About gemreptiles.com v3

We are a boutique hobbyist reptile breeding operation owned and operated by the love of my life, Becky and myself. I do all the web/business development, some of the husbandry and all of the marketing. Becky does most of the husbandry, even though I have the most experience with reptiles, we have more than 60 years of combined experience keeping and breeding exotics.

The primary marketplace for selling reptiles online is morphmarket.com. This site imports our inventory from MorphMarket and presents it on gemreptiles.com with our own branding and storefront experience.

Tech Stack

• Backend: Laravel 11, PHP 8.2+
• Frontend: Tailwind CSS 3, Alpine.js
• Database: PostgreSQL
• Cache: Redis (predis)
• Mail: SendGrid
• Assets: Vite

Application Architecture

The app has two data paths for animal listings:

1. JSON import (public storefront): An admin uploads a JSON export from MorphMarket via the dashboard. This file is saved to storage/app/public/animals.json and drives the homepage, category pages, and category filtering. These routes are cached by file mtime (30 min TTL) and never hit the database.

2. Database-backed listings: The /animals and /classifieds routes serve Eloquent-backed records with full search, filtering, and pagination. Animals are synced from the JSON import into the animals table via AnimalImportController.

Auth uses a simple is_admin flag on User. Only admins can create, edit, or delete animal records. The Seller model is a profile linked 1:1 to a User and is editable from the profile page.

MorphMarket JSON Import Format

The dashboard import accepts the JSON export from MorphMarket. Key fields used by the importer:

Field	Description
Animal_Id*	Used as the slug
Title*	Pet name
Category*	Species category (e.g. Ball Pythons, Corn Snakes)
State	For Sale, Breeder, Sold, Not For Sale
Enabled	Active or inactive
Visibility	Public or private
Price	Listing price
Dob	Date of birth (supports n/j/Y, n/Y, Y formats)
Sex	male / female
Photo_Urls	Space-separated image URLs
Mm_Url**	Link back to the MorphMarket listing
Desc	Description

Feature Flags

Flag	Default (prod)	Default (non-prod)	Description
FEATURE_CLASSIFIEDS	false	true	Enables the classifieds marketplace feature

Setup

composer install
npm install
cp .env.example .env
php artisan key:generate
php artisan migrate
php artisan storage:link
npm run build

For local development:

npm run dev        # Vite dev server with HMR
php artisan serve  # or use Laravel Herd / Valet

Notes

The source is open. Our content, logos, and UI designs are © All Rights Reserved 2024–2026.

Everything except secrets belongs in git. Blobs go in S3 or similar — not in the database, not in git.

Changelog

2026-05-08 (featured image selection for Animals and Species)

• Added is_featured boolean to media table; one featured media per mediable entity
• Admin star-picker UI on Animal and Species show pages via Livewire v3 FeaturedMediaPicker component + Alpine.js
• Star is mutually exclusive per entity (radio-style): clicking sets featured, clears others for same mediable
• PATCH /dashboard/media/{media}/feature route via DashboardMediaModerationController::setFeatured
• Animal and Species index thumbnails now prefer featured media over first media
• featuredMedia() + featuredApprovedMedia() relationships added to HasMedia trait and Species model

2026-05-08 (automated backups: spatie/laravel-backup)

• Installed spatie/laravel-backup v9 with spatie/db-dumper (PostgreSQL support)
• DB-only backup every 6 hours; full backup daily at 02:00; cleanup daily at 01:00
• All backups push to private_s3 disk → DO Spaces "privates" bucket under backups/ prefix
• Retention: all backups kept for 12 days (~48 DB backups), then 1/day until 30 days, then pruned
• Backups only run in the production environment; gzip compression on DB dumps
• Added BACKUP_ARCHIVE_PASSWORD and BACKUP_NOTIFICATION_EMAIL to .env.example
• Note: manually create backups/db and backups/laravel subdirectories in DO Spaces if desired

2026-05-08 (IncomingEmailController + Email CRM)

• POST /email/inbound (email.inbound) handles SendGrid Inbound Parse webhooks
• Parses from/to/subject/text/html + SendGrid envelope JSON; deduplicates by Message-ID header
• Threads messages by contact email + base subject (Re:/Fwd: stripped); reopens closed conversations on new mail
• Forwards to all is_admin users via queued ForwardedInboundEmailMail (was hardcoded; now dynamic)
• Admin CRM at /dashboard/conversations: list with status tabs (open/closed/spam/all), thread view, reply, status update, delete
• ConversationReplyMail queues reply to contact with Re: {subject} threading
• Fixed broken VerifyCsrfToken::$except (was array-of-array with route name; changed to URI string email/inbound)
• email_conversations + email_messages tables; EmailConversation / EmailMessage models

2026-05-08 (support ticket: fix email verification flow)

• New support ticket users now also receive a password reset link so they can log in before clicking the verification link
• Without a known password, users had no path to authenticate and complete email verification

2026-05-08 (rebrand Seller/Vendor → Social Media / Profiles)

• Routes /sellers → /profiles, names sellers.index/show → profiles.index/show
• Nav label "Sellers" → "Profiles"
• Profile section heading "Vendor Profile" → "Social Media & Profile"; description and save button updated
• Onboarding checklist item "Set up your vendor profile…" → "Add your social media links to your profile"
• Sellers index/show: visible text and empty-state copy updated to "Profiles" / "breeder"

2026-05-07 (Open Graph / link previews)

• Both layouts: default og:site_name, og:type, og:url, og:title, og:description, og:image + Twitter card tags using og-default.jpg from CDN
• animals/show: overrides og:type=product, og:title, og:description, og:image with animal's first media photo
• species/show: overrides og:title, og:description, og:image with first approved species media photo
• Any page can add @section('og_image', ...) / @section('og_title', ...) / @section('og_description', ...) to customize

2026-05-07 (support ticket form)

• New support_tickets table (name, email, type, message, user_id FK nullable)
• SupportTicketController: validates input, creates a new User with a secure random password when the email is unknown, fires Registered event to send verification email, links existing users without re-registering, creates ticket, queues SupportTicketAdminMail to all admin users
• SupportTicketAdminMail + emails/support-ticket-admin.blade.php admin notification email
• GET /support / POST /support (support.create / support.store)
• Support link added to site footer under Legal column

2026-05-07 (performance: SSR species init + CDN preconnect)

• Species search index: SpeciesController::index() pre-fetches page-1 results using same Redis cache key as search(); results injected as window.__speciesInitial__; Alpine init() consumes SSR data directly when in default state (no query/taxon/hasMedia/page=1), skipping the initial XHR entirely
• Both layouts: <link rel="preconnect" href="https://gemx.sfo3.digitaloceanspaces.com" crossorigin> — browser opens TLS connection before first image request
• welcome.blade.php: decoding="async" on all non-LCP lazy-loaded animal card images
• Species thumbnails sized to 100×100px (up from 40×40px); fetchpriority="high" on first result row image; loading="lazy" bound per-row via Alpine $index
• Axios removed from JS bundle (bootstrap.js); all XHR uses native fetch() — bundle 84KB → 45KB
• CSS bundle: removed invalid @tailwind forms; directive and redundant compiled-views glob from tailwind.config.js — bundle 81KB → 71KB
• Async CSS loading in production via rel="preload" as="style" onload swap with <noscript> fallback; inline critical background-color prevents FOUC in guest layout
• welcome.blade.php: LCP image computed server-side, <link rel="preload" as="image"> injected in <head>, first card image fetchpriority="high" (no loading="lazy"), rest lazy+decoding=async; width="800" height="800" on all card images (CLS)
• Removed x-transition from Alpine spinner and clear button (prevented forced reflow on layout-property reads)
• 301 redirects for all favicon/manifest paths via routes/web.php loop; scripts/nginx-favicons.conf added for Forge nginx config to pass favicon paths through to PHP

2026-05-07 (animal media pipeline)

• media:process-animals command: syncs animals/ prefix from DO Spaces, generates 400×400 square JPEG thumbnails via Intervention/Image, recompresses originals at Q85, syncs optimized originals and thumbs/animals/ back to DO Spaces, updates media.thumbnail_url per record
• animals:sync now calls media:process-animals after animals:mirror-media; JSON rewrite adds Thumbnail_Url field (first media record's thumbnail) alongside existing Photo_Urls
• welcome.blade.php: non-LCP animal cards use Thumbnail_Url ?? Photo_Urls[0]; LCP card always uses full-size original
• animals/index.blade.php: card img src uses thumbnail_url ?? url; LCP preload hint uses thumbnail

2026-05-07 (species media pipeline)

• media:process-species command: syncs species/ prefix from DO Spaces to storage/app/spaces/, generates 100×100 square JPEG thumbnails via Intervention/Image (Imagick driver), recompresses JPEG originals at Q85, syncs optimized originals and new thumbs/species/ prefix back to DO Spaces, and updates media.thumbnail_url idempotently
• media.thumbnail_url column added; SpeciesController::format() serves thumbnail_url ?? url so the species search index immediately uses 100px thumbnails once generated
• config/image.php created; Intervention/Image configured to use Imagick driver
• Command options: --dry-run, --force, --no-sync, --skip-optimize, --batch=N

2026-05-07 (welcome page + favicon)

• welcome.blade.php: title attributes on all sort/action/external links; richer alt text on animal images (name + category + sex + traits); loading="lazy" on animal images; rel="noopener noreferrer" on MorphMarket external links; invalid <h2 href> element corrected
• Favicon: both layouts now identical — removed redundant non-standard 100x100/192x192/256x256 rel="icon" links from guest layout (webmanifest covers Android/PWA sizes); added msapplication-TileColor to both layouts

2026-05-07 (SEO + performance)

• Meta descriptions added to all public-facing routes: homepage, category pages, animals index/show, species index/show, subspecies show, sellers index/show, classifieds index/show
• @stack('meta') added to both layouts (app.blade.php, guest.blade.php) for per-page meta injection
• Page titles updated to use Gem Reptiles consistently in both layouts; typo in guest layout title fixed
• Bunny.net font loading changed from render-blocking <link rel="stylesheet"> to non-blocking rel="preload" + onload swap with <noscript> fallback; eliminates font-induced paint delay
• Species search thumbnails: added width="40" height="40" (prevents CLS) and loading="lazy" (defers off-screen image fetches)

2026-05-07

• Cloudflare Turnstile rewritten: widget now uses data-execution="execute" + data-appearance="interaction-only" — challenge runs on form submit (via submitWithTurnstile()), not on page load; eliminates Safari password-save dialog race condition that was invalidating tokens
• Turnstile component owns the cf-turnstile-response hidden input (data-response-field="false"); widget resets can no longer clear the token
• UserSeeder: 32 verified + 1 unverified non-admin users via Faker safeEmail
• CI assets job: declare environment: Production to unlock environment secrets/vars; use vars.ASSET_URL (not secrets)
• Forge deploy: fetch manifest.json from DO Spaces CDN via curl (public-read, no-cache headers); optimize:clear before optimize to nuke stale view cache
• Frontend assets build and CDN sync moved to CI assets job (runs after tests pass on main); Forge no longer builds JS
• manifest.json synced with no-cache headers separately from hashed assets (immutable); deploy hook fires only after sync completes
• Vite base only applied in production mode; stale public/hot file deleted

2026-05-06

• "Back to Dashboard" link moved into Pulse header via anonymous component override (resources/views/vendor/pulse/components/pulse.blade.php); registered via prependNamespace in AppServiceProvider to override Pulse's anonymousComponentPath
• Vite config fixed: base (CDN asset URL) now only applied in production mode; dev mode no longer pollutes public/hot with CDN path, preventing stale hot file asset routing errors
• Nav renamed "Photos" → "Media" (admin-only nav link)
• Animal slug added to all inquiry email subjects (AnimalInquiryMail, InquiryConfirmationMail, InquiryAdminNotificationMail)

2026-05-06 (species-admin)

• Admin can edit all fields on species (species, common_name, author, higher_taxa, species_number, changes, description) via dashboard/species/{id}/edit; same for subspecies (genus, species, subspecies, author, description) via dashboard/subspecies/{id}/edit
• Admin can detach photos from species/subspecies (hard-deletes DB row, file stays on S3) via hover-reveal Detach button on show pages and edit pages
• Authenticated (non-admin) users can submit description proposals on species and subspecies detail pages; submissions go to species_content_submissions table with pending status
• Admin moderation queue at dashboard/submissions: approve (writes proposed_value to description) or reject; reviewer and timestamp recorded
• media.deleted_at column added (scaffolding for future soft-delete media library feature)

2026-05-06

• species:import-checklist command imports data from the Reptile Database XLSX checklist: 1,288 new species, 219 new subspecies, 9 taxonomy change notes, 53 type_species flag corrections; joins on sp_id/species_number; --dry-run and --task options (species|changes|type_species|subspecies|all)
• Species index back button now restores previous page number via species_page sessionStorage key
• Deploy hook: CI triggers POST to DEPLOY_HOOK GitHub Secret after tests pass on main

2026-05-06 (feature/content-updates)

• Species biography generation: species:generate-bios queues GenerateSpeciesBiographyJob per record; sources Wikipedia, iNaturalist, GBIF; outputs structured Markdown; skips existing bios unless --force; --limit, --model, --id, --dry-run options
• species:work-bios queue worker command for the species-bios queue
• species:normalize-bios command converts Wikipedia == heading == markup to Markdown headings in stored descriptions; --dry-run and --model options
• Wikipedia markup normalization applied directly to 7,084 existing species/subspecies rows via SQL regexp_replace
• Biography job now normalizes Wikipedia markup at generation time (headings, bold/italic, links, templates)
• Bios rendered as Markdown HTML (Str::markdown) in species and subspecies detail views, positioned below the photo gallery
• species:sync-taxonomy cross-checks all species against GBIF backbone; flags synonyms in the changes field (GBIF-SYNONYM:date:accepted_name); fills empty common_name from GBIF vernacular names; subspecies-safe (skips missing columns); options: --dry-run, --model, --limit, --min-confidence, --force, --genus, --family
• Taxonomy sync run: 68 species flagged as synonyms (notable: Trimeresurus→Craspedocephalus, Lygosoma→Riopa/Subdoluseps, Lobulia→Alpinoscincus/Nubeoscincus); 480 common names filled
• Species index pagination: links now appear both above and below the results table; results per page reduced from 100 to 80
• species:export-bios fixed: chunkById now includes id in select; writes to local disk via file_put_contents (default disk is S3)
• database/sql/ added to .gitignore (scratch SQL scripts)

2026-05-06

• Installed keepsuit/laravel-opentelemetry package for OpenTelemetry integration, enabling tracing and metrics collection
• Installed laravel/pulse package for real-time application performance monitoring and analytics dashboard, compatible with Alpine.js and Blade templates
• Configured Pulse with database migrations and published configuration/assets
• Published OpenTelemetry configuration files for further customization
• Added admin-only "Monitoring" navigation link in dashboard Quick Actions to access Laravel Pulse dashboard
• Added "Back to Dashboard" link in Pulse monitoring page header for easy navigation back to admin dashboard

2026-05-04 (continued)

• species:fetch-images source chain expanded to 7 sources: added Reptile Database (HTML scrape, genus/species/subspecies URL params), ARMI USGS gallery (public domain government images), BioLib.cz (3-step HTML scrape, 2 s rate-limit between requests)
• logs:upload now truncates each log file after successful S3 upload so subsequent runs start clean
• Species search: default alphabetical browse (100/page, paginated); text search returns flat results (no limit, cached 5 min); browse/taxa results cached 1 hour; Redis cache key prefix species_search:; sessionStorage LRU result cache (species_rc_ prefix, 20-entry max)
• Species search filter: "Has photos" checkbox + 6 mutually exclusive taxon pill buttons (Lizards, Snakes, Geckos, Turtles & Tortoises, Amphisbaenia, Crocodilians); taxon state persisted in sessionStorage
• Species detail view: reactive attribution bar below gallery updates on thumbnail click (title, author, license, source link, "Full attribution" link)
• Added media attribution page at /media/{id}/attribution; linked from species/subspecies detail views
• Clear search button always visible; resets query, taxon filter, and sessionStorage state

2026-05-04 (continued)

• Added logs:upload Artisan command — streams storage/logs/*.log to private_s3 disk; scheduled monthly on the 15th at midnight
• Added private_s3 filesystem disk (PRIVATE_S3_KEY, PRIVATE_S3_SECRET, PRIVATE_S3_BUCKET, PRIVATE_S3_REGION, PRIVATE_S3_ENDPOINT) — separate credentials from public media S3; visibility private
• Added FetchTaxonImageJob queued job — dispatches one job per species/subspecies record on the species-images queue; species:fetch-images --model=all --queue enqueues all unprocessed taxa
• Scheduled species:fetch-images --model=all --queue weekly, Sundays at 03:00 America/Boise
• species:fetch-images now accepts --queue flag to dispatch FetchTaxonImageJob per record instead of processing inline; --max option controls images per taxon (default 1); buildQuery orders 0-image records first so re-runs always make forward progress; default --limit bumped to 1,000
• Image license filter relaxed: null/unspecified licenses accepted; only "all rights reserved" explicitly rejected
• Photo column moved to leftmost position in species search results table

2026-05-04 (continued)

• Production database migrated from MySQL 8 to PostgreSQL
• Added species:fetch-images Artisan command — fetches free CC-licensed images for species and subspecies from a four-source chain: Wikipedia REST API → Wikimedia Commons direct file search → iNaturalist taxa API → GBIF species media API; resumable batched runs (--limit, --model, --id), --dry-run and --force options, 500ms rate-limiting delay between requests
• Added media:export-species — exports approved species/subspecies media (with attribution) to a portable JSON file for production import
• Added media:import-species — idempotent JSON import for production; matches records by scientific name (not ID) for cross-environment safety
• Added source_url and license_url columns to media table; existing license, author, copyright columns reused for full attribution storage
• Scientific names now displayed (italic, linked to species record) on animal index cards, animal detail pages, and the homepage welcome cards
• Species search results now show a thumbnail of the most recent approved photo instead of the type-specimen badge column
• Species search query now persists in sessionStorage; random seed only fires on first visit or after the field is cleared
• Added latestApprovedMedia() morphOne relationship to Species model; used by the search API to return thumbnails in a single eager-loaded query

2026-05-04 (continued)

• Applied patches to production server for kmod vulnerability
• Applied patches to production server for kmod vulnerability

2026-05-04 (continued)

• Added species:import Artisan command — imports Reptile Database CSV into species table, supports --dry-run and --csv= options, deduplicates on species_number
• Imported 11,440 species records from reptile_checklist_2020_12.csv
• Added SpeciesType enum (Syntype, Holotype, Lost, Paratype, Lectotype, Neotype)
• Added SpeciesTypeCast — parses space-delimited type tokens to SpeciesType[]; empty → "null"
• Updated species table: type_species bool → varchar(10), unique index on species_number
• Added MeiliSearch full-text species search (laravel/scout, meilisearch/meilisearch-php)
• Species search view with real-time Alpine.js UI, debounced input, dual-layer cache (client session + server 5-min TTL)
• Species detail view (/species/{id}) — taxonomy info card, approved photo gallery, user photo upload form
• Photo moderation pipeline: user uploads set moderation_status = pending; admin dashboard reviews/approves/rejects
• Added moderation_status column to media table (default approved — existing media unaffected)
• Admin nav "Photos" badge shows count of pending species photos
• Social auth buttons hidden on login and register views (pending re-enable)
• Added animals:backfill-species Artisan command — matches Animal.category to species via common_name LIKE, supports --dry-run and --force-first; hardcoded overrides for Western Hognose (Heterodon nasicus) and Coastal Carpet Pythons (Morelia spilota)
• Added species_id FK column to animals table (nullable, nullOnDelete)
• Added Animal→Species belongsTo relationship and Species→Animal hasMany
• Embedded Laravel 11 docs in storage/docs/laravel/ for local Claude inference reference
• Added Cloudflare Turnstile bot protection to animal and classified inquiry forms; server-side verification via ValidatesTurnstile trait; disabled automatically when TURNSTILE_SITE_KEY not set
• Species photo uploads use Digital Ocean Spaces (S3-compatible, sfo3); admin uploads bypass moderation and publish immediately
• Configured league/flysystem-aws-s3-v3 S3 adapter; visibility: public default on s3 disk

2026-04-30

• Added social auth (Google, Facebook, Twitch) via Laravel Socialite
• Added SocialAccount model and social_accounts migration

2026-04-27

• Added FEATURE_CLASSIFIEDS feature flag (disabled on production by default)
• Removed HTMX and Hyperscript; Alpine.js only

2024-04-24 Major overhaul and launch preparation

• Updated to Laravel 11
• Removed UUIDs from models
• Removed Dyrynda's deprecated UUID packages
• Updated package.json and composer.json dependencies
• Removed Daisy UI
• Added @tailwindcss/typography
• Updated root .gitignore