borrar nthreads explicito

pipeline o abrir archivo con open() o algo genera memory leaks...

.gitignore

@@ -3,3 +3,4 @@ dataJsons/
 log
 prueba
 datos.gob.ar*
+data/

Containerfile

@@ -0,0 +1,19 @@
+
+FROM docker.io/alpine:3.18 as build
+RUN apk add --no-cache npm esbuild
+RUN npm install -g esbuild
+WORKDIR /tmp/build
+
+COPY package.json .
+RUN npm install
+
+COPY download_json.js .
+RUN esbuild --bundle --format=cjs --platform=node --outfile=build.js download_json.js
+
+FROM docker.io/alpine:3.18
+RUN apk add --no-cache nodejs-current tini
+COPY pki/ca_intermediate_root_bundle.pem /usr/lib/ca_intermediate_root_bundle.pem
+COPY --from=build /tmp/build/build.js /usr/local/bin/download_json.js
+ENV NODE_EXTRA_CA_CERTS=/usr/lib/ca_intermediate_root_bundle.pem
+WORKDIR /data
+CMD ["/sbin/tini", "node", "/usr/local/bin/download_json.js", "https://datos.gob.ar/data.json", "http://datos.energia.gob.ar/data.json", "https://datos.magyp.gob.ar/data.json", "https://datos.acumar.gov.ar/data.json", "https://datasets.datos.mincyt.gob.ar/data.json", "https://datos.arsat.com.ar/data.json", "https://datos.cultura.gob.ar/data.json", "https://datos.mininterior.gob.ar/data.json", "https://datos.produccion.gob.ar/data.json", "https://datos.salud.gob.ar/data.json", "https://datos.transporte.gob.ar/data.json", "https://ckan.ciudaddemendoza.gov.ar/data.json", "https://datos.santafe.gob.ar/data.json", "https://datosabiertos.chaco.gob.ar/data.json", "https://datosabiertos.gualeguaychu.gov.ar/data.json", "https://datosabiertos.mercedes.gob.ar/data.json", "http://luj-bue-datos.paisdigital.innovacion.gob.ar/data.json", "https://datosabiertos.desarrollosocial.gob.ar", "http://datos.mindef.gov.ar/data.json"]

download_json.js

@@ -2,7 +2,7 @@
 import { mkdir, open, writeFile } from "node:fs/promises";
 import { Agent, fetch, request, setGlobalDispatcher } from "undici";
 import { join, normalize } from "node:path";
-import { pipeline } from "node:stream/promises";
+import pLimit from "p-limit";
 
 // FYI: al menos los siguientes dominios no tienen la cadena completa de certificados en HTTPS. tenemos que usar un hack (node_extra_ca_certs_mozilla_bundle) para conectarnos a estos sitios. (se puede ver con ssllabs.com) ojalá lxs administradorxs de estos servidores lo arreglen.
 // www.enargas.gov.ar, transparencia.enargas.gov.ar, www.energia.gob.ar, www.economia.gob.ar, datos.yvera.gob.ar
@@ -13,6 +13,11 @@ setGlobalDispatcher(
   })
 );
 
+/** key es host
+ * @type {Map<string, import("p-limit").LimitFunction>} */
+const limiters = new Map();
+const nThreads = process.env.N_THREADS ? parseInt(process.env.N_THREADS) : 8;
+
 class StatusCodeError extends Error {
   /**
    * @param {number} code
@@ -23,84 +28,102 @@ class StatusCodeError extends Error {
   }
 }
 class TooManyRedirectsError extends Error {}
-
-let jsonUrlString = process.argv[2];
-if (!jsonUrlString) {
+const jsonUrls = process.argv.slice(2);
+if (jsonUrls.length < 1) {
   console.error("Especificamente el url al json porfa");
   process.exit(1);
 }
-const jsonUrl = new URL(jsonUrlString);
-const outputPath = jsonUrl.host;
-await mkdir(outputPath, { recursive: true });
-const errorFile = await open(join(outputPath, "errors.jsonl"), "w");
+for (const url of jsonUrls)
+  downloadFromData(url).catch((error) =>
+    console.error(`${url} FALLÓ CON`, error)
+  );
 
-const jsonRes = await fetch(jsonUrl);
-// prettier-ignore
-const parsed = /** @type {{ dataset: Dataset[] }} */(await jsonRes.json())
-await writeFile(join(outputPath, "data.json"), JSON.stringify(parsed));
+/**
+ * @param {string} jsonUrlString
+ */
+async function downloadFromData(jsonUrlString) {
+  const jsonUrl = new URL(jsonUrlString);
+  const outputPath = jsonUrl.host;
+  await mkdir(outputPath, { recursive: true });
+  const errorFile = (
+    await open(join(outputPath, "errors.jsonl"), "w")
+  ).createWriteStream();
 
-const jobs = parsed.dataset.flatMap((dataset) =>
-  dataset.distribution.map((dist) => ({
+  try {
+    const jsonRes = await fetch(jsonUrl);
+    // prettier-ignore
+    const parsed = /** @type {{ dataset: Dataset[] }} */(await jsonRes.json())
+    await writeFile(join(outputPath, "data.json"), JSON.stringify(parsed));
+
+    /** @type {DownloadJob[]} */
+    const jobs = parsed.dataset.flatMap((dataset) =>
+      dataset.distribution
+        .filter((dist) => {
+          try {
+            patchUrl(new URL(dist.downloadURL));
+            return true;
+          } catch (error) {
+            errorFile.write(
+              JSON.stringify(encodeError({ dataset, dist }, error)) + "\n"
+            );
+            return false;
+          }
+        })
+        .map((dist) => ({
           dataset,
           dist,
           url: patchUrl(new URL(dist.downloadURL)),
+          outputPath,
+          attempts: 0,
         }))
-);
-const totalJobs = jobs.length;
-let nFinished = 0;
-let nErrors = 0;
+    );
+    const totalJobs = jobs.length;
+    let nFinished = 0;
+    let nErrors = 0;
 
-// por las dudas verificar que no hayan archivos duplicados
-chequearIdsDuplicados();
+    // por las dudas verificar que no hayan archivos duplicados
+    chequearIdsDuplicados(jobs);
 
-/** @type {Map< string, DownloadJob[] >} */
-let jobsPerHost = new Map();
-for (const job of jobs) {
-  jobsPerHost.set(job.url.host, [
-    ...(jobsPerHost.get(job.url.host) || []),
-    job,
-  ]);
-}
+    shuffleArray(jobs);
 
-const greens = [...jobsPerHost.entries()].flatMap(([host, jobs]) => {
-  const nThreads = 8;
-  return Array(nThreads)
-    .fill(0)
-    .map(() =>
-      (async () => {
-        let job;
-        while ((job = jobs.pop())) {
+    const promises = jobs.map((job) => {
+      let limit = limiters.get(job.url.host);
+      if (!limit) {
+        limit = pLimit(nThreads);
+        limiters.set(job.url.host, limit);
+      }
+      return limit(async () => {
         try {
           await downloadDistWithRetries(job);
         } catch (error) {
-            await errorFile.write(
-              JSON.stringify({
-                url: job.url.toString(),
-                ...encodeError(error),
-              }) + "\n"
-            );
+          await errorFile.write(JSON.stringify(job, encodeError(error)) + "\n");
           nErrors++;
         } finally {
           nFinished++;
         }
-        }
-      })()
-    );
-});
-process.stderr.write(`greens: ${greens.length}\n`);
+      });
+    });
 
-const interval = setInterval(() => {
-  process.stderr.write(`info: ${nFinished}/${totalJobs} done\n`);
-}, 30000);
-await Promise.all(greens);
-clearInterval(interval);
-if (nErrors > 0) console.error(`Finished with ${nErrors} errors`);
+    process.stderr.write(`info[${jsonUrl.host}]: 0/${totalJobs} done\n`);
+    const interval = setInterval(() => {
+      process.stderr.write(
+        `info[${jsonUrl.host}]: ${nFinished}/${totalJobs} done\n`
+      );
+    }, 30000);
+    await Promise.all(promises);
+    clearInterval(interval);
+    if (nErrors > 0)
+      console.error(`${jsonUrl.host}: Finished with ${nErrors} errors`);
+  } finally {
+    errorFile.close();
+  }
+}
 
 /**
  * @argument {DownloadJob} job
- * @argument {number} tries
+ * @argument {number} attempts
  */
-async function downloadDistWithRetries(job, tries = 0) {
+async function downloadDistWithRetries(job, attempts = 0) {
   const { url } = job;
   try {
     await downloadDist(job);
@@ -111,19 +134,19 @@ async function downloadDistWithRetries(job, tries = 0) {
       error instanceof StatusCodeError &&
       error.code === 403 &&
       url.host === "minsegar-my.sharepoint.com" &&
-      tries < 15
+      attempts < 15
     ) {
       await wait(15000);
-      return await downloadDistWithRetries(job, tries + 1);
+      return await downloadDistWithRetries(job, attempts + 1);
     }
     // si no fue un error de http, reintentar hasta 5 veces con 5 segundos de por medio
     else if (
       !(error instanceof StatusCodeError) &&
       !(error instanceof TooManyRedirectsError) &&
-      tries < 5
+      attempts < 10
     ) {
       await wait(5000);
-      return await downloadDistWithRetries(job, tries + 1);
+      return await downloadDistWithRetries(job, attempts + 1);
     } else throw error;
   }
 }
@@ -131,7 +154,7 @@ async function downloadDistWithRetries(job, tries = 0) {
 /**
  * @argument {DownloadJob} job
  */
-async function downloadDist({ dist, dataset, url }) {
+async function downloadDist({ dist, dataset, url, outputPath }) {
   // sharepoint no le gusta compartir a bots lol
   const spoofUserAgent = url.host.endsWith("sharepoint.com");
 
@@ -159,16 +182,18 @@ async function downloadDist({ dist, dataset, url }) {
     fileDirPath,
     sanitizeSuffix(dist.fileName || dist.identifier)
   );
-  const outputFile = await open(filePath, "w");
 
   if (!res.body) throw new Error("no body");
-  await pipeline(res.body, outputFile.createWriteStream());
+  await writeFile(filePath, res.body);
 }
 
 /** @typedef DownloadJob
  * @prop {Dataset} dataset
  * @prop {Distribution} dist
  * @prop {URL} url
+ * @prop {string} outputPath
+ * @prop {number} attempts
+ * @prop {Date=} waitUntil
  */
 /** @typedef Dataset
  * @prop {string} identifier
@@ -188,7 +213,10 @@ function sanitizeSuffix(path) {
   return normalize(path).replace(/^(\.\.(\/|\\|$))+/, "");
 }
 
-function chequearIdsDuplicados() {
+/**
+ * @param {DownloadJob[]} jobs
+ */
+function chequearIdsDuplicados(jobs) {
   const duplicated = hasDuplicates(
     jobs.map((j) => `${j.dataset.identifier}/${j.dist.identifier}`)
   );
@@ -206,17 +234,29 @@ function hasDuplicates(array) {
 
 /** @argument {number} ms */
 function wait(ms) {
-  if (ms < 0) return Promise.resolve();
   return new Promise((resolve) => setTimeout(resolve, ms));
 }
 
-function encodeError(error) {
+/**
+ * @param {{ dataset: Dataset, dist: Distribution, url?: URL }} job
+ * @param {any} error
+ */
+function encodeError(job, error) {
+  const always = {
+    url: job.url?.toString || job.dist.downloadURL,
+    datasetIdentifier: job.dataset.identifier,
+    distributionIdentifier: job.dist.identifier,
+  };
   if (error instanceof StatusCodeError)
-    return { kind: "http_error", status_code: error.code };
+    return { ...always, kind: "http_error", status_code: error.code };
   else if (error instanceof TooManyRedirectsError)
-    return { kind: "infinite_redirect" };
+    return { ...always, kind: "infinite_redirect" };
   else {
-    return { kind: "generic_error", error: error.code || error.message };
+    return {
+      ...always,
+      kind: "generic_error",
+      error: error.code || error.message,
+    };
   }
 }
 
@@ -231,3 +271,12 @@ function patchUrl(url) {
   }
   return url;
 }
+
+// https://stackoverflow.com/a/12646864
+/** @param {any[]} array */
+function shuffleArray(array) {
+  for (let i = array.length - 1; i > 0; i--) {
+    const j = Math.floor(Math.random() * (i + 1));
+    [array[i], array[j]] = [array[j], array[i]];
+  }
+}

package.json

@@ -5,13 +5,13 @@
   "description": "",
   "main": "index.js",
   "scripts": {
-    "run": "env NODE_EXTRA_CA_CERTS=node_modules/node_extra_ca_certs_mozilla_bundle/ca_bundle/ca_intermediate_root_bundle.pem node download_json.js"
+    "run": "env NODE_EXTRA_CA_CERTS=pki/ca_intermediate_root_bundle.pem node download_json.js"
   },
   "keywords": [],
   "author": "",
   "license": "ISC",
   "dependencies": {
-    "node_extra_ca_certs_mozilla_bundle": "^1.0.5",
+    "p-limit": "^5.0.0",
     "undici": "^5.28.0"
   },
   "devDependencies": {

pki/readme.md

@@ -0,0 +1 @@
+Generado por [node_extra_ca_certs_mozilla_bundle](https://www.npmjs.com/package/node_extra_ca_certs_mozilla_bundle), lo copiamos acá para que sea más fácil bundlearlo.

pnpm-lock.yaml

@@ -5,9 +5,9 @@ settings:
   excludeLinksFromLockfile: false
 
 dependencies:
-  node_extra_ca_certs_mozilla_bundle:
-    specifier: ^1.0.5
-    version: 1.0.5
+  p-limit:
+    specifier: ^5.0.0
+    version: 5.0.0
   undici:
     specifier: ^5.28.0
     version: 5.28.0
@@ -30,138 +30,11 @@ packages:
       undici-types: 5.26.5
     dev: true
 
-  /asynckit@0.4.0:
-    resolution: {integrity: sha512-Oei9OH4tRh0YqU3GxhX79dM/mwVgvbZJaSNaRk+bshkj0S5cfHcgYakreBjrHwatXKbz+IoIdYLxrKim2MjW0Q==}
-    dev: false
-
-  /axios@0.27.2:
-    resolution: {integrity: sha512-t+yRIyySRTp/wua5xEr+z1q60QmLq8ABsS5O9Me1AsE5dfKqgnCFzwiCZZ/cGNd1lq4/7akDWMxdhVlucjmnOQ==}
+  /p-limit@5.0.0:
+    resolution: {integrity: sha512-/Eaoq+QyLSiXQ4lyYV23f14mZRQcXnxfHrN0vCai+ak9G0pp9iEQukIIZq5NccEvwRB8PUnZT0KsOoDCINS1qQ==}
+    engines: {node: '>=18'}
     dependencies:
-      follow-redirects: 1.15.3
-      form-data: 4.0.0
-    transitivePeerDependencies:
-      - debug
-    dev: false
-
-  /bluebird@3.7.2:
-    resolution: {integrity: sha512-XpNj6GDQzdfW+r2Wnn7xiSAd7TM3jzkxGXBGTtWKuSXv1xUV+azxAm8jdWZN06QTQk+2N2XB9jRDkvbmQmcRtg==}
-    dev: false
-
-  /combined-stream@1.0.8:
-    resolution: {integrity: sha512-FQN4MRfuJeHf7cBbBMJFXhKSDq+2kAArBlmRBvcvFE5BB1HZKXtSFASDhdlz9zOYwxh8lDdnvmMOe/+5cdoEdg==}
-    engines: {node: '>= 0.8'}
-    dependencies:
-      delayed-stream: 1.0.0
-    dev: false
-
-  /cross-env@6.0.3:
-    resolution: {integrity: sha512-+KqxF6LCvfhWvADcDPqo64yVIB31gv/jQulX2NGzKS/g3GEVz6/pt4wjHFtFWsHMddebWD/sDthJemzM4MaAag==}
-    engines: {node: '>=8.0'}
-    hasBin: true
-    dependencies:
-      cross-spawn: 7.0.3
-    dev: false
-
-  /cross-spawn@7.0.3:
-    resolution: {integrity: sha512-iRDPJKUPVEND7dHPO8rkbOnPpyDygcDFtWjpeWNCgy8WP2rXcxXL8TskReQl6OrB2G7+UJrags1q15Fudc7G6w==}
-    engines: {node: '>= 8'}
-    dependencies:
-      path-key: 3.1.1
-      shebang-command: 2.0.0
-      which: 2.0.2
-    dev: false
-
-  /csvtojson@2.0.10:
-    resolution: {integrity: sha512-lUWFxGKyhraKCW8Qghz6Z0f2l/PqB1W3AO0HKJzGIQ5JRSlR651ekJDiGJbBT4sRNNv5ddnSGVEnsxP9XRCVpQ==}
-    engines: {node: '>=4.0.0'}
-    hasBin: true
-    dependencies:
-      bluebird: 3.7.2
-      lodash: 4.17.21
-      strip-bom: 2.0.0
-    dev: false
-
-  /delayed-stream@1.0.0:
-    resolution: {integrity: sha512-ZySD7Nf91aLB0RxL4KGrKHBXl7Eds1DAmEdcoVawXnLD7SDhpNgtuII2aAkg7a7QS41jxPSZ17p4VdGnMHk3MQ==}
-    engines: {node: '>=0.4.0'}
-    dev: false
-
-  /follow-redirects@1.15.3:
-    resolution: {integrity: sha512-1VzOtuEM8pC9SFU1E+8KfTjZyMztRsgEfwQl44z8A25uy13jSzTj6dyK2Df52iV0vgHCfBwLhDWevLn95w5v6Q==}
-    engines: {node: '>=4.0'}
-    peerDependencies:
-      debug: '*'
-    peerDependenciesMeta:
-      debug:
-        optional: true
-    dev: false
-
-  /form-data@4.0.0:
-    resolution: {integrity: sha512-ETEklSGi5t0QMZuiXoA/Q6vcnxcLQP5vdugSpuAyi6SVGi2clPPp+xgEhuMaHC+zGgn31Kd235W35f7Hykkaww==}
-    engines: {node: '>= 6'}
-    dependencies:
-      asynckit: 0.4.0
-      combined-stream: 1.0.8
-      mime-types: 2.1.35
-    dev: false
-
-  /is-utf8@0.2.1:
-    resolution: {integrity: sha512-rMYPYvCzsXywIsldgLaSoPlw5PfoB/ssr7hY4pLfcodrA5M/eArza1a9VmTiNIBNMjOGr1Ow9mTyU2o69U6U9Q==}
-    dev: false
-
-  /isexe@2.0.0:
-    resolution: {integrity: sha512-RHxMLp9lnKHGHRng9QFhRCMbYAcVpn69smSGcq3f36xjgVVWThj4qqLbTLlq7Ssj8B+fIQ1EuCEGI2lKsyQeIw==}
-    dev: false
-
-  /lodash@4.17.21:
-    resolution: {integrity: sha512-v2kDEe57lecTulaDIuNTPy3Ry4gLGJ6Z1O3vE1krgXZNrsQ+LFTGHVxVjcXPs17LhbZVGedAJv8XZ1tvj5FvSg==}
-    dev: false
-
-  /mime-db@1.52.0:
-    resolution: {integrity: sha512-sPU4uV7dYlvtWJxwwxHD0PuihVNiE7TyAbQ5SWxDCB9mUYvOgroQOwYQQOKPJ8CIbE+1ETVlOoK1UC2nU3gYvg==}
-    engines: {node: '>= 0.6'}
-    dev: false
-
-  /mime-types@2.1.35:
-    resolution: {integrity: sha512-ZDY+bPm5zTTF+YpCrAU9nK0UgICYPT0QtT1NZWFv4s++TNkcgVaT0g6+4R2uI4MjQjzysHB1zxuWL50hzaeXiw==}
-    engines: {node: '>= 0.6'}
-    dependencies:
-      mime-db: 1.52.0
-    dev: false
-
-  /node_extra_ca_certs_mozilla_bundle@1.0.5:
-    resolution: {integrity: sha512-Y+wek3qK8WYybCIxArGTmCEJCJ/6uGud/HCJECBZPIgagF9ba90nhnQMxBcMUAwQaR53iphGYp0JzlVPpUBsjg==}
-    requiresBuild: true
-    dependencies:
-      axios: 0.27.2
-      cross-env: 6.0.3
-      csvtojson: 2.0.10
-    transitivePeerDependencies:
-      - debug
-    dev: false
-
-  /path-key@3.1.1:
-    resolution: {integrity: sha512-ojmeN0qd+y0jszEtoY48r0Peq5dwMEkIlCOu6Q5f41lfkswXuKtYrhgoTpLnyIcHm24Uhqx+5Tqm2InSwLhE6Q==}
-    engines: {node: '>=8'}
-    dev: false
-
-  /shebang-command@2.0.0:
-    resolution: {integrity: sha512-kHxr2zZpYtdmrN1qDjrrX/Z1rR1kG8Dx+gkpK1G4eXmvXswmcE1hTWBWYUzlraYw1/yZp6YuDY77YtvbN0dmDA==}
-    engines: {node: '>=8'}
-    dependencies:
-      shebang-regex: 3.0.0
-    dev: false
-
-  /shebang-regex@3.0.0:
-    resolution: {integrity: sha512-7++dFhtcx3353uBaq8DDR4NuxBetBzC7ZQOhmTQInHEd6bSrXdiEyzCvG07Z44UYdLShWUyXt5M/yhz8ekcb1A==}
-    engines: {node: '>=8'}
-    dev: false
-
-  /strip-bom@2.0.0:
-    resolution: {integrity: sha512-kwrX1y7czp1E69n2ajbG65mIo9dqvJ+8aBQXOGVxqwvNbsXdFM6Lq37dLAY3mknUwru8CfcCbfOLL/gMo+fi3g==}
-    engines: {node: '>=0.10.0'}
-    dependencies:
-      is-utf8: 0.2.1
+      yocto-queue: 1.0.0
     dev: false
 
   /undici-types@5.26.5:
@@ -175,10 +48,7 @@ packages:
       '@fastify/busboy': 2.1.0
     dev: false
 
-  /which@2.0.2:
-    resolution: {integrity: sha512-BLI3Tl1TW3Pvl70l3yq3Y64i+awpwXqsGBYWkkqMtnbXgrMD+yj7rhW0kuEDxzJaYXGjEW5ogapKNMEKNMjibA==}
-    engines: {node: '>= 8'}
-    hasBin: true
-    dependencies:
-      isexe: 2.0.0
+  /yocto-queue@1.0.0:
+    resolution: {integrity: sha512-9bnSc/HEW2uRy67wc+T8UwauLuPJVn28jb+GtJY16iiKWyvmYJRXVT4UamsAEGQfPohgr2q4Tq0sQbQlxTfi1g==}
+    engines: {node: '>=12.20'}
     dev: false

readme.md

@@ -13,11 +13,18 @@ pnpm run run download_json.js https://datos.gob.ar/data.json
 # guarda en ./datos.gob.ar
 ```
 
+## contenedor
+
+```
+docker run --rm -it -v ./data:/data gitea.nulo.in/nulo/transicion-desordenada-diablo/downloader
+# descarga datos.gob.ar
+```
+
 ## formato de repo guardado
 
 - `{dominio de repo}`
   - `data.json`
-  - `errors.jsonl`
+  - `errors.jsonl`: archivo con todos los errores que se obtuvieron al intentar descargar todo.
   - `{identifier de dataset}`
     - `{identifier de distribution}`
       - `{fileName (o, si no existe, identifier de distribution)}`