import os import streamlit as st import json import matplotlib.pyplot as plt import matplotlib.gridspec as gridspec import matplotlib.animation as animation import time from PIL import Image from streamlit_image_comparison import image_comparison import numpy as np import re #import chromadb from textwrap import dedent import google.generativeai as genai api_key = os.environ["OPENAI_API_KEY"] from openai import OpenAI # Initialize OpenAI client and create embeddings oai_client = OpenAI() import numpy as np # Assuming chromadb and TruLens are correctly installed and configured #from chromadb.utils.embedding_functions import # Google Langchain from langchain_google_genai import GoogleGenerativeAI #Crew imports from crewai import Agent, Task, Crew, Process # Retrieve API Key from Environment Variable GOOGLE_AI_STUDIO = os.environ.get('GOOGLE_API_KEY') # Ensure the API key is available if not GOOGLE_AI_STUDIO: raise ValueError("API key not found. Please set the GOOGLE_AI_STUDIO2 environment variable.") # Set gemini_llm gemini_llm = GoogleGenerativeAI(model="gemini-pro", google_api_key=GOOGLE_AI_STUDIO) # CrewAI ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ # Tool import from crewai.tools.gemini_tools import GeminiSearchTools from crewai import Agent, Task, Crew, Process def crewai_process_gemini(research_topic): # Define your agents with roles and goals GeminiAgent = Agent( role='Story Writer', goal='To create a story from bullet points.', backstory="""You are an expert writer that understands how to make the average extraordinary on paper """, verbose=True, allow_delegation=False, llm = gemini_llm, tools=[ GeminiSearchTools.gemini_search ] ) # Create tasks for your agents task1 = Task( description=f"""From {research_topic} create your story by writing at least one sentence about each bullet point and make sure you have a transitional statement between scenes . BE VERBOSE.""", agent=GeminiAgent ) # Instantiate your crew with a sequential process crew = Crew( agents=[GeminiAgent], tasks=[task1], verbose=2, process=Process.sequential ) # Get your crew to work! result = crew.kickoff() return result st.set_page_config(layout="wide") # Animation Code +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ # HIN Number +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ from SPARQLWrapper import SPARQLWrapper, JSON from streamlit_agraph import agraph, TripleStore, Node, Edge, Config import json # Function to load JSON data def load_data(filename): with open(filename, 'r') as file: data = json.load(file) return data # Dictionary for color codes color_codes = { "residential": "#ADD8E6", "commercial": "#90EE90", "community_facilities": "#FFFF00", "school": "#FFFF00", "healthcare_facility": "#FFFF00", "green_space": "#90EE90", "utility_infrastructure": "#90EE90", "emergency_services": "#FF0000", "cultural_facilities": "#D8BFD8", "recreational_facilities": "#D8BFD8", "innovation_center": "#90EE90", "elderly_care_home": "#FFFF00", "childcare_centers": "#FFFF00", "places_of_worship": "#D8BFD8", "event_spaces": "#D8BFD8", "guest_housing": "#FFA500", "pet_care_facilities": "#FFA500", "public_sanitation_facilities": "#A0A0A0", "environmental_monitoring_stations": "#90EE90", "disaster_preparedness_center": "#A0A0A0", "outdoor_community_spaces": "#90EE90", # Add other types with their corresponding colors } #text +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ query = """ ***Introduction*** On his first day at Quantum Data Institute in Green Open Data City Aya, Elian marveled at the city’s harmonious blend of technology and nature. Guided to his mentor, Dr. Maya Lior, a pioneer in urban data ecosystems, their discussion quickly centered on Aya’s innovative design. Dr. Lior explained data analytics and green technologies were intricately woven into the city's infrastructure, and how they used a Custom GPT called Green Data City to create the design. To interact with the Custon GPT Green Data City design tool click the button below. Additionally, to see how it was built toggle the Explanation of Custom GPT "Create Green Data City" button. """ query2 = """ ***Global Citizen*** Elian and Dr. Maya Lior's journey to the Cultural Center,a beacon of sustainability and technological integration. Equipped with cutting-edge environmental monitoring sensors, occupancy detectors, and smart lighting systems, the center is a hub for innovation in resource management and climate action. There, they were greeted by Mohammad, a dedicated environmental scientist who, despite the language barrier, shared their passion for creating a sustainable future. Utilizing the Cohere translator, they engaged in a profound dialogue, seamlessly bridging the gap between languages. Their conversation, rich with ideas and insights on global citizenship and collaborative efforts to tackle climate change and resource scarcity, underscored the imperative of unity and innovation in facing the challenges of our time. This meeting, a melting pot of cultures and disciplines, symbolized the global commitment required to sustain our planet. As Elain is using the Cohere translator he wonders how to best utilize its efficiently. He studies a Custom GPT called Conversation Analzer. It translates a small portion of the message your sending so you can be comfortable that the essescene of what your are saying is being sent and aides in learning the language. It's mantra is language is not taught but caught. To try out the Custon GPT Conversation Anzylizer click the button below. Additionally, to see how it was built toggle the Explanation of Custom GPT "Conversation Analyzer" button. """ # Function to draw the grid with optional highlighting def draw_grid(data, highlight_coords=None): fig, ax = plt.subplots(figsize=(12, 12)) nrows, ncols = data['size']['rows'], data['size']['columns'] ax.set_xlim(0, ncols) ax.set_ylim(0, nrows) ax.set_xticks(range(ncols+1)) ax.set_yticks(range(nrows+1)) ax.grid(True) # Draw roads with a specified grey color road_color = "#606060" # Light grey; change to "#505050" for dark grey for road in data.get('roads', []): # Check for roads in the data start, end = road['start'], road['end'] # Determine if the road is vertical or horizontal based on start and end coordinates if start[0] == end[0]: # Vertical road for y in range(min(start[1], end[1]), max(start[1], end[1]) + 1): ax.add_patch(plt.Rectangle((start[0], nrows-y-1), 1, 1, color=road['color'])) else: # Horizontal road for x in range(min(start[0], end[0]), max(start[0], end[0]) + 1): ax.add_patch(plt.Rectangle((x, nrows-start[1]-1), 1, 1, color=road['color'])) # Draw buildings for building in data['buildings']: coords = building['coords'] b_type = building['type'] size = building['size'] color = color_codes.get(b_type, '#FFFFFF') # Default color is white if not specified if highlight_coords and (coords[0], coords[1]) == tuple(highlight_coords): highlighted_color = "#FFD700" # Gold for highlighting ax.add_patch(plt.Rectangle((coords[1], nrows-coords[0]-size), size, size, color=highlighted_color, edgecolor='black', linewidth=2)) else: ax.add_patch(plt.Rectangle((coords[1], nrows-coords[0]-size), size, size, color=color, edgecolor='black', linewidth=1)) ax.text(coords[1]+0.5*size, nrows-coords[0]-0.5*size, b_type, ha='center', va='center', fontsize=8, color='black') ax.set_xlabel('Columns') ax.set_ylabel('Rows') ax.set_title('Village Layout with Color Coding') return fig # Title ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ # Tabs +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ # Create the main app with three tabs tab1, tab2, tab3, tab4 = st.tabs(["Introduction","Global Citizen", "Green Village", "Control Room"]) with tab1: st.header("A day in the Life of Aya Green Data City") # Creating columns for the layout col1, col2 = st.columns([1, 2]) # Displaying the image in the left column with col1: image = Image.open('./data/intro_image.jpg') st.image(image, caption='Aya Green Data City') # Displaying the text above on the right with col2: st.markdown(query) # Displaying the audio player below the text voice_option = st.selectbox( 'Choose a voice:', ['alloy', 'echo', 'fable', 'onyx', 'nova', 'shimmer'] ) if st.button('Convert to Speech'): if query: try: response = oai_client.audio.speech.create( model="tts-1", voice=voice_option, input=query, ) # Stream or save the response as needed # For demonstration, let's assume we save then provide a link for downloading audio_file_path = "output.mp3" response.stream_to_file(audio_file_path) # Display audio file to download st.audio(audio_file_path, format='audio/mp3') st.success("Conversion successful!") except Exception as e: st.error(f"An error occurred: {e}") else: st.error("Please enter some text to convert.") st.header("Custom GPT Engineering Tools") st.link_button("Custom GPT Green Data City Creation Tool (Population 10,000 to 50,000)", "https://chat.openai.com/g/g-4bPJUaHS8-create-a-green-data-village") if st.button('Show/Hide Explanation of "Custom GPT Create Green Data City"'): # Toggle visibility st.session_state.show_instructions = not st.session_state.get('show_instructions', False) # Check if the instructions should be shown if st.session_state.get('show_instructions', False): st.write(""" On clicking "Create Data Village" create a Green Data Village following the 5 Steps below. Output a JSON file similar to the Example by completing the five Steps. To generate the provided JSON code, I would instruct a custom GPT to create a detailed description of a hypothetical smart city layout, named "Green Smart Village", starting with a population of 10,000 designed to grow to 50,000. This layout should include a grid size of 21x21, a list of buildings and roads, each with specific attributes: **Step 1:** General Instructions: Generate a smart city layout for "Green Smart Village" with a 21x21 grid. Include a population of 10,000 designed to grow to 50,000. **Step 2:** Buildings: For each building, specify its coordinates on the grid, type (e.g., residential, commercial, healthcare facility), size (in terms of the grid), color, and equipped sensors (e.g., smart meters, water flow sensors). Types of buildings should vary and include residential, commercial, community facilities, school, healthcare facility, green space, utility infrastructure, emergency services, cultural facilities, recreational facilities, innovation center, elderly care home, childcare centers, places of worship, event spaces, guest housing, pet care facilities, public sanitation facilities, environmental monitoring stations, disaster preparedness center, outdoor community spaces, typical road, and typical road crossing. **Step 3:** Assign each building unique sensors based on its type, ensuring a mix of technology like smart meters, occupancy sensors, smart lighting systems, and environmental monitoring sensors. **Step 4:** Roads: Detail the roads' start and end coordinates, color, and sensors installed. Ensure roads connect significant areas of the city, providing access to all buildings. Equip roads with sensors for traffic flow, smart streetlights, and pollution monitoring. MAKE SURE ALL BUILDINGS HAVE ACCESS TO A ROAD. This test scenario would evaluate the model's ability to creatively assemble a smart city plan with diverse infrastructure and technology implementations, reflecting real-world urban planning challenges and the integration of smart technologies for sustainable and efficient city management. Example: { "city": "City Name", "population": "Population Size", "size": { "rows": "Number of Rows", "columns": "Number of Columns" }, "buildings": [ { "coords": ["X", "Y"], "type": "Building Type", "size": "Building Size", "color": "Building Color", "sensors": ["Sensor Types"] } ], "roads": [ { "start": ["X Start", "Y Start"], "end": ["X End", "Y End"], "color": "Road Color", "sensors": ["Sensor Types"] } ] } **Step 5:** Finally create a Dalle image FOR EACH BUILDING in the JSON file depicting what a user will experience there in this green open data city including sensors. LABEL EACH IMAGE. """) with tab2: st.header("Becoming a Global Citizen") # Creating columns for the layout col1, col2 = st.columns([1, 2]) # Displaying the image in the left column with col1: image = Image.open('./data/global_image.jpg') st.image(image, caption='Cultural Center Cohere Translator') # Displaying the text above on the right with col2: st.markdown(query2) # Displaying the audio player below the text voice_option = st.selectbox( 'Choose a voice:', ['alloy', 'echo', 'fable', 'onyx', 'nova', 'shimmer'] ) if st.button('Convert to Speech'): if query2: try: response = oai_client.audio.speech.create( model="tts-1", voice=voice_option, input=query2, ) # Stream or save the response as needed # For demonstration, let's assume we save then provide a link for downloading audio_file_path = "output.mp3" response.stream_to_file(audio_file_path) # Display audio file to download st.audio(audio_file_path, format='audio/mp3') st.success("Conversion successful!") except Exception as e: st.error(f"An error occurred: {e}") else: st.error("Please enter some text to convert.") st.header("Custom GPT Engineering Tools") st.link_button("Conversation Analyzer", "https://chat.openai.com/g/g-XARuyBgpL-conversation-analyzer") if st.button('Show/Hide Explanation of "Conversation Analyzer"'): # Toggle visibility st.session_state.show_instructions = not st.session_state.get('show_instructions', False) # Check if the instructions should be shown if st.session_state.get('show_instructions', False): st.write(""" Upon click "Input Your Conversation" complete the following 8 steps 1. Input Acquisition: Ask the user to input the text they would like analyzed. 2. Key Word Identification: Analyze the text and advise the user on the number of words they would need in order to ensure the purpose of the text is conveyed. This involves processing the text using natural language processing (NLP) techniques to detect words that are crucial to understanding the essence of the conversation. FRIST give the number of words needed and SECOND the words in a bulleted list 3. Ask the user if they would like to use your number of words or reduce to a smaller optimized list designed to convey the most accurate amount of information possible given the reduced set. 4. Ask the user what language they would like to translate the input into. 5. For the newly optimized list of words give the translated words FIRST and the original "language of the input" SECOND . Don't give the definition of the word. 6. Show the translated input and highlight the keywords by bolding them. 7. Give a distinct 100x100 image of each keyword. Try to put them in a single image so they can be cropped out when needed. 8 Allow the user to provide feedback on the analysis and the outputs, allowing for additional or reduction of words. 9. Give the final translation with highlighted words and provide an efficiency score. Number of words chosen versus suggested words x 100 """) with tab3: st.header("Green Smart Village Application") # Divide the page into three columns col1, col2, col3 = st.columns(3) with col1: st.header("Today's Agenda") st.write("1. Morning Meeting\n2. Review Project Plans\n3. Lunch Break\n4. Site Visit\n5. Evening Wrap-up") st.header("Agent Advisors") st.write("Would you like to optimize your HIN number?") # Selection box for the function to execute process_selection = st.selectbox( 'Choose the process to run:', ('crewai_process_gemini', 'crewai_process_mixtral_crazy', 'crewai_process_mixtral_normal', 'crewai_process_zephyr_normal', 'crewai_process_phi2') ) # Button to execute the chosen function if st.button('Run Process'): if research_topic: # Ensure there's a topic provided if process_selection == 'crewai_process_gemini': result = crewai_process_gemini(research_topic) elif process_selection == 'crewai_process_mixtral_crazy': result = crewai_process_mixtral_crazy(research_topic) elif process_selection == 'crewai_process_mixtral_normal': result = crewai_process_mixtral_normal(research_topic) elif process_selection == 'crewai_process_zephyr_normal': result = crewai_process_zephyr_normal(research_topic) elif process_selection == 'crewai_process_phi2': result = crewai_process_phi2(research_topic) st.write(result) else: st.warning('Please enter a research topic.') st.header("My Incentive") st.write("Total incentive for HIN optimization") with col2: st.header("Green Smart Village Layout") data = load_data('grid.json') # Ensure this path is correct # Dropdown for selecting a building building_options = [f"{bld['type']} at ({bld['coords'][0]}, {bld['coords'][1]})" for bld in data['buildings']] selected_building = st.selectbox("Select a building to highlight:", options=building_options) selected_index = building_options.index(selected_building) selected_building_coords = data['buildings'][selected_index]['coords'] # Draw the grid with the selected building highlighted fig = draw_grid(data, highlight_coords=selected_building_coords) st.pyplot(fig) # Assuming sensors are defined in your data, display them sensors = data['buildings'][selected_index].get('sensors', []) st.write(f"Sensors in selected building: {', '.join(sensors)}") with col3: st.header("Check Your HIN Number") # config = Config(height=400, width=400, nodeHighlightBehavior=True, highlightColor="#F7A7A6", directed=True, collapsible=True) if sensors: # Check if there are sensors to display graph_store = TripleStore() building_name = f"{data['buildings'][selected_index]['type']} ({selected_building_coords[0]}, {selected_building_coords[1]})" # Iterate through each sensor and create a triple linking it to the building for sensor in sensors: sensor_id = f"Sensor: {sensor}" # Label for sensor nodes # Correctly add the triple without named arguments graph_store.add_triple(building_name, "has_sensor", sensor_id) # Configuration for the graph visualization agraph_config = Config(height=300, width=300, nodeHighlightBehavior=True, highlightColor="#F7A7A6", directed=True, collapsible=True) # Display the graph agraph(nodes=graph_store.getNodes(), edges=graph_store.getEdges(), config=agraph_config) hin_number = st.text_input("Enter your HIN number:") if hin_number: st.write("HIN number details...") # Placeholder for actual HIN number check with tab4: st.header("Control Room") st.write("Synthetic data should be used to drive control room") """ Smart meters Water flow sensors Temperature and humidity sensors Occupancy sensors HVAC control systems Smart lighting Security cameras Indoor air quality sensors Smart lighting systems Energy consumption monitors Patient monitoring systems Environmental monitoring sensors Energy management systems Soil moisture sensors Smart irrigation systems Leak detection sensors Grid monitoring sensors GPS tracking for vehicles Smart building sensors Dispatch management systems High-speed internet connectivity Energy consumption monitoring Smart security systems Environmental control systems Security systems Smart HVAC systems Smart locks Water usage monitoring Smart inventory management systems Waste level sensors Fleet management systems for sanitation vehicles Air quality sensors Weather stations Pollution monitors Early warning systems Communication networks Adaptive lighting systems Traffic flow sensors Smart streetlights Residential Building - Light Blue, with smart meters, water flow sensors, and temperature and humidity sensors. Commercial Building - Green, equipped with occupancy sensors, smart meters, and HVAC control systems. Community Facilities - Yellow, featuring smart lighting, security cameras, and occupancy sensors. School - Yellow, with indoor air quality sensors, smart lighting systems, and energy consumption monitors. Healthcare Facility - Yellow, having patient monitoring systems, environmental monitoring sensors, and energy management systems. Green Space - Dark Green, with soil moisture sensors, smart irrigation systems, and environmental monitoring sensors. Utility Infrastructure - Dark Green, including smart meters, leak detection sensors, and grid monitoring sensors. Emergency Services - Red, with GPS tracking for vehicles, smart building sensors, and dispatch management systems. Cultural Facilities - Purple, equipped with environmental monitoring sensors, occupancy sensors, and smart lighting. Recreational Facilities - Purple, featuring air quality sensors, smart equipment maintenance sensors, and energy management systems. Innovation Center - Green, with high-speed internet connectivity, energy consumption monitoring, and smart security systems. Elderly Care Home - Yellow, including patient monitoring sensors, environmental control systems, and security systems. Childcare Centers - Yellow, with indoor air quality sensors, security cameras, and occupancy sensors. Places of Worship - Purple, featuring smart lighting, energy consumption monitoring, and security cameras. Event Spaces - Purple, with smart HVAC systems, occupancy sensors, and smart lighting. Guest Housing - Orange, including smart locks, energy management systems, and water usage monitoring. Pet Care Facilities - Orange, equipped with environmental monitoring sensors, security systems, and smart inventory management systems. Public Sanitation Facilities - Grey, with waste level sensors, fleet management systems for sanitation vehicles, and air quality sensors. Environmental Monitoring Stations - Dark Green, featuring air quality sensors, weather stations, and pollution monitors. Disaster Preparedness Center - Grey, including early warning systems, communication networks, and environmental sensors. Outdoor Community Spaces - Dark Green, with environmental sensors, smart irrigation systems, and adaptive lighting systems. Typical Road - Dark Grey, equipped with traffic flow sensors, smart streetlights, and pollution monitoring sensors. Typical Road Crossing - Dark Grey, featuring traffic flow sensors, smart streetlights, and pollution monitoring sensors. """