Fix early-year growth estimation and make integration tests year-independent

[nicklegan]

Summary

Fixes growth estimation logic when running early in the year (< 60 days) and resolves CI failures caused by year-dependent fixture comparisons.

Changes

Growth estimation fix

• When early in the year, CalculateNewEstimate now projects current year totals using last year's growth delta instead of copying previous year's values unchanged
• This ensures displayed totals are consistent with displayed deltas

Test improvements

• Added unit tests for CalculateNewEstimate covering early-year, mid-year, and fallback scenarios
• Made integration tests deterministic by using dynamic year for test commits and filtering year-based output rows

Before

Early-year estimate showed previous year's totals with a delta suggesting growth (inconsistent).

After

Early-year estimate shows projected totals that match the applied delta (consistent).

[Copilot]

Pull request overview

Fixes growth estimation behavior early in the year so projected totals and deltas stay consistent, and updates integration testing to be year-independent/deterministic across calendar years.

Changes:

• Update CalculateNewEstimate early-year logic to project using the prior year’s growth delta (when available).
• Add unit tests covering early-year, later-year, and fallback estimation scenarios.
• Make integration outputs deterministic by normalizing/removing year-dependent rows and updating fixtures accordingly.

Reviewed changes

Copilot reviewed 5 out of 5 changed files in this pull request and generated 1 comment.

Show a summary per file

File	Description
script/remove-non-deterministic-rows	Filters additional non-deterministic rows and normalizes leading year values to YYYY for stable fixture comparisons.
script/build-test-repository	Uses the current year when creating test-repo commits so integration tests don’t go stale year-to-year.
pkg/display/sections/growth_tables_test.go	Adds unit tests for CalculateNewEstimate across early/mid-year and fallback cases.
pkg/display/sections/growth_estimates.go	Adjusts early-year estimation to apply last year’s delta (using two-years-ago data when present).
fixtures/git-metrics.txt	Updates fixture output to match normalized year placeholders and removed commit-date rows.

Comments suppressed due to low confidence (1)

script/remove-non-deterministic-rows:50

• The success check at the end now keys off the exit status of the last rm -f (or sed/rm), not the grep/filtering step (and even previously it didn’t validate grep). This means the script will report success and overwrite the input even if grep/sed fails. Capture and validate grep/sed exit codes explicitly (treating grep exit 1 as acceptable when it matches all lines), or enable set -e and restructure with a trap for temp-file cleanup.

grep -Ev -f "$pattern_file" "$input_file" > "$temp_file"
rm -f "$pattern_file"

# Normalize year references to YYYY for deterministic comparison
sed -i.bak -E 's/^([0-9]{4})([\^~\* ]+)/YYYY\2/g' "$temp_file"
rm -f "$temp_file.bak"

# Check if the temporary file was created successfully
if [ $? -eq 0 ]; then
  mv "$temp_file" "$input_file"
  echo "Non-deterministic rows removed successfully. Output written to '$input_file'."