Refresh the fleet from its published sites, and say ten everywhere

_fleet/README.md

@@ -42,7 +42,7 @@ python3 scripts/fleet/assemble_page.py --check
 
 Membership updates do not require rescanning the record corpora. The current
 September 2026 vocabulary census covers nine Mechs; TaxonMech is shown in the
-graph and cards with its published 625,960-taxon total (checked September 15, 2026), but its vocabulary counts have not
+graph and cards with its published 625,960-taxon total (checked September 20, 2026), but its vocabulary counts have not
 been measured by this pipeline. The heatmap uses the measured `fleet_data.json`
 order, and the page states this limitation. Add TaxonMech to the census roots
 and scanners before publishing measured vocabulary cells or overlap counts for it.
@@ -80,13 +80,38 @@ NaturalProductMech and TaxonMech both publish browse sites linked from their
 cards; these links are separate from the historical census's record-link routes.
 
 MIBiG and NPAtlas are carried through the whole pipeline alongside the
-ontologies, because they are how NaturalProductMech grounds its corpus.
+ontologies, because they are how NaturalProductMech cites its corpus. MIBiG is a
+seeded grounding source; NPAtlas is a cross-reference target only, since its
+licence bars ingestion into a CC BY 4.0 corpus.
 
-## Published-site refresh (September 15, 2026)
+`prefix_census.py`'s prefix alternation is a hand-maintained list and is known to
+be incomplete: TOGO, UTEX and CCAP are absent although comparable registries
+(MediaDive, DSMZ, ATCC, GOLD) are present. TOGO is CultureMech's second-largest
+structured namespace at 2,833 occurrences, so the heatmap currently understates
+it. Adding a prefix changes the heatmap's columns, so it needs a full rescan.
+
+## Published-site refresh (September 20, 2026)
 
 Cards and graph details use the live sites linked by each card, following their
-landing-page redirects. MediaIngredientMech counts come from its published
-`data/ingredients.json` (2,951 ingredients, 2,616 MAPPED), as on its landing page.
-CultureMech reports 10,657 published-browser recipes, not the older merged-corpus
-count. The vocabulary census and overlap assets remain their earlier snapshot;
-the page explicitly separates these from the refreshed site totals.
+landing-page redirects. Those redirects are client-side meta refreshes, not HTTP
+3xx, so `curl -L` stops at the 200-response shell and never reaches the real
+browser: fetch the `pages/` (or `app/`) URL directly when checking a card.
+
+Read the number the page *renders*, not the one in its HTML source. Two landing
+pages ship a stale static figure that JavaScript overwrites at runtime from the
+Mech's own index: ProteinTraitsMech's source says 408,978 records and 29 sources
+while `data/facets.json` serves 429,291 and 34, and MediaIngredientMech's source
+carries literal ellipsis placeholders filled from `data/ingredients.json`
+(2,951 ingredients, 2,616 MAPPED).
+
+CultureMech is the one card that does not take its Mech's headline tile. That
+tile reads 10,657 "recipes", which matches no current data layer: it is the
+March 2026 normalized count, recorded as `input_recipes` in the Mech's tracked
+`data/merge_yaml/merge_stats_2026.json` (2026-03-15), and it was already four
+months stale when the tile was written on 2026-06-30, by which time the corpus
+was 15,878 normalized and 6,103 merged. Nothing regenerates it. The card states 6,286 canonical media instead, which
+is what CultureMech's own `pages/index.html` publishes and what the fleet census
+counts; the merged/normalized split is explained in the card's detail text.
+
+The vocabulary census and overlap assets are a separate, dated snapshot; the page
+explicitly separates these from the site totals.

_fleet/data/mech_stats.json

@@ -1,6 +1,6 @@
 {
- "as_of": "2026-09-16",
- "merged_prs_total": 2778,
+ "as_of": "2026-09-20",
+ "merged_prs_total": 2994,
  "mechs": [
   {
    "mech": "CultureMech",
@@ -13,58 +13,58 @@
   {
    "mech": "MediaIngredientMech",
    "repo": "MediaIngredientMech",
-   "records": 2957,
+   "records": 2951,
    "reviewed": null,
    "status_field": null,
-   "merged_prs": 359
+   "merged_prs": 361
   },
   {
    "mech": "CommunityMech",
    "repo": "CommunityMech",
    "records": 364,
    "reviewed": null,
    "status_field": null,
-   "merged_prs": 528
+   "merged_prs": 534
   },
   {
    "mech": "TraitMech",
    "repo": "TraitMech",
-   "records": 618,
+   "records": 694,
    "reviewed": 427,
    "status_field": "mapping_status",
-   "merged_prs": 517
+   "merged_prs": 585
   },
   {
    "mech": "ProteinTraitsMech",
    "repo": "proteintraitsmech",
    "records": 429291,
    "reviewed": 39591,
    "status_field": "mapping_status",
-   "merged_prs": 314
+   "merged_prs": 316
   },
   {
    "mech": "AntibioticMech",
    "repo": "AntibioticMech",
    "records": 2939,
    "reviewed": 80,
    "status_field": "curation_status",
-   "merged_prs": 160
+   "merged_prs": 243
   },
   {
    "mech": "CellStructureMech",
    "repo": "CellStructureMech",
-   "records": 338,
+   "records": 421,
    "reviewed": 0,
    "status_field": "mapping_status",
-   "merged_prs": 425
+   "merged_prs": 465
   },
   {
    "mech": "HabitatMech",
    "repo": "HabitatMech",
    "records": 3206,
-   "reviewed": 684,
+   "reviewed": 686,
    "status_field": "mapping_status",
-   "merged_prs": 130
+   "merged_prs": 144
   },
   {
    "mech": "NaturalProductMech",
@@ -80,7 +80,7 @@
    "records": 625960,
    "reviewed": 0,
    "status_field": "mapping_status",
-   "merged_prs": 15
+   "merged_prs": 16
   }
  ]
 }

_fleet/data/prefix_census.json

@@ -2,21 +2,21 @@
  "HabitatMech": {
   "files": 3206,
   "prefixes": {
-   "NCBITaxon": 13226,
-   "ENVO": 3319,
+   "NCBITaxon": 13227,
+   "ENVO": 3368,
    "GOLD": 2562,
-   "BTO": 1363,
-   "UBERON": 1115,
-   "PMID": 513,
-   "FOODON": 200,
+   "BTO": 1367,
+   "UBERON": 1123,
+   "PMID": 516,
+   "FOODON": 198,
    "BacDive": 167,
    "PO": 117,
-   "DOI": 112,
+   "DOI": 116,
    "MESH": 83,
-   "NCIT": 37,
-   "CHEBI": 22,
-   "PATO": 4,
-   "GO": 2
+   "NCIT": 38,
+   "CHEBI": 24,
+   "GO": 7,
+   "PATO": 4
   }
  },
  "CommunityMech": {
@@ -35,12 +35,12 @@
   }
  },
  "TraitMech": {
-  "files": 618,
+  "files": 694,
   "prefixes": {
-   "DOI": 6088,
-   "METPO": 3453,
-   "RO": 2416,
-   "NCBITaxon": 955,
+   "DOI": 6749,
+   "METPO": 3541,
+   "RO": 2559,
+   "NCBITaxon": 975,
    "GO": 886,
    "CHEBI": 673,
    "PMID": 412,
@@ -54,44 +54,44 @@
   }
  },
  "CellStructureMech": {
-  "files": 338,
+  "files": 421,
   "prefixes": {
-   "GO": 5994,
-   "DOI": 4800,
-   "NCBITaxon": 1040,
+   "GO": 8623,
+   "DOI": 5931,
+   "NCBITaxon": 1211,
+   "PMID": 308,
    "UniProt": 284,
-   "PMID": 274,
    "ECO": 145,
    "PDB": 67,
-   "InterPro": 56,
-   "CHEBI": 45,
+   "InterPro": 58,
+   "CHEBI": 48,
    "Pfam": 22,
+   "ComplexPortal": 16,
    "METPO": 14,
    "RO": 7,
-   "ComplexPortal": 5,
+   "SO": 5,
    "MICRO": 4,
-   "SO": 4,
    "EMDB": 3,
    "BFO": 1
   }
  },
  "ProteinTraitsMech": {
   "files": 429291,
   "prefixes": {
-   "InterPro": 1478104,
-   "UniProt": 657598,
+   "InterPro": 1558488,
+   "UniProt": 770276,
    "PMID": 657527,
    "RHEA": 613456,
-   "Pfam": 521966,
+   "Pfam": 558504,
    "GO": 421586,
    "ARO": 397717,
    "CHEBI": 374424,
-   "NCBITaxon": 358314,
+   "NCBITaxon": 371580,
    "RO": 278407,
    "PDB": 235957,
    "DOI": 193845,
-   "NCBIfam": 139525,
-   "EC": 119999,
+   "NCBIfam": 141317,
+   "EC": 120016,
    "BFO": 74488,
    "ComplexPortal": 20753,
    "SO": 19302,
@@ -126,7 +126,7 @@
    "ARO": 16653,
    "CAS": 1823,
    "PMID": 1207,
-   "NCBITaxon": 956,
+   "NCBITaxon": 1057,
    "PubChem": 592,
    "DrugBank": 508,
    "UniProt": 278,
@@ -137,10 +137,10 @@
   }
  },
  "MediaIngredientMech": {
-  "files": 2957,
+  "files": 2951,
   "prefixes": {
-   "CHEBI": 7914,
-   "CAS": 692,
+   "CHEBI": 7895,
+   "CAS": 689,
    "MICRO": 345,
    "MESH": 301,
    "NCIT": 287,
@@ -174,5 +174,5 @@
    "CAS": 1
   }
  },
- "_as_of": "2026-09-15"
+ "_as_of": "2026-09-20"
 }