feat: import manual maintainers (CM-1299)

backend/src/osspckgs/migrations/V1782978597__package_maintainers_ingestion_source.sql

@@ -0,0 +1,2 @@
+ALTER TABLE package_maintainers
+    ADD COLUMN IF NOT EXISTS ingestion_source text;

services/apps/packages_worker/package.json

@@ -51,6 +51,8 @@
     "dev:nuget-worker:local": "set -a && . ../../../backend/.env.dist.local && . ../../../backend/.env.override.local && set +a && CROWD_TEMPORAL_TASKQUEUE=nuget-worker SERVICE=nuget-worker LOG_LEVEL=trace nodemon --watch src --watch ../../libs --ext ts --exec tsx --inspect=0.0.0.0:9242 src/bin/nuget-worker.ts",
     "backfill:maven": "SERVICE=maven tsx src/bin/maven-backfill.ts",
     "backfill:maven:local": "set -a && . ../../../backend/.env.dist.local && . ../../../backend/.env.override.local && set +a && SERVICE=maven LOG_LEVEL=info tsx src/bin/maven-backfill.ts",
+    "import:maven-maintainers": "SERVICE=maven tsx src/maven/scripts/importMaintainersFromCsv.ts",
+    "import:maven-maintainers:local": "set -a && . ../../../backend/.env.dist.local && . ../../../backend/.env.override.local && set +a && SERVICE=maven tsx src/maven/scripts/importMaintainersFromCsv.ts",
     "backfill:stewardship": "SERVICE=stewardship-backfill tsx src/bin/stewardship-backfill.ts",
     "backfill:stewardship:local": "set -a && . ../../../backend/.env.dist.local && . ../../../backend/.env.override.local && set +a && SERVICE=stewardship-backfill LOG_LEVEL=info tsx src/bin/stewardship-backfill.ts",
     "monitor:osspckgs": "SERVICE=bq-dataset-ingest tsx src/scripts/monitorOsspckgs.ts",

services/apps/packages_worker/src/maven/scripts/csv.ts

@@ -0,0 +1,56 @@
+export function parseCsv(content: string): Record<string, string>[] {
+  const rows: string[][] = []
+  let cur = ''
+  let inQuote = false
+  let row: string[] = []
+
+  for (let i = 0; i < content.length; i++) {
+    const ch = content[i]
+    const next = content[i + 1]
+
+    if (inQuote) {
+      if (ch === '"' && next === '"') {
+        cur += '"'
+        i++
+      } else if (ch === '"') {
+        inQuote = false
+      } else {
+        cur += ch
+      }
+    } else {
+      if (ch === '"') {
+        inQuote = true
+      } else if (ch === ',') {
+        row.push(cur)
+        cur = ''
+      } else if (ch === '\r' && next === '\n') {
+        row.push(cur)
+        cur = ''
+        rows.push(row)
+        row = []
+        i++
+      } else if (ch === '\n') {
+        row.push(cur)
+        cur = ''
+        rows.push(row)
+        row = []
+      } else {
+        cur += ch
+      }
+    }
+  }
+  if (cur || row.length) {
+    row.push(cur)
+    rows.push(row)
+  }
+
+  if (rows.length === 0) return []
+  const headers = rows[0]
+  return rows.slice(1).map((r) => {
+    const obj: Record<string, string> = {}
+    headers.forEach((h, i) => {
+      obj[h.trim()] = (r[i] ?? '').trim()
+    })
+    return obj
+  })
+}