mejorar contenedor de scraper

poder importar warcs especificos

db-datos/migrate.ts

@@ -1,4 +1,5 @@
 import Database from "bun:sqlite";
+import { join } from "node:path";
 import { drizzle } from "drizzle-orm/bun-sqlite";
 import { migrate } from "drizzle-orm/bun-sqlite/migrator";
 import * as schema from "./schema.js";
@@ -8,7 +9,7 @@ export function migrateDb() {
   const sqlite = new Database(DB_PATH);
   const db = drizzle(sqlite, { schema });
 
-  migrate(db, { migrationsFolder: "./drizzle" });
+  migrate(db, { migrationsFolder: join(import.meta.dir, "drizzle") });
   sqlite.run(`
 pragma journal_mode = WAL;
 PRAGMA synchronous = NORMAL;

scraper/Containerfile

@@ -5,22 +5,23 @@ FROM base AS builder
 ENV NODE_ENV=production
 COPY . .
 RUN bun install --frozen-lockfile \
- && bun build scraper/auto.ts --target=bun --outfile=/tmp/auto.build.js \
+ && bun build scraper/cli.ts --target=bun --outfile=/tmp/cli.build.js \
  && rm -rf node_modules/
 
 FROM base
 RUN apk add --no-cache wget zstd cronie tini
-RUN printf "#!/bin/sh\nexec bun /app/built.js" > /etc/periodic/daily/scraper \
+RUN printf "#!/bin/sh\nexec bun /bin/scraper auto" > /etc/periodic/daily/scraper \
  && chmod +x /etc/periodic/daily/scraper
 
-COPY --from=builder /tmp/auto.build.js /app/built.js
-COPY --from=builder /usr/src/app/db-datos/drizzle /app/drizzle
+COPY --from=builder /tmp/cli.build.js /bin/scraper
+COPY --from=builder /usr/src/app/db-datos/drizzle /bin/drizzle
 COPY --from=builder /usr/src/app/data /listas
 WORKDIR /app
 
 VOLUME /db
+ENV NODE_ENV=production
 ENV DB_PATH=/db/db.db
 ENV LISTS_DIR=/listas/
 
 CMD ["tini", "/usr/sbin/crond", "-n"]
-# CMD ["bun", "/app/built.js"]
+# CMD ["bun", "/bin/scraper"]

scraper/auto.ts

@@ -10,28 +10,6 @@ import { S3Client } from "@aws-sdk/client-s3";
 import { Upload } from "@aws-sdk/lib-storage";
 import { BunFile } from "bun";
 
-if (
-  !process.env.S3_ACCESS_KEY_ID ||
-  !process.env.S3_SECRET_ACCESS_KEY ||
-  !process.env.S3_BUCKET_NAME
-)
-  throw new Error("missing s3 creds");
-if (!process.env.TELEGRAM_BOT_TOKEN)
-  console.warn("no hay TELEGRAM_BOT_TOKEN, no voy a loggear por allá");
-if (!process.env.TELEGRAM_BOT_CHAT_ID)
-  console.warn("no hay TELEGRAM_BOT_CHAT_ID, no voy a loggear por allá");
-const { S3_BUCKET_NAME, S3_ACCESS_KEY_ID, S3_SECRET_ACCESS_KEY } = process.env;
-
-// https://www.backblaze.com/docs/cloud-storage-use-the-aws-sdk-for-javascript-v3-with-backblaze-b2
-const s3 = new S3Client({
-  endpoint: "https://s3.us-west-004.backblazeb2.com",
-  region: "us-west-004",
-  credentials: {
-    accessKeyId: S3_ACCESS_KEY_ID,
-    secretAccessKey: S3_SECRET_ACCESS_KEY,
-  },
-});
-
 const supermercados: Supermercado[] = [
   Supermercado.Carrefour,
   Supermercado.Coto,
@@ -44,10 +22,53 @@ const compressionQueue = new PQueue({ concurrency: 1 });
 // hacemos una cola para el scrapeo para no tener varios writers a la BD y no sobrecargar la CPU
 const scrapQueue = new PQueue({ concurrency: 1 });
 
-supermercados.forEach(downloadList);
-// await recompress("sqlite.db.gz", "sqlite.db.zst");
+export async function auto() {
+  const a = new Auto();
+  await Promise.all(supermercados.map((supr) => a.downloadList(supr)));
+}
 
-async function downloadList(supermercado: Supermercado) {
+class Auto {
+  s3Config?: { s3: S3Client; bucketName: string };
+  telegramConfig?: { token: string; chatId: string };
+
+  constructor() {
+    if (
+      !process.env.S3_ACCESS_KEY_ID ||
+      !process.env.S3_SECRET_ACCESS_KEY ||
+      !process.env.S3_BUCKET_NAME
+    ) {
+      if (process.env.NODE_ENV === "development") {
+        console.warn("faltan creds de s3, no voy a subir a s3");
+      } else {
+        throw new Error("faltan creds de s3");
+      }
+    } else {
+      this.s3Config = {
+        // https://www.backblaze.com/docs/cloud-storage-use-the-aws-sdk-for-javascript-v3-with-backblaze-b2
+        s3: new S3Client({
+          endpoint: "https://s3.us-west-004.backblazeb2.com",
+          region: "us-west-004",
+          credentials: {
+            accessKeyId: process.env.S3_ACCESS_KEY_ID,
+            secretAccessKey: process.env.S3_SECRET_ACCESS_KEY,
+          },
+        }),
+        bucketName: process.env.S3_BUCKET_NAME,
+      };
+    }
+
+    if (!process.env.TELEGRAM_BOT_TOKEN)
+      console.warn("no hay TELEGRAM_BOT_TOKEN, no voy a loggear por allá");
+    else if (!process.env.TELEGRAM_BOT_CHAT_ID)
+      console.warn("no hay TELEGRAM_BOT_CHAT_ID, no voy a loggear por allá");
+    else
+      this.telegramConfig = {
+        token: process.env.TELEGRAM_BOT_TOKEN,
+        chatId: process.env.TELEGRAM_BOT_CHAT_ID,
+      };
+  }
+
+  async downloadList(supermercado: Supermercado) {
     const listPath = resolve(
       join(process.env.LISTS_DIR ?? "../data", `${supermercado}.txt`)
     );
@@ -74,36 +95,40 @@ async function downloadList(supermercado: Supermercado) {
     });
     const t0 = performance.now();
     await subproc.exited;
-  inform(`wget para ${zstdWarcName} tardó ${formatMs(performance.now() - t0)}`);
+    this.inform(
+      `wget para ${zstdWarcName} tardó ${formatMs(performance.now() - t0)}`
+    );
 
     const gzippedWarcPath = join(ctxPath, "temp.warc.gz");
     if (!(await fileExists(gzippedWarcPath))) {
-    const err = report(`no encontré el ${gzippedWarcPath}`);
+      const err = this.report(`no encontré el ${gzippedWarcPath}`);
       throw err;
     }
 
-  await compressionQueue.add(() => recompress(gzippedWarcPath, zstdWarcPath));
+    await compressionQueue.add(() =>
+      this.recompress(gzippedWarcPath, zstdWarcPath)
+    );
     if (!(await fileExists(zstdWarcPath))) {
-    const err = report(`no encontré el ${zstdWarcPath}`);
+      const err = this.report(`no encontré el ${zstdWarcPath}`);
       throw err;
     }
 
-  scrapAndInform({ zstdWarcPath, zstdWarcName });
+    this.scrapAndInform({ zstdWarcPath, zstdWarcName });
 
     try {
-    await uploadToBucket({
+      await this.uploadToBucket({
         fileName: zstdWarcName,
         file: Bun.file(zstdWarcPath),
       });
     } catch (error) {
-    inform(`Falló subir ${zstdWarcName} a S3; ${error}`);
+      this.inform(`Falló subir ${zstdWarcName} a S3; ${error}`);
       console.error(error);
     }
 
     // TODO: borrar archivos temporales
   }
 
-async function scrapAndInform({
+  async scrapAndInform({
     zstdWarcPath,
     zstdWarcName,
   }: {
@@ -118,13 +143,13 @@ async function scrapAndInform({
 
     if (res) {
       const { took, progress } = res;
-    inform(
+      this.inform(
         `Procesado ${zstdWarcName} (${progress.done} ok, ${
           progress.errors.length
         } errores) (tardó ${formatMs(took)})`
       );
     } else {
-    inform(`Algo falló en ${zstdWarcName}`);
+      this.inform(`Algo falló en ${zstdWarcName}`);
     }
   }
 
@@ -132,7 +157,7 @@ async function scrapAndInform({
    * toma un archivo gzippeado y lo recomprime con zstd.
    * borra el archivo original.
    */
-function recompress(inputPath: string, outputPath: string) {
+  recompress(inputPath: string, outputPath: string) {
     // XXX: por alguna razón no funciona en Bun 1.0.20
     // const decompressor = Bun.spawn({
     //   cmd: ["gzip", "-dc", inputPath],
@@ -164,7 +189,7 @@ function recompress(inputPath: string, outputPath: string) {
       decompressor.stdout.pipe(compressor.stdin);
       compressor.on("close", (code) => {
         if (code !== 0) {
-        const err = report(`zstd threw error code ${code}`);
+          const err = this.report(`zstd threw error code ${code}`);
           reject(err);
         }
         resolve(void 0);
@@ -172,17 +197,23 @@ function recompress(inputPath: string, outputPath: string) {
     });
   }
 
-async function uploadToBucket({
+  async uploadToBucket({
     fileName,
     file,
   }: {
     fileName: string;
     file: BunFile;
   }) {
+    if (!this.s3Config) {
+      this.inform(
+        `[s3] Se intentó subir ${fileName} pero no tenemos creds de S3`
+      );
+      return;
+    }
     const parallelUploads3 = new Upload({
-    client: s3,
+      client: this.s3Config.s3,
       params: {
-      Bucket: S3_BUCKET_NAME,
+        Bucket: this.s3Config.bucketName,
         Key: fileName,
         Body: file,
       },
@@ -190,17 +221,29 @@ async function uploadToBucket({
     await parallelUploads3.done();
   }
 
-function inform(msg: string) {
-  sendTelegramMsg(msg);
+  inform(msg: string) {
+    this.sendTelegramMsg(msg);
     console.info(msg);
   }
-function report(msg: string) {
-  inform(msg);
+  report(msg: string) {
+    this.inform(msg);
     const error = new Error(msg);
 
     return error;
   }
 
+  async sendTelegramMsg(text: string) {
+    if (!this.telegramConfig) return;
+    const url = new URL(
+      `https://api.telegram.org/bot${this.telegramConfig.token}/sendMessage`
+    );
+    url.searchParams.set("chat_id", this.telegramConfig.chatId);
+    url.searchParams.set("text", text);
+    await fetch(url);
+  }
+}
+// await recompress("sqlite.db.gz", "sqlite.db.zst");
+
 // no se llama exists porque bun tiene un bug en el que usa fs.exists por mas que exista una funcion llamada exists
 async function fileExists(path: string) {
   try {
@@ -211,17 +254,6 @@ async function fileExists(path: string) {
   }
 }
 
-async function sendTelegramMsg(text: string) {
-  if (!process.env.TELEGRAM_BOT_TOKEN || !process.env.TELEGRAM_BOT_CHAT_ID)
-    return;
-  const url = new URL(
-    `https://api.telegram.org/bot${process.env.TELEGRAM_BOT_TOKEN}/sendMessage`
-  );
-  url.searchParams.set("chat_id", process.env.TELEGRAM_BOT_CHAT_ID);
-  url.searchParams.set("text", text);
-  await fetch(url);
-}
-
 function formatMs(ms: number) {
   return formatDuration(intervalToDuration({ start: 0, end: Math.round(ms) }));
 }

scraper/cli.ts

@@ -0,0 +1,19 @@
+import { auto } from "./auto.js";
+import { parseWarc } from "./scrap.js";
+
+if (process.argv[2] === "auto") {
+  await auto();
+} else if (process.argv[2] === "scrap") {
+  const warcPaths = process.argv.slice(3);
+  if (warcPaths.length > 0) {
+    for (const path of warcPaths) {
+      await parseWarc(path);
+    }
+  } else {
+    console.error("Especificá WARCs para scrapear.");
+    process.exit(1);
+  }
+} else {
+  console.error("Especificá una acción (tipo `auto` o `scrap`) para hacer.");
+  process.exit(1);
+}

scraper/scrap.ts

@@ -32,12 +32,6 @@ const getPrevPrecio = db
   .limit(1)
   .prepare();
 
-if (process.argv[1].endsWith("/scrap.ts")) {
-  for (const path of process.argv.slice(2)) {
-    await parseWarc(path);
-  }
-}
-
 export type Precio = typeof schema.precios.$inferInsert;
 export type Precioish = Omit<
   Precio,