SPB Git forge
38commits 1branches 0releases
338.7 MBsize
maindefault branch
3 h agolast push
HTML 53.9% TypeScript 44.5% JavaScript 0.6% SQL 0.5%
1.3 KB · 24 lines yaml
Raw Blame History
1id: teraco2name: Teraco3domain: www.teraco.co.za4kind: operator5mode: hybrid6priority: 27license: "Publicly accessible operator pages; factual data only; © Teraco (a Digital Realty company)"8attribution: "Teraco"9homepage: https://www.teraco.co.za/data-centre-locations/10notes: "Metro pages /data-centre-locations/<metro>/ contain one block per facility (code JB1+JB3+JB5, campus, white space SQM, IT load MW, street address) → one record per block."11coverage: { operators: ["Teraco"], countries: ["ZA"] }12fetch: { level: 1, maxLevel: 2, rpm: 15, concurrency: 2, respectRobots: true, maxCreditsPerRun: 20 }13discovery:14  sitemap: ["https://www.teraco.co.za/page-sitemap.xml", "https://www.teraco.co.za/post-sitemap.xml"]15  include: ["/data-centre-locations/[a-z-]+/$", "/news/", "/press-releases?/"]16  exclude: ["\\?", "#"]17  classify:18    - { pattern: "/data-centre-locations/[a-z-]+/$", group: facility_pages, pageType: facility_page, priority: 60 }19    - { pattern: "/(news|press-releases?)/.", group: newsroom, pageType: press_release, priority: 50 }20schedule: { facility_pages: weekly, newsroom: daily, sitemap: weekly }21extractors:22  facility: { match: ["/data-centre-locations/[a-z-]+/$"], kind: facility, parser: teraco_facility_v1 }23defaults: { operatorName: Teraco, facilityType: colocation, countryIso2: ZA }24