Modernize pipe operator and recode() usage

[jt14den]

Summary

• Converts %>% to the native pipe |> throughout episodes 1, 3, 4, and 5, matching the current tidyverse style guide recommendation for R 4.1+ (this lesson already requires R 4.5+ per renv.lock).
• Replaces dplyr::recode(), which is marked [Superseded] on the current dplyr reference page, with case_match() in episodes 3 and 5. Each conversion includes an explicit .default argument to preserve recode()'s pass-through behavior for any value not in the mapping list, since case_match() returns NA for unmatched values by default.
• Updates the accompanying prose (objectives, section headings, exercise text, keypoints) to match.

Test plan

• sandpaper::validate_lesson() passes with no errors
• All four modified episodes knit cleanly from a fresh session with no rendered R errors
• Verified case_match() output is byte-identical to recode() output when run against the actual episodes/data/books.csv for both subCollection and format
• Confirmed every raw code value in the dataset is covered by the mapping list, so .default is a safety net rather than currently masking a data problem
• Would appreciate a second read on the case_match() conversions in episode 3 (lines ~203, ~221, ~551) and episode 5 (lines ~43, ~59, ~156) before merge, since this is a real behavioral change to double-check, not just a rename

[github-actions]

🆗 Pre-flight checks passed 😃

This pull request has been checked and contains no modified workflow files, spoofing, or invalid commits.

Results of any additional workflows will appear here when they are done.

[nickyjgarland]

The changes in syntax all seem fine but while checking case_match() as a replacement I noted in the documentation that this function was also depreciated. Is there a benefit of using this function or should we switch to a suggested alternative such as recode_values()?