Enable autocomplete for undeployed seqera-config nodes

[ewels]

Summary

• Enable autocomplete for launchpad names and data link names to work immediately after the seqera-config node is saved, without requiring deployment
• Use RED.nodes.node() in the frontend to get config values from editor state
• Use RED.nodes.getCredentials() in the backend to look up saved API tokens for undeployed config nodes

Problem

Previously, when configuring nodes like workflow-launch or datalink-poll, the autocomplete functionality would only work after the seqera-config node had been deployed. This meant users had to:

1. Create/edit the seqera-config node
2. Deploy the entire flow
3. Only then could they use autocomplete in other nodes

This was frustrating for new users setting up their first flow.

Solution

The autocomplete now works as soon as the seqera-config node is saved (clicking "Done" or "Update" in the config dialog). No deployment required.

Frontend changes:

• Use RED.nodes.node(configId) to access config node properties from editor state
• Pass baseUrl and workspaceId as query parameters to the backend

Backend changes:

• Accept baseUrl/workspaceId from query params when config node isn't deployed
• Use RED.nodes.getCredentials(configId) to look up saved API token
• This works because credentials are saved immediately when the config dialog is closed

Test plan

• Test autocomplete works immediately after selecting an existing (deployed) config
• Test autocomplete works after creating a new config, saving it (Done), but not deploying
• Test autocomplete with workspace ID override
• Test graceful failure when no config is selected
• Test graceful failure when config has no saved credentials
• Verify all 165 existing tests still pass

🤖 Generated with Claude Code

[github-actions]

🐳 Docker PR Build

Status	✅ Built successfully
Commit	6364970b7efd4f8d0ef3c48c72d757610c55a5a0
Standard	ghcr.io/seqeralabs/node-red-seqera:pr-31
Studios	ghcr.io/seqeralabs/node-red-seqera-studios:pr-31

View workflow run