Merge pull request #3 from sbslee/0.4.0-dev

0.4.0 dev

CHANGELOG.md

@@ -1,13 +1,16 @@
 # CHANGELOG
 
-## 0.3.0 (2023-03-30)
+## 0.4.0 (2023-06-02)
+* Allow users to modify chatbot parameters (e.g. prompt, temperature, and chunk size).
+
+## 0.3.0 (2023-05-30)
 * Add multi-document chat functionality.
 * Add chat functionality for additional document formats (.pdf, .doc, and .docx).
 
-## 0.2.0 (2023-03-27)
+## 0.2.0 (2023-05-27)
 * Enable users to initiate a new chat session.
 * Enable users to choose their preferred GPT model.
 * Enable users to chat with their text (.txt) documents.
 
-## 0.1.0 (2023-03-24)
+## 0.1.0 (2023-05-24)
 * Initial release.

README.md

@@ -1,5 +1,7 @@
 # README
 
+[![PyPI](https://badge.fury.io/py/kanu.svg)](https://badge.fury.io/py/kanu)
+
 KANU is a minimalistic Python-based GUI for various chatbots.
 
 ## Chatbots
@@ -26,4 +28,4 @@ $ kanu
 
 ## Changelog
 
-See the [CHANGELOG.md](./CHANGELOG.md) file for details.
+See the [CHANGELOG.md](https://github.com/sbslee/kanu/blob/main/CHANGELOG.md) file for details.

kanu/__main__.py

@@ -4,12 +4,20 @@
 from .version import __version__
 from .utils import Tooltip
 
+CHATGPT_PROMPT = """You are a helpful assistant."""
+DOCGPT_PROMPT = """Use the following pieces of context to answer the question at the end. If you don't know the answer, just say that you don't know, don't try to make up an answer.
+
+{context}
+
+Question: {question}
+Helpful Answer:"""
+
 class KANU:
     def __init__(self, root):
         self.container = None
         self.root = root
         self.root.title(f"KANU ({__version__})")
-        self.root.geometry("600x400")
+        self.root.geometry("600x500")
         self.homepage()
 
     def homepage(self):
@@ -40,14 +48,29 @@ def config_chatgpt(self):
         b.grid(row=4, column=0)
         b = tk.Radiobutton(self.container, variable=self.model, text="gpt-4", value="gpt-4")
         b.grid(row=4, column=1)
-        l = tk.Label(self.container, text="OpenAI API key:")
+        l = tk.Label(self.container, text="System message ⓘ:")
+        Tooltip(l, "The system message helps set the behavior of the chatbot.")
         l.grid(row=5, column=0, columnspan=2)
+        self.prompt = tk.Text(self.container, height=9, width=42)
+        sb = tk.Scrollbar(self.container, command=self.prompt.yview)
+        self.prompt.insert("1.0", CHATGPT_PROMPT)
+        self.prompt.grid(row=6, column=0, columnspan=2, sticky="nsew")
+        sb.grid(row=6, column=2, sticky="ns")
+        self.prompt["yscrollcommand"] = sb.set
+        l = tk.Label(self.container, text="Temperature ⓘ:")
+        Tooltip(l, "The randomness in generating responses, which ranges between 0 and 1, with 0 indicating almost deterministic behavior.")
+        l.grid(row=7, column=0, columnspan=2)
+        self.temperature = tk.DoubleVar(self.container, value=0.5)
+        e = tk.Entry(self.container, textvariable=self.temperature)
+        e.grid(row=8, column=0, columnspan=2)
+        l = tk.Label(self.container, text="OpenAI API key:")
+        l.grid(row=9, column=0, columnspan=2)
         e = tk.Entry(self.container)
-        e.grid(row=6, column=0, columnspan=2)
-        b = tk.Button(self.container, text="Submit", command=lambda: self.deploy_agent("ChatGPT", e.get(), self.model.get()))
-        b.grid(row=7, column=0)
+        e.grid(row=10, column=0, columnspan=2)
+        b = tk.Button(self.container, text="Submit", command=lambda: self.deploy_agent("ChatGPT", e.get(), self.model.get(), self.temperature.get(), self.prompt.get("1.0", "end-1c")))
+        b.grid(row=11, column=0)
         b = tk.Button(self.container, text="Go back", command=lambda: self.homepage())
-        b.grid(row=7, column=1)
+        b.grid(row=11, column=1)
 
     def config_docgpt(self):
         self.container.pack_forget()
@@ -68,18 +91,27 @@ def config_docgpt(self):
         self.model = tk.StringVar(self.container, value="gpt-3.5-turbo")
         l = tk.Label(self.container, text="Model:")
         l.grid(row=9, column=0, columnspan=2)
-        b = tk.Radiobutton(self.container, variable=self.model, text="gpt-3.5-turbo", value="gpt-3.5-turbo")
-        b.grid(row=10, column=0)
-        b = tk.Radiobutton(self.container, variable=self.model, text="gpt-4", value="gpt-4")
-        b.grid(row=10, column=1)
+        rb = tk.Radiobutton(self.container, variable=self.model, text="gpt-3.5-turbo", value="gpt-3.5-turbo")
+        rb.grid(row=10, column=0)
+        rb = tk.Radiobutton(self.container, variable=self.model, text="gpt-4", value="gpt-4")
+        rb.grid(row=10, column=1)
+        l = tk.Label(self.container, text="System message ⓘ:")
+        Tooltip(l, "The system message helps set the behavior of the chatbot.")
+        l.grid(row=11, column=0, columnspan=3)
+        self.prompt = tk.Text(self.container, height=9, width=42)
+        sb = tk.Scrollbar(self.container, command=self.prompt.yview)
+        self.prompt.insert("1.0", DOCGPT_PROMPT)
+        self.prompt.grid(row=12, column=0, columnspan=3, sticky="nsew")
+        sb.grid(row=12, column=3, sticky="ns")
+        self.prompt["yscrollcommand"] = sb.set
         l = tk.Label(self.container, text="OpenAI API key:")
-        l.grid(row=11, column=0, columnspan=2)
+        l.grid(row=13, column=0, columnspan=2)
         e = tk.Entry(self.container)
-        e.grid(row=12, column=0, columnspan=2)
-        b = tk.Button(self.container, text="Submit", command=lambda: self.deploy_agent("DocGPT", e.get(), self.model.get()))
-        b.grid(row=13, column=0)
+        e.grid(row=14, column=0, columnspan=2)
+        b = tk.Button(self.container, text="Submit", command=lambda: self.deploy_agent("DocGPT", e.get(), self.model.get(), self.prompt.get("1.0", "end-1c")))
+        b.grid(row=15, column=0)
         b = tk.Button(self.container, text="Go back", command=lambda: self.homepage())
-        b.grid(row=13, column=1)
+        b.grid(row=15, column=1)
 
     def deploy_agent(self, agent, *args, **kwargs):
         if agent == "ChatGPT":

kanu/chatgpt.py

@@ -3,9 +3,11 @@
 import openai
 
 class ChatGPT:
-    def __init__(self, kanu, openai_key, model):
+    def __init__(self, kanu, openai_key, model, temperature, prompt):
         self.kanu = kanu
         self.model = model
+        self.temperature = temperature
+        self.prompt = prompt
         openai.api_key = openai_key
 
     def run(self):
@@ -28,11 +30,12 @@ def run(self):
 
     def send_message(self, entry):
         if not self.messages:
-            self.messages.append({"role": "system", "content": "You are a helpful assistant."})
+            self.messages.append({"role": "system", "content": self.prompt})
         self.messages += [{"role": "user", "content": entry.get()}]
         bot_response = openai.ChatCompletion.create(
             model=self.model,
             messages=self.messages,
+            temperature=self.temperature,
         )
         response = bot_response["choices"][0]["message"]["content"]
         self.messages += [{"role": "assistant", "content": response}]

kanu/docgpt.py

@@ -7,6 +7,7 @@
 from langchain.vectorstores import Chroma
 from langchain.chat_models import ChatOpenAI
 from langchain.chains import RetrievalQA
+from langchain.prompts import PromptTemplate
 
 from langchain.document_loaders import (
     TextLoader,
@@ -22,11 +23,11 @@
     ".doc": UnstructuredWordDocumentLoader,
     ".docx": UnstructuredWordDocumentLoader,
 }
-
 class DocGPT:
-    def __init__(self, kanu, openai_key, model):
+    def __init__(self, kanu, openai_key, model, prompt):
         self.kanu = kanu
         self.model = model
+        self.prompt = prompt
         os.environ["OPENAI_API_KEY"] = openai_key
 
     def run(self):
@@ -39,8 +40,8 @@ def run(self):
         b.grid(row=1, column=0)
         b = tk.Button(self.kanu.container, text="Reload", command=lambda: self.run())
         b.grid(row=1, column=2)
-        l = tk.Message(self.kanu.container, width=300, text="Option 1. Create a new database")
-        l.grid(row=2, column=0, columnspan=3)
+        m = tk.Message(self.kanu.container, width=300, text="Option 1. Create a new database")
+        m.grid(row=2, column=0, columnspan=3)
         l = tk.Label(self.kanu.container, text="Document ⓘ:")
         Tooltip(l, "Directory containing documents for the database.")
         l.grid(row=3, column=0) 
@@ -55,25 +56,42 @@ def run(self):
         self.new_database_label.grid(row=4, column=1)
         b = tk.Button(self.kanu.container, text="Browse", command=self.specify_new_database_directory)
         b.grid(row=4, column=2)
+        l = tk.Label(self.kanu.container, text="Chunk size ⓘ:")
+        Tooltip(l, "The maximum number of characters in each chunk.")
+        l.grid(row=5, column=0)
+        self.chunk_size = tk.IntVar(self.kanu.container, value=1000)
+        e = tk.Entry(self.kanu.container, textvariable=self.chunk_size)
+        e.grid(row=5, column=1, columnspan=2)
+        l = tk.Label(self.kanu.container, text="Chunk overlap ⓘ:")
+        Tooltip(l, "The number of overlapping characters between adjacent chunks.")
+        l.grid(row=6, column=0)
+        self.chunk_overlap = tk.IntVar(self.kanu.container, value=50)
+        e = tk.Entry(self.kanu.container, textvariable=self.chunk_overlap)
+        e.grid(row=6, column=1, columnspan=2)
         self.option1_button = tk.Button(self.kanu.container, text="Go with Option 1", command=self.go_with_option1)
-        self.option1_button.grid(row=5, column=0, columnspan=3)
+        self.option1_button.grid(row=7, column=0, columnspan=3)
         self.option1_button["state"] = tk.DISABLED
-        l = tk.Message(self.kanu.container, width=300, text="Option 2. Use an existing database")
-        l.grid(row=6, column=0, columnspan=3)
+        m = tk.Message(self.kanu.container, width=300, text="Option 2. Use an existing database")
+        m.grid(row=8, column=0, columnspan=3)
         l = tk.Label(self.kanu.container, text="Database ⓘ:")
         Tooltip(l, "Directory where the database is stored.")
-        l.grid(row=7, column=0)
+        l.grid(row=9, column=0)
         self.old_database_label = tk.Label(self.kanu.container, text="Not selected", fg="red")
-        self.old_database_label.grid(row=7, column=1)
+        self.old_database_label.grid(row=9, column=1)
         b = tk.Button(self.kanu.container, text="Browse", command=self.specify_old_database_directory)
-        b.grid(row=7, column=2)
+        b.grid(row=9, column=2)
         self.option2_button = tk.Button(self.kanu.container, text="Go with Option 2", command=self.go_with_option2)
-        self.option2_button.grid(row=8, column=0, columnspan=3)
+        self.option2_button.grid(row=10, column=0, columnspan=3)
         self.option2_button["state"] = tk.DISABLED
 
     def query(self):
         self.db = Chroma(persist_directory=self.database_directory, embedding_function=OpenAIEmbeddings())
-        self.qa = RetrievalQA.from_chain_type(llm=ChatOpenAI(model_name=self.model), chain_type="stuff", retriever=self.db.as_retriever())
+        self.qa = RetrievalQA.from_chain_type(
+            llm=ChatOpenAI(model_name=self.model),
+            chain_type="stuff",
+            retriever=self.db.as_retriever(),
+            chain_type_kwargs={"prompt": PromptTemplate(template=self.prompt, input_variables=["context", "question"])}
+        )
         self.kanu.container.pack_forget()
         self.kanu.container = tk.Frame(self.kanu.root)
         self.kanu.container.pack()
@@ -107,7 +125,7 @@ def go_with_option1(self):
                 loader = DOCUMENT_LOADERS[file_ext](file_path)
                 document = loader.load()[0]
                 documents.append(document)
-        text_splitter = RecursiveCharacterTextSplitter(chunk_size=1000, chunk_overlap=50)
+        text_splitter = RecursiveCharacterTextSplitter(chunk_size=self.chunk_size.get(), chunk_overlap=self.chunk_overlap.get())
         texts = text_splitter.split_documents(documents)
         db = Chroma.from_documents(texts, OpenAIEmbeddings(), persist_directory=self.database_directory)
         db.add_documents(texts)    

kanu/utils.py

@@ -15,7 +15,7 @@ def show_tooltip(self, event):
         self.tooltip = tk.Toplevel(self.widget)
         self.tooltip.wm_overrideredirect(True)
         self.tooltip.wm_geometry(f"+{x}+{y}")
-        l = tk.Label(self.tooltip, text=self.text, background="#ffffe0", relief="solid", borderwidth=1)
+        l = tk.Label(self.tooltip, text=self.text, background="#ffffe0", relief="solid", borderwidth=1, wraplength=400)
         l.pack()
 
     def hide_tooltip(self, event):

kanu/version.py

@@ -1 +1 @@
-__version__ = "0.3.0"
+__version__ = "0.4.0"