feat: rename to paperspec and add the website

- Rename manuspec to paperspec everywhere outside pixi.toml: prose, the
  tools, provenance tool names and the fixture key derivations, so every
  fixture and vector is rebuilt.
- Move the namespace addresses under the site, mirroring the repository:
  schema $id https://paperspec.scient.ing/schemas/<file>, the JSON-LD
  context https://paperspec.scient.ing/context/scienting-1.jsonld, and
  term IRIs https://paperspec.scient.ing/vocab#<term>. Format
  identifiers (scienting-*) and URNs (urn:scienting:) are unchanged.
- tools/site.py (standard library only) builds _site/: every file of the
  repository at its own path, each Markdown file rendered as a page with
  sidebar navigation, heading anchors matching the forge's, folder
  listings, a vocabulary page for the context's IRIs, a heading search
  and a 404 page; light and dark themes.
- check.py builds the site and checks that every Markdown file has a
  page, every internal link and fragment resolves, no Markdown is left
  unrendered, and every $id, the context address and the vocabulary IRIs
  are files of the site.
- deploy/: a Caddyfile with the media types and CORS the data needs, and
  a guide; .gitea/workflows/site.yml builds on main and publishes with
  rsync when the SITE_* secrets are set.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

tools/check.py

@@ -37,7 +37,7 @@ from tools.oracle.tree import Refusal  # noqa: E402
 
 FAILS = []
 # Folders that are not part of the repository: version control, environments and caches.
-SKIP_DIRS = {".git", ".pixi", "node_modules", "__pycache__"}
+SKIP_DIRS = {".git", ".pixi", "node_modules", "__pycache__", "_site", ".claude"}
 COUNTS = {}
 
 
@@ -852,7 +852,7 @@ COUNT_PATTERNS = [
     (NUM + r" cases in all\b", "cases"),
 ]
 # Where the prose names the specification's own version; other standards' versions are not checked.
-VERSION_PATTERN = r"\b(?:manuspec|the formats are|the specification is at|this is|it is|is at|at) version (\d+\.\d+\.\d+)\b|\bmanuspec (\d+\.\d+\.\d+)\b"
+VERSION_PATTERN = r"\b(?:paperspec|the formats are|the specification is at|this is|it is|is at|at) version (\d+\.\d+\.\d+)\b|\bpaperspec (\d+\.\d+\.\d+)\b"
 
 
 def check_counts():
@@ -907,6 +907,88 @@ def check_counts():
                        % (rel, i, word, what, have[what]))
 
 
+def check_site():
+    """The website builds, every Markdown file has a page, every internal link and fragment resolves, no Markdown
+    is left unrendered, and the namespace addresses the formats name are files the site serves."""
+    from html.parser import HTMLParser
+    from tools import site
+
+    class Page(HTMLParser):
+        def __init__(self):
+            super().__init__(convert_charrefs=True)
+            self.ids, self.links, self.text, self.skip = set(), [], [], 0
+
+        def handle_starttag(self, tag, attrs):
+            a = dict(attrs)
+            if "id" in a:
+                self.ids.add(a["id"])
+            for k in ("href", "src"):
+                if a.get(k):
+                    self.links.append(a[k])
+            if tag in ("code", "pre", "script", "style"):
+                self.skip += 1
+
+        def handle_endtag(self, tag):
+            if tag in ("code", "pre", "script", "style"):
+                self.skip -= 1
+
+        def handle_data(self, data):
+            if not self.skip:
+                self.text.append(data)
+
+    with tempfile.TemporaryDirectory() as tmp:
+        out = os.path.join(tmp, "site")
+        site.build(out)
+        pages = {}
+        for d, _, fs in os.walk(out):
+            for f in fs:
+                if f.endswith(".html"):
+                    pg = Page()
+                    with open(os.path.join(d, f), encoding="utf-8") as fh:
+                        pg.feed(fh.read())
+                    pages[os.path.relpath(os.path.join(d, f), out)] = pg
+        for d, dirs, fs in os.walk(ROOT):
+            dirs[:] = [x for x in dirs if x not in SKIP_DIRS]
+            for f in fs:
+                if f.endswith(".md"):
+                    md = os.path.relpath(os.path.join(d, f), ROOT)
+                    ok("site", site.page_path(md) in pages, "%s has no page" % md)
+        for name, pg in sorted(pages.items()):
+            if name == "404.html":
+                continue
+            # Markdown syntax left in a single text node means a construct the renderer did not recognize.
+            bad = [m.group(0) for chunk in pg.text
+                   for m in re.finditer(r"\]\(|\*\*|`|\n *(?:\||:   |#{1,6} |[-*] |\d+\. )", chunk)]
+            ok("site", not bad, "%s: unrendered Markdown %r" % (name, sorted(set(bad))[:3]))
+            for link in pg.links:
+                if re.match(r"^[a-z][a-z0-9+.-]*:", link) or link.startswith("//"):
+                    continue
+                path, _, frag = link.partition("#")
+                target = os.path.normpath(os.path.join(os.path.dirname(name), path)) if path else name
+                full = os.path.join(out, target)
+                if os.path.isdir(full):
+                    target = os.path.join(target, "index.html")
+                    full = os.path.join(out, target)
+                ok("site", os.path.exists(full), "%s: link to a missing file %s" % (name, link))
+                if frag and target in pages:
+                    ok("site", frag in pages[target].ids, "%s: no #%s in %s" % (name, frag, target))
+        # The addresses the formats name are files of the site.
+        for f in SCHEMA_FILES:
+            sid = STORE.load(f)["$id"]
+            ok("site", sid.startswith(site.BASE_URL + "/") and
+               os.path.exists(os.path.join(out, sid[len(site.BASE_URL) + 1:])), "%s is not served at its $id" % f)
+        ctx_url = jl(os.path.join(p("fixtures", "records", "layer"), "annotations.jsonl"))[0]["@context"][1]
+        ok("site", ctx_url.startswith(site.BASE_URL + "/") and
+           os.path.exists(os.path.join(out, ctx_url[len(site.BASE_URL) + 1:])), "the context is not served at %s" % ctx_url)
+        vocab = js(p("context", "scienting-1.jsonld"))["@context"]["sct"]
+        ok("site", vocab == site.BASE_URL + "/vocab#", "the vocabulary IRI %s is not the site's vocab page" % vocab)
+        terms = [v if isinstance(v, str) else v["@id"] for k, v in js(p("context", "scienting-1.jsonld"))["@context"].items()
+                 if not k.startswith("@")]
+        for t in terms:
+            if t.startswith("sct:"):
+                ok("site", t[4:] in pages["vocab/index.html"].ids, "the vocabulary page has no #%s" % t[4:])
+
+
 def main():
     import argparse
     ap = argparse.ArgumentParser(
@@ -932,6 +1014,7 @@ def main():
     check_counts()
     check_prose()
     check_links()
+    check_site()
     for k in sorted(COUNTS):
         print("%-17s %5d checks" % (k, COUNTS[k]))
     if FAILS: