index + createchatchain

doc_generator/Utils/createChatChain.py

@@ -0,0 +1,52 @@
+from langchain.chains import LLMChain, ChatVectorDBChain, loadQAChain
+from langchain.llms import OpenAIChat
+from langchain.prompts import PromptTemplate
+
+
+# Define the prompt template for condensing the follow-up question
+condense_prompt = PromptTemplate(
+    template="Given the following conversation and a follow up question, rephrase the follow up question to be a standalone question.\n\n"
+             "Chat History:\n{chat_history}\nFollow Up Input: {question}\nStandalone question:"
+)
+
+def make_qa_prompt(project_name, repository_url, content_type, chat_prompt, target_audience):
+    additional_instructions = f"\nHere are some additional instructions for answering questions about {content_type}:\n{chat_prompt}" if chat_prompt else ""
+    return PromptTemplate(
+        template=f"You are an AI assistant for a software project called {project_name}. You are trained on all the {content_type} that makes up this project.\n"
+                 f"The {content_type} for the project is located at {repository_url}.\n"
+                 "You are given the following extracted parts of a technical summary of files in a {content_type} and a question. "
+                 "Provide a conversational answer with hyperlinks back to GitHub.\n"
+                 "You should only use hyperlinks that are explicitly listed in the context. Do NOT make up a hyperlink that is not listed.\n"
+                 "Include lots of {content_type} examples and links to the {content_type} examples, where appropriate.\n"
+                 "Assume the reader is a {target_audience} but is not deeply familiar with {project_name}.\n"
+                 "Assume the reader does not know anything about how the project is structured or which folders/files are provided in the context.\n"
+                 "Do not reference the context in your answer. Instead use the context to inform your answer.\n"
+                 "If you don't know the answer, just say \"Hmm, I'm not sure.\" Don't try to make up an answer.\n"
+                 "If the question is not about the {project_name}, politely inform them that you are tuned to only answer questions about the {project_name}.\n"
+                 "Your answer should be at least 100 words and no more than 300 words.\n"
+                 "Do not include information that is not directly relevant to the question, even if the context includes it.\n"
+                 "Always include a list of reference links to GitHub from the context. Links should ONLY come from the context.\n"
+                 f"{additional_instructions}\n"
+                 "Question: {question}\n\n"
+                 "Context:\n{context}\n\n"
+                 "Answer in Markdown:"
+    )
+
+def make_chain(project_name, repository_url, content_type, chat_prompt, target_audience, vectorstore, llms, on_token_stream=None):
+    llm = llms[1] if len(llms) > 1 else llms[0]
+    question_generator = LLMChain(
+        llm=OpenAIChat(temperature=0.1, model_name=llm),
+        prompt=condense_prompt
+    )
+
+    qa_prompt = make_qa_prompt(project_name, repository_url, content_type, chat_prompt, target_audience)
+    doc_chain = loadQAChain(
+        llm=OpenAIChat(temperature=0.2, frequency_penalty=0, presence_penalty=0, model_name=llm, streaming=bool(on_token_stream)),
+        prompt=qa_prompt
+    )
+
+    return ChatVectorDBChain(
+        vectorstore=vectorstore,
+        combine_documents_chain=doc_chain,
+        question_generator_chain=question_generator
+    )

doc_generator/Utils/index.py

@@ -0,0 +1,50 @@
+from prompt_toolkit import prompt
+from prompt_toolkit.shortcuts import clear
+import os
+from langchain.embeddings import OpenAIEmbeddings
+import hnswlib
+from markdown2 import markdown
+from createChatChain import make_chain
+
+chat_history = []
+
+def display_welcome_message(project_name):
+    print(f"Welcome to the {project_name} chatbot.")
+    print(f"Ask any questions related to the {project_name} codebase, and I'll try to help. Type 'exit' to quit.\n")
+
+def query(name, repository_url, output, content_type, chat_prompt, target_audience, llms):
+    data_path = os.path.join(output, 'docs', 'data')
+    vector_store = hnswlib.load(data_path, OpenAIEmbeddings())  # Assuming a synchronous load or pre-loaded
+    chain = make_chain(name, repository_url, content_type, chat_prompt, target_audience, vector_store, llms)
+
+    clear()
+    display_welcome_message(name)
+
+    while True:
+        question = prompt(f"How can I help with {name}?\n")
+        if question.strip().lower() == 'exit':
+            break
+
+        print('Thinking...')
+        try:
+            response = chain.call({'question': question, 'chat_history': chat_history})
+            chat_history.append((question, response['text']))
+            print('\n\nMarkdown:\n')
+            print(markdown(response['text']))
+        except Exception as error:
+            print(f"Something went wrong: {error}")
+
+if __name__ == "__main__":
+    # Example config objects, these need to be defined or imported properly
+    repo_config = {
+        "name": "ProjectName",
+        "repository_url": "https://github.com/yourproject",
+        "output": "/path/to/output",
+        "content_type": "docs",
+        "chat_prompt": "Additional instructions here",
+        "target_audience": "developers"
+    }
+    user_config = {
+        "llms": ["gpt-3.5-turbo"]
+    }
+    query(**repo_config, **user_config)