[Epic E] Code & Test Quality Audit

[chris-prener]

[Epic E] Code & Test Quality Audit

Goal

The zippeR package has undergone a comprehensive audit of its test suite and source code, producing a prioritized catalog of coverage gaps and improvement opportunities. Follow-on issues have been filed and scoped for each actionable finding.

Why now?

Epics A–D addressed the initial wave of quality and infrastructure work, but the analysis was scoped to known issues at the time. A fresh, systematic audit ensures nothing was missed and surfaces second-order improvements — edge-case coverage, performance opportunities, and code patterns that emerged during the first round of refactoring.

Acceptance criteria (epic-level)

• Test suite analysis complete: every exported function mapped to its test coverage status (covered / partial / uncovered).
• Source code analysis complete: improvement opportunities identified and categorized (refactor, performance, readability, correctness).
• Follow-on issues filed for each actionable finding, linked as sub-issues to this epic.
• All filed sub-issues triaged with priority and WSJF.

Decisions and constraints

• No ADR needed — this epic is an analysis/audit effort, not an architectural decision. No structural choices that future sessions might re-litigate.
• Findings that do require architectural decisions will trigger ADR filing at sub-issue pickup time.

Sub-issues

Documentation quality

• Replace \donttest{} / \dontrun{} with @examplesIf across all roxygen2 examples #77 — Replace \donttest{} / \dontrun{} with @examplesIf across all roxygen2 examples

Analysis tasks

• [Capture] Analyze testing suite for coverage gaps #43 — Analyze testing suite for coverage gaps
• [Capture] Deep analysis of source code for improvement opportunities #44 — Deep analysis of source code for improvement opportunities
• [Capture] Analyze CI pipeline for speed improvements #50 — Analyze CI pipeline for speed improvements
• research: evaluate opportunities to reduce package dependencies #37 — research: evaluate opportunities to reduce package dependencies

Dependency reduction (from #37)

• refactor: drop purrr dependency (replace map_dfr with Base R) #80 — Drop purrr (Phase 1)
• refactor: drop spatstat.univar dependency (internal weighted.median) #81 — Drop spatstat.univar (Phase 1)
• refactor: drop stringr dependency (replace with Base R string functions) #82 — Drop stringr (Phase 1)
• refactor: replace httr with httr2 or base R HTTP #83 — Replace httr (Phase 2)
• refactor: replace readr with utils::read.csv() for CSV reading #84 — Replace readr (Phase 2)
• refactor: replace tidyr::pivot_wider with stats::reshape() #85 — Replace tidyr (Phase 2)
• research: evaluate dplyr removal feasibility (108 call sites) #86 — Evaluate dplyr removal (Phase 3, deferred)
• enhancement: support bare (unquoted) column name inputs via base R NSE #88 — Support bare (unquoted) column inputs via base R NSE (Phase 3, deferred)

High-priority bugs (from #44)

• fix: zi_aggregate() weighting produces cartesian expansion for intensive vars #55 — fix: zi_aggregate() weighting produces cartesian expansion for intensive vars
• fix: zi_load_labels_list() filter compares column to itself #52 — fix: zi_load_labels_list() filter compares column to itself
• fix: zi_load_crosswalk() COUNTYSUB branch is dead code #56 — fix: zi_load_crosswalk() COUNTYSUB branch is dead code
• fix: zi_load_labels() USPS vintage filter uses wrong column name #53 — fix: zi_load_labels() USPS vintage filter uses wrong column name
• fix: zi_repair() strips leading zeros via numeric coercion #57 — fix: zi_repair() strips leading zeros via numeric coercion
• fix: zi_get_geometry() county download failure path references undefined variable #54 — fix: zi_get_geometry() county download failure path references undefined variable

Test coverage gaps (from #43)

• test: add coverage for HUD crosswalk/loading paths #58 — test: add coverage for HUD crosswalk/loading paths
• test: add positive-path assertions for geometry, demographics, and label functions #62 — test: add positive-path assertions for geometry, demographics, and label functions
• test: add coverage for aggregation internals (intensive, weights, decennial) #59 — test: add coverage for aggregation internals (intensive, weights, decennial)

Medium-priority improvements (from #44)

• fix: input validation gaps across multiple functions #63 — fix: input validation gaps across multiple functions
• fix: error handling improvements for remote/network operations #61 — fix: error handling improvements for remote/network operations
• fix: API consistency issues (docs vs implementation drift) #64 — fix: API consistency issues (docs vs implementation drift)
• refactor: DRY and performance improvements #60 — refactor: DRY and performance improvements

Dependencies

None.

Related work

• Requirements doc: none — tiny-epic bypass
• Supporting objective: O1 (CRAN-ready package quality) in docs/OBJECTIVES.md
• ADRs: none — no-adr-needed: audit/analysis epic with no architectural choices
• Predecessors: [Epic A] [Epic A] Test Coverage & Quality #22 (Test Coverage), [Epic B] [Epic B] Code Style & Error Handling #23 (Code Style)
• Successors: Sub-issues to be filed from audit findings

Out of scope

• Implementing the fixes — this epic produces the analysis and files the issues; implementation is follow-on work.
• Re-running Epic A or B work — this is additive analysis, not a redo.
• CI/CD pipeline changes — infrastructure work is out of scope.