update proteinpaint-client to 2.170.22

[siosonel]

Description

Fixes

• Lollipop in OncoMatrix does not use the current cohort (https://gdc-ctds.atlassian.net/browse/SV-2775)
• Correlation Plot: Wrong data can be rendered when switching cohorts (https://gdc-ctds.atlassian.net/browse/SV-2774)

Also, use node 20.11.0 in proteinpaint/dev.sh script.

Checklist

• Added proper unit tests
• Left proper TODO messages for any remaining tasks
• Scanned for web accessibility with aXe, and mitigated or documented
  flagged issues

Screenshots/Screen Recordings (if Appropriate)

[siosonel]

@craigrbarnes @philtom-ctds I can't seem to get rid of the "peer": true changes to package-lock.json. Please let me know if that is an issue and/or if you'd like me to open a new PR with cleaner commit history. I've tried different approaches, including downloading a package-lock.json copy from the develop branch and re-running npm run dev to regenerate the package-lock while using node 20.11.0 and npm 10.2.4 .

EDIT: There was a clean package-lock diff for PR # 1673. I forked the sjpp-client-2.170.13 branch not from develop branch, but from sjpp-client-2.170.12 branch. Maybe I should do something similar in this PR?

[philtom-ctds]

@craigrbarnes @philtom-ctds I can't seem to get rid of the "peer": true changes to package-lock.json. Please let me know if that is an issue and/or if you'd like me to open a new PR with cleaner commit history. I've tried different approaches, including downloading a package-lock.json copy from the develop branch and re-running npm run dev to regenerate the package-lock while using node 20.11.0 and npm 10.2.4 .

EDIT: There was a clean package-lock diff for PR # 1673. I forked the sjpp-client-2.170.13 branch not from develop branch, but from sjpp-client-2.170.12 branch. Maybe I should do something similar in this PR?

(i clicked the wrong button and accidentally close this pr. i reopened it.)

the peer property itself is not erroneous. the issue was with npm 11 changing how it resolved and identified peer dependencies. there were multiple bugs in npm 11 that caused dependency resolution issue. for example
npm/cli#8645
and
npm/cli#8535 (comment)

i checked out this branch and ran npm i with npm v10. there were no changes to the lock file. the peer changes in this pr should be bringing the lock file back in sync with npm v10. (i believe the last pr used npm v11.)

[craigrbarnes]

Please remove the additions to dev.sh or provide justification for it's use.

[craigrbarnes]

Does this need to be added? Anyone who runs this will have their nvm environment changed

[siosonel]

I can change to not use source and only use the nvm binary. This proterinpaint/dev.sh script is only for developers who have a local proteinpaint repo, it's not connected to the gff npm run dev. It's both a convenience script and better way to not forget PP dev setup requirements within GFF.

export NVM_DIR="$HOME/.nvm"
[ -s "$NVM_DIR/nvm.sh" ] && \. "$NVM_DIR/nvm.sh"  # This loads nvm
[ -s "$NVM_DIR/bash_completion" ] && \. "$NVM_DIR/bash_completion"  # This loads nvm bash_completion

[paribartandhakal]

I don't think this is correct. We don't need to add this in root level package.json. It's already inside portal-proto's package.json where it is being used.

cc. @craigrbarnes

[craigrbarnes]

Yes I agree

[siosonel]

I'll fix

[siosonel]

@craigrbarnes @philtom-ctds I'll close this PR to generate a cleaner update. We'll also revert the fix for the race condition and implement a simpler fix.