learning-website-framework1-10  Exercise 2: Check the Build Against the Content
==========================================================================

During a migration, the question "did every page survive?" must be answered by
a script. Compare the set of pages the content tree says should exist with the
set of pages that were actually built, in both directions.

Needs sitemaps.py (Chapter 7).

Save as parity.py:

import os, sys
from sitemaps import collect

def built_pages(dist):
    out = set()
    for dirpath, dirnames, filenames in os.walk(dist):
        if "index.html" in filenames:
            rel = os.path.relpath(dirpath, dist).replace(os.sep, "/")
            out.add("/" if rel == "." else "/" + rel + "/")
    return out

if __name__ == "__main__":
    content, dist = sys.argv[1], sys.argv[2]
    expected = set(collect(content))
    built = built_pages(dist)
    missing = sorted(expected - built)
    extra = sorted(built - expected)
    print(f"expected from content/: {len(expected)}   built in dist/: {len(built)}")
    print(f"expected but NOT built: {len(missing)}")
    for u in missing[:6]:
        print("   " + u)
    print(f"built but not expected: {len(extra)}")
    tops = {}
    for u in extra:
        tops.setdefault(u.strip("/").split("/")[0] or "(root)", []).append(u)
    for k, v in sorted(tops.items(), key=lambda kv: -len(kv[1]))[:8]:
        print(f"   {k:<24}{len(v):>5}   e.g. {v[0]}")

Run it against the current site's build (astro-site/dist):

  python parity.py "<debserver>/claude-projects/website-content/content" "<debserver>/astro-site/dist"

Output (checked by running it):

expected from content/: 4925   built in dist/: 5014
expected but NOT built: 0
built but not expected: 89
   sidebar                    43   e.g. /sidebar/ai/cheat_sheet_claude_code_commands_print/
   resources                  20   e.g. /resources/csset/project_files/Exercises/Ch1/01_04/starter files/
   programming                14   e.g. /programming/general-purpose-languages/python/python-lessons/python_lesson_boolean_logic_truthy_falsy_print/
   japan                       5   e.g. /japan/japanese-language/japanese-lessons/japanese_lesson_checking_into_the_hotel_print/
   hungary                     2   e.g. /hungary/hungarian-lessons/hungarian_lesson_possessive_suffixes_1st_2nd_person_print/
   (root)                      1   e.g. /
   admin                       1   e.g. /admin/
   ai                          1   e.g. /ai/claude-tools/claude-lessons/

What it shows
-------------
- Expected but not built: 0. Every page the content tree defines was built.
- Built but not expected: 89. These are pages the live site publishes that the
  content walk in Chapter 7 deliberately skips: mainly the print variants of
  cheat sheets and lessons (names ending in _print), a number of
  resource folders, the home page, the admin dashboard and the course index.
- So the print variants are real, public pages. Decide what you want for them:
  keep them, give each a canonical link to the main page, or ask search engines
  not to index them (Learning Website: Framework & Architecture 7).

Use the same script during the migration: expected = the area's pages from the
content tree; built = the new site's output folder. The result you want is
0 missing, and an extra list you can explain page by page.

WHY THIS WORKS AS AN ANSWER
---------------------------
A two-way comparison catches both losses (missing pages) and surprises (pages
nobody planned). Running it on today's build first proves the method on a site
you already trust, and it turned up the print pages as a decision still to make.