[Bug]: Color Coding Metadata in Dupe Groups is Inconsistent #115

New issue

Closed

opened 2026-06-21 23:34:10 +00:00 by ValleyGeek · 2 comments

ValleyGeek commented

2026-06-21 23:34:10 +00:00

Owner

App version

0.1.54 Beta

Platform

Both

Steps to reproduce

Open a directory with visual duplicates
Scan and review duplicates of different attributes (resolution or size)

Expected behavior

The highest value set in any category is always highlighted in green. Matching, but not the highest/lowest are in yellow. The lowest value set in any category is highlighted in red

Actual behavior

Some top values sets are yellow, no green is displayed or sometimes they are given no highlight.

Additional context

No response

### App version 0.1.54 Beta ### Platform Both ### Steps to reproduce 1. Open a directory with visual duplicates 2. Scan and review duplicates of different attributes (resolution or size) ### Expected behavior The highest value set in any category is always highlighted in green. Matching, but not the highest/lowest are in yellow. The lowest value set in any category is highlighted in red ### Actual behavior Some top values sets are yellow, no green is displayed or sometimes they are given no highlight. ### Additional context _No response_

ValleyGeek added the

Priority/Medium

Kind/Bug

labels

2026-06-21 23:34:10 +00:00

Claude referenced this issue from a commit

2026-06-24 02:01:04 +00:00

fix(dupe): highlight partial-tie metadata values consistently (#115)

Claude commented

2026-06-24 02:05:12 +00:00

Member

Root cause: compare_tier() in core/dupe_metadata_compare.py only returned "tie" (yellow) when every value in a duplicate group was identical. A value shared by some — but not all — images, without itself being the group's overall highest or lowest (e.g. two images tied at a middling resolution with a third strictly lower), fell through to "none" and got no highlight instead of yellow. This is the "some top sets are yellow, no green, or no highlight" inconsistency described.

Fix (branch 0.2.2/issue-115-dupe-color-coding, stacked on 0.2.2/issue-114-soft-lock-after-scan): check the overall max/min first (so ties at the extremes always render green/red), then fall back to a count-based check for matching non-extreme values (yellow). Added regression tests for the partial-tie case. Work continues on the branch; PR to main pending user request.

Root cause: `compare_tier()` in `core/dupe_metadata_compare.py` only returned `"tie"` (yellow) when *every* value in a duplicate group was identical. A value shared by some — but not all — images, without itself being the group's overall highest or lowest (e.g. two images tied at a middling resolution with a third strictly lower), fell through to `"none"` and got no highlight instead of yellow. This is the "some top sets are yellow, no green, or no highlight" inconsistency described. Fix (branch `0.2.2/issue-115-dupe-color-coding`, stacked on `0.2.2/issue-114-soft-lock-after-scan`): check the overall max/min first (so ties at the extremes always render green/red), then fall back to a count-based check for matching non-extreme values (yellow). Added regression tests for the partial-tie case. Work continues on the branch; PR to `main` pending user request.

Claude referenced this issue from a commit

2026-06-24 04:21:28 +00:00

fix(dupe): render a full-group metadata tie as green, not yellow (#115)

Claude commented

2026-06-24 04:21:37 +00:00

Member

Follow-up refinement per direction: when every image in the group shares the same value for a metric (a full tie), that value is now rendered green ("high") rather than yellow ("tie") — since a full tie is simultaneously the group's highest and lowest. Partial ties (some, not all, images matching on a non-extreme value) still render yellow. Pushed to 0.2.2/issue-115-dupe-color-coding.

Claude referenced this issue from a commit

2026-06-24 05:31:14 +00:00

docs: PR-prep — bump to 0.2.4 for issue #115 dupe color-coding fix

Claude closed this issue

2026-06-24 05:31:21 +00:00

Claude referenced this issue from a commit

2026-06-24 05:37:44 +00:00

fix(dupe): consistent green/yellow/red metadata highlighting (#115)

Claude referenced this issue from a commit

2026-06-24 05:37:44 +00:00

docs: PR-prep — bump to 0.2.4 for issue #115 dupe color-coding fix

Claude referenced this issue from a pull request that will close it,