From a0650ae7c0e0a6245472f64204b1d1b0a0791e46 Mon Sep 17 00:00:00 2001
From: na-trium-144 <100704180+na-trium-144@users.noreply.github.com>
Date: Mon, 8 Sep 2025 21:10:20 +0900
Subject: [PATCH] =?UTF-8?q?gemini=E3=81=B8=E3=81=AE=E3=83=AA=E3=82=AF?=
 =?UTF-8?q?=E3=82=A8=E3=82=B9=E3=83=88=E3=81=AB=E5=A4=B1=E6=95=97=E3=81=97?=
 =?UTF-8?q?=E3=81=9F=E5=A0=B4=E5=90=88proxy=E3=82=92=E4=BD=BF=E3=81=86?=
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

---
 app/actions/chatActions.ts     |  8 +++-----
 app/actions/gemini.ts          | 24 ++++++++++++++++++++++++
 app/actions/questionExample.ts |  7 ++-----
 3 files changed, 29 insertions(+), 10 deletions(-)
 create mode 100644 app/actions/gemini.ts

diff --git a/app/actions/chatActions.ts b/app/actions/chatActions.ts
index 1fc2e6f..bd2c14b 100644
--- a/app/actions/chatActions.ts
+++ b/app/actions/chatActions.ts
@@ -1,7 +1,7 @@
 'use server';
 
-import { GoogleGenerativeAI } from "@google/generative-ai";
 import { z } from "zod";
+import { generateContent } from "./gemini";
 
 interface FormState {
   response: string;
@@ -13,8 +13,6 @@ const ChatSchema = z.object({
   documentContent: z.string().min(1, { message: "コンテキストとなるドキュメントがありません。"}),
 });
 
-const genAI = new GoogleGenerativeAI(process.env.API_KEY!);
-
 type ChatParams = z.input<typeof ChatSchema>;
 
 export async function askAI(params: ChatParams): Promise<FormState> {
@@ -30,7 +28,7 @@ export async function askAI(params: ChatParams): Promise<FormState> {
   const { userQuestion, documentContent } = parseResult.data;
 
   try {
-    const model = genAI.getGenerativeModel({ model: "gemini-1.5-flash" });
+    
     const prompt = `
 以下のPythonチュートリアルのドキュメントの内容を正確に理解し、ユーザーからの質問に対して、初心者にも分かりやすく、丁寧な解説を提供してください。
 
@@ -48,7 +46,7 @@ ${userQuestion}
 - 
 
 `;
-    const result = await model.generateContent(prompt);
+    const result = await generateContent(prompt);
     const response = result.response;
     const text = response.text();
     return { response: text, error: null };
diff --git a/app/actions/gemini.ts b/app/actions/gemini.ts
new file mode 100644
index 0000000..3af3a16
--- /dev/null
+++ b/app/actions/gemini.ts
@@ -0,0 +1,24 @@
+"use server";
+
+import { GoogleGenerativeAI, ModelParams } from "@google/generative-ai";
+
+export async function generateContent(prompt: string) {
+  const params: ModelParams = {
+    model: "gemini-1.5-flash",
+  };
+
+  const genAI = new GoogleGenerativeAI(process.env.API_KEY!);
+  try {
+    const model = genAI.getGenerativeModel(params);
+    return await model.generateContent(prompt);
+  } catch (e: unknown) {
+    if (String(e).includes("User location is not supported")) {
+      const model = genAI.getGenerativeModel(params, {
+        baseUrl: "https://gemini-proxy.utcode.net",
+      });
+      return await model.generateContent(prompt);
+    } else {
+      throw e;
+    }
+  }
+}
diff --git a/app/actions/questionExample.ts b/app/actions/questionExample.ts
index 846a13a..6b0bf87 100644
--- a/app/actions/questionExample.ts
+++ b/app/actions/questionExample.ts
@@ -1,15 +1,13 @@
 "use server";
 
-import { GoogleGenerativeAI } from "@google/generative-ai";
 import { z } from "zod";
+import { generateContent } from "./gemini";
 
 const QuestionExampleSchema = z.object({
   lang: z.string().min(1),
   documentContent: z.string().min(1),
 });
 
-const genAI = new GoogleGenerativeAI(process.env.API_KEY!);
-
 type QuestionExampleParams = z.input<typeof QuestionExampleSchema>;
 
 export async function getQuestionExample(
@@ -28,7 +26,6 @@ export async function getQuestionExample(
 
   const { lang, documentContent } = parseResult.data;
 
-  const model = genAI.getGenerativeModel({ model: "gemini-1.5-flash" });
   const prompt = `
 以下の${lang}チュートリアルのドキュメントに対して、想定される初心者のユーザーからの質問の例を箇条書きで複数挙げてください。
 強調などはせずテキストだけで1行ごとに1つ出力してください。
@@ -36,7 +33,7 @@ export async function getQuestionExample(
 # ドキュメント
 ${documentContent}
 `;
-  const result = await model.generateContent(prompt);
+  const result = await generateContent(prompt);
   const response = result.response;
   const text = response.text();
   return text.trim().split("\n");