the file with the same name, and hence eliminates the need to explicitly and redundantly defines a name for a data.

By submitting this pull request, I confirm that you can use, modify, copy, and redistribute this contribution, under the terms of your choice.