Y26-080 - Pool status indiciators

[BenTopping]

Closes #2661

Changes proposed in this pull request

• Adds pool status indicators (tick for valid, cross for invalid)

Instructions for Reviewers

This was done using antigravity IDE and requires some thorough testing before merge. For those interested this was the process:

• Start new agent conversation (using Gemini 3.1 pro (High) and planning mode).
• Paste in story (raw text).
• Confirm implementation plan.
• Review changes.
• I suggested: 'Move the icon SVG's to separate vue components in src/components/shares/icons'
  • Originally the agent had added the svg's inline in both components. This isnt best practice in traction so I requested it move them to a separate component.
• I suggested: 'Rename the icons to TractionCrossIcon and TractionTickIcon'
  • Originally the agent had named them WellPass and WellFail. This isn't ideal because these are reusable icons and don't need to relate to the wells.
• I suggested: 'Add unit tests to check the icon appears when expected. TickIcon appears when pool is valid, crossicon appears when pool is invalid, no icon appears when pool is empty'
  • Originally the agent had no unit tests. This was likely caused by there being no criteria for them in the story description.
• I suggested: 'Change icon data-attribute to be more descriptive, 'pass' should be 'tick-icon' and 'fail' should be 'cross-icon'.
  • Originally the agent used 'pass' and 'fail' data attributes. Not very descriptive.
• I suggested: 'in pacbioRunWell can we check against something better than the status classes, can we abstract the check in a similar way to flexiblepoolwell'
  • Originally the agent was checking for css classes inline based on the status function. Seemed flakey.

This took roughly an hour total but that was likely because the criteria was a bit vague. The total processing time was much less but each change requires approval. The agent also ran unit tests autonomously to check its changes were okay but this tripped up a few times with errors about node versions etc. It did not spin up a server and screenshot the changes, not sure if thats possible. It also did not lint the code, I did that manually.

I think the implementation would have been okay without any of the suggestions, it correctly identified the location for the changes, but the code would have been untested and poorly abstracted. I suspect this would have been quicker to build manually but it was an interesting experiment.