mostrar link de donde viene cada portal

downloader/download_json.js

@@ -9,7 +9,7 @@ import { zData } from "common/schema.js";
 setGlobalDispatcher(
   new Agent({
     pipelining: 0,
-  })
+  }),
 );
 
 /** key es host
@@ -43,7 +43,7 @@ const targets = urls.map((url) => {
 });
 for (const target of targets)
   downloadFromData(target).catch((error) =>
-    console.error(`${target.type}+${target.url} FALLÓ CON`, error)
+    console.error(`${target.type}+${target.url} FALLÓ CON`, error),
   );
 
 /**
@@ -68,6 +68,8 @@ async function downloadFromData(target) {
     await open(join(outputPath, "errors.jsonl"), "w")
   ).createWriteStream();
   try {
+    let nFinished = 0;
+    let nErrors = 0;
     /** @type {DownloadJob[]} */
     const jobs = parsed.dataset.flatMap((dataset) =>
       dataset.distribution
@@ -82,11 +84,12 @@ async function downloadFromData(target) {
               return true;
             } catch (error) {
               errorFile.write(
-                JSON.stringify(encodeError({ dataset, dist }, error)) + "\n"
+                JSON.stringify(encodeError({ dataset, dist }, error)) + "\n",
               );
+              nErrors++;
               return false;
             }
-          }
+          },
         )
         .map((dist) => ({
           dataset,
@@ -94,11 +97,9 @@ async function downloadFromData(target) {
           url: patchUrl(new URL(dist.downloadURL)),
           outputPath,
           attempts: 0,
-        }))
+        })),
     );
     const totalJobs = jobs.length;
-    let nFinished = 0;
-    let nErrors = 0;
 
     // por las dudas verificar que no hayan archivos duplicados
     chequearIdsDuplicados(jobs, outputPath);
@@ -126,7 +127,7 @@ async function downloadFromData(target) {
     process.stderr.write(`info[${outputPath}]: 0/${totalJobs} done\n`);
     const interval = setInterval(() => {
       process.stderr.write(
-        `info[${outputPath}]: ${nFinished}/${totalJobs} done\n`
+        `info[${outputPath}]: ${nFinished}/${totalJobs} done\n`,
       );
     }, 30000);
     await Promise.all(promises);
@@ -161,8 +162,8 @@ async function downloadDistWithRetries(job, attempts = 0) {
     // intentar hasta 15 veces con 15 segundos de por medio
     if (
       error instanceof StatusCodeError &&
-      error.code === 403 &&
+      ((error.code === 403 && url.host === "minsegar-my.sharepoint.com") ||
-      url.host === "minsegar-my.sharepoint.com" &&
+        (error.code === 503 && url.host === "cdn.buenosaires.gob.ar")) &&
       attempts < 15
     ) {
       await wait(15000);
@@ -204,12 +205,12 @@ async function downloadDist({ dist, dataset, url, outputPath }) {
   const fileDirPath = join(
     outputPath,
     sanitizeSuffix(dataset.identifier),
-    sanitizeSuffix(dist.identifier)
+    sanitizeSuffix(dist.identifier),
   );
   await mkdir(fileDirPath, { recursive: true });
   const filePath = join(
     fileDirPath,
-    sanitizeSuffix(dist.fileName || dist.identifier)
+    sanitizeSuffix(dist.fileName || dist.identifier),
   );
 
   if (!res.body) throw new Error("no body");
@@ -239,11 +240,11 @@ function sanitizeSuffix(path) {
  */
 function chequearIdsDuplicados(jobs, id) {
   const duplicated = hasDuplicates(
-    jobs.map((j) => `${j.dataset.identifier}/${j.dist.identifier}`)
+    jobs.map((j) => `${j.dataset.identifier}/${j.dist.identifier}`),
   );
   if (duplicated) {
     console.error(
-      `ADVERTENCIA[${id}]: ¡encontré duplicados! es posible que se pisen archivos entre si`
+      `ADVERTENCIA[${id}]: ¡encontré duplicados! es posible que se pisen archivos entre si`,
     );
   }
 }

frontend/src/lib/components/SourceLink.svelte

@@ -5,11 +5,13 @@
 </script>
 
 <a
-  class="flex items-center leading-none text-gray-600 dark:text-gray-300 gap-1 pt-2"
+  class="flex items-center gap-1 leading-none text-gray-600 dark:text-gray-300"
   {href}
   target="_blank"
   rel="noopener"
 >
   <ExternalLink fill="currentColor" class="h-4" />
-  <span class="underline">Fuente</span>
+  <span class="underline">
+    <slot>Fuente</slot>
+  </span>
 </a>

frontend/src/lib/routes/Dataset.svelte

@@ -81,7 +81,7 @@
                   <small>{dist.fileName}</small>
                 {/if}
               </div>
-              <div class="flex flex-col items-center">
+              <div class="flex flex-col items-center gap-2">
                 {#if !error}
                   <button
                     type="button"

frontend/src/lib/routes/Dump.svelte

@@ -41,8 +41,9 @@
           })}
           <li>
             <div class="flex justify-between gap-3 px-6 py-5">
-              <div class="flex flex-col">
+              <div class="flex flex-col gap-2">
-                <h3 class="text-lg">{site.title}</h3>
+                <h3 class="text-lg leading-none">{site.title}</h3>
+                <SourceLink href={site.url}>{site.url}</SourceLink>
                 <p class="text-sm">{site.description}</p>
               </div>
               <div class="flex shrink-0 flex-col items-center justify-center">
@@ -51,14 +52,16 @@
                   class="focus:shadow-outline inline-flex items-center justify-center rounded-md bg-blue-600 px-4 py-2 text-sm font-medium tracking-wide text-white transition-colors duration-200 hover:bg-blue-700 focus:outline-none focus:ring-2 focus:ring-blue-700 focus:ring-offset-2"
                   >Ver portal</a
                 >
-                <SourceLink href={site.url} />
               </div>
             </div>
           </li>
         {/each}
       </ul>
     {:catch error}
-      Hubo un error intenando cargar este archivo. <pre>{error}</pre>
+      <div class="p-6">
+        <p>Hubo un error intenando cargar este archivo.</p>
+        <p class="text-red-700 dark:text-red-500">{error}</p>
+      </div>
     {/await}
   </Container>
 </main>

frontend/src/lib/routes/Home.svelte

@@ -1,6 +1,6 @@
 <script lang="ts">
   import Container from "../components/Container.svelte";
-  import Portal from "./Home/Portal.svelte";
+  import Dump from "./Home/Dump.svelte";
 </script>
 
 <main class="prose mx-auto dark:prose-invert">
@@ -27,7 +27,22 @@
       <ul
         class="divide-y divide-gray-100 border-y border-y-gray-100 dark:divide-gray-700 dark:border-y-gray-700"
       >
-        <Portal />
+        <!-- nPortales: find . -maxdepth 1 -mindepth 1 -type d | wc -l -->
+        <!-- nDatasets: jq '.dataset | length' */data.json | awk '{s+=$1} END {print s}' -->
+        <!-- size: du -sh -->
+        <Dump
+          dumpUrl="https://archivos.nulo.ar/portales-de-datos/dump-2023-12-09/"
+          nPortales={50}
+          nDatasets={4098}
+          size="147 GB"
+          glow>9 de diciembre de 2023</Dump
+        >
+        <Dump
+          dumpUrl="https://archivos.nulo.ar/dump-2023-12-08/"
+          nPortales={43}
+          nDatasets={3277}
+          size="100 GB">8 de diciembre de 2023</Dump
+        >
       </ul>
     </div>
 

frontend/src/lib/routes/Home/Dump.svelte

@@ -0,0 +1,34 @@
+<script lang="ts">
+  import ArrowForward from "eva-icons/outline/svg/arrow-forward-outline.svg?component";
+  import { inject } from "regexparam";
+  import { routes } from "../../router";
+  import DumpBadge from "./DumpBadge.svelte";
+
+  export let dumpUrl: string;
+  export let nPortales: number;
+  export let nDatasets: number;
+  export let size: string;
+  export let glow: boolean = false;
+</script>
+
+<li>
+  <a
+    class="flex justify-between px-6 py-5 shadow-blue-300 hover:bg-gray-50 dark:hover:bg-gray-700"
+    class:shadow-glow={glow}
+    href={inject(routes.Dump, {
+      dumpUrl: encodeURIComponent(dumpUrl),
+    })}
+  >
+    <div>
+      <h3 class="text-lg"><slot /></h3>
+      <DumpBadge>{nPortales} portales</DumpBadge>
+      <DumpBadge>{nDatasets} datasets</DumpBadge>
+      <DumpBadge>{size}</DumpBadge>
+    </div>
+    <ArrowForward
+      fill="currentColor"
+      aria-hidden="true"
+      class="w-6 shrink-0 text-gray-600 dark:text-gray-400"
+    />
+  </a>
+</li>

frontend/src/lib/routes/Home/DumpBadge.svelte

@@ -0,0 +1,5 @@
+<span
+  class="relative ml-1 inline-flex items-center rounded-full border border-current px-2 py-1 text-xs font-semibold text-blue-800 dark:text-blue-400"
+>
+  <span><slot /></span>
+</span>

frontend/src/lib/routes/Home/Portal.svelte

@@ -1,33 +0,0 @@
-<script lang="ts">
-  import ArrowForward from "eva-icons/outline/svg/arrow-forward-outline.svg?component";
-  import { inject } from "regexparam";
-  import { routes } from "../../router";
-</script>
-
-<li>
-  <a
-    class="shadow-glow flex justify-between px-6 py-5 shadow-blue-300 hover:bg-gray-50 dark:hover:bg-gray-700"
-    href={inject(routes.Dump, {
-      dumpUrl: encodeURIComponent("https://archivos.nulo.ar/dump-2023-12-08/"),
-    })}
-  >
-    <div>
-      <h3 class="text-lg">8 de diciembre de 2023</h3>
-      <!-- <span
-        class="relative ml-1 inline-flex items-center rounded-full border border-current px-2 py-1 text-xs font-semibold text-blue-800 dark:text-blue-400"
-      >
-        <span>portales</span>
-      </span>
-      <span
-        class="relative ml-1 inline-flex items-center rounded-full border border-current px-2 py-1 text-xs font-semibold text-blue-800 dark:text-blue-400"
-      >
-        <span>datasets</span>
-      </span> -->
-    </div>
-    <ArrowForward
-      fill="currentColor"
-      aria-hidden="true"
-      class="w-6 shrink-0 text-gray-600 dark:text-gray-400  "
-    />
-  </a>
-</li>

frontend/src/lib/routes/Portal.svelte

@@ -51,7 +51,7 @@
       <p class="p-6">Cargando..</p>
     {:then { data, errors }}
       <header
-        class="border-b border-b-gray-200 px-6 py-5 leading-none dark:border-b-gray-700"
+        class="flex flex-col gap-1 border-b border-b-gray-200 px-6 py-5 leading-none dark:border-b-gray-700"
       >
         <small>
           Viendo portal archivado de