| 1 | # Spec: Social link unfurls and SEO for public repos |
| 2 | |
| 3 | Status: implemented (branch `feature/social-unfurls-seo`, merged to `development`). |
| 4 | Reverse-engineered from the shipped code so the behavior is documented and |
| 5 | testable. Scope is meta tags, Open Graph card images, and crawlability for |
| 6 | public repositories. |
| 7 | |
| 8 | ## Problem |
| 9 | |
| 10 | A pasted siGit repo link produced no preview card in Slack, X, LinkedIn, and |
| 11 | Discord, and public repos were not cleanly indexable. Those are the channels |
| 12 | where developer tools spread, so a shared link should render a preview card and |
| 13 | leave a backlink. |
| 14 | |
| 15 | ## Goals |
| 16 | |
| 17 | 1. Every public repo page serves complete SEO and social meta tags in the |
| 18 | initial server HTML, because crawlers do not run JavaScript. |
| 19 | 2. Each public repo has a generated 1200x630 preview image. |
| 20 | 3. Public repos are crawlable and listed in the sitemap; private and |
| 21 | auth-gated pages are not, and leak nothing. |
| 22 | |
| 23 | ## Non-goals |
| 24 | |
| 25 | Profile pages, stars/forks/explore/trending UI, full-text search, and |
| 26 | analytics. This work is meta tags, card images, and crawlability only. |
| 27 | |
| 28 | ## Rendering model |
| 29 | |
| 30 | The app is server-rendered Rails (ERB + Hotwire), so meta tags and a crawlable |
| 31 | HTML body already appear in the initial response. No client-render workaround is |
| 32 | needed. |
| 33 | |
| 34 | ## Routes |
| 35 | |
| 36 | | Method | Path | Controller | Purpose | |
| 37 | |--------|------|------------|---------| |
| 38 | | GET | `/og.png` | `og_images#default` | Generic sitewide card (default `og:image`). | |
| 39 | | GET | `/og/:username/:repository.png` | `og_images#show` | Per-repo card. Dotted names allowed (`Qwen2.5-GGUF`), `format: false`. | |
| 40 | | GET | `/sitemap.xml` | `sitemaps#index` | Public repos plus their owners (pre-existing). | |
| 41 | |
| 42 | Both `/og` routes are declared before the `/:username` matcher so they are not |
| 43 | read as profiles. |
| 44 | |
| 45 | ## Behavior |
| 46 | |
| 47 | ### Meta tags (all pages) |
| 48 | |
| 49 | Driven by `SeoHelper` and the `shared/_seo` partial. Defaults degrade so any |
| 50 | page emits valid, non-empty tags. Pages override with `content_for`: |
| 51 | `:title`, `:description`, `:og_image`, `:og_image_alt`, `:og_type`, `:robots`, |
| 52 | `:structured_data`. |
| 53 | |
| 54 | Emitted: `<title>`, `meta description`, `link canonical` (path only, no query or |
| 55 | fragment), Open Graph (`og:type`, `og:site_name` = `siGit`, `og:title`, |
| 56 | `og:description`, `og:url`, `og:image`, `og:image:type`, `og:image:width` = |
| 57 | 1200, `og:image:height` = 630, `og:image:alt`, `og:locale`), Twitter |
| 58 | (`summary_large_image`, title, description, image, image:alt), and JSON-LD |
| 59 | (`Organization` + `WebSite` sitewide). |
| 60 | |
| 61 | ### Public repo pages |
| 62 | |
| 63 | `repositories#show`, `#model`, `#tree`, `#cicd`, and `blobs#show` render |
| 64 | `shared/_repository_social`, which sets the description, the per-repo |
| 65 | `og:image` (`/og/:owner/:repo.png`), and the image alt text. The two canonical |
| 66 | repo pages (`show`, `model`) also emit `SoftwareSourceCode` JSON-LD with name, |
| 67 | description, repo URL, author, dates, image, and `programmingLanguage` when a |
| 68 | primary language is detected. |
| 69 | |
| 70 | `Repository#seo_description` returns the description when present, otherwise a |
| 71 | fallback that names the owner and the kind (code repo or open-weights model). |
| 72 | This guarantees non-empty social text. |
| 73 | |
| 74 | ### OG card images |
| 75 | |
| 76 | `OgImageService` builds an SVG from the brand template and rasterizes it to PNG |
| 77 | with libvips. Card content: repo name, `@owner`, wrapped description (up to |
| 78 | three lines), primary-language dot with color, star count when above zero, and |
| 79 | the siGit wordmark. The generic card carries the tagline only. |
| 80 | |
| 81 | All user-controlled text is XML-escaped before going into the SVG. |
| 82 | |
| 83 | Primary language is guessed from file extensions on the default branch |
| 84 | (`Repository#primary_language`), using a small Linguist-style color map. Model |
| 85 | repos that carry weight files surface the weight format instead of a code |
| 86 | language. |
| 87 | |
| 88 | ### Caching and performance |
| 89 | |
| 90 | `OgImagesController` answers conditional GETs with `stale?(etag:, public:)`. The |
| 91 | ETag is `OgImageService::TEMPLATE_VERSION` plus `Repository#og_version`, a hash |
| 92 | of repo id, name, description, star count, default-branch tip commit, and |
| 93 | `updated_at`. A repeat fetch with a matching `If-None-Match` returns 304. Bytes |
| 94 | are cached in `Rails.cache` keyed by the same version, so the SVG is rasterized |
| 95 | only on a cache miss, never on the hot path when a card exists. Responses set |
| 96 | `Cache-Control: public` with a long max-age. |
| 97 | |
| 98 | The template version bumps invalidate every cached card. The content hash |
| 99 | invalidates a single repo's card when its content changes. |
| 100 | |
| 101 | ### Fallback |
| 102 | |
| 103 | If rasterization is unavailable (no libvips, no SVG or font support), the |
| 104 | endpoint logs a warning and serves the static brand icon |
| 105 | (`public/favicon/web-app-manifest-512x512.png`) with a short TTL, so a transient |
| 106 | failure self-heals and the endpoint never returns 500. |
| 107 | |
| 108 | ### Privacy guards |
| 109 | |
| 110 | - A private repo returns 404 on `/og/:owner/:repo.png` for everyone, including |
| 111 | the owner, so no card can leak. |
| 112 | - Visibility is decided server-side from the repo's `is_private` flag, not a |
| 113 | client signal. A logged-out request to a private repo page renders the static |
| 114 | 404 with no metadata. |
| 115 | - Private repos and users who own only private repos are excluded from |
| 116 | `sitemap.xml`. |
| 117 | - Auth and transactional pages (`sessions`, `registrations`, `passwords`, |
| 118 | `confirmations`, `billing`, `users#settings`, `repositories#new`) call |
| 119 | `noindex!` in the controller, which makes `SeoHelper#meta_robots` emit |
| 120 | `noindex, nofollow`. |
| 121 | |
| 122 | ### robots.txt |
| 123 | |
| 124 | `public/robots.txt` allows content, disallows `/auth`, `/settings`, `/billing`, |
| 125 | `/new`, `/api/`, and `/*/raw/`, and points at `/sitemap.xml`. |
| 126 | |
| 127 | ## Acceptance criteria |
| 128 | |
| 129 | 1. A public repo URL pasted into Slack, X, LinkedIn, or Discord shows a |
| 130 | `summary_large_image` card with the correct title, description, and image. |
| 131 | 2. `view-source` on a public repo page shows the OG, Twitter, and canonical |
| 132 | tags in the initial HTML. |
| 133 | 3. `/og/:owner/:repo.png` returns a 1200x630 PNG, is a 304 on the second hit |
| 134 | with a matching ETag, and has a fallback for missing metadata. |
| 135 | 4. A private repo URL returns no preview, carries `noindex`, and is absent from |
| 136 | `sitemap.xml`. |
| 137 | 5. `robots.txt` and `sitemap.xml` validate; public repos appear in the sitemap, |
| 138 | auth and non-content routes do not. |
| 139 | |
| 140 | ## Production dependency |
| 141 | |
| 142 | The runtime image installs `libvips`, `librsvg2-2`, and `fonts-dejavu-core` so |
| 143 | libvips can rasterize SVG text. macOS dev hosts usually lack libvips, so the |
| 144 | endpoint returns the static fallback locally; the card renders in CI and |
| 145 | production. |
| 146 | |
| 147 | ## Tests |
| 148 | |
| 149 | RSpec, the project framework. |
| 150 | |
| 151 | - `spec/requests/repository_seo_spec.rb`: meta output for public vs private, |
| 152 | JSON-LD, indexable flag, empty-description fallback, noindex on auth pages. |
| 153 | - `spec/requests/og_images_spec.rb`: PNG content type, cache headers, 304 on |
| 154 | conditional GET, dotted names, private 404, unknown 404, default card. |
| 155 | - `spec/requests/sitemap_robots_spec.rb`: sitemap inclusion and exclusion, |
| 156 | robots directives. |
| 157 | - `spec/models/repository_seo_spec.rb`: `seo_description`, `og_version`, |
| 158 | `primary_language`. |
| 159 | - `spec/services/og_image_service_spec.rb`: SVG structure, XML-escaping, default |
| 160 | card, render-or-typed-error. |
| 161 | |
| 162 | ## Key files |
| 163 | |
| 164 | - `app/services/og_image_service.rb`, `app/controllers/og_images_controller.rb` |
| 165 | - `app/helpers/seo_helper.rb`, `app/views/shared/_seo.html.erb`, |
| 166 | `app/views/shared/_repository_social.html.erb` |
| 167 | - `app/models/repository.rb` (`seo_description`, `primary_language`, |
| 168 | `og_version`), `app/services/git_repository_service.rb` (`head_sha`, |
| 169 | `tree_filenames`) |
| 170 | - `app/controllers/application_controller.rb` (`noindex!`), `config/routes.rb`, |
| 171 | `public/robots.txt`, `Dockerfile` |
| 172 | - `docs/seo-and-social-unfurls.md` (how to verify) |