Initial commit: funeral provider discovery pipeline

Python crawlers for VIC Register, Funerals Australia, NFDA n8n workflows for scheduled discovery and enrichment SQLite schema and seeded dev database (1,463 providers) End-to-end process documentation in n8n/PROCESS.md
2026-04-24 10:27:08 +10:00
commit cc91427789
30 changed files with 4706 additions and 0 deletions
--- a/database/IMAGE-MAPPING.md
+++ b/database/IMAGE-MAPPING.md
@@ -0,0 +1,69 @@
+# Image Assets & Verified Provider Mapping
+
+## Image Directory Structure
+
+All images are downloaded locally in `images/` with the following structure:
+
+```
+images/
+├── manifest.json              # Full index mapping CMS IDs → local paths
+├── providers/{slug}/          # 12 verified brands
+│   ├── logo.{ext}             # Rectangular/stacked logo
+│   └── badge.{ext}            # Circular/square badge (for cards)
+├── funeral-homes/{slug}/      # 7 parent organisations
+│   └── logo.{ext}
+├── locations/{slug}/          # 20 physical offices
+│   └── photo.{ext}            # Building/staff hero photo
+├── coffins/{category}/        # 201 coffins by range
+│   └── {slug}/01.{ext}       # 1-4 images per coffin
+├── venues/{slug}/             # 1,678 service venues
+│   └── 01.{ext}
+└── crematoriums/{slug}/       # 38 crematoriums
+    └── 01.{ext}
+```
+
+## Verified Brand → Image Mapping
+
+These are the 12 existing verified brands from the CMS, with their image paths:
+
+| CMS ID | Brand | Logo | Badge |
+|--------|-------|------|-------|
+| 1 | H.Parsons Funeral Directors | `providers/hparsons-funeral-directors/logo.png` | `providers/hparsons-funeral-directors/badge.png` |
+| 3 | Rankins Funerals | `providers/rankins-funerals/logo.webp` | `providers/rankins-funerals/badge.png` |
+| 4 | Parsons Ladies Funeral Directors | `providers/parsons-ladies-funeral-directors/logo.png` | `providers/parsons-ladies-funeral-directors/badge.png` |
+| 5 | Wollongong City Funerals | `providers/wollongong-city-funerals/logo.webp` | `providers/wollongong-city-funerals/badge.png` |
+| 6 | Easy Funerals | `providers/easy-funerals/logo.webp` | `providers/easy-funerals/badge.png` |
+| 7 | Mackay Family Funerals | `providers/mackay-family-funerals/logo.webp` | `providers/mackay-family-funerals/badge.png` |
+| 8 | H.Parsons Shoalhaven | `providers/hparsons-funeral-directors-shoalhaven/logo.png` | `providers/hparsons-funeral-directors-shoalhaven/badge.png` |
+| 9 | Killick Family Funerals | `providers/killick-family-funerals/logo.webp` | `providers/killick-family-funerals/badge.png` |
+| 10 | Kenneally's Funerals | `providers/kenneallys-funerals/logo.webp` | `providers/kenneallys-funerals/badge.png` |
+| 11 | Lady Anne Funerals | `providers/lady-anne-funerals/logo.webp` | `providers/lady-anne-funerals/badge.png` |
+| 12 | Mannings Funerals | `providers/mannings-funerals/logo.webp` | `providers/mannings-funerals/badge.png` |
+| 13 | Botanical Funerals | `providers/botanical-funerals-by-ian-allison/logo.webp` | `providers/botanical-funerals-by-ian-allison/badge.png` |
+
+## How to Use on the Demo Site
+
+### For verified providers:
+- Serve images from `images/providers/{slug}/` for logos and badges
+- Serve location photos from `images/locations/{slug}/`
+- Serve product images from `images/coffins/`, `images/venues/`, `images/crematoriums/`
+- The `manifest.json` contains the full mapping from CMS record IDs to local file paths
+
+### For unverified providers:
+- **No images** — they have no logo, badge, or photos
+- Use a generic placeholder or text-based display (business name initials, etc.)
+- Images are only added when a provider signs up to become verified
+
+### Importing verified brands:
+The 12 verified brands need to be imported into the database with their full data from
+`schemas/brands-full.json` (brand details, locations, packages, inclusions) and linked
+to their images. Some of these brands were also discovered by the crawler and already
+exist in `providers.db` as unverified — they should be **upgraded** (set `verified = true`,
+add images) rather than duplicated.
+
+### Product images:
+- 201 coffins with 1-4 images each, organised by range (solid-timber, custom-board, etc.)
+- 1,678 venue photos
+- 38 crematorium photos
+- These are only relevant for verified provider flows (arrangement booking)
+- The `manifest.json` maps each product's CMS ID to its local image path
--- a/database/PROVIDER-SCHEMA-SPEC.md
+++ b/database/PROVIDER-SCHEMA-SPEC.md
@@ -0,0 +1,209 @@
+# Provider Data Model — Verified & Unverified Providers
+
+This document extends the CMS schema (`schemas/cms-schema-spec.md`) with support for
+unverified (auto-discovered) providers alongside the existing verified (signed-up) providers.
+
+---
+
+## Overview
+
+The platform lists funeral directors in two categories:
+
+- **Verified providers** — Signed up to the platform. Full branding (logo, badge, colours),
+  complete package configuration, and online arrangement booking enabled.
+- **Unverified providers** — Auto-discovered from public registries and their own websites.
+  Listed with whatever public information is available. Can apply to become verified.
+
+All providers share the same `funeral_brand` table and schema. The difference is driven
+by data completeness and the `verified` / `listing_tier` fields.
+
+---
+
+## Schema Changes to FuneralBrand
+
+These fields are **added** to the existing FuneralBrand collection from `cms-schema-spec.md`:
+
+| Field | Type | Default | Purpose |
+|-------|------|---------|---------|
+| `verified` | Boolean | `false` | `true` for signed-up partners, `false` for auto-discovered |
+| `listing_tier` | Enum | `'listed'` | Display tier, computed from data quality (see below) |
+| `hidden` | Boolean | `true` | Unverified providers start hidden until admin-reviewed |
+| `source_key` | String (unique) | `null` | Provenance identifier, e.g. `"nfda:1234"` |
+| `source_url` | String (URL) | `null` | Where this record was discovered |
+| `last_enriched_at` | DateTime | `null` | When data was last refreshed from provider's website |
+| `enrichment_status` | Enum | `'pending'` | `pending` / `partial` / `complete` / `failed` |
+
+### Fields that become optional for unverified providers
+
+These fields are **required** for verified providers but **nullable** for unverified:
+
+| Field | Verified | Unverified |
+|-------|----------|------------|
+| `logo` | Required (brand logo image) | `null` — no images until they sign up |
+| `badge` | Required (card badge image) | `null` — no images until they sign up |
+| `description` | Required | Optional (extracted from their website if available) |
+| `backgroundColour` | Set (brand theme) | `null` — use platform default |
+| `foregroundColour` | Set (brand theme) | `null` — use platform default |
+| `modalDescription` | Set | `null` |
+| `code` | Set (URL slug) | Auto-generated from business name |
+
+### Fields present for both verified and unverified
+
+| Field | Notes |
+|-------|-------|
+| `title` | Business name (always present) |
+| `phone` | Contact phone (present for ~94% of providers) |
+| `email` | Contact email (present for ~66%) |
+| `website` | External website URL (present for ~68%) |
+| `abn` | Australian Business Number (strongest dedup key) |
+| `businessAddress/Suburb/State/Postcode` | Business location |
+| `availableFuneralTypes` | Comma-separated funeral type IDs |
+
+---
+
+## Listing Tiers
+
+Every provider is assigned a `listing_tier` that determines how they appear on the platform.
+The tier is **computed from data quality** — specifically from what package/pricing data exists.
+
+| Tier | Value | Criteria | UI Treatment |
+|------|-------|----------|-------------|
+| **Verified** | `'verified'` | `verified = true` | Full branding, package selection, online arrangements, custom images |
+| **Priced** | `'priced'` | Unverified + 2 or more packages with itemized inclusion prices | Show packages with line-item breakdowns, no arrangements |
+| **Estimated** | `'estimated'` | Unverified + at least 1 package with a total price | Show package prices, "Contact for full details" on breakdowns |
+| **Listed** | `'listed'` | Unverified + no pricing data | Show contact info only, "Contact for pricing" CTA |
+
+### Tier computation logic
+
+```
+if brand.verified:
+    tier = 'verified'
+elif brand has 2+ packages, each with 2+ priced inclusions:
+    tier = 'priced'
+elif brand has 1+ packages with any price:
+    tier = 'estimated'
+else:
+    tier = 'listed'
+```
+
+### Upgrade incentive
+
+Each tier below verified creates a natural CTA for the provider:
+- `listed` → "Publish your pricing to help families compare"
+- `estimated` → "Add detailed breakdowns to stand out"  
+- `priced` → "Sign up to enable online arrangements and add your branding"
+
+---
+
+## Data Relationships (unchanged from CMS spec, but applied to both tiers)
+
+```
+FuneralBrand (verified or unverified)
+  ├── Location[] (physical offices — at least 1 per provider)
+  ├── Package[] (funeral plan bundles — 0 for 'listed' tier)
+  │     └── PackageInclusion[] (fee line items — 0 for 'estimated' tier)
+  ├── KnownFor[] (feature badges — verified only typically)
+  └── FuneralArea[] (service regions — M:N)
+```
+
+### Package (same schema as CMS spec, with additions)
+
+| Field | Type | Notes |
+|-------|------|-------|
+| `id` | PK | |
+| `title` | String | e.g. "Direct Cremation", "Chapel Service" |
+| `description` | Text | What's included |
+| `funeral_type` | Enum | `Service & Cremation`, `Service & Burial`, `Cremation Only`, `Graveside Burial`, `Water Cremation` |
+| `brand_id` | FK → FuneralBrand | |
+| `source_url` | String | Where this pricing was found (provider's website) |
+| `extraction_confidence` | Float 0-1 | How reliable the extracted data is (0.7 = HTML, 0.6 = PDF) |
+| `sort` | Integer | Display order |
+| `hidden` | Boolean | |
+
+### PackageInclusion (same schema as CMS spec)
+
+| Field | Type | Notes |
+|-------|------|-------|
+| `id` | PK | |
+| `price` | Decimal | Dollar amount |
+| `optional` | Boolean | User can opt in/out |
+| `complimentary` | Boolean | Included free |
+| `display` | Boolean | Whether shown to user |
+| `inclusion_type_title` | String | Category label (see standard types below) |
+| `package_id` | FK → Package | |
+
+### Standard inclusion type names
+
+These are the consistent labels used across all providers:
+
+**Standard fees:** Professional Service Fee, Transportation Service Fee, Professional Mortuary Care, Death Registration Certificate, Cremation Certificate/Permit, Government Levy, Accommodation
+
+**Products:** Coffin, Cremation Fee, Cemetery Fee, Celebrant Fee
+
+**Optional extras:** Saturday Service Fee, Twilight Service Surcharge, Viewing Fee, After Hours Transfer Surcharge, Dressing Fee, Embalming, Digital Recording, Webstreaming, Coffin Bearing by Funeral Directors
+
+---
+
+## Current Data
+
+The database (`database/providers.db`, SQLite) contains:
+
+| Metric | Count |
+|--------|-------|
+| Total providers | 1,463 |
+| With phone | 1,380 (94%) |
+| With email | 972 (66%) |
+| With website | 994 (68%) |
+| With description | 618 (42%) |
+| Total packages | 416 |
+| Total inclusions | 388 |
+
+### Tier distribution
+
+| Tier | Providers |
+|------|-----------|
+| Verified | 0 (existing 12 brands not yet imported as verified) |
+| Priced | 10 |
+| Estimated | 111 |
+| Listed | 1,342 |
+
+### State distribution
+
+| State | Providers | With Pricing |
+|-------|-----------|-------------|
+| VIC | 701 | 77 |
+| NSW | 269 | 8 |
+| QLD | 151 | 21 |
+| SA | 85 | 1 |
+| WA | 79 | 12 |
+| TAS | 25 | 0 |
+| NT | 7 | 0 |
+| ACT | 9 | 0 |
+
+---
+
+## Database Schema Files
+
+- **`database/schema.sql`** — Full Postgres schema (production-ready)
+- **`database/schema_sqlite.sql`** — SQLite schema (dev/demo)
+- **`database/providers.db`** — Live SQLite database with 1,463 providers
+- **`database/seed_verified.sql`** — Script to mark imported CMS brands as verified
+
+The schema is designed to be **additive** to the existing CMS schema from `schemas/cms-schema-spec.md`.
+The original 12 verified brands and their packages/products should be imported first, then
+`seed_verified.sql` marks them as `verified = true, listing_tier = 'verified'`.
+
+---
+
+## Verified Provider Upgrade Path
+
+When an unverified provider applies to become verified:
+
+1. They claim their listing (email verification or ABN match)
+2. They fill in missing fields: description, logo, badge, brand colours
+3. They configure packages with full inclusion breakdowns
+4. They enable arrangement booking
+5. Admin approves → `verified = true, listing_tier = 'verified'`
+
+The backend should support this flow — updating an existing unverified brand
+record rather than creating a new one.
--- a/database/providers.db
+++ b/database/providers.db
--- a/database/schema.sql
+++ b/database/schema.sql
@@ -0,0 +1,285 @@
+-- Provider Discovery Pipeline - Database Schema
+-- Designed for Postgres. Compatible with SilverStripe CMS adaptation.
+--
+-- This schema covers the provider-facing tables needed for both
+-- verified (signed-up) and unverified (auto-discovered) providers.
+-- Product catalog tables (coffins, venues, etc.) are NOT included here —
+-- those only apply to verified providers and live in the main CMS.
+
+BEGIN;
+
+-- ============================================================
+-- ENUMS
+-- ============================================================
+
+CREATE TYPE enrichment_status AS ENUM ('pending', 'partial', 'complete', 'failed');
+
+-- Listing tier determines how a provider appears on the platform.
+-- Computed from data quality: verified status + packages + inclusions.
+CREATE TYPE listing_tier AS ENUM (
+    'verified',    -- Tier 1: Signed up, full branding, arrangements enabled
+    'priced',      -- Tier 2: Unverified, 2+ packages with itemized inclusion prices
+    'estimated',   -- Tier 3: Unverified, at least one total package price
+    'listed'       -- Tier 4: Unverified, contact info only, no pricing
+);
+
+CREATE TYPE funeral_type_enum AS ENUM (
+    'Service & Cremation',
+    'Service & Burial',
+    'Cremation Only',
+    'Graveside Burial',
+    'Water Cremation'
+);
+
+-- ============================================================
+-- 1. FUNERAL HOME (parent organisation)
+-- ============================================================
+
+CREATE TABLE funeral_home (
+    id              SERIAL PRIMARY KEY,
+    title           TEXT NOT NULL,
+    website         TEXT,
+    created_at      TIMESTAMPTZ NOT NULL DEFAULT NOW(),
+    updated_at      TIMESTAMPTZ NOT NULL DEFAULT NOW()
+);
+
+-- ============================================================
+-- 2. FUNERAL BRAND (customer-facing provider)
+-- ============================================================
+
+CREATE TABLE funeral_brand (
+    id                  SERIAL PRIMARY KEY,
+    title               TEXT NOT NULL,
+    description         TEXT,
+    modal_description   TEXT,
+    email               TEXT,
+    phone               TEXT,
+    website             TEXT,
+    abn                 TEXT,
+    code                TEXT UNIQUE,           -- URL slug (e.g. "hparsons")
+    sort                INTEGER DEFAULT 0,
+    hidden              BOOLEAN NOT NULL DEFAULT TRUE,  -- unverified start hidden
+
+    -- Address
+    business_address    TEXT,
+    business_suburb     TEXT,
+    business_state      TEXT,
+    business_postcode   TEXT,
+
+    -- Branding (nullable — unverified providers have no images)
+    background_colour   TEXT,
+    foreground_colour    TEXT,
+
+    -- Organisation
+    funeral_home_id     INTEGER REFERENCES funeral_home(id) ON DELETE SET NULL,
+
+    -- Verified vs auto-discovered
+    verified            BOOLEAN NOT NULL DEFAULT FALSE,
+
+    -- Provenance tracking
+    source_key          TEXT UNIQUE,           -- "{source}:{externalId}" for dedup
+    source_url          TEXT,                  -- where this record was found
+    last_enriched_at    TIMESTAMPTZ,
+    enrichment_status   enrichment_status NOT NULL DEFAULT 'pending',
+
+    -- Listing tier (computed from data quality)
+    listing_tier        listing_tier NOT NULL DEFAULT 'listed',
+
+    -- Funeral types offered (comma-separated IDs, same as existing CMS)
+    available_funeral_types TEXT,
+
+    created_at          TIMESTAMPTZ NOT NULL DEFAULT NOW(),
+    updated_at          TIMESTAMPTZ NOT NULL DEFAULT NOW()
+);
+
+-- Deduplication indexes
+CREATE INDEX idx_brand_abn ON funeral_brand(abn) WHERE abn IS NOT NULL;
+CREATE INDEX idx_brand_listing_tier ON funeral_brand(listing_tier);
+CREATE INDEX idx_brand_source_key ON funeral_brand(source_key) WHERE source_key IS NOT NULL;
+CREATE INDEX idx_brand_name_postcode ON funeral_brand(title, business_postcode);
+CREATE INDEX idx_brand_verified ON funeral_brand(verified);
+CREATE INDEX idx_brand_hidden ON funeral_brand(hidden);
+CREATE INDEX idx_brand_enrichment ON funeral_brand(enrichment_status) WHERE verified = FALSE;
+
+-- ============================================================
+-- 3. LOCATION (physical office/chapel)
+-- ============================================================
+
+CREATE TABLE location (
+    id              SERIAL PRIMARY KEY,
+    title           TEXT NOT NULL,           -- display name (e.g. "Kingaroy, QLD")
+    address         TEXT,
+    suburb          TEXT,
+    state           TEXT,
+    postcode        TEXT,
+    country         TEXT DEFAULT 'Australia',
+    lat             DOUBLE PRECISION,
+    lng             DOUBLE PRECISION,
+    rating          REAL,                    -- Google rating 0-5
+    rating_num      INTEGER,                 -- number of Google reviews
+    google_place_key TEXT,                   -- Google Places ID
+
+    brand_id        INTEGER NOT NULL REFERENCES funeral_brand(id) ON DELETE CASCADE,
+
+    created_at      TIMESTAMPTZ NOT NULL DEFAULT NOW(),
+    updated_at      TIMESTAMPTZ NOT NULL DEFAULT NOW()
+);
+
+CREATE INDEX idx_location_brand ON location(brand_id);
+CREATE INDEX idx_location_state ON location(state);
+CREATE INDEX idx_location_postcode ON location(postcode);
+CREATE INDEX idx_location_coords ON location(lat, lng);
+CREATE INDEX idx_location_google ON location(google_place_key) WHERE google_place_key IS NOT NULL;
+
+-- ============================================================
+-- 4. FUNERAL AREA (service region)
+-- ============================================================
+
+CREATE TABLE funeral_area (
+    id              SERIAL PRIMARY KEY,
+    title           TEXT NOT NULL,
+    code            TEXT,
+    description     TEXT,
+    postcodes       TEXT,                    -- comma-separated postcode list
+    sort            INTEGER DEFAULT 0,
+    hidden          BOOLEAN DEFAULT FALSE,
+    created_at      TIMESTAMPTZ NOT NULL DEFAULT NOW(),
+    updated_at      TIMESTAMPTZ NOT NULL DEFAULT NOW()
+);
+
+-- Junction: brand <-> funeral_area
+CREATE TABLE brand_funeral_area (
+    brand_id        INTEGER NOT NULL REFERENCES funeral_brand(id) ON DELETE CASCADE,
+    funeral_area_id INTEGER NOT NULL REFERENCES funeral_area(id) ON DELETE CASCADE,
+    PRIMARY KEY (brand_id, funeral_area_id)
+);
+
+-- ============================================================
+-- 5. PACKAGE (funeral plan bundle)
+-- ============================================================
+
+CREATE TABLE package (
+    id              SERIAL PRIMARY KEY,
+    title           TEXT NOT NULL,
+    description     TEXT,
+    sort            INTEGER DEFAULT 0,
+    hidden          BOOLEAN DEFAULT FALSE,
+    for_whom        TEXT,                    -- 'myself' / 'someone' / null (both)
+    religion        TEXT,                    -- comma-separated supported religions
+    funeral_type    funeral_type_enum,
+
+    brand_id        INTEGER NOT NULL REFERENCES funeral_brand(id) ON DELETE CASCADE,
+
+    -- Provenance (for AI-extracted packages)
+    source_url      TEXT,                    -- page this was extracted from
+    extraction_confidence REAL,              -- 0-1 confidence score from AI
+
+    created_at      TIMESTAMPTZ NOT NULL DEFAULT NOW(),
+    updated_at      TIMESTAMPTZ NOT NULL DEFAULT NOW()
+);
+
+CREATE INDEX idx_package_brand ON package(brand_id);
+CREATE INDEX idx_package_type ON package(funeral_type);
+
+-- Junction: package <-> funeral_area
+CREATE TABLE package_funeral_area (
+    package_id      INTEGER NOT NULL REFERENCES package(id) ON DELETE CASCADE,
+    funeral_area_id INTEGER NOT NULL REFERENCES funeral_area(id) ON DELETE CASCADE,
+    PRIMARY KEY (package_id, funeral_area_id)
+);
+
+-- ============================================================
+-- 6. PACKAGE INCLUSION (fee line item within a package)
+-- ============================================================
+
+CREATE TABLE package_inclusion (
+    id                  SERIAL PRIMARY KEY,
+    price               NUMERIC(10,2) NOT NULL,
+    optional            BOOLEAN NOT NULL DEFAULT FALSE,
+    complimentary       BOOLEAN NOT NULL DEFAULT FALSE,
+    display             BOOLEAN NOT NULL DEFAULT TRUE,
+    description         TEXT,
+    sort                INTEGER DEFAULT 0,
+    inclusion_type_title TEXT NOT NULL,       -- category label (e.g. "Professional Service Fee")
+
+    package_id          INTEGER NOT NULL REFERENCES package(id) ON DELETE CASCADE,
+
+    created_at          TIMESTAMPTZ NOT NULL DEFAULT NOW(),
+    updated_at          TIMESTAMPTZ NOT NULL DEFAULT NOW()
+);
+
+CREATE INDEX idx_inclusion_package ON package_inclusion(package_id);
+
+-- ============================================================
+-- 7. KNOWN FOR (feature badges on provider cards)
+-- ============================================================
+
+CREATE TABLE known_for (
+    id              SERIAL PRIMARY KEY,
+    title           TEXT NOT NULL,
+    brand_id        INTEGER NOT NULL REFERENCES funeral_brand(id) ON DELETE CASCADE
+);
+
+CREATE INDEX idx_known_for_brand ON known_for(brand_id);
+
+-- ============================================================
+-- 8. SOURCE LOG (audit trail of scrape runs)
+-- ============================================================
+
+CREATE TABLE source_log (
+    id              SERIAL PRIMARY KEY,
+    source_name     TEXT NOT NULL,            -- 'vic_register', 'gathered_here', 'nfda', 'funerals_australia'
+    run_started_at  TIMESTAMPTZ NOT NULL DEFAULT NOW(),
+    run_finished_at TIMESTAMPTZ,
+    records_found   INTEGER DEFAULT 0,
+    records_new     INTEGER DEFAULT 0,
+    records_updated INTEGER DEFAULT 0,
+    records_skipped INTEGER DEFAULT 0,
+    status          TEXT DEFAULT 'running',   -- 'running', 'completed', 'failed'
+    error_message   TEXT,
+    metadata        JSONB                     -- any extra run info
+);
+
+-- ============================================================
+-- 9. SOURCE RECORD (raw scraped data, kept for audit)
+-- ============================================================
+
+CREATE TABLE source_record (
+    id              SERIAL PRIMARY KEY,
+    source_name     TEXT NOT NULL,
+    source_id       TEXT NOT NULL,            -- external ID from the source
+    source_url      TEXT,
+    raw_data        JSONB NOT NULL,           -- original scraped data
+    normalized_data JSONB,                    -- mapped to intermediate format
+    matched_brand_id INTEGER REFERENCES funeral_brand(id) ON DELETE SET NULL,
+    match_type      TEXT,                     -- 'source_key', 'abn', 'name_postcode', 'fuzzy', 'new'
+    processed_at    TIMESTAMPTZ,
+    log_id          INTEGER REFERENCES source_log(id) ON DELETE SET NULL,
+    created_at      TIMESTAMPTZ NOT NULL DEFAULT NOW(),
+
+    UNIQUE(source_name, source_id)
+);
+
+CREATE INDEX idx_source_record_source ON source_record(source_name, source_id);
+CREATE INDEX idx_source_record_brand ON source_record(matched_brand_id) WHERE matched_brand_id IS NOT NULL;
+
+-- ============================================================
+-- UPDATED_AT TRIGGER
+-- ============================================================
+
+CREATE OR REPLACE FUNCTION update_updated_at()
+RETURNS TRIGGER AS $$
+BEGIN
+    NEW.updated_at = NOW();
+    RETURN NEW;
+END;
+$$ LANGUAGE plpgsql;
+
+CREATE TRIGGER trg_funeral_home_updated BEFORE UPDATE ON funeral_home FOR EACH ROW EXECUTE FUNCTION update_updated_at();
+CREATE TRIGGER trg_funeral_brand_updated BEFORE UPDATE ON funeral_brand FOR EACH ROW EXECUTE FUNCTION update_updated_at();
+CREATE TRIGGER trg_location_updated BEFORE UPDATE ON location FOR EACH ROW EXECUTE FUNCTION update_updated_at();
+CREATE TRIGGER trg_funeral_area_updated BEFORE UPDATE ON funeral_area FOR EACH ROW EXECUTE FUNCTION update_updated_at();
+CREATE TRIGGER trg_package_updated BEFORE UPDATE ON package FOR EACH ROW EXECUTE FUNCTION update_updated_at();
+CREATE TRIGGER trg_package_inclusion_updated BEFORE UPDATE ON package_inclusion FOR EACH ROW EXECUTE FUNCTION update_updated_at();
+
+COMMIT;
--- a/database/schema_sqlite.sql
+++ b/database/schema_sqlite.sql
@@ -0,0 +1,221 @@
+-- Provider Discovery Pipeline - SQLite Schema (for local dev/testing)
+-- Production uses Postgres (see schema.sql)
+
+-- ============================================================
+-- FUNERAL HOME
+-- ============================================================
+
+CREATE TABLE IF NOT EXISTS funeral_home (
+    id              INTEGER PRIMARY KEY AUTOINCREMENT,
+    title           TEXT NOT NULL,
+    website         TEXT,
+    created_at      TEXT NOT NULL DEFAULT (datetime('now')),
+    updated_at      TEXT NOT NULL DEFAULT (datetime('now'))
+);
+
+-- ============================================================
+-- FUNERAL BRAND
+-- ============================================================
+
+CREATE TABLE IF NOT EXISTS funeral_brand (
+    id                  INTEGER PRIMARY KEY AUTOINCREMENT,
+    title               TEXT NOT NULL,
+    description         TEXT,
+    modal_description   TEXT,
+    email               TEXT,
+    phone               TEXT,
+    website             TEXT,
+    abn                 TEXT,
+    code                TEXT UNIQUE,
+    sort                INTEGER DEFAULT 0,
+    hidden              INTEGER NOT NULL DEFAULT 1,
+
+    business_address    TEXT,
+    business_suburb     TEXT,
+    business_state      TEXT,
+    business_postcode   TEXT,
+
+    background_colour   TEXT,
+    foreground_colour   TEXT,
+
+    funeral_home_id     INTEGER REFERENCES funeral_home(id) ON DELETE SET NULL,
+
+    verified            INTEGER NOT NULL DEFAULT 0,
+    source_key          TEXT UNIQUE,
+    source_url          TEXT,
+    last_enriched_at    TEXT,
+    enrichment_status   TEXT NOT NULL DEFAULT 'pending' CHECK(enrichment_status IN ('pending','partial','complete','failed')),
+
+    -- Listing tier: verified | priced | estimated | listed
+    listing_tier        TEXT NOT NULL DEFAULT 'listed'
+        CHECK(listing_tier IN ('verified','priced','estimated','listed')),
+
+    available_funeral_types TEXT,
+
+    created_at          TEXT NOT NULL DEFAULT (datetime('now')),
+    updated_at          TEXT NOT NULL DEFAULT (datetime('now'))
+);
+
+CREATE INDEX IF NOT EXISTS idx_brand_abn ON funeral_brand(abn);
+CREATE INDEX IF NOT EXISTS idx_brand_source_key ON funeral_brand(source_key);
+CREATE INDEX IF NOT EXISTS idx_brand_listing_tier ON funeral_brand(listing_tier);
+CREATE INDEX IF NOT EXISTS idx_brand_name_postcode ON funeral_brand(title, business_postcode);
+CREATE INDEX IF NOT EXISTS idx_brand_verified ON funeral_brand(verified);
+CREATE INDEX IF NOT EXISTS idx_brand_hidden ON funeral_brand(hidden);
+
+-- ============================================================
+-- LOCATION
+-- ============================================================
+
+CREATE TABLE IF NOT EXISTS location (
+    id              INTEGER PRIMARY KEY AUTOINCREMENT,
+    title           TEXT NOT NULL,
+    address         TEXT,
+    suburb          TEXT,
+    state           TEXT,
+    postcode        TEXT,
+    country         TEXT DEFAULT 'Australia',
+    lat             REAL,
+    lng             REAL,
+    rating          REAL,
+    rating_num      INTEGER,
+    google_place_key TEXT,
+
+    brand_id        INTEGER NOT NULL REFERENCES funeral_brand(id) ON DELETE CASCADE,
+
+    created_at      TEXT NOT NULL DEFAULT (datetime('now')),
+    updated_at      TEXT NOT NULL DEFAULT (datetime('now'))
+);
+
+CREATE INDEX IF NOT EXISTS idx_location_brand ON location(brand_id);
+CREATE INDEX IF NOT EXISTS idx_location_postcode ON location(postcode);
+
+-- ============================================================
+-- FUNERAL AREA
+-- ============================================================
+
+CREATE TABLE IF NOT EXISTS funeral_area (
+    id              INTEGER PRIMARY KEY AUTOINCREMENT,
+    title           TEXT NOT NULL,
+    code            TEXT,
+    description     TEXT,
+    postcodes       TEXT,
+    sort            INTEGER DEFAULT 0,
+    hidden          INTEGER DEFAULT 0,
+    created_at      TEXT NOT NULL DEFAULT (datetime('now')),
+    updated_at      TEXT NOT NULL DEFAULT (datetime('now'))
+);
+
+CREATE TABLE IF NOT EXISTS brand_funeral_area (
+    brand_id        INTEGER NOT NULL REFERENCES funeral_brand(id) ON DELETE CASCADE,
+    funeral_area_id INTEGER NOT NULL REFERENCES funeral_area(id) ON DELETE CASCADE,
+    PRIMARY KEY (brand_id, funeral_area_id)
+);
+
+-- ============================================================
+-- PACKAGE
+-- ============================================================
+
+CREATE TABLE IF NOT EXISTS package (
+    id              INTEGER PRIMARY KEY AUTOINCREMENT,
+    title           TEXT NOT NULL,
+    description     TEXT,
+    sort            INTEGER DEFAULT 0,
+    hidden          INTEGER DEFAULT 0,
+    for_whom        TEXT,
+    religion        TEXT,
+    funeral_type    TEXT CHECK(funeral_type IN (
+        'Service & Cremation','Service & Burial','Cremation Only',
+        'Graveside Burial','Water Cremation'
+    )),
+
+    brand_id        INTEGER NOT NULL REFERENCES funeral_brand(id) ON DELETE CASCADE,
+
+    source_url      TEXT,
+    extraction_confidence REAL,
+
+    created_at      TEXT NOT NULL DEFAULT (datetime('now')),
+    updated_at      TEXT NOT NULL DEFAULT (datetime('now'))
+);
+
+CREATE INDEX IF NOT EXISTS idx_package_brand ON package(brand_id);
+
+CREATE TABLE IF NOT EXISTS package_funeral_area (
+    package_id      INTEGER NOT NULL REFERENCES package(id) ON DELETE CASCADE,
+    funeral_area_id INTEGER NOT NULL REFERENCES funeral_area(id) ON DELETE CASCADE,
+    PRIMARY KEY (package_id, funeral_area_id)
+);
+
+-- ============================================================
+-- PACKAGE INCLUSION
+-- ============================================================
+
+CREATE TABLE IF NOT EXISTS package_inclusion (
+    id                  INTEGER PRIMARY KEY AUTOINCREMENT,
+    price               REAL NOT NULL,
+    optional            INTEGER NOT NULL DEFAULT 0,
+    complimentary       INTEGER NOT NULL DEFAULT 0,
+    display             INTEGER NOT NULL DEFAULT 1,
+    description         TEXT,
+    sort                INTEGER DEFAULT 0,
+    inclusion_type_title TEXT NOT NULL,
+
+    package_id          INTEGER NOT NULL REFERENCES package(id) ON DELETE CASCADE,
+
+    created_at          TEXT NOT NULL DEFAULT (datetime('now')),
+    updated_at          TEXT NOT NULL DEFAULT (datetime('now'))
+);
+
+CREATE INDEX IF NOT EXISTS idx_inclusion_package ON package_inclusion(package_id);
+
+-- ============================================================
+-- KNOWN FOR
+-- ============================================================
+
+CREATE TABLE IF NOT EXISTS known_for (
+    id              INTEGER PRIMARY KEY AUTOINCREMENT,
+    title           TEXT NOT NULL,
+    brand_id        INTEGER NOT NULL REFERENCES funeral_brand(id) ON DELETE CASCADE
+);
+
+CREATE INDEX IF NOT EXISTS idx_known_for_brand ON known_for(brand_id);
+
+-- ============================================================
+-- SOURCE LOG
+-- ============================================================
+
+CREATE TABLE IF NOT EXISTS source_log (
+    id              INTEGER PRIMARY KEY AUTOINCREMENT,
+    source_name     TEXT NOT NULL,
+    run_started_at  TEXT NOT NULL DEFAULT (datetime('now')),
+    run_finished_at TEXT,
+    records_found   INTEGER DEFAULT 0,
+    records_new     INTEGER DEFAULT 0,
+    records_updated INTEGER DEFAULT 0,
+    records_skipped INTEGER DEFAULT 0,
+    status          TEXT DEFAULT 'running',
+    error_message   TEXT,
+    metadata        TEXT  -- JSON string
+);
+
+-- ============================================================
+-- SOURCE RECORD
+-- ============================================================
+
+CREATE TABLE IF NOT EXISTS source_record (
+    id              INTEGER PRIMARY KEY AUTOINCREMENT,
+    source_name     TEXT NOT NULL,
+    source_id       TEXT NOT NULL,
+    source_url      TEXT,
+    raw_data        TEXT NOT NULL,     -- JSON string
+    normalized_data TEXT,              -- JSON string
+    matched_brand_id INTEGER REFERENCES funeral_brand(id) ON DELETE SET NULL,
+    match_type      TEXT,
+    processed_at    TEXT,
+    log_id          INTEGER REFERENCES source_log(id) ON DELETE SET NULL,
+    created_at      TEXT NOT NULL DEFAULT (datetime('now')),
+
+    UNIQUE(source_name, source_id)
+);
+
+CREATE INDEX IF NOT EXISTS idx_source_record_source ON source_record(source_name, source_id);
--- a/database/seed_verified.sql
+++ b/database/seed_verified.sql
@@ -0,0 +1,24 @@
+-- Seed script: Mark existing brands as verified
+-- Run after importing existing CMS data into the new schema.
+--
+-- This updates all pre-existing brands (imported from brands-full.json)
+-- to verified=true, hidden=false, enrichment_status='complete'.
+
+UPDATE funeral_brand
+SET verified = TRUE,
+    hidden = FALSE,
+    enrichment_status = 'complete',
+    listing_tier = 'verified',
+    updated_at = NOW()
+WHERE id IN (
+    -- IDs from the existing 12 brands in brands-full.json
+    -- These will be populated during the initial CMS data import.
+    -- Update this list to match actual imported IDs.
+    SELECT id FROM funeral_brand WHERE source_key IS NULL
+);
+
+-- Alternatively, if importing with known codes:
+-- UPDATE funeral_brand SET verified = TRUE, hidden = FALSE, enrichment_status = 'complete'
+-- WHERE code IN ('hparsons', 'parsons-ladies', 'rankins', 'killick', 'botanical',
+--                'easy', 'wollongong-city', 'kenneallys', 'lady-anne',
+--                'mackay', 'mannings', 'guardian');