Allow deploy by artifact_id

[RadxaYuntian]

Sometimes my workflow failed and I need to rerun it. I have both upload-pages-artifact and deploy-pages in the same job, and during the 2nd run deploy-pages will fail with:

Run actions/deploy-pages@v4
  with:
    token: ***
    timeout: 600000
    error_count: 10
    reporting_interval: 5000
    artifact_name: github-pages
    preview: false
  env:
    GITHUB_PAGES: true
Fetching artifact metadata for "github-pages" in this workflow run
Found 2 artifact(s)
Error: Fetching artifact metadata failed. Is githubstatus.com reporting issues with API requests, Pages, or Actions? Please re-run the deployment at a later time.
Error: Error: Multiple artifacts named "github-pages" were unexpectedly found for this workflow run. Artifact count is 2.
    at getArtifactMetadata (/home/runner/work/_actions/actions/deploy-pages/v4/src/internal/api-client.js:89:1)
    at processTicksAndRejections (node:internal/process/task_queues:95:5)
    at Deployment.create (/home/runner/work/_actions/actions/deploy-pages/v4/src/internal/deployment.js:66:1)
    at main (/home/runner/work/_actions/actions/deploy-pages/v4/src/index.js:30:1)
Error: Error: Multiple artifacts named "github-pages" were unexpectedly found for this workflow run. Artifact count is 2.

upload-pages-artifact already outputs artifact_id which should be unique between runs, and #251 seems already have the plumbing laid out. If we can pass this artifact_id to deploy-pages there will be no ambiguity for which artifact we want to upload.

[rajsite]

Hit this pretty regularly unfortunately. Workaround is to manually go in and delete artifacts from the previous builds before re-running jobs. Which is a bummer when I might want to come back and reference those again to learn more about the failures.