Code review of R analysis

[mbrotos]

No description provided.

[Copilot]

Pull Request Overview

This PR updates the data-loading function’s default behavior and adds a new end-to-end test script for the analysis functions.

• Changed load_parquet() to use filter_tail_ns = -1 and dropped the hardcoded "warehouse" filter.
• Added analysis/acc_cum_test.R with a synthetic dataset and sanity checks for is_correct(), cumulative accuracy, and performance functions.

Reviewed Changes

Copilot reviewed 2 out of 2 changed files in this pull request and generated 1 comment.

File	Description
analysis/common_functions.R	Updated load_parquet() signature and removed the warehouse filter
analysis/acc_cum_test.R	New test script with sample data, function calls, and printouts

Comments suppressed due to low confidence (3)

analysis/acc_cum_test.R:1

• Use library(tidyverse) instead of require() so the script fails immediately if the package isn’t installed, rather than continuing silently.

require(tidyverse)

analysis/acc_cum_test.R:3

• Since tidyverse already loads dplyr, this library(dplyr) call is redundant and could be removed to avoid duplication.

library(dplyr)

analysis/acc_cum_test.R:6

• [nitpick] Consider adding unit tests for load_parquet(), particularly around filter_tail_ns behavior and iteration filtering, to ensure coverage of the updated defaults.

source("analysis/common_functions.R")

[Copilot]

The parameter filter_tail_ns is no longer referenced inside load_parquet(). Consider removing it or adding filtering logic so it isn’t a redundant argument.

Suggested change

	load_parquet <- function(path, is_table_lock=FALSE, filter_tail_ns=-1) {
	load_parquet <- function(path, is_table_lock=FALSE, filter_tail_ns=-1) {
	# Load a parquet file and optionally filter rows based on the `filter_tail_ns` parameter.
	# `filter_tail_ns` specifies a time threshold for filtering rows. If set to -1 (default), no filtering is applied.