Creationshop/cannaiq
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Remove domain-based shortcuts for Curaleaf/Sol detection

Browse Source 
Menu detection now always crawls websites to find actual embedded menu
providers instead of marking stores as proprietary based on domain alone.
This fixes detection for stores like Curaleaf that may use Dutchie embeds.

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude <noreply@anthropic.com>
This commit is contained in:
Kelly
2025-12-03 21:55:31 -07:00
parent 95fc8bb4cc
commit 33e12de3f2
1 changed files with 11 additions and 42 deletions
Download Patch File Download Diff File Expand all files Collapse all files

53
backend/src/dutchie-az/services/menu-detection.ts
View File

@@ -29,8 +29,6 @@ const DISPENSARY_COLUMNS = `
export type MenuProvider =
| 'dutchie'
| 'curaleaf' // Curaleaf proprietary platform (not crawlable via Dutchie)
| 'sol' // Sol Flower proprietary platform (not crawlable via Dutchie)
| 'treez'
| 'jane'
| 'iheartjane'
@@ -68,22 +66,7 @@ export interface BulkDetectionResult {
// ============================================================
const PROVIDER_URL_PATTERNS: Array<{ provider: MenuProvider; patterns: RegExp[] }> = [
// IMPORTANT: Curaleaf and Sol must come BEFORE dutchie to take precedence
// These stores have their own proprietary menu systems (not crawlable via Dutchie)
{
provider: 'curaleaf',
patterns: [
/curaleaf\.com\/stores\//i, // e.g., https://curaleaf.com/stores/curaleaf-az-glendale-east
/curaleaf\.com\/dispensary\//i, // e.g., https://curaleaf.com/dispensary/arizona
],
},
{
provider: 'sol',
patterns: [
/livewithsol\.com/i, // e.g., https://www.livewithsol.com/locations/sun-city/
/solflower\.com/i, // alternate domain if any
],
},
// We detect provider based on the actual menu link we find, not just the site domain.
{
provider: 'dutchie',
patterns: [
@@ -199,19 +182,6 @@ function isCuraleafUrl(url: string | null | undefined): boolean {
return /curaleaf\.com\/(stores|dispensary)\//i.test(url);
}
/**
* Extract the Curaleaf store URL from a website URL
* Handles both /stores/ and /dispensary/ formats
*/
function extractCuraleafStoreUrl(url: string): string | null {
if (!url) return null;
// If it's already a Curaleaf stores/dispensary URL, use it
if (isCuraleafUrl(url)) {
return url;
}
return null;
}
/**
* Fetch a page and extract all links
*/
@@ -653,30 +623,29 @@ export async function detectAndResolveDispensary(dispensaryId: number): Promise<
return result;
}
} else {
// Website crawl didn't find any menu provider - mark as proprietary
const notCrawlableReason = `${detectedProvider} proprietary menu - no embedded menu provider found`;
console.log(`[MenuDetection] ${dispensary.name}: Website crawl found no menu provider - marking as ${detectedProvider}`);
// Website crawl didn't find any menu provider - mark unknown with reason
const notCrawlableReason = `No embedded menu provider found`;
console.log(`[MenuDetection] ${dispensary.name}: Website crawl found no menu provider - marking as unknown`);
await query(
`
UPDATE dispensaries SET
menu_type = $1,
menu_type = 'unknown',
platform_dispensary_id = NULL,
provider_detection_data = COALESCE(provider_detection_data, '{}'::jsonb) ||
jsonb_build_object(
'detected_provider', $1::text,
'detected_provider', 'unknown'::text,
'detection_method', 'url_pattern_with_crawl'::text,
'detected_at', NOW(),
'website_crawled', $2::text,
'website_crawl_pages', $3::jsonb,
'website_crawled', $1::text,
'website_crawl_pages', $2::jsonb,
'not_crawlable', true,
'not_crawlable_reason', $4::text
'not_crawlable_reason', $3::text
),
updated_at = NOW()
WHERE id = $5
WHERE id = $4
`,
[
detectedProvider,
website,
JSON.stringify(crawlResult.crawledPages),
notCrawlableReason,
@@ -688,7 +657,7 @@ export async function detectAndResolveDispensary(dispensaryId: number): Promise<
}
}
// If not dutchie and not a proprietary domain we need to crawl, just update menu_type
// If not dutchie, just update menu_type (non-dutchie providers)
if (detectedProvider !== 'dutchie') {
await query(
`