Correct the sub-cluster counts stated on cluster pages
Four cluster pages stated a number of sub-clusters that differed from the list beneath it.
| Cluster | Stated | Listed | Fix |
|---|---|---|---|
| 2 | 8 | 7 | Description now says 7 |
| 7 | 6 | 7 | The 7th was a citation parsed as a heading; it moves back to "Rights retention strategies", so 6 are listed |
| 8 | 6 | 7 | Description now says 7, as the Google Doc does |
| 9 | 17 | 19 | Description now says 19, as the Google Doc does |
Cluster 7 wasn't in #964; it uses different wording ("This cluster has 6").
Why this edits the JSON directly: a full re-sync from the doc would currently empty Cluster 11 and add a false sub-cluster to Cluster 2 (#978). So this is a scoped edit, as in #826, and content/clusters/cluster-2/index.md comes from re-running generate_cluster_pages.py.
Needs a content decision: the doc still says 8 for Cluster 2 and names "Survey Design & Analysis", which has no section there. If that sub-cluster is restored, the count goes back to 8. This PR describes what the site lists today.
Timing: data/clusters_v4.json is in the data artifact, so the new text reaches forrt.org after the next data-processing run on main.
Verified: in a local build (Hugo 0.164) the stated count equals the number of sections on clusters 2, 7, 8 and 9. scripts/tests/test_clusters_data.py is new and fails on all four clusters without the data change. check_seo_html.py exits cleanly.
Fixes #964
Created with Claude Opus 5.5 in Claude Code.
Summary by cubic
Corrects the stated sub-cluster counts on cluster 2, 7, 8, and 9 pages so they match the sub-clusters actually listed beneath them.
- Moves a citation misparsed as a sub-cluster heading in cluster 7 back into "Rights retention strategies", reducing its listed sub-clusters from 7 to 6.
- Updates the descriptions of clusters 2, 8, and 9 to state the counts currently listed on the site.
- Adds
scripts/tests/test_clusters_data.py, which fails on all four clusters without the data change.
Migration
data/clusters_v4.json is in the data artifact, so the corrected text reaches forrt.org after the next data-processing run on main. The Google Doc still states 8 sub-clusters for cluster 2 and names "Survey Design & Analysis", which has no section there; if that sub-cluster is restored later, the count should go back to 8.