querying the backend with websockets

.env.example

@@ -1,3 +1,3 @@
 OPENAI_API_KEY=
 TAVILY_API_KEY=
-DOC_PATH=./docs/my-docs
+DOC_PATH=./my-docs

README.md

@@ -184,7 +184,7 @@ You can instruct the GPT Researcher to run research tasks based on your local do
 Step 1: Add the env variable `DOC_PATH` pointing to the folder where your documents are located.
 
 ```bash
-export DOC_PATH="./docs/my-docs"
+export DOC_PATH="./my-docs"
 ```
 
 Step 2: 

docs/docs/gpt-researcher/context/local-docs.md

@@ -5,7 +5,7 @@ You can instruct the GPT Researcher to run research tasks based on your local do
 Step 1: Add the env variable `DOC_PATH` pointing to the folder where your documents are located.
 
 ```bash
-export DOC_PATH="./docs/my-docs"
+export DOC_PATH="./my-docs"
 ```
 
 Step 2: 

docs/docs/gpt-researcher/gptr/querying-the-backend.md

@@ -0,0 +1,106 @@
+# Querying the Backend
+
+## Introduction
+
+In this section, we will discuss how to query the GPTR backend server. The GPTR backend server is a Python server that runs the GPTR Python package. The server listens for WebSocket connections and processes incoming messages to generate reports, streaming back logs and results to the client.
+
+An example WebSocket client is implemented in the `gptr-webhook.js` file below.
+
+This function sends a Webhook Message to the GPTR Python backend running on localhost:8000, but this example can also be modified to query a [GPTR Server hosted on Linux](https://docs.gptr.dev/docs/gpt-researcher/getting-started/linux-deployment).
+
+// gptr-webhook.js
+
+```javascript
+
+const WebSocket = require('ws');
+
+let socket = null;
+let responseCallback = null;
+
+async function initializeWebSocket() {
+  if (!socket) {
+    const host = 'localhost:8000';
+    const ws_uri = `ws://${host}/ws`;
+
+    socket = new WebSocket(ws_uri);
+
+    socket.onopen = () => {
+      console.log('WebSocket connection established');
+    };
+
+    socket.onmessage = (event) => {
+      const data = JSON.parse(event.data);
+      console.log('WebSocket data received:', data);
+
+      if (data.content === 'dev_team_result' 
+          && data.output.rubber_ducker_thoughts != undefined
+          && data.output.tech_lead_review != undefined) {
+        if (responseCallback) {
+          responseCallback(data.output);
+          responseCallback = null; // Clear callback after use
+        }
+      } else {
+        console.log('Received data:', data);
+      }
+    };
+
+    socket.onclose = () => {
+      console.log('WebSocket connection closed');
+      socket = null;
+    };
+
+    socket.onerror = (error) => {
+      console.error('WebSocket error:', error);
+    };
+  }
+}
+
+async function sendWebhookMessage(message) {
+  return new Promise((resolve, reject) => {
+    if (!socket || socket.readyState !== WebSocket.OPEN) {
+      initializeWebSocket();
+    }
+
+    const data = {
+      task: message,
+      report_type: 'dev_team',
+      report_source: 'web',
+      tone: 'Objective',
+      headers: {},
+      repo_name: 'elishakay/gpt-researcher'
+    };
+
+    const payload = "start " + JSON.stringify(data);
+
+    responseCallback = (response) => {
+      resolve(response); // Resolve the promise with the WebSocket response
+    };
+
+    if (socket.readyState === WebSocket.OPEN) {
+      socket.send(payload);
+      console.log('Message sent:', payload);
+    } else {
+      socket.onopen = () => {
+        socket.send(payload);
+        console.log('Message sent after connection:', payload);
+      };
+    }
+  });
+}
+
+module.exports = {
+  sendWebhookMessage
+};
+```
+
+And here's how you can leverage this helper function:
+
+```javascript
+const { sendWebhookMessage } = require('./gptr-webhook');
+
+async function main() {
+  const message = 'What are the thoughts of the rubber duck?';
+  const response = await sendWebhookMessage(message);
+  console.log('Response:', response);
+}
+```

docs/sidebars.js

@@ -32,6 +32,7 @@
       items: [
         'gpt-researcher/gptr/pip-package',
         'gpt-researcher/gptr/example',
+        'gpt-researcher/gptr/querying-the-backend',
         'gpt-researcher/gptr/config',
         'gpt-researcher/gptr/automated-tests',
         'gpt-researcher/gptr/troubleshooting',