API conversation

application/app.py

@@ -1,5 +1,6 @@
 import os
 import json
+import traceback
 
 import dotenv
 import requests
@@ -86,7 +87,6 @@ def api_answer():
 
     # use try and except  to check for exception
     try:
-
         # check if the vectorstore is set
         if "active_docs" in data:
             vectorstore = "vectors/" + data["active_docs"]
@@ -144,6 +144,8 @@ def api_answer():
         # }
         return result
     except Exception as e:
+        # print whole traceback
+        traceback.print_exc()
         print(str(e))
         return bad_request(500,str(e))
 
@@ -185,4 +187,4 @@ def after_request(response):
 
 
 if __name__ == "__main__":
-    app.run(debug=True)
+    app.run(debug=True, port=5001)

frontend/src/conversation/Conversation.tsx

@@ -30,7 +30,7 @@ export default function Conversation() {
 
   return (
     <div className="flex justify-center p-6">
-      <div className="flex mt-20 w-10/12 flex-col transition-all md:w-1/2">
+      <div className="mt-20 flex w-10/12 flex-col transition-all md:w-1/2">
         {messages.map((message, index) => {
           return (
             <ConversationBubble

frontend/src/conversation/conversationApi.ts

@@ -3,26 +3,32 @@ import { Answer } from './conversationModels';
 export function fetchAnswerApi(
   question: string,
   apiKey: string,
+  selectedDocs: string,
 ): Promise<Answer> {
   // a mock answer generator, this is going to be replaced with real http call
   return new Promise((resolve, reject) => {
-    setTimeout(() => {
-      let result = '';
-      const characters =
-        'ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789';
-      const charactersLength = characters.length;
-      let counter = 0;
-      while (counter < 5) {
-        result += characters.charAt(
-          Math.floor(Math.random() * charactersLength),
-        );
-        counter += 1;
-      }
-      const randNum = getRandomInt(0, 10);
-      randNum < 5
-        ? reject()
-        : resolve({ answer: result, query: question, result });
-    }, 3000);
+    const activeDocs = 'default';
+    fetch('https://docsgpt.arc53.com/api/answer', {
+      method: 'POST',
+      headers: {
+        'Content-Type': 'application/json',
+      },
+      body: JSON.stringify({
+        question: question,
+        api_key: apiKey,
+        embeddings_key: apiKey,
+        history: localStorage.getItem('chatHistory'),
+        active_docs: selectedDocs,
+      }),
+    })
+      .then((response) => response.json())
+      .then((data) => {
+        const result = data.answer;
+        resolve({ answer: result, query: question, result });
+      })
+      .catch((error) => {
+        reject();
+      });
   });
 }
 

frontend/src/conversation/conversationSlice.ts

@@ -14,7 +14,24 @@ export const fetchAnswer = createAsyncThunk<
   { state: RootState }
 >('fetchAnswer', async ({ question }, { getState }) => {
   const state = getState();
-  const answer = await fetchAnswerApi(question, state.preference.apiKey);
+  let namePath = state.preference.selectedDocs?.name;
+  if (state.preference.selectedDocs?.language === namePath) {
+    namePath = '.project';
+  }
+
+  const docPath =
+    state.preference.selectedDocs?.language +
+    '/' +
+    namePath +
+    '/' +
+    state.preference.selectedDocs?.version +
+    '/' +
+    state.preference.selectedDocs?.model;
+  const answer = await fetchAnswerApi(
+    question,
+    state.preference.apiKey,
+    docPath,
+  );
   return answer;
 });