|
| 1 | +DESCRIPTION > |
| 2 | + Leaderboard ranking projects by package download counts in the last 30 days. |
| 3 | + Compares the current period (last 30 days) with the previous period (30-60 days ago). |
| 4 | + Higher download counts rank higher, showing the most downloaded packages. |
| 5 | + |
| 6 | +NODE leaderboards_package_downloads_projects |
| 7 | +DESCRIPTION > |
| 8 | + Retrieves all projects from the populated datasource |
| 9 | + |
| 10 | +SQL > |
| 11 | + SELECT id, name, slug, segmentId, logoUrl, collectionsSlugs, isLF |
| 12 | + FROM insights_projects_populated_ds |
| 13 | + GROUP BY id, name, slug, segmentId, logoUrl, collectionsSlugs, isLF |
| 14 | + |
| 15 | +NODE leaderboards_package_downloads_current_period |
| 16 | +DESCRIPTION > |
| 17 | + Calculates package downloads for the last 30 days using cumulative data with 30-day buffer windows to handle missing dates |
| 18 | + |
| 19 | +SQL > |
| 20 | + WITH |
| 21 | + recent_downloads AS ( |
| 22 | + SELECT insightsProjectId, argMax(downloadsCount, date) as recentDownloads |
| 23 | + FROM packageDownloads |
| 24 | + WHERE date >= now() - INTERVAL 60 DAY AND date <= now() AND insightsProjectId != '' |
| 25 | + GROUP BY insightsProjectId |
| 26 | + ), |
| 27 | + baseline_downloads AS ( |
| 28 | + SELECT insightsProjectId, argMax(downloadsCount, date) as baselineDownloads |
| 29 | + FROM packageDownloads |
| 30 | + WHERE |
| 31 | + date >= now() - INTERVAL 60 DAY |
| 32 | + AND date < now() - INTERVAL 30 DAY |
| 33 | + AND insightsProjectId != '' |
| 34 | + GROUP BY insightsProjectId |
| 35 | + ) |
| 36 | + SELECT r.insightsProjectId, r.recentDownloads - coalesce(b.baselineDownloads, 0) as downloads |
| 37 | + FROM recent_downloads r |
| 38 | + LEFT JOIN baseline_downloads b ON r.insightsProjectId = b.insightsProjectId |
| 39 | + |
| 40 | +NODE leaderboards_package_downloads_previous_period |
| 41 | +DESCRIPTION > |
| 42 | + Calculates package downloads for the previous 30 days (30-60 days ago) using cumulative data with 30-day buffer windows to handle missing dates |
| 43 | + |
| 44 | +SQL > |
| 45 | + WITH |
| 46 | + baseline_downloads AS ( |
| 47 | + SELECT insightsProjectId, argMax(downloadsCount, date) as baselineDownloads |
| 48 | + FROM packageDownloads |
| 49 | + WHERE |
| 50 | + date >= now() - INTERVAL 90 DAY |
| 51 | + AND date < now() - INTERVAL 30 DAY |
| 52 | + AND insightsProjectId != '' |
| 53 | + GROUP BY insightsProjectId |
| 54 | + ), |
| 55 | + older_downloads AS ( |
| 56 | + SELECT insightsProjectId, argMax(downloadsCount, date) as olderDownloads |
| 57 | + FROM packageDownloads |
| 58 | + WHERE |
| 59 | + date >= now() - INTERVAL 90 DAY |
| 60 | + AND date < now() - INTERVAL 60 DAY |
| 61 | + AND insightsProjectId != '' |
| 62 | + GROUP BY insightsProjectId |
| 63 | + ) |
| 64 | + SELECT b.insightsProjectId, b.baselineDownloads - coalesce(o.olderDownloads, 0) as downloads |
| 65 | + FROM baseline_downloads b |
| 66 | + LEFT JOIN older_downloads o ON b.insightsProjectId = o.insightsProjectId |
| 67 | + |
| 68 | +NODE leaderboards_copy_package_downloads |
| 69 | +DESCRIPTION > |
| 70 | + Joins project metadata with download counts and ranks by most downloaded packages |
| 71 | + |
| 72 | +SQL > |
| 73 | + SELECT |
| 74 | + row_number() OVER (ORDER BY coalesce(c.downloads, 0) DESC) as rank, |
| 75 | + p.id as id, |
| 76 | + p.segmentId as segmentId, |
| 77 | + p.name as name, |
| 78 | + p.slug as slug, |
| 79 | + p.logoUrl as logoUrl, |
| 80 | + p.collectionsSlugs as collectionsSlugs, |
| 81 | + p.isLF as isLF, |
| 82 | + cast(coalesce(c.downloads, 0) as Float64) as value, |
| 83 | + cast(coalesce(pp.downloads, 0) as Float64) as previousPeriodValue |
| 84 | + FROM leaderboards_package_downloads_projects p |
| 85 | + INNER JOIN leaderboards_package_downloads_current_period c ON p.id = c.insightsProjectId |
| 86 | + LEFT JOIN leaderboards_package_downloads_previous_period pp ON p.id = pp.insightsProjectId |
| 87 | + WHERE c.downloads > 0 |
| 88 | + ORDER BY value DESC |
0 commit comments