Create pages/Phase1.py

pages/Phase1.py

@@ -0,0 +1,109 @@
+import streamlit as st
+from huggingface_hub import InferenceClient
+import os
+import fitz  # PyMuPDF
+
+st.title("ChatGPT-like Chatbot")
+
+base_url = "https://api-inference.huggingface.co/models/"
+
+API_KEY = os.environ.get('HUGGINGFACE_API_KEY')
+headers = {"Authorization": "Bearer " + str(API_KEY)}
+
+model_links = {
+    "Mistral-7B": base_url + "mistralai/Mistral-7B-Instruct-v0.2",
+    "Mistral-22B": base_url + "mistral-community/Mixtral-8x22B-v0.1",
+}
+
+model_info = {
+    "Mistral-7B": {
+        'description': """The Mistral model is a **Large Language Model (LLM)** that's able to have question and answer interactions.\n \
+            \nIt was created by the [**Mistral AI**](https://mistral.ai/news/announcing-mistral-7b/) team as has over  **7 billion parameters.** \n""",
+        'logo': 'https://mistral.ai/images/logo_hubc88c4ece131b91c7cb753f40e9e1cc5_2589_256x0_resize_q97_h2_lanczos_3.webp'},
+    "Zephyr-7B": {
+        'description': """The Zephyr model is a **Large Language Model (LLM)** that's able to have question and answer interactions.\n \
+         \nFrom Huggingface: \n\
+         Zephyr is a series of language models that are trained to act as helpful assistants. \
+         [Zephyr 7B Gemma](https://huggingface.co/HuggingFaceH4/zephyr-7b-gemma-v0.1)\
+         is the third model in the series, and is a fine-tuned version of google/gemma-7b \
+         that was trained on on a mix of publicly available, synthetic datasets using Direct Preference Optimization (DPO)\n""",
+        'logo': 'https://huggingface.co/HuggingFaceH4/zephyr-7b-gemma-v0.1/resolve/main/thumbnail.png'}
+}
+
+def format_prompt(message, custom_instructions=None):
+    prompt = ""
+    if custom_instructions:
+        prompt += f"[INST] {custom_instructions} [/INST]"
+    prompt += f"[INST] {message} [/INST]"
+    return prompt
+
+def reset_conversation():
+    st.session_state.conversation = []
+    st.session_state.messages = []
+    return None
+
+def read_pdf(file_path):
+    doc = fitz.open(file_path)
+    text = ""
+    for page in doc:
+        text += page.get_text()
+    return text
+
+models = [key for key in model_links.keys()]
+
+selected_model = st.sidebar.selectbox("Select Model", models)
+temp_values = st.sidebar.slider('Select a temperature value', 0.0, 1.0, (0.5))
+st.sidebar.button('Reset Chat', on_click=reset_conversation)
+st.sidebar.write(f"You're now chatting with **{selected_model}**")
+st.sidebar.markdown(model_info[selected_model]['description'])
+st.sidebar.image(model_info[selected_model]['logo'])
+st.sidebar.markdown("*Generated content may be inaccurate or false.*")
+
+if "prev_option" not in st.session_state:
+    st.session_state.prev_option = selected_model
+
+if st.session_state.prev_option != selected_model:
+    st.session_state.messages = []
+    st.session_state.prev_option = selected_model
+    reset_conversation()
+
+repo_id = model_links[selected_model]
+st.subheader(f'AI - {selected_model}')
+
+if "messages" not in st.session_state:
+    st.session_state.messages = []
+
+for message in st.session_state.messages:
+    with st.chat_message(message["role"]):
+        st.markdown(message["content"])
+
+pdf_path = st.file_uploader("Upload a PDF file", type="pdf")
+
+if pdf_path:
+    pdf_text = read_pdf(pdf_path)
+
+if prompt := st.chat_input(f"Hi I'm {selected_model}, ask me a question"):
+    custom_instruction = "Act like a Human in conversation"
+
+    with st.chat_message("user"):
+        st.markdown(prompt st.session_state.messages.append({"role": "user", "content": prompt})
+
+    formatted_prompt = format_prompt(prompt, custom_instruction)
+
+    with st.chat_message("assistant"):
+        with st.spinner("Thinking..."):
+            client = InferenceClient(model=repo_id, token=API_KEY)
+
+            if pdf_path:
+                # If a PDF is uploaded, use its text for answering
+                formatted_prompt = f"{pdf_text}\n\n{formatted_prompt}"
+
+            response = client.text_generation(
+                formatted_prompt, 
+                max_length=500, 
+                temperature=temp_values
+            )
+            
+            response_content = response["generated_text"]
+            st.markdown(response_content)
+            st.session_state.messages.append({"role": "assistant", "content": response_content})