6.0.0 initial commit | built on LangChain v1.0 pre-release

.streamlit/template.secrets.toml

@@ -1,3 +1,6 @@
+[DEEPSEEK]
+DEEPSEEK_API_KEY = ""
+
 [OPENAI]
 openai_api_key = ""
 

README.md

@@ -1,10 +1,10 @@
 # FreeStream
 
-Build your own personal chatbot interface with Streamlit!
+Streamlit Chatbots - Work in Progress
 
 ***TLDR***:
-- A repository to help you get started building your own chatbots using LangChain
-- All conversation content is traced via LangSmith for developer evaluation
+- Build on top of an already robust chatbot that reasons before responding, and has access to tools
+- Customize a chatbot to respond exactly how you want
 - Hostable for free through Streamlit Community Cloud
 - API keys required
 - Pay-per-use ChatGPT style interface
@@ -14,10 +14,6 @@ Build your own personal chatbot interface with Streamlit!
 - [Quickstart](#quickstart)
   - [Installation](#installation)
 - [Description](#description)
-  - [Key Concepts](#key-concepts)
-  - [What can I do with FreeStream?](#what-can-i-do-with-freestream)
-- [Functional Requirements](#functional-requirements)
-- [Non-Functional Requirements](#non-functional-requirements)
 - [License](./LICENSE)
 - [LLM Providers' Privacy Policies](#llm-providers-privacy-policies)
 
@@ -42,16 +38,9 @@ Run:
 poetry install
 ```
 
-You will need to set all required secrets, which require their own respective accounts.
+You will need to set all required secrets(API keys), which require their own respective accounts.
 Make a copy of "template.secrets.toml" and rename it to "secrets.toml" in the root of the project. Fill out each field in the file.
 
-**Need API Keys?**
-| **API Platform** | **Link** |
-| ---- | ---------- |
-| Claude | https://console.anthropic.com/ |
-| Google | https://aistudio.google.com/app/apikey |
-| Langchain | https://smith.langchain.com/ |
-| OpenAI | https://platform.openai.com/api-keys |
 
 You can then start the development server with hot reloading by running:
 
@@ -62,11 +51,9 @@ poetry run streamlit run ./freestream/🏡_Home.py
 ---
 
 ## Description
-I originally created this project as a chatbot for law and medical professionals, but I quickly realized a more flexible system would benefit everyone.
-
-#### **Key Concepts**
+Just a freaking robot.
 
-*Related to extending the capabilities of generative AI.*
+*Here's some papers I found interesting when first learning about generative AI and "augmented generation."*
 | **Concept** | **Definition** |
 | ---- | ---------- |
 | [Large Language Model](https://en.wikipedia.org/wiki/Large_language_model "Wikipedia: Large language model") | A model that can generate text. |
@@ -76,34 +63,12 @@ I originally created this project as a chatbot for law and medical professionals
 | [ColBERT](https://arxiv.org/abs/2004.12832 "Arxiv: 2004.12832") | Efficient BERT-Based Document Search |
 | [RAPTOR](https://arxiv.org/abs/2401.18059 "Arxiv: 2401.18059") | Recursive Abstractive Processing for Tree-Organized Retrieval |
 
-### What can I do with FreeStream?
-
-FreeStream has two chatbots where you can interact with an LLM of your choosing, for example, GPT-4o or Claude Opus. You can very easily add more LLMs to the chatbot dictionary, like Llama 3 via [Ollama](https://ollama.com/ "Ollama home page"), or Gemini-Pro through LangChain's [`ChatGoogleGenerativeAI`](https://api.python.langchain.com/en/latest/chat_models/langchain_google_genai.chat_models.ChatGoogleGenerativeAI.html "LangChain API Docs"). The original chatbot for this project was "RAGbot," which allows you to ask questions about your upload file(s). Curie, is a more for programming and self-learning purposes.
-
-#### Functional Requirements
-
-The application **MUST**...
-1. Provide a user interface for chatting with large language models.
-2. Have a retrieval augmented generative chatbot.
-3. Provide a range of chatbot pages, differentiated by their prompt engineering.
-4. Let the user "drop-in" their choice of LLM at any time during a conversation.
-5. ~~Allow users to perform image upscaling (PDF, JPEG, PNG) without limits.~~
-
-#### Non-Functional Requirements
-
-The application **SHOULD**...
-1. Aim for 24/7 availability.
-2. Prioritize ease of navigation
-3. Feature a visually appealing, seamless interface.
-
 ---
 
 # [License](./LICENSE)
 
 # LLM Providers' Privacy Policies
 
-- [OpenAI Privacy Policy](https://openai.com/policies/privacy-policy)
-- [Google](https://transparency.google/our-policies/privacy-policy-terms-of-service/ "Was unable to find a privacy policy specific to Google AI Studio.")
-- [Anthropic](https://support.anthropic.com/en/articles/7996866-how-long-do-you-store-personal-data "Support forum response that may suddenly be obsoleted.")
+- [DeepSeek Privacy Policy](https://cdn.deepseek.com/policies/en-US/deepseek-privacy-policy.html)
 - [Streamlit](https://streamlit.io/privacy-policy/)
 

freestream/pages/1_💬_Curie.py → freestream/pages/.archive/1_💬_Curie.py

@@ -4,7 +4,6 @@
 import streamlit as st
 from langchain.chains import ConversationalRetrievalChain
 from langchain.memory import ConversationBufferMemory
-from langchain_anthropic import ChatAnthropic
 from langchain_community.chat_message_histories import \
     StreamlitChatMessageHistory
 from langchain_core.prompts import ChatPromptTemplate, MessagesPlaceholder
@@ -108,38 +107,38 @@
     # )
 }
 
-anthropic_models = {
-    "Claude: Haiku": ChatAnthropic(
-        model="claude-3-haiku-20240307",
-        anthropic_api_key=anthropic_api_key,
-        temperature=temperature_slider,
-        streaming=True,
-        max_tokens=4096,
-    ),
-    "Claude 3.5: Sonnet": ChatAnthropic(
-        model="claude-3-5-sonnet-20240620",
-        anthropic_api_key=anthropic_api_key,
-        temperature=temperature_slider,
-        streaming=True,
-        max_tokens=4096,
-    ),
-    # "Claude: Opus": ChatAnthropic(
-    #     model="claude-3-opus-20240229",
-    #     anthropic_api_key=anthropic_api_key,
-    #     temperature=temperature_slider,
-    #     streaming=True,
-    #     max_tokens=4096,
-    # ),
-}
+# anthropic_models = {
+#     "Claude: Haiku": ChatAnthropic(
+#         model="claude-3-haiku-20240307",
+#         anthropic_api_key=anthropic_api_key,
+#         temperature=temperature_slider,
+#         streaming=True,
+#         max_tokens=4096,
+#     ),
+#     "Claude 3.5: Sonnet": ChatAnthropic(
+#         model="claude-3-5-sonnet-20240620",
+#         anthropic_api_key=anthropic_api_key,
+#         temperature=temperature_slider,
+#         streaming=True,
+#         max_tokens=4096,
+#     ),
+#     "Claude: Opus": ChatAnthropic(
+#         model="claude-3-opus-20240229",
+#         anthropic_api_key=anthropic_api_key,
+#         temperature=temperature_slider,
+#         streaming=True,
+#         max_tokens=4096,
+#     ),
+# }
 
 # Master dictionary
 model_names = {}
 
 # Update model master dictionary based on present API keys
 if openai_api_key:
     model_names.update(openai_models)
-if anthropic_api_key:
-    model_names.update(anthropic_models)
+# if anthropic_api_key:
+#     model_names.update(anthropic_models)
 
 # Create a dropdown menu for selecting a chat model
 selected_model = st.selectbox(

freestream/pages/.archive/🏡_Home.py

@@ -0,0 +1,89 @@
+import datetime
+import streamlit as st
+from langchain_deepseek import ChatDeepSeek
+from langchain_core.messages import HumanMessage, AIMessage, SystemMessage
+from freestream import footer
+from pages import save_conversation_history
+
+st.set_page_config(
+    page_title="FreeStream: A basic chatbot to build on.", page_icon="🏡"
+)
+
+st.title("FreeStream")
+st.header(":green[_A basic chatbot to build on._]", divider="red")
+st.markdown(footer, unsafe_allow_html=True)
+
+DEEPSEEK_API_KEY = st.sidebar.text_input("DeepSeek API Key", type="password")
+# Button to clear conversation history
+if st.sidebar.button("Clear message history", use_container_width=True):
+    st.session_state.clear()
+
+
+# Initialize chat model (but only if API key is provided)
+if DEEPSEEK_API_KEY:
+    chat_model = ChatDeepSeek(
+        temperature=0.5,
+        api_key=DEEPSEEK_API_KEY,
+        model="deepseek-chat",
+        max_tokens=128,
+        streaming=True
+    )
+
+# Initialize session state for messages
+if "messages" not in st.session_state:
+    st.session_state.messages = []
+
+
+# Display chat history
+for message in st.session_state.messages:
+    with st.chat_message(message["role"]):
+        st.markdown(message["content"])
+
+# Save the formatted conversation history to a variable
+formatted_history = save_conversation_history(st.session_state.messages)
+current_time = datetime.datetime.now().strftime("%Y-%m-%d %H:%M:%S")
+# Create a sidebar button to download the conversation history
+st.sidebar.download_button(
+    label="Download conversation history",
+    data=formatted_history,
+    file_name=f"conversation_history {current_time}.md",
+    mime="text/markdown",
+    key="download_conversation_history_button",
+    help="Download the conversation history as a text file with some formatting.",
+    use_container_width=True,   
+)
+
+# Handle user input
+if user_input := st.chat_input("type here<3"):
+    # Display user message
+    with st.chat_message("user"):
+        st.markdown(user_input)
+
+    # Add user message to history
+    st.session_state.messages.append({"role": "user", "content": user_input})
+
+    # Display assistant response
+    with st.chat_message("assistant"):
+        message_placeholder = st.empty()
+        full_response = ""
+
+        # Convert messages to LangChain format
+        lc_messages = []
+        for msg in st.session_state.messages:
+            if msg["role"] == "user":
+                lc_messages.append(HumanMessage(content=msg["content"]))
+            elif msg["role"] == "assistant":
+                lc_messages.append(AIMessage(content=msg["content"]))
+            elif msg["role"] == "system":
+                lc_messages.append(SystemMessage(content=msg["content"]))
+
+        # Stream the response
+        for chunk in chat_model.stream(lc_messages):
+            if hasattr(chunk, 'content'):
+                full_response += chunk.content
+                message_placeholder.markdown(full_response + "▌")
+
+        message_placeholder.markdown(full_response)
+
+    # Add assistant response to history
+    st.session_state.messages.append({"role": "assistant", "content": full_response})

freestream/pages/utils/chatbot_operators.py

@@ -5,7 +5,6 @@
 from typing import List
 
 import streamlit as st
-import torch
 from langchain.text_splitter import RecursiveCharacterTextSplitter
 from langchain_community.document_loaders import UnstructuredFileLoader
 from langchain_community.embeddings import HuggingFaceEmbeddings
@@ -47,7 +46,8 @@ def __init__(self):
         )
         self.embeddings = HuggingFaceEmbeddings(
             model_name="all-MiniLM-L6-v2",
-            model_kwargs={"device": "cuda" if torch.cuda.is_available() else "cpu"},
+            # model_kwargs={"device": "cuda" if torch.cuda.is_available() else "cpu"},
+            model_kwargs={"device": "cpu"},
         )
 
     @st.cache_resource(ttl="1h")

freestream/pages/utils/streamlit_operators.py

@@ -4,7 +4,7 @@
 import streamlit as st
 
 # Define a function to change the background to an image via URL
-# https://discuss.streamlit.io/t/how-do-i-use-a-background-image-on-streamlit/5067/19?u=daethyra
+# https://discuss.streamlit.io/t/how-do-i-use-a-background-image-on-streamlit/5067/19
 def set_bg_url():
     """
     A function to unpack an image from url and set as bg.
@@ -27,7 +27,7 @@ def set_bg_url():
 
 
 # Define a function to change the background to a local image
-# https://discuss.streamlit.io/t/how-do-i-use-a-background-image-on-streamlit/5067/16?u=daethyra
+# https://discuss.streamlit.io/t/how-do-i-use-a-background-image-on-streamlit/5067/16
 def set_bg_local(main_bg):
     """
     A function to unpack an image from root folder and set as bg.