Use 3 levels of scope of data files (#36)

Data can now be in a separate namespace (in this case, the file name
without extension) so that data coming from files like bibtex are better
scoped. The updated rules around data files would be:

* `{page}.{ext}`: added to the namespace of only `{page}.md`
* `index.{ext}`: added to the namespace of each page in a folder.
* `{data}.{ext}`: added to each page in a folder but in its own
  namespace `{data}`

Previously, the yaml/json files put things under one or two variables,
like "education" or "publications". Now, that level can be removed and
set by the data file name instead.

Author: leouieda

doc/manual/data.json doc/manual/index.jsondoc/manual/data.json renamed to doc/manual/index.json

@@ -96,16 +96,19 @@ def build(config_file, console=None, style=""):
         for path in sorted(data_files):
             console.print(f"   {str(path)}")
             data = load_data(path)
-            if data["id"] in site:
+            if data["id"] in site and not data["id"].endswith("/index"):
                 propagate_to_page = [site[data["id"]]]
             else:
                 propagate_to_page = [
                     page for page in site.values() if page["parent"] == data["parent"]
                 ]
             for page in propagate_to_page:
                 console.print(f"     ↳ {page['id']}")
-                # Merge the two data dictionaries with 'page' taking precedence
-                page.update({**data["content"], **page})
+                if data["id"].endswith("/index") or page["id"] == data["id"]:
+                    # Merge the two data dictionaries with 'page' taking precedence
+                    page.update({**data["content"], **page})
+                else:
+                    page[data["id"].split("/")[-1]] = data["content"]
     else:
         console.print("   None found.")