check point, notebook basic structure defined

backend/prisma/migrations/20250409231956_add_image_url/migration.sql

@@ -0,0 +1,2 @@
+-- AlterTable
+ALTER TABLE "Article" ADD COLUMN     "imageUrl" TEXT;

backend/prisma/schema.prisma

@@ -35,6 +35,7 @@ model Article {
   updatedAt   DateTime     @updatedAt
   categories  String[]
   authors     String[]
+  imageUrl    String?
   Note        Note[]
 
   @@index([url])

backend/src/articles/articles.controller.ts

@@ -1,4 +1,4 @@
-import { Controller, Get, Query } from '@nestjs/common';
+import { Controller, Get, Query, Param } from '@nestjs/common';
 import { ArticlesService } from './articles.service';
 
 @Controller('api/articles')
@@ -20,4 +20,9 @@ export class ArticlesController {
   async getCount() {
     return { count: await this.articlesService.count() };
   }
+
+  @Get(':id')
+  async findOne(@Param('id') id: string) {
+    return this.articlesService.findOne(id);
+  }
 }

backend/src/articles/articles.service.ts

@@ -1,4 +1,4 @@
-import { Injectable } from '@nestjs/common';
+import { Injectable, NotFoundException } from '@nestjs/common';
 import { PrismaService } from '../prisma/prisma.service';
 import { PaginatedArticles } from './types';
 
@@ -46,6 +46,28 @@ export class ArticlesService {
     };
   }
 
+  async findOne(id: string) {
+    const article = await this.prisma.article.findUnique({
+      where: { id },
+      select: {
+        id: true,
+        title: true,
+        content: true,
+        source: true,
+        url: true,
+        publishedAt: true,
+        categories: true,
+        authors: true,
+      },
+    });
+
+    if (!article) {
+      throw new NotFoundException(`Article with ID ${id} not found`);
+    }
+
+    return article;
+  }
+
   async count(): Promise<number> {
     return this.prisma.article.count();
   }

backend/src/rss/rss.controller.ts

@@ -6,7 +6,9 @@ export class RssController {
   constructor(private readonly rssService: RssService) {}
 
   @Post('fetch')
-  async fetchAndStore(@Body('feedUrl') feedUrl: string): Promise<{ message: string }> {
+  async fetchAndStore(
+    @Body('feedUrl') feedUrl: string,
+  ): Promise<{ message: string }> {
     await this.rssService.fetchAndStoreArticles(feedUrl);
     return { message: 'RSS feed processed and articles stored successfully.' };
   }
@@ -15,4 +17,4 @@ export class RssController {
   async getAllArticles() {
     return this.rssService.getAllArticles();
   }
-}
+}

backend/src/rss/rss.service.ts

@@ -5,25 +5,76 @@ import * as Parser from 'rss-parser';
 @Injectable()
 export class RssService {
   constructor(private prisma: PrismaService) {}
-  feddUrl: string = 'https://www.rt.com/rss/';
-  async fetchAndStoreArticles(feedUrl: string): Promise<void> {
-    const parser = new Parser();
-    const feed = await parser.parseURL(feedUrl);
 
-    for (const item of feed.items) {
-      await this.prisma.article.upsert({
-        where: { url: item.link },
-        update: {},
-        create: {
-          title: item.title || 'Untitled',
-          content: item.contentSnippet || '',
-          source: feed.title || 'Unknown Source',
-          url: item.link || '',
-          publishedAt: item.isoDate ? new Date(item.isoDate) : new Date(),
-          categories: item.categories || [],
-          authors: item.creator ? [item.creator] : [],
-        },
-      });
+  private parser: Parser;
+
+  onModuleInit() {
+    // Configure parser with custom fields
+    this.parser = new Parser({
+      customFields: {
+        item: [
+          ['content:encoded', 'contentEncoded'],
+          ['dc:creator', 'creator'],
+        ],
+      },
+    });
+  }
+
+  private cleanHtml(html: string): string {
+    // Remove HTML tags and decode entities
+    return html
+      .replace(/<[^>]*>/g, '') // Remove HTML tags
+      .replace(/&nbsp;/g, ' ') // Replace &nbsp; with space
+      .replace(/&amp;/g, '&') // Replace &amp; with &
+      .replace(/&quot;/g, '"') // Replace &quot; with "
+      .replace(/&#39;/g, "'"); // Replace &#39; with '
+  }
+
+  async fetchAndStoreArticles(feedUrl: string): Promise<void> {
+    try {
+      const feed = await this.parser.parseURL(feedUrl);
+
+      for (const item of feed.items) {
+        // Extract content from either contentEncoded or description
+        const rawContent = item.contentEncoded || item.content || item.description || '';
+        const cleanContent = this.cleanHtml(rawContent);
+
+        // Get the first image URL from content if available
+        const imageMatch = rawContent.match(/<img[^>]+src="([^">]+)"/);
+        const imageUrl = imageMatch ? imageMatch[1] : null;
+
+        // Extract categories and authors
+        const categories = Array.isArray(item.categories) ? item.categories : [];
+        const authors = item.creator ? [item.creator] : [];
+
+        await this.prisma.article.upsert({
+          where: { url: item.link || '' },
+          update: {
+            title: item.title || 'Untitled',
+            content: cleanContent,
+            source: feed.title || 'RT',
+            url: item.link || '',
+            publishedAt: item.isoDate ? new Date(item.isoDate) : new Date(),
+            categories,
+            authors,
+            imageUrl,
+            updatedAt: new Date(),
+          },
+          create: {
+            title: item.title || 'Untitled',
+            content: cleanContent,
+            source: feed.title || 'RT',
+            url: item.link || '',
+            publishedAt: item.isoDate ? new Date(item.isoDate) : new Date(),
+            categories,
+            authors,
+            imageUrl,
+          },
+        });
+      }
+    } catch (error) {
+      console.error('Error fetching RSS feed:', error);
+      throw error;
     }
   }