Update app.py

app.py

@@ -1,50 +1,270 @@
-import streamlit as st
+import os
+import time
+import uuid
+from typing import List, Tuple, Optional, Dict, Union
+
 import google.generativeai as genai
-from streamlit import file_uploader
+import streamlit as st
+from PIL import Image
+
+print("google-generativeai:", genai.__version__)
+
+GOOGLE_API_KEY = os.environ.get("GOOGLE_API_KEY")
+
+TITLE = """<h1 align="center">Gemini Playground 💬</h1>"""
+SUBTITLE = """<h2 align="center">Play with Gemini Pro and Gemini Pro Vision API</h2>"""
+DUPLICATE = """
+<div style="text-align: center; display: flex; justify-content: center; align-items: center;">
+    <a href="https://huggingface.co/spaces/SkalskiP/ChatGemini?duplicate=true">
+        <img src="https://bit.ly/3gLdBN6" alt="Duplicate Space" style="margin-right: 10px;">
+    </a>
+    <span>Duplicate the Space and run securely with your 
+        <a href="https://makersuite.google.com/app/apikey">GOOGLE API KEY</a>.
+    </span>
+</div>
+"""
+
+AVATAR_IMAGES = (
+    None,
+    "https://media.roboflow.com/spaces/gemini-icon.png"
+)
+
+IMAGE_CACHE_DIRECTORY = "/tmp"
+IMAGE_WIDTH = 512
+CHAT_HISTORY = List[Tuple[Optional[Union[Tuple[str], str]], Optional[str]]]
+
+
+def preprocess_stop_sequences(stop_sequences: str) -> Optional[List[str]]:
+    if not stop_sequences:
+        return None
+    return [sequence.strip() for sequence in stop_sequences.split(",")]
+
+
+def preprocess_image(image: Image.Image) -> Optional[Image.Image]:
+    image_height = int(image.height * IMAGE_WIDTH / image.width)
+    return image.resize((IMAGE_WIDTH, image_height))
+
+
+def cache_pil_image(image: Image.Image) -> str:
+    image_filename = f"{uuid.uuid4()}.jpeg"
+    os.makedirs(IMAGE_CACHE_DIRECTORY, exist_ok=True)
+    image_path = os.path.join(IMAGE_CACHE_DIRECTORY, image_filename)
+    image.save(image_path, "JPEG")
+    return image_path
+
+
+def preprocess_chat_history(
+    history: CHAT_HISTORY
+) -> List[Dict[str, Union[str, List[str]]]]:
+    messages = []
+    for user_message, model_message in history:
+        if isinstance(user_message, tuple):
+            pass
+        elif user_message is not None:
+            messages.append({'role': 'user', 'parts': [user_message]})
+        if model_message is not None:
+            messages.append({'role': 'model', 'parts': [model_message]})
+    return messages
+
 
-# Generative AI setup
-api_key = "AIzaSyC70u1sN87IkoxOoIj4XCAPw97ae2LZwNM"   
-genai.configure(api_key=api_key)
+def upload(files: Optional[List[str]], chatbot: CHAT_HISTORY) -> CHAT_HISTORY:
+    for file in files:
+        image = Image.open(file).convert('RGB')
+        image = preprocess_image(image)
+        image_path = cache_pil_image(image)
+        chatbot.append(((image_path,), None))
+    return chatbot
 
-generation_config = {
-  "temperature": 0.9,
-  "max_output_tokens": 3000  
-}
 
-safety_settings = []
+def user(text_prompt: str, chatbot: CHAT_HISTORY):
+    if text_prompt:
+        chatbot.append((text_prompt, None))
+    return "", chatbot
+
+
+def bot(
+    google_key: str,
+    files: Optional[List[str]],
+    temperature: float,
+    max_output_tokens: int,
+    stop_sequences: str,
+    top_k: int,
+    top_p: float,
+    chatbot: CHAT_HISTORY
+):
+    if len(chatbot) == 0:
+        return chatbot
+
+    google_key = google_key if google_key else GOOGLE_API_KEY
+    if not google_key:
+        raise ValueError(
+            "GOOGLE_API_KEY is not set. "
+            "Please follow the instructions in the README to set it up.")
+
+    genai.configure(api_key=google_key)
+    generation_config = genai.types.GenerationConfig(
+        temperature=temperature,
+        max_output_tokens=max_output_tokens,
+        stop_sequences=preprocess_stop_sequences(stop_sequences=stop_sequences),
+        top_k=top_k,
+        top_p=top_p)
+
+    if files:
+        text_prompt = [chatbot[-1][0]] \
+            if chatbot[-1][0] and isinstance(chatbot[-1][0], str) \
+            else []
+        image_prompt = [Image.open(file).convert('RGB') for file in files]
+        model = genai.GenerativeModel('gemini-pro-vision')
+        response = model.generate_content(
+            text_prompt + image_prompt,
+            stream=True,
+            generation_config=generation_config)
+    else:
+        messages = preprocess_chat_history(chatbot)
+        model = genai.GenerativeModel('gemini-pro')
+        response = model.generate_content(
+            messages,
+            stream=True,
+            generation_config=generation_config)
+
+    # streaming effect
+    chatbot[-1][1] = ""
+    for chunk in response:
+        for i in range(0, len(chunk.text), 10):
+            section = chunk.text[i:i + 10]
+            chatbot[-1][1] += section
+            time.sleep(0.01)
+            yield chatbot
+
 
 # Streamlit UI
-st.title("Chatbot")
+st.markdown(TITLE, unsafe_allow_html=True)
+st.markdown(SUBTITLE, unsafe_allow_html=True)
+st.markdown(DUPLICATE, unsafe_allow_html=True)
+
+# Sidebar for parameters
+st.sidebar.header("Parameters")
+google_key = st.sidebar.text_input(
+    label="GOOGLE API KEY",
+    value="",
+    type="password",
+    help="You have to provide your own GOOGLE_API_KEY for this app to function properly",
+    key="google_key"
+)
+temperature = st.sidebar.slider(
+    label="Temperature",
+    min_value=0.0,
+    max_value=1.0,
+    value=0.4,
+    step=0.05,
+    help=(
+        "Temperature controls the degree of randomness in token selection. Lower "
+        "temperatures are good for prompts that expect a true or correct response, "
+        "while higher temperatures can lead to more diverse or unexpected results. "
+    ),
+    key="temperature"
+)
+max_output_tokens = st.sidebar.slider(
+    label="Token limit",
+    min_value=1,
+    max_value=2048,
+    value=1024,
+    step=1,
+    help=(
+        "Token limit determines the maximum amount of text output from one prompt. A "
+        "token is approximately four characters. The default value is 2048."
+    ),
+    key="max_output_tokens"
+)
+stop_sequences = st.sidebar.text_input(
+    label="Add stop sequence",
+    value="",
+    help=(
+        "A stop sequence is a series of characters (including spaces) that stops "
+        "response generation if the model encounters it. The sequence is not included "
+        "as part of the response. You can add up to five stop sequences."
+    ),
+    key="stop_sequences"
+)
+top_k = st.sidebar.slider(
+    label="Top-K",
+    min_value=1,
+    max_value=40,
+    value=32,
+    step=1,
+    help=(
+        "Top-k changes how the model selects tokens for output. A top-k of 1 means the "
+        "selected token is the most probable among all tokens in the model’s "
+        "vocabulary (also called greedy decoding), while a top-k of 3 means that the "
+        "next token is selected from among the 3 most probable tokens (using "
+        "temperature)."
+    ),
+    key="top_k"
+)
+top_p = st.sidebar.slider(
+    label="Top-P",
+    min_value=0.0,
+    max_value=1.0,
+    value=1.0,
+    step=0.01,
+    help=(
+        "Top-p changes how the model selects tokens for output. Tokens are selected "
+        "from most probable to least until the sum of their probabilities equals the "
+        "top-p value. For example, if tokens A, B, and C have a probability of .3, .2, "
+        "and .1 and the top-p value is .5, then the model will select either A or B as "
+        "the next token (using temperature). "
+    ),
+    key="top_p"
+)
 
-# Use text_input for text input by typing
-user_text = st.text_input("Type your text here:")
+# Main area for chatbot
+st.header("Chatbot")
+chatbot = st.session_state.get("chatbot", [])
+if len(chatbot) % 2 == 0:
+    role = "user"
+else:
+    role = "model"
 
-# Use file_uploader for image input
-user_image = st.file_uploader("Upload an image file here", type=["png", "jpg", "jpeg"])
+for user_message, model_message in chatbot:
+    if isinstance(user_message, tuple):
+        st.image(user_message[0], use_column_width=True)
+    elif user_message is not None:
+        st.markdown(f"**User:** {user_message}")
+    if model_message is not None:
+        st.markdown(f"**Model:** {model_message}")
 
-# Check if the user has entered text or uploaded an image
-if user_text or user_image:
-    # Create the prompt parts accordingly
+# Text input for user message
+user_text = st.text_input("Type your text here:", key="user_text")
+
+# File uploader for user image
+user_image = st.file_uploader("Upload an image file here", type=["png", "jpg", "jpeg"], key="user_image")
+
+# Button for running the bot
+run_button = st.button("Run", key="run_button")
+
+# Logic for handling user input and bot response
+if run_button or user_text or user_image:
+    # Append user input to chatbot history
     if user_text:
-        prompt_parts = [user_text]
-        model_name = "gemini-pro" # Use the text-only model
-    else:
-        prompt_parts = [{
-            "mime_type": user_image.type,
-            "data": user_image.read() 
-        }]
-        model_name = "gemini-pro-vision" # Use the multimodal model
-    
-    # Model code
-    model = genai.GenerativeModel(
-        model_name=model_name,   
-        generation_config=generation_config,
-        safety_settings=safety_settings
-    )
-    
-    response = model.generate_content(prompt_parts) 
-    response_text = response.text
+        chatbot.append((user_text, None))
+    elif user_image:
+        image = Image.open(user_image).convert('RGB')
+        image = preprocess_image(image)
+        image_path = cache_pil_image(image)
+        chatbot.append(((image_path,), None))
     
+    # Call the bot function with parameters and chatbot history
+    bot_response = bot(
+        google_key=google_key,
+        files=None,
+        temperature
     # Display the user input and the model response
-    st.markdown(f"**User:** {prompt_parts[0]['text']}")
+    if user_text:
+        st.markdown(f"**User:** {user_text}")
+    elif user_image:
+        st.image(user_image, use_column_width=True)
     st.markdown(f"**Model:** {response_text}")
+    
+    # Update the chatbot history with the model response
+    chatbot[-1][1] = response_text
+    st.session_state["chatbot"] = chatbot