DEV: Node color repeats

[ikb6]

I thought we had at least 12 unique node colors before it starts to repeat. I tried the Guam tree from the nCOV SPHERES dataset. Colored by rare mutations. Colors repeat on the seventh color.

[billswitzer2]

Right. I also suggest we re-evaluate this limitation. From: ikb6 ***@***.***> Sent: Friday, July 17, 2026 9:38 AM To: CDCgov/MicrobeTrace ***@***.***> Cc: Subscribed ***@***.***> Subject: [CDCgov/MicrobeTrace] DEV: Node color repeats (Issue #1736) CAUTION: This email originated from outside of the organization. Do not click links or open attachments unless you recognize the sender and know the content is safe.

…

________________________________ [https://avatars.githubusercontent.com/u/33426295?s=20&v=4]ikb6 created an issue (CDCgov/MicrobeTrace#1736)<#1736> I thought we had at least 12 unique node colors before it starts to repeat. I tried the Guam tree from the nCOV SPHERES dataset. Colored by rare mutations. Colors repeat on the seventh color. image.png (view on web)<https://github.com/user-attachments/assets/9887f6ac-d3ce-4a7c-ac4e-e9bb4441a4cd> — Reply to this email directly, view it on GitHub<#1736?email_source=notifications&email_token=AHMTDBMERI6ONSKFFFJFGAD5FITZZA5CNFSL4Z3JMQ5C6L3HNF2C22DVMIXUS43TOVSS6NBZGEYTQMJZGMZTJJTSMVQXG33OVJZXKYTTMNZGSYTFMSSWK5TFNZ2KYZTPN52GK4S7MNWGSY3L>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/AHMTDBJF3UGY2U6JGSMQSNT5FITZZAVCNFSNUABFKJSXA33TNF2G64TZHMYTIOJXHA3DEMJSHNEXG43VMU5TIOJRGE4DCOJTGM2KC5QC>. Triage notifications, keep track of coding agent tasks and review pull requests on the go with GitHub Mobile for iOS<https://github.com/notifications/mobile/ios/AHMTDBJRNXYSLK4N47VUWLT5FITZZA5CNFSL4Z3JMQ5C6L3HNF2C22DVMIXUS43TOVSS6NBZGEYTQMJZGMZTJJTSMVQXG33OVJZXKYTTMNZGSYTFMSSWK5TFNZ2KUZTPN52GK4S7NFXXG> and Android<https://github.com/notifications/mobile/android/AHMTDBJR6IFVUUUYDD2QMAL5FITZZA5CNFSL4Z3JMQ5C6L3HNF2C22DVMIXUS43TOVSS6NBZGEYTQMJZGMZTJJTSMVQXG33OVJZXKYTTMNZGSYTFMSSWK5TFNZ2K4ZTPN52GK4S7MFXGI4TPNFSA>. Download it today! You are receiving this because you are subscribed to this thread.Message ID: ***@***.******@***.***>>

[glstott]

It looks like this bug was caused by saved sessions containing a per-field color cache with just seven colors. The color-mapping code then treated the cached values as the full palette.

Suspected cause:
These lines from commit d3913625:

// Lines 88–94: use the saved palette without checking its size
if (!updatedColorsTable[nodeColorVariable]) {
  updatedColorsTable[nodeColorVariable] = updatedNodeColors;
} else {
  updatedNodeColors = [...updatedColorsTable[nodeColorVariable]];
}

// Lines 109–116: repeat whichever palette was loaded
if (distinctValues.length > updatedNodeColors.length) {
  let expandedColors: string[] = [];
  let neededTimes = Math.ceil(distinctValues.length / updatedNodeColors.length);
  while (neededTimes-- > 0) {
    expandedColors = expandedColors.concat(updatedNodeColors);
  }
  updatedNodeColors = expandedColors;
}

// Lines 131–137: reapply stale cyclic colors from history
if (indexInHistory !== -1) {
  updatedNodeColors[i] = updatedColorsTableHistory[val];
} else {
  updatedColorsTableHistory[val] = updatedNodeColors[i];
}

Doing some additional tests before submitting a pull request.

[glstott]

Added a fix and submitted pull request to resolve. It appears to be leakage between variables if you swap between fields of different sizes. Review PR #1737 for more details and for a synthetic dataset with instructions to reproduce reliably.