Merge pull request #18 from toshiakit/dev

Update: Token usage changes

README.md

@@ -12,8 +12,10 @@ The app and class simply serve as an interface to the ChatGPT API. You should be
 
 * adds support for GPT-4 models [available to all API users who have a history of successful payments](https://openai.com/blog/gpt-4-api-general-availability)
 * adds Connection Timeout settings. You can add proxy via [Web Preferences](https://www.mathworks.com/help/matlab/ref/preferences.html) in MATLAB.
-* adds support for stop sequences (max 4). 
+* adds support for stop sequences (max 4).
+* adds support for 3 follow-up questions. 
 * improves the description of the API error messages
+* removed fee calculation from the usage panel due to frequent updates to the API pricing. 
 
 ## Requirements
 
@@ -51,7 +53,8 @@ setenv("OPENAI_API_KEY","your key here")
 
 ## How to use: chatGPT class
 
-chatGPT class creates an instance to access OpenAI API using your own API key, and optionally `max_tokens` that determine the length of the response. Once you create an instance, then you can use its chat method to post prompt via OpenAPI ChatGPT API released on March 1, 2023.
+chatGPT class creates an instance to access OpenAI API using your own API key, and optionally `max_tokens` that determine the length of the response. Once you create an instance, then you can use its chat method to post prompt via OpenAPI ChatGPT API updated on June 13, 2023.
+Please note that this class doesn't support function calling.
 
 ### Where to find it
 [chatGPT](helpers/chatGPT.m) class is located in `helpers` folder. 
@@ -70,7 +73,7 @@ You can also set `max_tokens` parameter, which determines the length of the resp
 myBot = chatGPT(max_tokens=50);
 ```
 You can also specify other parameters, such as `model` and `temperature`.
-* models: gpt-3.5-turbo (default - stable release), gpt-3.5-turbo-0301 (latest release) 
+* models: gpt-3.5-turbo (default - stable release), gpt-4 (requires paid subsription to the API) 
 * temperature: 0 = more strict, 1 =  balanced, 2 = more creative
 
 You can pass a `role` to prime the chatbot for a specific use case. By default the chatbot is primed to act as an AI assistant, using the prompt "You are a helpful assistant." You can customize this for specific use cases. 
@@ -89,16 +92,17 @@ answer = chat(myBot,"your prompt")
 You will get an error if the API didn't return response successfully.
 
 ### Check the token usage
-Use `usage` method to obtain the total number of tokens used in the current session. It returns both the number of tokens as well as the cost. 
+Use `usage` method to obtain the number of tokens used in the current session. It returns the completion tokens used, prompt tokens used, and total tokens used.
+Different API rate applies depending on the model and type of tokens. Please consult [OpenAI pricing page](https://openai.com/pricing) for details. 
 
 ```matlab
-[tokens,cost] = usage(myBot)
+[completion_tokens,prompt_tokens,total_tokens] = usage(myBot)
 ```
 ### Save the chat history
 Use `saveChat` method to save the chat history to a file. The supported format includes .mat, .xlsx, and .json. 
 
 ```matlab
-[tokens,cost] = saveChat(myBot,format=".xlsx")
+saveChat(myBot,format=".xlsx")
 ```
 
 ## chatGPT class example

helpers/chatGPT.m

@@ -3,7 +3,7 @@
     %   Create an instance using your own API key, and optionally
     %   max_tokens that determine the length of the response
     %
-    %   ask method lets you send a prompt text to the API as HTTP request
+    %   chat method lets you send a prompt text to the API as HTTP request
     %   and parses the response.
     %
     %   Before using, set an environment variable with your OpenAI API key
@@ -14,7 +14,7 @@
     properties (Access = public)
         % the API endpoint
         api_endpoint = "https://api.openai.com/v1/chat/completions";
-        % ChatGPT model to use - gpt-3.5-turbo or gpt-3.5-turbo-0301
+        % ChatGPT model to use - gpt-3.5-turbo, etc.
         model;
         % what role the bot should play
         role;
@@ -25,6 +25,8 @@
         % store chat log in messages object
         messages;
         % store usage
+        completion_tokens = 0;
+        prompt_tokens = 0;
         total_tokens = 0;
     end
 
@@ -106,6 +108,10 @@
                 obj.messages = [obj.messages, ...
                     struct('role',"assistant",'content',responseText)];
                 % add the tokens used
+                obj.completion_tokens = obj.completion_tokens + ...
+                    response.Body.Data.usage.completion_tokens;
+                obj.prompt_tokens = obj.prompt_tokens + ...
+                    response.Body.Data.usage.prompt_tokens;
                 obj.total_tokens = obj.total_tokens + ...
                     response.Body.Data.usage.total_tokens;
             else
@@ -115,12 +121,11 @@
             end
         end
 
-        function [tokens_used, cost] = usage(obj)
-            %USAGE retunrs the number of tokens used and the cost
-            tokens_used = obj.total_tokens;
-            % as of March 1, 2023, it's priced at $0.002 per 1K tokens
-            price = 0.002/1000;
-            cost = round(tokens_used * price,4);
+        function [completion_tokensd, prompt_tokens, total_tokens] = usage(obj)
+            %USAGE retunrs the number of tokens used
+            completion_tokensd = obj.completion_tokens;
+            prompt_tokens = obj.prompt_tokens;
+            total_tokens = obj.total_tokens;
         end
 
         function saveChat(obj,options)

helpers/chatter.m

@@ -74,6 +74,10 @@
                 obj.messages = [obj.messages, ...
                     struct('role',"assistant",'content',responseText)];
                 % add the numbers of tokens used
+                obj.completion_tokens = obj.completion_tokens + ...
+                    response.Body.Data.usage.completion_tokens;
+                obj.prompt_tokens = obj.prompt_tokens + ...
+                    response.Body.Data.usage.prompt_tokens;
                 obj.total_tokens = obj.total_tokens + ...
                     response.Body.Data.usage.total_tokens;
             else
@@ -84,7 +88,8 @@
                     responseText = responseText + newline + "Check your API key.";
                     responseText = responseText + newline + "You may have an invalid API key.";
                 elseif string(response.StatusCode) == "404"
-                    responseText = responseText + newline + "You may not have access to the model.";
+                    responseText = responseText + newline + "You may not have access to the model: " + ...
+                        obj.model + ". Consider using another model.";
                 elseif string(response.StatusCode) == "429"
                     responseText = responseText + newline + "You exceeded the API limit. Your free trial for OpenAI API may have expired.";
                 end