Merge e3beb63182 into ee203df485

admin/app/services/download_service.ts

@@ -9,6 +9,9 @@ import { DownloadJobWithProgress, DownloadProgressData, RunDownloadJobParams } f
 import type { Job, Queue } from 'bullmq'
 import { normalize } from 'path'
 import { deleteFileIfExists } from '../utils/fs.js'
+import { refreshRetryParams } from '../utils/download_retry.js'
+import { CollectionManifestService } from './collection_manifest_service.js'
+import type { ZimCategoriesSpec } from '../../types/collections.js'
 import transmit from '@adonisjs/transmit/services/main'
 import { BROADCAST_CHANNELS } from '../../constants/broadcast.js'
 
@@ -200,11 +203,13 @@ export class DownloadService {
           return { success: true, message: `Retrying download for model ${modelName}` }
         }
 
-        // For file downloads (zim, map, etc.), re-dispatch with original params
-        const params = job.data as RunDownloadJobParams
-        if (!params.url || !params.filepath) {
+        // For file downloads (zim, map, etc.), re-dispatch with the stored params,
+        // refreshed from the current catalog for a curated ZIM
+        const stored = job.data as RunDownloadJobParams
+        if (!stored.url || !stored.filepath) {
           return { success: false, message: 'Cannot retry: missing URL or filepath in job data' }
         }
+        const params = refreshRetryParams(stored, await this.currentZimCatalog())
 
         // Remove the old failed job, then dispatch a fresh one
         await job.remove().catch(() => {})
@@ -216,6 +221,17 @@ export class DownloadService {
     return { success: false, message: 'Failed job not found. It may have already been dismissed.' }
   }
 
+  /** Null when the catalog cannot be read, so a retry falls back to its stored params. */
+  private async currentZimCatalog(): Promise<ZimCategoriesSpec | null> {
+    try {
+      return await new CollectionManifestService().getSpecWithFallback<ZimCategoriesSpec>(
+        'zim_categories'
+      )
+    } catch {
+      return null
+    }
+  }
+
   async cancelJob(jobId: string): Promise<{ success: boolean; message: string }> {
     const queue = this.queueService.getQueue(RunDownloadJob.queue)
     const job = await queue.getJob(jobId)

admin/app/utils/download_retry.ts

@@ -0,0 +1,38 @@
+import { dirname, join } from 'node:path'
+
+import type { RunDownloadJobParams } from '../../types/downloads.js'
+import type { ZimCategoriesSpec } from '../../types/collections.js'
+
+/**
+ * Params for retrying a failed download. A job's data is frozen when the job
+ * is created, so a curated ZIM whose catalog entry has since been corrected
+ * (a dead URL replaced, a new build date) is retried with what the catalog
+ * says now. The filename changes with the URL because ZIM filenames carry the
+ * build date.
+ *
+ * Anything else keeps its stored params: a download that came from no
+ * collection, a non-ZIM download, and a resource the catalog no longer lists
+ * (removed or renamed), so a retry is never worse than re-sending the original.
+ */
+export function refreshRetryParams(
+  params: RunDownloadJobParams,
+  spec: ZimCategoriesSpec | null
+): RunDownloadJobParams {
+  const metadata = params.resourceMetadata
+  if (!spec || params.filetype !== 'zim' || !metadata?.collection_ref) return params
+
+  const current = spec.categories
+    .flatMap((category) => category.tiers)
+    .flatMap((tier) => tier.resources)
+    .find((resource) => resource.id === metadata.resource_id && resource.type !== 'dataset')
+  const filename = current?.url.split('/').pop()
+  if (!current || !filename) return params
+
+  return {
+    ...params,
+    url: current.url,
+    filepath: join(dirname(params.filepath), filename),
+    totalBytes: current.size_mb > 0 ? current.size_mb * 1024 * 1024 : params.totalBytes,
+    resourceMetadata: { ...metadata, version: current.version },
+  }
+}

admin/tests/unit/download_retry.spec.ts

@@ -0,0 +1,115 @@
+import * as assert from 'node:assert/strict'
+import { test } from 'node:test'
+
+import { refreshRetryParams } from '../../app/utils/download_retry.js'
+import type { RunDownloadJobParams } from '../../types/downloads.js'
+import type { ZimCategoriesSpec } from '../../types/collections.js'
+
+const failed: RunDownloadJobParams = {
+  url: 'https://download.kiwix.org/zim/other/prepper_en_2025-01.zim',
+  filepath: '/app/storage/zim/prepper_en_2025-01.zim',
+  timeout: 30000,
+  allowedMimeTypes: ['application/x-zim'],
+  filetype: 'zim',
+  title: 'Prepper',
+  totalBytes: 100 * 1024 * 1024,
+  resourceMetadata: {
+    resource_id: 'prepper_en',
+    version: '2025-01',
+    collection_ref: 'survival',
+  },
+}
+
+function specWith(resource: { id: string; version: string; url: string; size_mb: number }) {
+  return {
+    spec_version: '1',
+    categories: [
+      {
+        name: 'Survival',
+        slug: 'survival',
+        icon: 'tent',
+        description: '',
+        language: 'en',
+        tiers: [
+          {
+            name: 'Essential',
+            slug: 'essential',
+            description: '',
+            resources: [{ title: 'Prepper', description: '', ...resource }],
+          },
+        ],
+      },
+    ],
+  } satisfies ZimCategoriesSpec
+}
+
+test('retry of a curated download uses the url, version and filename the catalog has now', () => {
+  const retried = refreshRetryParams(
+    failed,
+    specWith({
+      id: 'prepper_en',
+      version: '2026-09',
+      url: 'https://download.kiwix.org/zim/other/prepper_en_2026-09.zim',
+      size_mb: 120,
+    })
+  )
+
+  assert.equal(retried.url, 'https://download.kiwix.org/zim/other/prepper_en_2026-09.zim')
+  assert.equal(retried.filepath, '/app/storage/zim/prepper_en_2026-09.zim')
+  assert.equal(retried.totalBytes, 120 * 1024 * 1024)
+  assert.deepEqual(retried.resourceMetadata, {
+    resource_id: 'prepper_en',
+    version: '2026-09',
+    collection_ref: 'survival',
+  })
+  assert.equal(retried.title, 'Prepper')
+})
+
+test('retry keeps the stored params when the catalog no longer has the resource', () => {
+  const retried = refreshRetryParams(
+    failed,
+    specWith({
+      id: 'prepper_en_renamed',
+      version: '2026-09',
+      url: 'https://download.kiwix.org/zim/other/prepper_en_renamed_2026-09.zim',
+      size_mb: 120,
+    })
+  )
+
+  assert.deepEqual(retried, failed)
+})
+
+test('retry keeps the stored params when no catalog is available', () => {
+  assert.deepEqual(refreshRetryParams(failed, null), failed)
+})
+
+test('retry of a download that came from no collection keeps the stored params', () => {
+  const manual: RunDownloadJobParams = {
+    ...failed,
+    resourceMetadata: { resource_id: 'prepper_en', version: '2025-01', collection_ref: null },
+  }
+  const spec = specWith({
+    id: 'prepper_en',
+    version: '2026-09',
+    url: 'https://download.kiwix.org/zim/other/prepper_en_2026-09.zim',
+    size_mb: 120,
+  })
+
+  assert.deepEqual(refreshRetryParams(manual, spec), manual)
+  assert.deepEqual(refreshRetryParams({ ...failed, resourceMetadata: undefined }, spec), {
+    ...failed,
+    resourceMetadata: undefined,
+  })
+})
+
+test('retry of a map download keeps the stored params', () => {
+  const map: RunDownloadJobParams = { ...failed, filetype: 'map' }
+  const spec = specWith({
+    id: 'prepper_en',
+    version: '2026-09',
+    url: 'https://download.kiwix.org/zim/other/prepper_en_2026-09.zim',
+    size_mb: 120,
+  })
+
+  assert.deepEqual(refreshRetryParams(map, spec), map)
+})