From fe93de79ae0da45a6c3cb5e37fd4a9f43cbd24dd Mon Sep 17 00:00:00 2001
From: Nikhil-Doye <nikhil.doye@gmail.com>
Date: Mon, 20 Oct 2025 14:18:06 -0400
Subject: [PATCH] Enhance OpenAI service with proxy support and error handling

- Introduced a method to retrieve an optional proxy base URL for production use, improving security by allowing API calls to be routed through a backend.
- Updated error handling to provide clear messages when the API key is not configured or when the proxy is not set in production.
- Refactored the web scraping processor to handle undefined configuration values more gracefully, ensuring default behaviors are maintained.
---
 src/services/openaiService.ts                 | 100 +++++++++++++-----
 .../processors/webScrapingProcessor.ts        |   3 +-
 2 files changed, 74 insertions(+), 29 deletions(-)

diff --git a/src/services/openaiService.ts b/src/services/openaiService.ts
index 7130b7b..5e38885 100644
--- a/src/services/openaiService.ts
+++ b/src/services/openaiService.ts
@@ -9,11 +9,19 @@ const getApiKey = (): string => {
   );
 };
 
-// Initialize OpenAI client for DeepSeek
+// Optional proxy base URL (recommended for production)
+const getProxyBaseUrl = (): string | null => {
+  const fromStorage = localStorage.getItem("api_base_url");
+  const fromEnv = process.env.REACT_APP_API_BASE;
+  return fromStorage || fromEnv || null;
+};
+
+// Initialize OpenAI client for DeepSeek (fallback-only; avoid in production)
 const openai = new OpenAI({
   apiKey: getApiKey(),
-  baseURL: "https://api.deepseek.com/v1", // DeepSeek API endpoint
-  dangerouslyAllowBrowser: true, // Only for development - in production, use a backend
+  baseURL: "https://api.deepseek.com/v1",
+  // Keep browser use only as a last-resort fallback during local dev
+  dangerouslyAllowBrowser: true,
 });
 
 export interface OpenAIConfig {
@@ -36,29 +44,47 @@ export const callOpenAI = async (
   config: OpenAIConfig
 ): Promise<OpenAIResponse> => {
   try {
-    // Check if API key is configured
-    const apiKey = getApiKey();
-    if (
-      !apiKey ||
-      apiKey === "your_openai_api_key_here" ||
-      apiKey === "your_deepseek_api_key_here"
-    ) {
+    // Prefer secure proxy in production
+    const proxyBase = getProxyBaseUrl();
+    if (proxyBase) {
+      const resp = await fetch(`${proxyBase.replace(/\/$/, "")}/llm/chat`, {
+        method: "POST",
+        headers: {
+          "Content-Type": "application/json",
+        },
+        body: JSON.stringify({ prompt, config }),
+      });
+
+      if (!resp.ok) {
+        const text = await resp.text();
+        throw new Error(`Proxy error: ${resp.status} ${text}`);
+      }
+
+      const data = await resp.json();
+      return {
+        content: data.content || "",
+        usage: data.usage,
+      };
+    }
+
+    // Fallback (dev only): direct browser call (keys are exposed!)
+    if (process.env.NODE_ENV === "production") {
       throw new Error(
-        "DeepSeek API key not configured. Please configure your API key in the settings."
+        "LLM proxy not configured. In production, configure REACT_APP_API_BASE or localStorage 'api_base_url' to a secure backend."
       );
     }
 
-    // Update the OpenAI client with the current API key
-    openai.apiKey = apiKey;
+    const apiKey = getApiKey();
+    if (!apiKey) {
+      throw new Error(
+        "DeepSeek API key not configured. Set it in settings for local development, or configure a proxy for production."
+      );
+    }
 
+    openai.apiKey = apiKey;
     const response = await openai.chat.completions.create({
       model: config.model,
-      messages: [
-        {
-          role: "user",
-          content: prompt,
-        },
-      ],
+      messages: [{ role: "user", content: prompt }],
       temperature: config.temperature,
       max_tokens: config.maxTokens || 1000,
     });
@@ -90,20 +116,38 @@ export const generateEmbedding = async (
   text: string,
   model: string = "text-embedding-ada-002"
 ): Promise<number[]> => {
-  const apiKey = getApiKey();
-  if (!apiKey || apiKey === "your_openai_api_key_here") {
+  // Prefer secure proxy in production
+  const proxyBase = getProxyBaseUrl();
+  if (proxyBase) {
+    const resp = await fetch(`${proxyBase.replace(/\/$/, "")}/llm/embeddings`, {
+      method: "POST",
+      headers: { "Content-Type": "application/json" },
+      body: JSON.stringify({ text, model }),
+    });
+    if (!resp.ok) {
+      const textBody = await resp.text();
+      throw new Error(`Proxy error: ${resp.status} ${textBody}`);
+    }
+    const data = await resp.json();
+    return data.embedding as number[];
+  }
+
+  // Fallback (dev only)
+  if (process.env.NODE_ENV === "production") {
     throw new Error(
-      "OpenAI API key not configured. Please configure your API key in the settings."
+      "Embedding proxy not configured. In production, configure REACT_APP_API_BASE or localStorage 'api_base_url' to a secure backend."
+    );
+  }
+
+  const apiKey = getApiKey();
+  if (!apiKey) {
+    throw new Error(
+      "OpenAI API key not configured. Set it in settings for local development, or configure a proxy for production."
     );
   }
 
   openai.apiKey = apiKey;
-
-  const response = await openai.embeddings.create({
-    model,
-    input: text,
-  });
-
+  const response = await openai.embeddings.create({ model, input: text });
   return response.data[0].embedding;
 };
 
diff --git a/src/services/processors/webScrapingProcessor.ts b/src/services/processors/webScrapingProcessor.ts
index b89639c..c78dd6c 100644
--- a/src/services/processors/webScrapingProcessor.ts
+++ b/src/services/processors/webScrapingProcessor.ts
@@ -11,7 +11,8 @@ export default async function webScrapingProcessor(
     const result = await scrapeWithFirecrawl({
       url: config.url,
       formats: config.formats || ["markdown"],
-      onlyMainContent: config.onlyMainContent || true,
+      onlyMainContent:
+        config.onlyMainContent !== undefined ? config.onlyMainContent : true,
       maxLength: config.maxLength || 10000,
     });