aia-website/www/reports/migration-summary.md

# Migration Summary

## Status

A maintainable Astro migration and substantial fidelity-first pass are
implemented. The live public site was used as the authority for the shared
shell, English and Spanish homepages, service-page family, team page, library
indexes, and blog-post family.

## Migrated Content and Assets

- 46 English, 12 Spanish, and 7 Arabic library posts.
- 21 English, 9 Spanish, and 2 Arabic page records.
- One author record per language.
- 167 self-hosted images, 20 font files, and two PDF downloads.
- Current June and May 2026 English library articles are included.

HubSpot-generated wrappers, analytics, scripts, CSS, query-language duplicates,
AMP variants, pagination, and author archive variants are not carried into the
Astro implementation.

## Live-Source Fidelity Pass

Live pages inspected during the pass include:

- English and Spanish homepages
- ABA Therapy, Autism Evaluations, and Learner Social Club
- About, Team, Client Consultation, and Giveback & Donate
- English and Arabic Library indexes
- Representative current library articles

Implemented fidelity work includes:

- Rebuilt the live two-tier header, navigation hierarchy, language controls,
  utility links, footer columns, and contact information.
- Rebuilt the complete English and Spanish homepages in their live section
  order using one reusable Astro component and self-hosted source assets.
- Moved the English and Spanish homepage copy into section-oriented `home:`
  frontmatter blocks so the page is easier to edit in Front Matter CMS.
- Rebuilt service pages around the live compact title banner and editorial
  presentation, with current visible headings and working calls to action.
- Recovered and restored the live decorative hero imagery for all mapped
  English and Spanish page routes plus every language's library index.
- Restored the consultation form section's injected background image.
- Added a dedicated live-derived team card grid rather than presenting the
  extracted team content as a generic article.
- Rebuilt the contact page around the live two-column content/form section and
  full-width Leaflet map while retaining local clickable obfuscated phone
  links.
- Rebuilt the giveback page as four source-faithful full-width sections with
  its original image/copy proportions, fundraiser color band, nonprofit logo,
  emphasis, and external donation calls to action.
- Rebuilt library indexes and blog-post presentation around the live sidebar,
  article list, byline, featured-image, and counter patterns.
- Converted FAQ-bearing posts to MDX and routed their accordions, the
  standalone FAQ page, and the insurance FAQ through one reusable component.
  Each accordion uses the same animated, single-open interaction and a
  component-scoped caret color override; blog and insurance instances also
  emit matching `FAQPage` JSON-LD.
- Restored 32 live-source blog blockquotes with their original emphasis,
  links, fawn backgrounds, spacing, rounded presentation, and quote-mark SVG.
- Replaced generic oversized cards, rounded controls, and marketing heroes
  with live-derived typography, palette, widths, spacing, and compact controls.
- Corrected material live-source discrepancies found during the pass,
  including current homepage ESA copy, testimonial content, and ABA copy.

### Intentional visual deviations

- The homepage and `/insurance` carrier-logo rows intentionally share one
  responsive layout. The first and last logos sit flush with the row edges,
  and the Blue Cross Blue Shield mark is emphasized at 125% of the standard
  desktop logo width, following its treatment on the live homepage rather
  than preserving the live pages' differing row spacing.
- The Community 4 Autism logo on `/donate-autism-giveback` is horizontally
  centered on mobile instead of retaining the live page's left alignment.
- The empty `<h6>` retained by the live `/schedule-consultation` page was
  removed. It had no visible content and generated an empty heading ID in the
  migrated Markdown, so omitting it improves heading semantics and fragment
  integrity without removing page content.

## URLs and Redirects

Clean public URLs are preserved, including:

- `/aba-therapy`
- `/autism-evaluations`
- `/learner-social-club`
- `/client-consultation`
- `/library`
- `/library/post-slug`
- available `/es/...` and `/ar/library/...` routes

Phase 3A replaces the earlier 97-canonical sitemap with publication-policy
output: 92 local URLs in an indexing-enabled build and none in staging.
Nginx rewrites remain limited to the brief aliases `/aba`, `/autismevaluations`,
and `/learnersocialclub`.

## Forms

Visible consultation/contact-style pages render styled static HTML forms with
backend and spam-protection TODO comments. Submission is intentionally
disabled. Appointment and enrollment calls to action route to the static
consultation page rather than retaining the production Jotform backend.

The three English service pages also reproduce the live site's delayed,
right-edge inquiry drawers, including their page-specific enrollment copy,
shared field set, modal overlay, and responsive layout. These drawer forms are
static and their submit controls remain disabled until a production backend
and spam protection are selected; this is an intentional functional and visual
deviation from the live HubSpot form.

## External Dependencies

- The contact page map intentionally matches the live Leaflet/OpenStreetMap
  implementation and loads Leaflet from `unpkg.com` plus map tiles from
  `tile.openstreetmap.org`.

## Multilingual

- Spanish has the full live-derived homepage, translated service pages,
  library index, and available translated posts.
- Arabic has RTL support, a clearly marked placeholder landing page, the live
  library family, and available Arabic posts.
- Language-switcher choices are emitted only when the corresponding Astro
  route exists, preventing dead translated-route links.
- Arabic pages use `lang="ar"` and `dir="rtl"`; English and Spanish use LTR.
- No large Arabic translations were invented.

## Likes and Views

The post counter uses `PUBLIC_AIA_API_BASE`, defaulting to
`https://api.azinstitute4autism.com`. Its contract was verified against the
current live-site script:

- `POST /stats/batch` for likes and views
- `POST /likes` to like
- `DELETE /likes/{slug}` to unlike

It preserves cookie-backed liked state and degrades gracefully if the API is
unavailable.

## SEO and Accessibility

- Canonicals, Open Graph tags, Twitter card tags, semantic titles, and global
  organization JSON-LD are emitted by shared layouts.
- Organization JSON-LD intentionally omits raw `telephone` and `email` fields;
  visible contact details use the Astro obfuscation component to avoid exposing
  plain email and phone targets in rendered HTML.
- Library-index canonicals and current visible page H1s are explicitly set.
- Extracted image alt text is retained where available.
- Semantic landmarks, skip link, labeled forms, keyboard-operable navigation,
  language/direction attributes, and responsive layouts are present.
- Sätteri heading attributes remain enabled. Markdown supports explicit
  fragment IDs such as `{#accepted-insurance-heading}`; the installed MDX
  parser requires HTML heading IDs instead. The earlier claim that MDX also
  supports the shorthand was corrected during Phase 2 (see below).
- Sätteri directive parsing was disabled during Phase 2 after confirming that
  there are no intentional consumers and that it silently removes literal
  colon-containing text. No content or dependency versions were changed.

## Validation Results

- `npm audit --offline`: passed with zero known vulnerabilities under Node
  `24.18.0`; `flake.nix` provides a compatible current Node release.
- `npm run audit:links`: passed with zero broken internal source links. The
  audit now checks generated routes, root-relative references, relative
  Markdown links, and public assets when rendered output is unavailable.
- `npm run audit:images`: passed with no missing mapped page banners, section
  backgrounds, or blog featured images.
- `npm run audit:blog`: passed with no article-header content duplicated in
  blog content and valid MDX FAQ components for 125 questions across 19 posts.
- `npm run generate:sitemap`: passed; generated 97 URLs.
- `npm run generate:redirects`: passed.
- All migration `.mjs` tools and the sandbox DNS helper pass `node --check`.
- The Astro compiler parsed all 49 `.astro` files successfully.
- `npm run build`: passed with zero Astro diagnostics and generated 97 static
  pages.
- `npm run build:sandbox`: passed; generated 97 static pages.
- Live contact-page structure was inspected in browser at a 1280px desktop
  viewport; local generated HTML was checked for the contact layout, Leaflet
  assets, map container, and obfuscated clickable phone script. Local browser
  rendering was blocked because the Playwright browser cannot reach shell
  loopback servers and blocks `file:` URLs in this environment.
- Nix shell verification and `npm audit` retrieval were blocked by sandbox
  proxy/cache network resets.

## Manual Review

- Perform full-page desktop and mobile visual comparisons for page-specific
  layouts beyond the completed homepage, team, service, library, and article
  families.
- Editorially review extracted long-form content for remaining extraction artifacts,
  stale phone/email references, and heading hierarchy.
- Review all source alt text, keyboard behavior, screen-reader output, and
  contrast with accessibility tooling.
- Have fluent reviewers approve Spanish content and Arabic RTL presentation.
- Select and connect a form backend with spam protection before launch.
- Validate like/view CORS, credentials, and production behavior after deployment.

## Phase 2: Markdown Rendering Repair — 2026-10-07 UTC

These are newly verified local rendering findings. The accepted Phase 1 audit
and route evidence remain unchanged. No production requests were made, no
source content was rewritten, and no dependencies were upgraded.

### Inspection and Cause

Before changing configuration, searched the Markdown/MDX corpus, documentation,
parser configuration, scripts, plugins, layouts, and rendering components for
container, leaf, and inline directives and their consumers. There are no
intentional directives or registered directive transformations. Documentation
examples in code fences, ordinary punctuation, URLs, times, ratios, frontmatter
fields, CSS selectors, and explicit heading IDs are not directive consumers.

For all 100 content files, stripped frontmatter and inspected the installed
Sätteri `markdownToMdast` / `mdxToMdast` output with the original
`headingAttributes: true, directive: true` features. This found 24 unintended
`textDirective` nodes in seven files, with zero container or leaf directives.
Sätteri treats colon suffixes such as `:1`, `:00`, and `:Plan` as inline
directives; without a transformation, the renderer drops those nodes. Astro's
MDX integration inherits the same processor features, so both formats were
affected. The original full build and the failing regression build confirmed
the resulting text loss.

Only `directive` was changed to `false`; `headingAttributes: true` is retained.

| Source under `src/content/` | Restored literal text | Unintended nodes |
| --- | --- | ---: |
| `pages/en/aba-therapy.mdx` | `1:1`, `1:2` in headings and prose | 4 |
| `pages/es/aba-therapy.md` | `1:1`, `1:2` in headings and prose | 4 |
| `blog/en/aba-school-readiness-arizona.mdx` | `1:1`, `1:10` in prose and tables | 4 |
| `blog/es/aba-school-readiness-arizona.mdx` | `1:1`, `1:10` in prose and tables | 3 |
| `pages/en/learner-social-club.mdx` | `4:00 PM – 6:00 PM`, twice | 4 |
| `pages/es/learner-social-club.mdx` | `4:00 p. m. – 6:00 p. m.`, twice | 4 |
| `pages/en/faqs.mdx` | `Opportunities:Plan breaks` | 1 |

### Regression and Generated-Output Verification

`npm run test:markdown` creates a temporary Astro project, loads the real
`astro.config.mjs`, and uses the existing dependencies and copied source
components/content. It builds representative `.md` and `.mdx` fixtures plus
all seven affected collection entries through `astro:content`'s `getEntry` and
`render`. The fixtures never enter the production route tree, and all temporary
files are removed after the run.

The 20 tests cover `1:1`, `1:2`, `1:10`, ordinary spaced and adjacent colon
punctuation, English/Spanish times, plain URLs and link destinations (including
port/query/fragment), inline/fenced code examples, MDX embedded markup, explicit
heading IDs and fragment links, and the affected real content. Before the
configuration change, 14 tests failed and six passed. Afterward, all 20 passed.

The full build still generates 97 pages. Comparing the before/after `<main>`
HTML on all 97 pages found changes only in the seven routes corresponding to
the table above. This comparison normalizes the existing random
`ContactObfuscation` IDs in span attributes and script variables; that component
was not changed. Generated `/aba-therapy` and `/es/aba-therapy` each now contain
both ratios in their headings and prose (two occurrences of each ratio).

All existing explicit IDs remain intact. Four automatically generated service
heading IDs now reflect the restored text:

| Route | Previous generated ID | Corrected generated ID |
| --- | --- | --- |
| `/aba-therapy` | `aba-therapy-1-program` | `aba-therapy-11-program` |
| `/aba-therapy` | `aba-therapy-1-academic-readiness-program` | `aba-therapy-12-academic-readiness-program` |
| `/es/aba-therapy` | `programa-de-terapia-aba-1` | `programa-de-terapia-aba-11` |
| `/es/aba-therapy` | `terapia-aba-1-programa-de-preparación-académica` | `terapia-aba-12-programa-de-preparación-académica` |

No source references to the old IDs were found. Compatibility with production
fragment links remains part of the later URL reconciliation; this repair does
not introduce redirects or edit service content.

The installed MDX parser rejects Markdown's `{#id}` shorthand with directive
parsing either enabled or disabled. This is a pre-existing limitation, not a
regression from the repair. No current content uses that shorthand. Tests
verify Markdown shorthand and MDX's supported `<h2 id="...">` syntax. README
guidance was corrected; a broader parser change is outside this phase.

### Validation

Run from `www/`:

```sh
npm run test:markdown
npm run build
npm run audit:links
npm run audit:blog
git diff --check
```

- Markdown regression tests: 20 passed, zero failures.
- Build: passed; 74 checked files, zero errors/warnings/hints, 97 static pages.
- Link audit: zero broken internal links.
- Blog audit: zero content audit failures.
- Diff whitespace check: passed.
- No changes to application content, dependencies, lockfile, Phase 1 evidence,
  or `merge-plan.md`. No Phase 3 work was started.

## Phase 3A: Route Eligibility, Canonicals, and Indexing — 2026-10-07 UTC

This section records new implementation and verification. Earlier Phase 1
evidence and the Phase 2 checkpoint are preserved. No deployment, form
submission, backend activation, dependency change, content reconciliation,
hreflang, translation-alternate graph, article schema, or FAQ schema work was
performed.

### Publication Architecture

`src/utils/publication-policy.ts` is the shared, framework-independent
TypeScript policy. It derives local routes from collection/language/slug,
normalizes public paths, rejects route collisions (including drafts and utility
paths), validates canonicals, and decides route and sitemap eligibility.
`src/utils/publication.ts` adapts the Astro collections and reads the existing
`PUBLIC_ALLOW_INDEXING` switch. The canonical site origin comes from `site.json`,
also used by Astro configuration.

All content routes now pass through `src/pages/[...slug].astro`. The former
fixed home and Library routes could bypass `draft`; they now use the same
manifest as other pages. The existing Arabic home and three Library index
templates were moved to presentation components, retaining their visible
content. The English/Spanish home component receives its selected entry.
Shared layouts obtain canonical/indexing policy by the actual local route;
individual templates no longer invent or pass canonical strings.

The sitemap is an Astro endpoint generated alongside HTML. Its locations come
from eligible local route URLs, never canonical strings. The old public
sitemap copy was removed. `generate:sitemap` delegates to the Astro build
instead of maintaining a separate frontmatter-regex generator. Route discovery
and the source-link audit fallback reuse the shared publication policy; route
evidence now records eligibility, noindex, policy canonical, external-canonical
status, and production sitemap eligibility separately from rendered metadata.

The sole new frontmatter/schema field is `noindex`, a boolean defaulting to
false. Existing `draft` and `canonical` fields cover the remaining rules.

### Verified Exceptions and Read-Only Production Evidence

Six successful GET requests were made between **02:42:02 and 02:42:08 UTC**:
`robots.txt` first, then the five routes below. Requests were sequential, at
least one second apart, with redirects not followed. Robots allowed these
paths. No publisher pages, forms, or production APIs were requested. The web
tool could not retrieve robots.txt; the direct GET succeeded. All five route
observations agree with the accepted saved production evidence; these are
new checks, not silent revisions to historical findings.

| Route | Fresh HTTP / robots evidence | Implemented exception | Production sitemap |
| --- | --- | --- | --- |
| `/ar` | 404 at 02:42:08 | Existing placeholder marked `draft: true`; no HTML route | Excluded |
| `/schedule-consultation` | 200 / `noindex` at 02:42:04 | `noindex: true`; accessible, self-canonical | Excluded |
| `/employee-portal` | 200 / `noindex` at 02:42:05 | `noindex: true`; accessible, self-canonical | Excluded |
| `/library/community-highlight-meet-rula-diab` | 200 / `noindex` at 02:42:06 | `noindex: true`; external canonical retained | Excluded |
| `/library/new-aia-scottsdale-office` | 200 / `noindex` at 02:42:07 | `noindex: true`; external canonical retained | Excluded |

The exact verified external canonical targets are:

- [VoyagePhoenix original](https://voyagephoenix.com/interview/community-highlights-meet-rula-diab-of-arizona-institute-for-autism/)
- [Scottsdale.org original](https://www.scottsdale.org/airpark/features/integrity-empowerment-and-excellence-arizona-institute-for-autism-expands-to-new-office/article_79c31a18-2f35-11ee-9459-6f29f5a08420.html)

Only five content files changed, all in frontmatter: the Arabic home draft
flag, four noindex flags, and the two external canonical values. Portal access
and replacement behavior remain unchanged and deferred to human decision.

### Before / After and Output Verification

| Behavior | Before Phase 3A | After Phase 3A |
| --- | --- | --- |
| Route publication | Collection routes filtered drafts; fixed indexes bypassed them | Every content route uses the same eligibility manifest |
| Generated HTML | 97 pages, including `/ar` placeholder | 96 pages; only `/ar` removed |
| Canonical handling | Optional unchecked values passed by templates; syndicated posts self-canonicalized | Exactly one normalized self canonical, or the verified external exception; missing ordinary values derive from the route |
| Indexing-enabled robots | All pages defaulted to indexable | 92 `index,follow`; four verified `noindex,follow` exceptions |
| Staging robots | Global `noindex,nofollow` | Global protection retained on all 96 pages; no page override |
| Sitemap | Static list of 97 frontmatter canonicals | 92 eligible local URLs when indexing is enabled; empty in staging |
| Robots sitemap advertisement | Environment dependent | Still advertised only in the indexing-enabled build |

Published noindex pages remain accessible. An external canonical independently
excludes a page from the sitemap; it does not automatically invent a noindex
directive. The two real syndicated articles explicitly retain their verified
noindex status. Invalid/relative/unsafe canonicals, local targets that disagree
with the route or public origin, duplicate published canonical targets, and
route collisions fail validation. Duplicate frontmatter keys also fail parsing.

The language menu now uses eligible routes without a hardcoded `/ar` entry.
The Arabic header logo falls back to the existing English homepage while its
localized homepage is unpublished. No hreflang or new translation relationships
were introduced. Main content text, titles, and H1s are unchanged on all 96
retained routes. There are no stylesheet or asset changes.

The indexing-enabled local artifact was checked at **02:54:56 UTC**: all 96
heads matched their manifest canonicals/robots; 92 local sitemap URLs; no `/ar`.
The final ordinary build restores staging output with 96 noindex pages and
zero sitemap locations. No environment files or deployment configuration changed.

### Regression Coverage and Reproduction

`npm run test:publication` passes **19 tests** covering normal publication,
draft exclusion, accessible noindex pages, external canonicals independently
of noindex, normalized/missing/malformed/duplicate canonicals, cross-collection
and cross-locale collisions, reserved routes, staging protection, and all five
verified exceptions. Generated-head checks reject absent, duplicate, and invalid
canonical tags. It builds the actual application in temporary directories in
both environment modes, then additionally marks a home and Library index as
drafts and confirms their routes disappear. The tests never change real source,
environment configuration, or `dist`.

Validation from `www/`, in the requested order (with evidence refresh steps):

```sh
npm run test:publication
npm run build
PUBLIC_ALLOW_INDEXING=true npm run build
npm run audit:routes -- --offline
npm run audit:routes -- --check
npm run audit:links
npm run build
npm run audit:routes -- --offline
npm run audit:routes -- --check
git diff --check
```

All required commands passed. Each build checked 71 files with zero errors,
warnings, or hints, and generated 96 HTML pages. The link audit found zero
broken internal links. Existing regression suites also passed: 20 Markdown
tests and nine route-audit tests. Dependencies and the lockfile are unchanged.
The source-only link fallback also passed in an isolated directory with zero
broken links. The `generate:sitemap` compatibility command was verified using
Astro's public build API, followed by another normal staging build.

The final durable local reconciliation is
[`route-reconciliation-2026-10-07-offline-04-11-25-151Z.json`](route-reconciliation-2026-10-07-offline-04-11-25-151Z.json).
It reuses the earlier 70 production request records with their original
observation dates; neither offline reconciliation made new production requests.
Only the final staging reconciliation is retained in the repository; the
intermediate indexing reconciliation was moved to temporary storage after its
checks. The six targeted requests described above are separate fresh evidence.

The normalized discovery inventory remains 129 routes, with 102 URLs in the
saved production sitemap. Local generated routes decrease from 97 to 96;
sitemap/local overlap stays 92, and the ten sitemap-listed live-only routes
remain outstanding. `/ar` remains visible in the inventory as draft,
ineligible, not generated, and verified absent on production. Earlier audit
files retain their historical facts.

No new human publication-policy decision is needed for Phase 3A. Human gates
for portal access, forms/backends, analytics/consent, landing-page ownership,
and conflicting service facts remain deferred. Phase 3B was not started.