rm coto-crawlee

rust/src/api/main.rs

@@ -266,6 +266,44 @@ async fn search(State(pool): State<SqlitePool>, Path(query): Path<String>) -> im
     Json(results)
 }
 
+#[derive(sqlx::FromRow, Debug, Serialize)]
+struct Metadata {
+    ean: String,
+    fetched_at: chrono::DateTime<Utc>,
+    precio_centavos: Option<i64>,
+    in_stock: Option<bool>,
+    url: String,
+    name: Option<String>,
+    image_url: Option<String>,
+}
+
+async fn dump_latest_metadata(State(pool): State<SqlitePool>) -> impl IntoResponse {
+    let precios = sqlx::query!("
+    SELECT p.id, p.ean, p.name, p.image_url, p.url, p.precio_centavos, p.in_stock, p.fetched_at
+    FROM precios p
+    INNER JOIN (
+        SELECT ean, MAX(fetched_at) as max_fetched_at
+        FROM precios
+        GROUP BY ean
+    ) latest ON p.ean = latest.ean AND p.fetched_at = latest.max_fetched_at
+    WHERE p.name IS NOT NULL")
+    .fetch_all(&pool)
+    .await
+    .unwrap()
+    .into_iter()
+    .map(|r| Metadata {
+        ean: r.ean,
+        fetched_at: DateTime::from_timestamp(r.fetched_at, 0).unwrap(),
+        image_url: r.image_url,
+        name: r.name,
+        in_stock: r.in_stock.map(|x| x == 1),
+        precio_centavos: r.precio_centavos,
+        url: r.url,
+    })
+    .collect_vec();
+    Json(precios)
+}
+
 async fn get_info(State(pool): State<SqlitePool>) -> impl IntoResponse {
     #[derive(Serialize)]
     struct Info {
@@ -321,6 +359,7 @@ async fn main() {
         .route("/api/0/ean/:ean/history", get(get_product_history))
         .route("/api/0/info", get(get_info))
         .route("/api/0/search/:query", get(search))
+        .route("/api/0/internal/latest-metadata", get(dump_latest_metadata))
         .with_state(pool);
 
     let listener = tokio::net::TcpListener::bind("0.0.0.0:8000").await.unwrap();

sepa/db/schema.ts

@@ -11,6 +11,8 @@ import {
   index,
   pgMaterializedView,
   pgView,
+  timestamp,
+  boolean,
 } from "drizzle-orm/pg-core";
 
 export const datasets = pgTable(
@@ -211,3 +213,22 @@ export const productos_descripcion_index = pgTable(
     ).using("gin", sql`to_tsvector('spanish', ${table.productos_descripcion})`),
   })
 );
+
+// véase scripts/refresh-scrapped-metadata.ts
+export const productos_metadata_scrapped = pgTable(
+  "productos_metadata_scrapped",
+  {
+    ean: bigint("ean", { mode: "bigint" }),
+    fetchedAt: timestamp("fetched_at").notNull(),
+    precioCentavos: integer("precio_centavos"),
+    inStock: boolean("in_stock"),
+    url: text("url").notNull(),
+    name: text("name"),
+    imageUrl: text("image_url"),
+  },
+  (table) => ({
+    productos_metadata_scrapped_ean_idx: index(
+      "productos_metadata_scrapped_ean_idx"
+    ).on(table.ean),
+  })
+);

sepa/package.json

@@ -3,7 +3,8 @@
   "private": true,
   "type": "module",
   "devDependencies": {
-    "@types/bun": "^1.1.7",
+    "@types/bun": "^1.1.11",
+    "bun-types": "^1.1.30",
     "@types/papaparse": "^5.3.14",
     "drizzle-kit": "^0.24.2"
   },

sepa/scripts/refresh-scrapped-metadata.ts

@@ -0,0 +1,43 @@
+/**
+ * este script actualiza la base de datos "nueva" a partir de una base de datos
+ * generada por el scraper "viejo" de preciazo, que scrapea los sitios de los supermercados.
+ *
+ * solo guarda los últimos metadatos de cada producto.
+ *
+ * se le pasa la base de datos SQLite del scraper como parametro.
+ */
+
+import { drizzle } from "drizzle-orm/postgres-js";
+import postgres from "postgres";
+import * as schema from "../db/schema";
+import { Database } from "bun:sqlite";
+
+if (!process.argv[2]) {
+  console.error("falta pasar la base de datos del scraper como parametro");
+  process.exit(1);
+}
+
+const db = drizzle(postgres(), {
+  schema,
+  logger: true,
+});
+using scraperDb = new Database(process.argv[2], {
+  strict: true,
+  readonly: true,
+});
+
+const precios = scraperDb.query(`
+  SELECT p.id, p.ean, p.name, p.image_url, p.url, p.precio_centavos, p.in_stock, p.fetched_at
+  FROM precios p
+  INNER JOIN (
+      SELECT ean, MAX(fetched_at) as max_fetched_at
+      FROM precios
+      GROUP BY ean
+  ) latest ON p.ean = latest.ean AND p.fetched_at = latest.max_fetched_at
+  WHERE p.name IS NOT NULL
+`);
+
+// @ts-expect-error bun 1.1.30 has outdated types, it's fixed in main branch
+for (const row of precios.iterate()) {
+  console.log(row);
+}

sepa/sitio2/package.json

@@ -51,12 +51,14 @@
 		"d3-scale": "^4.0.2",
 		"date-fns": "^4.1.0",
 		"drizzle-orm": "^0.33.0",
+		"geojson": "^0.5.0",
 		"layerchart": "^0.44.0",
 		"leaflet": "^1.9.4",
 		"leaflet.markercluster": "^1.5.3",
 		"lucide-svelte": "^0.441.0",
 		"maplibre-gl": "^4.7.1",
 		"postgres": "^3.4.4",
+		"svelte-maplibre": "^0.9.14",
 		"tailwind-merge": "^2.5.2",
 		"tailwind-variants": "^0.2.1"
 	}

sepa/sitio2/src/lib/components/ui/button/index.ts

@@ -16,8 +16,10 @@ const buttonVariants = tv({
 		size: {
 			default: 'h-10 px-4 py-2',
 			sm: 'h-9 rounded-md px-3',
+			xs: 'h-8 rounded-md px-2',
 			lg: 'h-11 rounded-md px-8',
-			icon: 'h-10 w-10'
+			icon: 'h-10 w-10',
+			icon_sm: 'h-8 w-8'
 		}
 	},
 	defaultVariants: {

sepa/sitio2/src/lib/sepa-utils.ts

@@ -29,6 +29,13 @@ export const pesosFormatter = new Intl.NumberFormat('es-AR', {
 	currency: 'ARS'
 });
 
+export const dateFormatter = Intl.DateTimeFormat('es-AR', {
+	year: 'numeric',
+	month: '2-digit',
+	day: '2-digit',
+	weekday: 'long'
+});
+
 export function parseMarcas(marcas: readonly string[]) {
 	const x = marcas
 		.map((m) => m.trim().replaceAll(/['`´]/g, ''))

sepa/sitio2/src/routes/id_producto/[id]/+page.svelte

@@ -1,15 +1,77 @@
 <script lang="ts">
 	import type { PageData } from './$types';
-	import { ArrowLeft } from 'lucide-svelte';
+	import { ArrowLeft, ArrowRight, MapPin } from 'lucide-svelte';
-	import Map from '$lib/components/Map.svelte';
 	import Badge from '$lib/components/ui/badge/badge.svelte';
 	import { goto } from '$app/navigation';
-	import { generateGoogleMapsLink, pesosFormatter, processBanderaNombre } from '$lib/sepa-utils';
+	import {
+		dateFormatter,
+		generateGoogleMapsLink,
+		pesosFormatter,
+		processBanderaNombre
+	} from '$lib/sepa-utils';
 	import { page } from '$app/stores';
+	import {
+		DefaultMarker,
+		MapLibre,
+		Popup,
+		GeoJSON,
+		CircleLayer,
+		SymbolLayer,
+		HeatmapLayer
+	} from 'svelte-maplibre';
+	import style from '$lib/components/map_style.json';
+	import type { GeoJSON as GeoJSONType } from 'geojson';
+	import type { DataDrivenPropertyValueSpecification } from 'maplibre-gl';
+	import Button from '$lib/components/ui/button/button.svelte';
 
 	export let data: PageData;
 
+	$: id_producto = $page.params.id;
 	const query = $page.url.searchParams.get('query');
+
+	function generateGeoJSON(precios: (typeof data)['precios']): GeoJSONType {
+		const prices = data.precios.map((p) => p.productos_precio_lista);
+		const sortedPrices = prices.sort((a, b) => a - b);
+		const q1Index = Math.floor(sortedPrices.length * 0.1);
+		const q3Index = Math.floor(sortedPrices.length * 0.9);
+		const iqr = sortedPrices[q3Index] - sortedPrices[q1Index];
+		const lowerBound = sortedPrices[q1Index] - 1.5 * iqr;
+		const upperBound = sortedPrices[q3Index] + 1.5 * iqr;
+		const filteredPrices = sortedPrices.filter((p) => p >= lowerBound && p <= upperBound);
+		const min = Math.min(...filteredPrices);
+		const max = Math.max(...filteredPrices);
+
+		return {
+			type: 'FeatureCollection',
+			features: data.precios.map((precio) => ({
+				type: 'Feature',
+				geometry: {
+					type: 'Point',
+					coordinates: [precio.sucursales_longitud, precio.sucursales_latitud]
+				},
+				properties: {
+					id: Math.random(),
+					id_comercio: precio.id_comercio,
+					id_sucursal: precio.id_sucursal,
+					precio: precio.productos_precio_lista,
+					nombre: precio.sucursales_nombre,
+					descripcion: precio.productos_descripcion,
+					direccion: `${precio.sucursales_calle} ${precio.sucursales_numero ?? ''}`,
+					comercio: processBanderaNombre(precio),
+					fecha: precio.dataset_date
+				}
+			}))
+		};
+	}
+
+	$: geoJSON = generateGeoJSON(data.precios);
+
+	function hoverStateFilter(
+		offValue: number,
+		onValue: number
+	): DataDrivenPropertyValueSpecification<number> {
+		return ['case', ['boolean', ['feature-state', 'hover'], false], onValue, offValue];
+	}
 </script>
 
 <svelte:head>
@@ -34,84 +96,202 @@
 			<Badge variant="outline">EAN {data.id_producto}</Badge>
 		</div>
 	</div>
-
+	<MapLibre
-	<Map
+		style={style as any}
-		mapMap={(map, L) => {
+		class="relative h-full max-h-full min-h-[50vh] w-full flex-1"
-			// var markers = L.MarkerClusterGroup();
+		standardControls
-			const myRenderer = L.canvas({ padding: 0.5 });
+		zoom={9}
-			const prices = data.precios.map((p) => p.productos_precio_lista);
+		center={[-58.381944444444, -34.599722222222]}
-			const sortedPrices = prices.sort((a, b) => a - b);
+	>
-			const q1Index = Math.floor(sortedPrices.length * 0.1);
+		<!-- cluster={{
-			const q3Index = Math.floor(sortedPrices.length * 0.9);
+				radius: 50,
-			const iqr = sortedPrices[q3Index] - sortedPrices[q1Index];
+				// maxZoom: 14,
-			const lowerBound = sortedPrices[q1Index] - 1.5 * iqr;
+				maxZoom: 14,
-			const upperBound = sortedPrices[q3Index] + 1.5 * iqr;
+				properties: {
-			const filteredPrices = sortedPrices.filter((p) => p >= lowerBound && p <= upperBound);
+					total_precio: ['+', ['get', 'precio']],
-			const min = Math.min(...filteredPrices);
+					precio_promedio: [
-			const max = Math.max(...filteredPrices);
+						'number',
-			console.log({ min, max, outliers: prices.length - filteredPrices.length });
+						['/', ['+', ['number', ['get', 'precio']]], ['get', 'point_count']]
-
+					]
-			// For each row in data, create a marker and add it to the map
-			// For each row, columns `Latitude`, `Longitude`, and `Title` are required
-			for (const precio of data.precios) {
-				const normalizedPrice = (precio.productos_precio_lista - min) / (max - min);
-				// Safari doesn't support color-mix, so we'll use a fallback
-				const color = getSafeColor(normalizedPrice);
-
-				const createElement = () => {
-					const div = document.createElement('div');
-
-					[
-						`fecha del precio: ${precio.dataset_date}`,
-						`precio: ${pesosFormatter.format(precio.productos_precio_lista)}`,
-						`comercio: ${processBanderaNombre(precio)} (${precio.comercio_razon_social} CUIT ${precio.comercio_cuit})`,
-						`sucursal: ${precio.sucursales_nombre}`,
-						`dirección: ${precio.sucursales_calle} ${precio.sucursales_numero}`,
-						() => {
-							const a = document.createElement('a');
-							if (precio.sucursales_calle) {
-								a.href = generateGoogleMapsLink({
-									sucursales_calle: precio.sucursales_calle,
-									sucursales_numero: precio.sucursales_numero
-								});
-							}
-							a.target = '_blank';
-							a.append('ver en Google Maps');
-							return a;
-						},
-						`descripcion del producto segun el comercio: ${precio.productos_descripcion}`,
-						() => {
-							const a = document.createElement('a');
-							a.href = `/id_producto/${data.id_producto}/sucursal/${precio.id_comercio}/${precio.id_sucursal}`;
-							a.append('ver precios historicos');
-							return a;
-						}
-					].forEach((el) => {
-						div.append(typeof el === 'function' ? el() : el);
-						div.append(document.createElement('br'));
-					});
-					return div;
-				};
-
-				var marker = L.circleMarker([precio.sucursales_latitud, precio.sucursales_longitud], {
-					opacity: 1,
-					renderer: myRenderer,
-					color,
-					radius: 5
-				})
-					.bindPopup(createElement)
-					.addTo(map);
-				marker.on('click', function (this: L.CircleMarker) {
-					this.openPopup();
-				});
-			}
-
-			// Helper function to get a color that works in Safari
-			function getSafeColor(normalizedPrice: number) {
-				const r = Math.round(255 * normalizedPrice);
-				const g = Math.round(255 * (1 - normalizedPrice));
-				return `rgb(${r}, ${g}, 0)`;
 				}
+			}} -->
+		<GeoJSON id="precios" data={geoJSON}>
+			<!-- <HeatmapLayer
+				paint={{
+					// Increase the heatmap weight based on price magnitude
+					'heatmap-weight': [
+						'interpolate',
+						['linear'],
+						['get', 'precio'], // Get precio from properties
+						Math.min(...data.precios.map((p) => p.productos_precio_lista)), // Start at 0 weight for minimum price
+						0,
+						Math.max(...data.precios.map((p) => p.productos_precio_lista)), // Adjust this max value based on your price range
+						1
+					],
+					// Increase the heatmap intensity by zoom level
+					'heatmap-intensity': ['interpolate', ['linear'], ['zoom'], 0, 1, 9, 3],
+					// Color ramp for heatmap. Domain is 0 (low) to 1 (high).
+					'heatmap-color': [
+						'interpolate',
+						['linear'],
+						['heatmap-density'],
+						0,
+						'rgba(0, 255, 0, 0)',
+						// 0.2,
+						// 'rgb(0, 204, 255)',
+						// 0.4,
+						// 'rgb(128, 255, 128)',
+						// 0.6,
+						// 'rgb(255, 255, 102)',
+						// 0.8,
+						// 'rgb(255, 128, 0)',
+						0.9,
+						'rgb(100, 255, 0)',
+						1,
+						'rgb(255, 0, 0)'
+					],
+					// Adjust the heatmap radius by zoom level
+					'heatmap-radius': [
+						'interpolate',
+						['linear'],
+						['get', 'precio'],
+						Math.min(...data.precios.map((p) => p.productos_precio_lista)),
+						2,
+						Math.max(...data.precios.map((p) => p.productos_precio_lista)),
+						10
+					],
+					'heatmap-opacity': 0.8
 				}}
-	/>
+			/> -->
+
+			<!-- <CircleLayer
+				id="cluster_circles"
+				applyToClusters
+				hoverCursor="pointer"
+				paint={{
+					'circle-color': [
+						'step',
+						['get', 'total_precio'],
+						'#51bbd6',
+						10,
+						'#f1f075',
+						30,
+						'#f28cb1'
+					],
+					// 'circle-radius': ['step', ['get', 'point_count'], 15, 10, 20, 30, 25],
+					'circle-radius': 5,
+					'circle-stroke-color': '#fff',
+					'circle-stroke-width': 1,
+					'circle-stroke-opacity': hoverStateFilter(0, 1)
+				}}
+				manageHoverState
+				on:click={(e) => {
+					console.log(e);
+				}}
+			>
+				<!-- <Popup openOn="click" closeOnClickInside let:data>
+					{#if data?.properties}
+						<div class="p-2">
+							<p class="font-bold">Grupo de {data.properties.point_count} precios</p>
+							<p>Precio promedio: ${data.properties.precio_promedio.toFixed(2)}</p>
 						</div>
+					{/if}
+				</Popup> --
+			</CircleLayer>
+
+			<SymbolLayer
+				id="cluster_labels"
+				interactive={false}
+				applyToClusters
+				layout={{
+					'text-field': [
+						'format',
+						['get', 'point_count_abbreviated'],
+						{},
+						'\n$',
+						{},
+						['number-format', ['get', 'total_precio'], { 'max-fraction-digits': 2 }],
+						{ 'font-scale': 0.8 }
+					],
+					'text-size': 12,
+					'text-offset': [0, -0.1]
+				}}
+			/>-->
+
+			<CircleLayer
+				id="precio_circle"
+				applyToClusters={false}
+				hoverCursor="pointer"
+				paint={{
+					'circle-color': [
+						'interpolate',
+						['linear'],
+						['get', 'precio'],
+						Math.min(...data.precios.map((p) => p.productos_precio_lista)),
+						'rgba(0,255,0,0)',
+						Math.max(...data.precios.map((p) => p.productos_precio_lista)),
+						'rgba(255,0,0,1)'
+					],
+					'circle-radius': ['interpolate', ['linear'], ['zoom'], 3, 4, 10, 6],
+					'circle-stroke-width': 1,
+					'circle-stroke-color': '#fff'
+					// 'circle-stroke-opacity': hoverStateFilter(0, 1)
+				}}
+			>
+				<Popup openOn="click" closeOnClickInside let:data>
+					{#if data?.properties}
+						<div class="flex flex-col gap-2 px-3 py-2">
+							<div class="flex flex-col gap-1">
+								<span class="text-xs uppercase leading-none text-neutral-500">
+									{dateFormatter.format(new Date(data.properties.fecha))}
+								</span>
+								<span class="text-xl font-bold leading-none">
+									{pesosFormatter.format(data.properties.precio)}
+								</span>
+							</div>
+
+							<div class="flex gap-2">
+								<div class="flex flex-col leading-none">
+									<span class="font-medium">{data.properties.comercio}</span>
+									<span class="text-sm">{data.properties.direccion}</span>
+								</div>
+
+								<Button
+									href={generateGoogleMapsLink({
+										sucursales_calle: data.properties.sucursales_calle,
+										sucursales_numero: data.properties.sucursales_numero
+									})}
+									target="_blank"
+									variant="outline"
+									size="icon_sm"
+									class="inline-flex items-center gap-1"
+								>
+									<MapPin class="size-4" />
+								</Button>
+							</div>
+
+							<div>
+								<Button
+									variant="default"
+									size="xs"
+									href={`/id_producto/${id_producto}/sucursal/${data.properties.id_comercio}/${data.properties.id_sucursal}`}
+									class="group"
+								>
+									Precios históricos
+									<ArrowRight class="mx-1 size-4 transition-transform group-hover:translate-x-1" />
+								</Button>
+							</div>
+						</div>
+					{/if}
+				</Popup>
+			</CircleLayer>
+		</GeoJSON>
+	</MapLibre>
+</div>
+
+<style>
+	:global(.maplibregl-popup-content) {
+		border-radius: 0.3rem;
+		padding: 0;
+	}
+</style>