feat: implement feed classification and podcast content generation

scripts/fetch_and_generate.ts

@@ -1,5 +1,8 @@
 import Parser from "rss-parser";
-import { openAI_GenerateScript } from "../services/llm";
+import { 
+  openAI_ClassifyFeed, 
+  openAI_GeneratePodcastContent 
+} from "../services/llm";
 import { generateTTS } from "../services/tts";
 import { saveEpisode, markAsProcessed } from "../services/database";
 import { updatePodcastRSS } from "../services/podcast";
@@ -36,68 +39,81 @@ async function main() {
     feedUrls = [];
   }
 
+  // フィードごとに処理
   for (const url of feedUrls) {
     const feed = await parser.parseURL(url);
-    for (const item of feed.items) {
+    
+    // フィードのカテゴリ分類
+    const feedTitle = feed.title || url;
+    const category = await openAI_ClassifyFeed(feedTitle);
+    console.log(`フィード分類完了: ${feedTitle} - ${category}`);
+    
+    // 昨日の記事のみフィルタリング
+    const yesterday = new Date();
+    yesterday.setDate(yesterday.getDate() - 1);
+    
+    const yesterdayItems = feed.items.filter(item => {
       const pub = new Date(item.pubDate || "");
-      const today = new Date();
-      const yesterday = new Date(today);
-      yesterday.setDate(today.getDate() - 1);
-
-      if (
+      return (
         pub.getFullYear() === yesterday.getFullYear() &&
         pub.getMonth() === yesterday.getMonth() &&
         pub.getDate() === yesterday.getDate()
-      ) {
-        // Use item.id if available, otherwise generate fallback ID from title or link
-        const itemId = item["id"] as string | undefined;
-        const fallbackId = item.link || item.title || JSON.stringify(item);
-        const finalItemId =
-          itemId && typeof itemId === "string" && itemId.trim() !== ""
-            ? itemId
-            : `fallback-${Buffer.from(fallbackId).toString("base64")}`;
-
-        // Skip if even the fallback ID is missing (should be rare)
-        if (!finalItemId || finalItemId.trim() === "") {
-          console.warn(`フィードアイテムのIDを生成できませんでした`, {
-            feedUrl: url,
-            itemTitle: item.title,
-            itemLink: item.link,
-          });
-          continue;
-        }
-
-        const already = await markAsProcessed(url, finalItemId);
-        if (already) {
-          console.log(`既に処理済み: ${finalItemId}`);
-          continue;
-        }
-
-        console.log(`スクリプト生成開始: ${finalItemId}`);
-        const scriptText = await openAI_GenerateScript({
-          title: item.title ?? "",
-          link: item.link ?? "",
-          contentSnippet: item.contentSnippet ?? "",
+      );
+    });
+    
+    if (yesterdayItems.length === 0) {
+      console.log(`昨日の記事が見つかりません: ${feedTitle}`);
+      continue;
+    }
+    
+    // ポッドキャスト原稿生成
+    console.log(`ポッドキャスト原稿生成開始: ${feedTitle}`);
+    const podcastContent = await openAI_GeneratePodcastContent(feedTitle, yesterdayItems);
+    
+    // トピックごとの統合音声生成
+    const feedUrlHash = crypto.createHash("md5").update(url).digest("hex");
+    const categoryHash = crypto.createHash("md5").update(category).digest("hex");
+    const uniqueFilename = `${feedUrlHash}-${categoryHash}.mp3`;
+    
+    const audioFilePath = await generateTTS(uniqueFilename, podcastContent);
+    console.log(`音声ファイル生成完了: ${audioFilePath}`);
+    
+    // エピソードとして保存（各フィードにつき1つの統合エピソード）
+    const firstItem = yesterdayItems[0];
+    const pub = new Date(firstItem.pubDate || "");
+    
+    await saveEpisode({
+      id: `topic-${categoryHash}`,
+      title: `${category}: ${feedTitle}`,
+      pubDate: pub.toISOString(),
+      audioPath: audioFilePath,
+      sourceLink: url,
+    });
+    
+    console.log(`エピソード保存完了: ${category} - ${feedTitle}`);
+    
+    // 個別記事の処理記録
+    for (const item of yesterdayItems) {
+      const itemId = item["id"] as string | undefined;
+      const fallbackId = item.link || item.title || JSON.stringify(item);
+      const finalItemId =
+        itemId && typeof itemId === "string" && itemId.trim() !== ""
+          ? itemId
+          : `fallback-${Buffer.from(fallbackId).toString("base64")}`;
+          
+      if (!finalItemId || finalItemId.trim() === "") {
+        console.warn(`フィードアイテムのIDを生成できませんでした`, {
+          feedUrl: url,
+          itemTitle: item.title,
+          itemLink: item.link,
         });
-        
-        // Generate a unique filename using the feed URL hash and item ID
-        const feedUrlHash = crypto.createHash("md5").update(url).digest("hex");
-        const itemIdHash = crypto.createHash("md5").update(finalItemId).digest("hex");
-        const uniqueFilename = `${feedUrlHash}-${itemIdHash}.mp3`;
-        
-        const audioFilePath = await generateTTS(uniqueFilename, scriptText);
-
-        console.log(`音声ファイル生成完了: ${audioFilePath}`);
-        
-        await saveEpisode({
-          id: finalItemId,
-          title: item.title ?? "",
-          pubDate: pub.toISOString(),
-          audioPath: audioFilePath,
-          sourceLink: item.link ?? "",
-        });
-        
-        console.log(`エピソード保存完了: ${finalItemId}`);
+        continue;
+      }
+      
+      const already = await markAsProcessed(url, finalItemId);
+      if (already) {
+        console.log(`既に処理済み: ${finalItemId}`);
+        continue;
       }
     }
   }