# Import libraries and references: import anthropic import base64 import glob import hashlib import json import os import pandas as pd import pytz import random import re import shutil import streamlit as st import time import traceback import uuid import zipfile from PIL import Image from azure.cosmos import CosmosClient, exceptions from datetime import datetime from git import Repo from github import Github from gradio_client import Client from urllib.parse import quote # 🎭 App Configuration - Because every app needs a good costume! Site_Name = 'πŸ™GitCosmos🌌 - AI Azure Cosmos DB and Github Agent' title = "πŸ™GitCosmos🌌 - AI Azure Cosmos DB and Github Agent" helpURL = 'https://huggingface.co/awacke1' bugURL = 'https://huggingface.co/spaces/awacke1/AzureCosmosDBUI/' icons = 'πŸ™πŸŒŒπŸ’«' st.set_page_config( page_title=title, page_icon=icons, layout="wide", initial_sidebar_state="auto", menu_items={ 'Get Help': helpURL, 'Report a bug': bugURL, 'About': title } ) # 🌌 Cosmos DB configuration - Where data goes to party! ENDPOINT = "https://acae-afd.documents.azure.com:443/" DATABASE_NAME = os.environ.get("COSMOS_DATABASE_NAME") CONTAINER_NAME = os.environ.get("COSMOS_CONTAINER_NAME") Key = os.environ.get("Key") # 🌐 Your local app URL - Home sweet home LOCAL_APP_URL = "https://huggingface.co/spaces/awacke1/AzureCosmosDBUI" CosmosDBUrl = 'https://portal.azure.com/#@AaronCWackergmail.onmicrosoft.com/resource/subscriptions/003fba60-5b3f-48f4-ab36-3ed11bc40816/resourceGroups/datasets/providers/Microsoft.DocumentDB/databaseAccounts/acae-afd/dataExplorer' # πŸ€– Anthropic configuration - Teaching machines to be more human (and funnier) anthropicclient = anthropic.Anthropic(api_key=os.environ.get("ANTHROPIC_API_KEY")) # 🧠 Initialize session state - Because even apps need a good memory if "chat_history" not in st.session_state: st.session_state.chat_history = [] # πŸ› οΈ Helper Functions - The unsung heroes of our code # πŸ“Ž Get a file download link - Making file sharing as easy as stealing candy from a baby def get_download_link(file_path): with open(file_path, "rb") as file: contents = file.read() b64 = base64.b64encode(contents).decode() file_name = os.path.basename(file_path) return f'Download {file_name}πŸ“‚' # 🎲 Generate a unique ID - Because being unique is important (just ask your mother) def generate_unique_id(): timestamp = datetime.utcnow().strftime('%Y%m%d%H%M%S%f') unique_uuid = str(uuid.uuid4()) returnValue = f"{timestamp}-{unique_uuid}" st.write('New Unique ID:' + returnValue) return # πŸ“ Generate a filename - Naming files like a pro (or a very confused librarian) def generate_filename(prompt, file_type): central = pytz.timezone('US/Central') safe_date_time = datetime.now(central).strftime("%m%d_%H%M") safe_prompt = re.sub(r'\W+', '', prompt)[:90] return f"{safe_date_time}{safe_prompt}.{file_type}" # πŸ’Ύ Create and save a file - Because data hoarding is a legitimate hobby def create_file(filename, prompt, response, should_save=True): if not should_save: return with open(filename, 'w', encoding='utf-8') as file: file.write(prompt + "\n\n" + response) # πŸ“– Load file content - Bringing words back from the digital grave def load_file(file_name): with open(file_name, "r", encoding='utf-8') as file: content = file.read() return content # πŸ” Display glossary entity - Making search fun again (as if it ever was) def display_glossary_entity(k): search_urls = { "πŸš€πŸŒŒArXiv": lambda k: f"/?q={quote(k)}", "πŸ“–": lambda k: f"https://en.wikipedia.org/wiki/{quote(k)}", "πŸ”": lambda k: f"https://www.google.com/search?q={quote(k)}", "πŸŽ₯": lambda k: f"https://www.youtube.com/results?search_query={quote(k)}", } links_md = ' '.join([f"{emoji}" for emoji, url in search_urls.items()]) st.markdown(f"{k} {links_md}", unsafe_allow_html=True) # πŸ—œοΈ Create zip of files - Squeezing files together like sardines in a can def create_zip_of_files(files): zip_name = "all_files.zip" with zipfile.ZipFile(zip_name, 'w') as zipf: for file in files: zipf.write(file) return zip_name # 🎬 Get video HTML - Making videos play nice (or at least trying to) def get_video_html(video_path, width="100%"): video_url = f"data:video/mp4;base64,{base64.b64encode(open(video_path, 'rb').read()).decode()}" return f''' ''' # 🎡 Get audio HTML - Let the music play (and hope it's not Baby Shark) def get_audio_html(audio_path, width="100%"): audio_url = f"data:audio/mpeg;base64,{base64.b64encode(open(audio_path, 'rb').read()).decode()}" return f''' ''' # 🌌 Cosmos DB functions - Where data goes to live its best life # πŸ“š Get databases - Collecting databases like Pokemon cards def get_databases(client): return [db['id'] for db in client.list_databases()] # πŸ“¦ Get containers - Finding where all the good stuff is hidden def get_containers(database): return [container['id'] for container in database.list_containers()] # πŸ“„ Get documents - Retrieving the sacred texts (or just some JSON) def get_documents(container, limit=None): query = "SELECT * FROM c ORDER BY c._ts DESC" items = list(container.query_items(query=query, enable_cross_partition_query=True, max_item_count=limit)) return items # πŸ“₯ Insert record - Adding new data (and crossing fingers it doesn't break anything) def insert_record(container, record): try: container.create_item(body=record) return True, "Record inserted successfully! πŸŽ‰" except exceptions.CosmosHttpResponseError as e: return False, f"HTTP error occurred: {str(e)} 🚨" except Exception as e: return False, f"An unexpected error occurred: {str(e)} 😱" # πŸ”„ Update record - Giving data a makeover def update_record(container, updated_record): try: container.upsert_item(body=updated_record) return True, f"Record with id {updated_record['id']} successfully updated. πŸ› οΈ" except exceptions.CosmosHttpResponseError as e: return False, f"HTTP error occurred: {str(e)} 🚨" except Exception as e: return False, f"An unexpected error occurred: {traceback.format_exc()} 😱" # πŸ—‘οΈ Delete record - Saying goodbye to data (it's not you, it's me) def delete_record(container, record): try: container.delete_item(item=record['id'], partition_key=record['id']) return True, f"Record with id {record['id']} successfully deleted. πŸ—‘οΈ" except exceptions.CosmosHttpResponseError as e: return False, f"HTTP error occurred: {str(e)} 🚨" except Exception as e: return False, f"An unexpected error occurred: {traceback.format_exc()} 😱" # πŸ’Ύ Save to Cosmos DB - Preserving data for future generations (or just until the next update) def save_to_cosmos_db(container, query, response1, response2): try: if container: # Generate a unique ID that includes a timestamp timestamp = datetime.utcnow().strftime('%Y%m%d%H%M%S%f') unique_uuid = str(uuid.uuid4()) new_id = f"{timestamp}-{unique_uuid}" # Create new document with proper name field record = { "id": new_id, "name": new_id, # Set name equal to ID to avoid null name error "query": query, "response1": response1, "response2": response2, "timestamp": datetime.utcnow().isoformat(), "type": "ai_response", # Add document type for better organization "version": "1.0" } try: # Create the new document container.create_item(body=record) st.success(f"Record saved successfully with ID: {record['id']}") # Refresh the documents display st.session_state.documents = get_documents(container) except exceptions.CosmosHttpResponseError as e: st.error(f"Error saving record to Cosmos DB: {e}") else: st.error("Cosmos DB container is not initialized.") except Exception as e: st.error(f"An unexpected error occurred: {str(e)}") def save_to_cosmos_db_old(container, query, response1, response2): try: if container: record = { "id": generate_unique_id(), "query": query, "response1": response1, "response2": response2, "timestamp": datetime.utcnow().isoformat() } try: container.create_item(body=record) st.success(f"Record saved successfully with ID: {record['id']}") # Refresh the documents display st.session_state.documents = get_documents(container) except exceptions.CosmosHttpResponseError as e: st.error(f"Error saving record to Cosmos DB: {e}") else: st.error("Cosmos DB container is not initialized.") except Exception as e: st.error(f"An unexpected error occurred: {str(e)}") # πŸ™ GitHub functions - Where code goes to socialize # πŸ“₯ Download GitHub repo - Cloning repos like it's going out of style def download_github_repo(url, local_path): if os.path.exists(local_path): shutil.rmtree(local_path) Repo.clone_from(url, local_path) # πŸ—œοΈ Create zip file - Squeezing files tighter than your budget def create_zip_file(source_dir, output_filename): shutil.make_archive(output_filename, 'zip', source_dir) # πŸ—οΈ Create repo - Building digital homes for lonely code def create_repo(g, repo_name): user = g.get_user() return user.create_repo(repo_name) # πŸš€ Push to GitHub - Sending code to the cloud (hopefully not the rainy kind) def push_to_github(local_path, repo, github_token): repo_url = f"https://{github_token}@github.com/{repo.full_name}.git" local_repo = Repo(local_path) if 'origin' in [remote.name for remote in local_repo.remotes]: origin = local_repo.remote('origin') origin.set_url(repo_url) else: origin = local_repo.create_remote('origin', repo_url) if not local_repo.heads: local_repo.git.checkout('-b', 'main') current_branch = 'main' else: current_branch = local_repo.active_branch.name local_repo.git.add(A=True) if local_repo.is_dirty(): local_repo.git.commit('-m', 'Initial commit') origin.push(refspec=f'{current_branch}:{current_branch}') def save_or_clone_to_cosmos_db(container, document=None, clone_id=None): def generate_complex_unique_id(): timestamp = datetime.utcnow().strftime('%Y%m%d%H%M%S%f') random_component = ''.join(random.choices('abcdefghijklmnopqrstuvwxyz0123456789', k=8)) return f"{timestamp}-{random_component}-{str(uuid.uuid4())}" max_retries = 10 base_delay = 0.1 for attempt in range(max_retries): try: new_id = generate_complex_unique_id() if clone_id: try: existing_doc = container.read_item(item=clone_id, partition_key=clone_id) new_doc = { 'id': new_id, 'originalText': existing_doc.get('originalText', ''), 'qtPrompts': existing_doc.get('qtPrompts', []), 'cloned_from': clone_id, 'cloned_at': datetime.utcnow().isoformat() } except exceptions.CosmosResourceNotFoundError: return False, f"Document with ID {clone_id} not found for cloning." else: if document is None: return False, "No document provided for saving" document['id'] = new_id document['created_at'] = datetime.utcnow().isoformat() new_doc = document response = container.create_item(body=new_doc) return True, f"{'Cloned' if clone_id else 'New'} document saved successfully with ID: {response['id']}" except exceptions.CosmosHttpResponseError as e: if e.status_code == 409: delay = base_delay * (2 ** attempt) + random.uniform(0, 0.1) time.sleep(delay) continue return False, f"Error saving to Cosmos DB: {str(e)}" except Exception as e: return False, f"An unexpected error occurred: {str(e)}" return False, "Failed to save document after maximum retries." # πŸ“¦ Archive current container - Packing up data like you're moving to a new digital house def archive_current_container(database_name, container_name, client): try: base_dir = "./cosmos_archive_current_container" if os.path.exists(base_dir): shutil.rmtree(base_dir) os.makedirs(base_dir) db_client = client.get_database_client(database_name) container_client = db_client.get_container_client(container_name) items = list(container_client.read_all_items()) container_dir = os.path.join(base_dir, container_name) os.makedirs(container_dir) for item in items: item_id = item.get('id', f"unknown_{datetime.now().strftime('%Y%m%d%H%M%S')}") with open(os.path.join(container_dir, f"{item_id}.json"), 'w') as f: json.dump(item, f, indent=2) archive_name = f"{container_name}_archive_{datetime.now().strftime('%Y%m%d%H%M%S')}" shutil.make_archive(archive_name, 'zip', base_dir) return get_download_link(f"{archive_name}.zip") except Exception as e: return f"An error occurred while archiving data: {str(e)} 😒" def gen_AI_IO_filename(display_query, output): # Get current time in Central Time Zone with milliseconds now_central = datetime.now(pytz.timezone("America/Chicago")) timestamp = now_central.strftime("%Y-%m-%d-%I-%M-%S-%f-%p") # Limit components to prevent excessive filename length display_query = display_query[:50] # Truncate display_query to 50 chars output_snippet = re.sub(r'[^A-Za-z0-9]+', '_', output[:100]) # Truncate output_snippet to 100 chars filename = f"{timestamp} - {display_query} - {output_snippet}.md" return filename # πŸ” Search glossary - Finding needles in digital haystacks def search_glossary(query): st.markdown(f"### πŸ” SearchGlossary for: {query}") model_options = ['mistralai/Mixtral-8x7B-Instruct-v0.1', 'mistralai/Mistral-7B-Instruct-v0.2'] model_choice = st.selectbox('🧠 Select LLM Model', options=model_options, index=1, key=f"model_choice_{id(query)}") database_options = ['Semantic Search', 'Arxiv Search - Latest - (EXPERIMENTAL)'] database_choice = st.selectbox('πŸ“š Select Database', options=database_options, index=0, key=f"database_choice_{id(query)}") # πŸ•΅οΈβ€β™‚οΈ Searching the glossary for: query all_results = "" # Limit the query display to 80 characters display_query = query[:80] + "..." if len(query) > 80 else query st.markdown(f"πŸ•΅οΈβ€β™‚οΈ Running ArXiV AI Analysis with Query: {display_query} - ML model: {model_choice} and Option: {database_options}") # πŸ” ArXiV RAG researcher expert ~-<>-~ Paper Summary & Ask LLM client = Client("awacke1/Arxiv-Paper-Search-And-QA-RAG-Pattern") # πŸ” ArXiv RAG researcher expert ~-<>-~ Paper Summary & Ask LLM - api_name: /ask_llm result = client.predict( prompt=query, llm_model_picked="mistralai/Mixtral-8x7B-Instruct-v0.1", stream_outputs=True, api_name="/ask_llm" ) st.markdown("# Mixtral-8x7B-Instruct-v0.1") st.markdown(result) #st.code(result, language="python", line_numbers=True) # πŸ” ArXiv RAG researcher expert ~-<>-~ Paper Summary & Ask LLM - api_name: /ask_llm result2 = client.predict( prompt=query, llm_model_picked="mistralai/Mistral-7B-Instruct-v0.2", stream_outputs=True, api_name="/ask_llm" ) st.markdown("# Mistral-7B-Instruct-v0.2") st.markdown(result2) #st.code(result2, language="python", line_numbers=True) # πŸ” ArXiv RAG researcher expert ~-<>-~ Paper Summary & Ask LLM - api_name: /update_with_rag_md response2 = client.predict( message=query, # str in 'parameter_13' Textbox component llm_results_use=10, database_choice="Semantic Search", llm_model_picked="mistralai/Mistral-7B-Instruct-v0.2", api_name="/update_with_rag_md" ) st.markdown("# Mistral-7B-Instruct-v0.2 update_with_rag_md 0") st.markdown(response2[0]) #st.code(response2[0], language="python", line_numbers=True, wrap_lines=True) st.markdown("# Mistral-7B-Instruct-v0.2 update_with_rag_md 1") st.markdown(response2[1]) #st.code(response2[1], language="python", line_numbers=True, wrap_lines=True) # βœ… Persist AI Results to Markdown Files filename = gen_AI_IO_filename(display_query, result) create_file(filename, query, result) st.markdown(f"βœ… File saved as: `{filename}`") filename = gen_AI_IO_filename(display_query, result2) create_file(filename, query, result2) st.markdown(f"βœ… File saved as: `{filename}`") filename = gen_AI_IO_filename(display_query, response2[0]) create_file(filename, query, response2[0]) st.markdown(f"βœ… File saved as: `{filename}`") filename = gen_AI_IO_filename(display_query, response2[1]) create_file(filename, query, response2[1]) st.markdown(f"βœ… File saved as: `{filename}`") return result, result2, response2 # πŸ“ Generate a safe filename from the first few lines of content def generate_filename_from_content(content, file_type="md"): # Extract the first few lines or sentences first_sentence = content.split('\n', 1)[0][:90] # Limit the length to 90 characters # Remove special characters to make it a valid filename safe_name = re.sub(r'[^\w\s-]', '', first_sentence) # Limit length to be compatible with Windows and Linux safe_name = safe_name[:50].strip() # Adjust length limit return f"{safe_name}.{file_type}" # πŸ’Ύ Create and save a file def create_file_from_content(content, should_save=True): if not should_save: return filename = generate_filename_from_content(content) with open(filename, 'w', encoding='utf-8') as file: file.write(content) return filename # πŸ“‚ Display list of saved .md files in the sidebar def display_saved_files_in_sidebar(): all_files = glob.glob("*.md") all_files.sort(reverse=True) all_files = [file for file in all_files if not file.lower().startswith('readme')] # Exclude README.md st.sidebar.markdown("## πŸ“ Saved Markdown Files") for file in all_files: col1, col2, col3 = st.sidebar.columns([6, 2, 1]) with col1: st.markdown(f"πŸ“„ {file}") with col2: st.sidebar.download_button( label="⬇️ Download", data=open(file, 'rb').read(), file_name=file ) with col3: if st.sidebar.button("πŸ—‘", key=f"delete_{file}"): os.remove(file) st.rerun() def clone_record(container, clone_id): try: existing_doc = container.read_item(item=clone_id, partition_key=clone_id) new_doc = existing_doc.copy() new_doc['id'] = generate_unique_id() # Generate new unique ID with timestamp new_doc['name'] = new_doc['id'] # Generate new unique ID with timestamp new_doc['createdAt'] = datetime.utcnow().isoformat() # Update the creation time new_doc['_rid'] = None # Reset _rid or any system-managed fields new_doc['_self'] = None new_doc['_etag'] = None new_doc['_attachments'] = None new_doc['_ts'] = None # Reset timestamp to be updated by Cosmos DB automatically # Insert the cloned document response = container.create_item(body=new_doc) st.success(f"Cloned document saved successfully with ID: {new_doc['id']} πŸŽ‰") # Refresh the documents in session state st.session_state.documents = list(container.query_items( query="SELECT * FROM c ORDER BY c._ts DESC", enable_cross_partition_query=True )) except exceptions.CosmosResourceNotFoundError: st.error(f"Document with ID {clone_id} not found for cloning.") except exceptions.CosmosHttpResponseError as e: st.error(f"HTTP error occurred: {str(e)} 🚨") except Exception as e: st.error(f"An unexpected error occurred: {str(e)} 😱") def create_new_blank_record(container): try: # Get the structure of the latest document (to preserve schema) latest_doc = container.query_items(query="SELECT * FROM c ORDER BY c._ts DESC", enable_cross_partition_query=True, max_item_count=1) if latest_doc: new_doc_structure = latest_doc[0].copy() else: new_doc_structure = {} new_doc = {key: "" for key in new_doc_structure.keys()} # Set all fields to blank new_doc['id'] = generate_unique_id() # Generate new unique ID new_doc['createdAt'] = datetime.utcnow().isoformat() # Set creation time # Insert the new blank document response = container.create_item(body=new_doc) st.success(f"New blank document saved successfully with ID: {new_doc['id']} πŸŽ‰") # Refresh the documents in session state st.session_state.documents = list(container.query_items( query="SELECT * FROM c ORDER BY c._ts DESC", enable_cross_partition_query=True )) except exceptions.CosmosHttpResponseError as e: st.error(f"HTTP error occurred: {str(e)} 🚨") except Exception as e: st.error(f"An unexpected error occurred: {str(e)} 😱") # Function to preprocess the pasted content def preprocess_text(text): # Replace CRLF and other newline variations with the JSON newline escape sequence text = text.replace('\r\n', '\\n') text = text.replace('\r', '\\n') text = text.replace('\n', '\\n') # Escape double quotes inside the text text = text.replace('"', '\\"') # Optionally remove or handle other special characters that might not be JSON-safe # Here, we remove characters like tabs or non-ASCII characters (as an example) text = re.sub(r'[\t]', ' ', text) # Replace tabs with spaces text = re.sub(r'[^\x00-\x7F]+', '', text) # Remove non-ASCII characters # Normalize spaces (strip leading/trailing whitespace) text = text.strip() return text def load_file_content(file_path): """Load and return file content with error handling""" try: with open(file_path, 'r', encoding='utf-8') as file: return file.read() except Exception as e: st.error(f"Error loading file: {str(e)}") return None def save_file_content(file_path, content): """Save file content with error handling""" try: with open(file_path, 'w', encoding='utf-8') as file: file.write(content) return True except Exception as e: st.error(f"Error saving file: {str(e)}") return False def display_file_viewer(file_path): """Display file content in markdown viewer""" content = load_file_content(file_path) if content: st.markdown("### πŸ“„ File Viewer") st.markdown(f"**Viewing:** {file_path}") # Add file metadata file_stats = os.stat(file_path) st.markdown(f"**Last modified:** {datetime.fromtimestamp(file_stats.st_mtime).strftime('%Y-%m-%d %H:%M:%S')}") st.markdown(f"**Size:** {file_stats.st_size} bytes") # Display content in markdown st.markdown("---") st.markdown(content) # Add download button st.download_button( label="⬇️ Download File", data=content, file_name=os.path.basename(file_path), mime="text/markdown" ) def display_file_editor(file_path): """Display file content in both Markdown and Code Editor views""" # Initialize file content in session state if not already present if 'file_content' not in st.session_state: st.session_state.file_content = {} # Load content if not in session state or if it's a different file if file_path not in st.session_state.file_content: content = load_file_content(file_path) if content is not None: st.session_state.file_content[file_path] = content else: return st.markdown("### ✏️ File Editor") st.markdown(f"**Editing:** {file_path}") # Create tabs for different views markdown_tab, code_tab = st.tabs(["Markdown View", "Code Editor"]) with markdown_tab: st.markdown("### πŸ“„ Markdown Preview") st.markdown(st.session_state.file_content[file_path]) with code_tab: st.markdown("### πŸ’» Code Editor") # Create a unique key for the text area editor_key = f"editor_{hash(file_path)}" # Editor with syntax highlighting for markdown new_content = st.text_area( "Edit content below:", value=st.session_state.file_content[file_path], height=400, key=editor_key ) # Add save and download buttons below both views col1, col2 = st.columns([1, 5]) with col1: if st.button("πŸ’Ύ Save Changes"): if save_file_content(file_path, new_content): st.session_state.file_content[file_path] = new_content st.success("File saved successfully! πŸŽ‰") time.sleep(1) st.rerun() with col2: st.download_button( label="⬇️ Download File", data=new_content, file_name=os.path.basename(file_path), mime="text/markdown" ) def display_file_editor_old(file_path): """Display file content in editor with save functionality""" # Initialize file content in session state if not already present if 'file_content' not in st.session_state: st.session_state.file_content = {} # Load content if not in session state or if it's a different file if file_path not in st.session_state.file_content: content = load_file_content(file_path) if content is not None: st.session_state.file_content[file_path] = content else: return st.markdown("### ✏️ File Editor") st.markdown(f"**Editing:** {file_path}") # Create a unique key for the text area editor_key = f"editor_{hash(file_path)}" # Editor with syntax highlighting for markdown new_content = st.text_area( "Edit content below:", value=st.session_state.file_content[file_path], height=400, key=editor_key ) col1, col2 = st.columns([1, 5]) with col1: if st.button("πŸ’Ύ Save Changes"): if save_file_content(file_path, new_content): st.session_state.file_content[file_path] = new_content st.success("File saved successfully! πŸŽ‰") time.sleep(1) st.rerun() with col2: st.download_button( label="⬇️ Download File", data=new_content, file_name=os.path.basename(file_path), mime="text/markdown" ) def update_file_management_section(): # Initialize session state variables if 'file_view_mode' not in st.session_state: st.session_state.file_view_mode = None if 'current_file' not in st.session_state: st.session_state.current_file = None if 'file_content' not in st.session_state: st.session_state.file_content = {} all_files = glob.glob("*.md") all_files.sort(reverse=True) # File management buttons in sidebar st.sidebar.title("πŸ“ File Management") if st.sidebar.button("πŸ—‘ Delete All Files"): for file in all_files: os.remove(file) st.session_state.file_content = {} # Clear the file content cache st.session_state.current_file = None st.session_state.file_view_mode = None st.rerun() if st.sidebar.button("⬇️ Download All Files"): zip_file = create_zip_of_files(all_files) st.sidebar.markdown(get_download_link(zip_file), unsafe_allow_html=True) # Display files in sidebar with action buttons for file in all_files: col1, col2, col3, col4 = st.sidebar.columns([1,3,1,1]) with col1: if st.button("🌐", key=f"view_{file}"): st.session_state.current_file = file st.session_state.file_view_mode = 'view' if file not in st.session_state.file_content: content = load_file_content(file) if content is not None: st.session_state.file_content[file] = content st.rerun() with col2: st.markdown(get_download_link(file), unsafe_allow_html=True) with col3: if st.button("πŸ“‚", key=f"edit_{file}"): st.session_state.current_file = file st.session_state.file_view_mode = 'edit' if file not in st.session_state.file_content: content = load_file_content(file) if content is not None: st.session_state.file_content[file] = content st.rerun() with col4: if st.button("πŸ—‘", key=f"delete_{file}"): os.remove(file) if file in st.session_state.file_content: del st.session_state.file_content[file] if st.session_state.current_file == file: st.session_state.current_file = None st.session_state.file_view_mode = None st.rerun() # Display viewer or editor in main area based on mode if st.session_state.current_file: if st.session_state.file_view_mode == 'view': display_file_viewer(st.session_state.current_file) elif st.session_state.file_view_mode == 'edit': display_file_editor(st.session_state.current_file) # 🎭 Main function - "All the world's a stage, and all the code merely players" -Shakespeare, probably def main(): st.markdown("### πŸ™Git🌌CosmosπŸ’« - Azure Cosmos DB and Github Agent") # 🎲 Session state vars - "Life is like a session state, you never know what you're gonna get" if 'logged_in' not in st.session_state: st.session_state.logged_in = False if 'selected_records' not in st.session_state: st.session_state.selected_records = [] if 'client' not in st.session_state: st.session_state.client = None if 'selected_database' not in st.session_state: st.session_state.selected_database = None if 'selected_container' not in st.session_state: st.session_state.selected_container = None if 'selected_document_id' not in st.session_state: st.session_state.selected_document_id = None if 'current_index' not in st.session_state: st.session_state.current_index = 0 if 'cloned_doc' not in st.session_state: st.session_state.cloned_doc = None # πŸ” Query processing - "To search or not to search, that is the query" try: query_params = st.query_params query = query_params.get('q') or query_params.get('query') or '' if query: result, result2, result3, response2 = search_glossary(query) # πŸ’Ύ Save results - "Every file you save is a future you pave" try: if st.button("Save AI Output"): filename = create_file_from_content(result) st.success(f"File saved: {filename}") filename = create_file_from_content(result2) st.success(f"File saved: {filename}") filename = create_file_from_content(result3) st.success(f"File saved: {filename}") filename = create_file_from_content(response2) st.success(f"File saved: {filename}") display_saved_files_in_sidebar() except Exception as e: st.error(f"An unexpected error occurred: {str(e)} 😱") # 🌟 Cosmos DB operations - "In Cosmos DB we trust, but we still handle errors we must" try: save_to_cosmos_db(st.session_state.cosmos_container, query, result, result) save_to_cosmos_db(st.session_state.cosmos_container, query, result2, result2) save_to_cosmos_db(st.session_state.cosmos_container, query, result3, result3) save_to_cosmos_db(st.session_state.cosmos_container, query, response2[0], response2[0]) save_to_cosmos_db(st.session_state.cosmos_container, query, response2[1], response2[1]) except exceptions.CosmosHttpResponseError as e: st.error(f"HTTP error occurred: {str(e)} 🚨") except Exception as e: st.error(f"An unexpected error occurred: {str(e)} 😱") st.stop() except Exception as e: st.markdown(' ') # πŸ” Auth check - "With great keys come great connectivity" if Key: st.session_state.primary_key = Key st.session_state.logged_in = True else: st.error("Cosmos DB Key is not set in environment variables. πŸ”‘βŒ") return if st.session_state.logged_in: # 🌌 DB initialization - "In the beginning, there was connection string..." try: if st.session_state.client is None: st.session_state.client = CosmosClient(ENDPOINT, credential=st.session_state.primary_key) # πŸ“š Navigation setup - "Navigation is not about where you are, but where you're going" st.sidebar.title("πŸ™Git🌌CosmosπŸ’«πŸ—„οΈNavigator") databases = get_databases(st.session_state.client) selected_db = st.sidebar.selectbox("πŸ—ƒοΈ Select Database", databases) st.markdown(CosmosDBUrl) # πŸ”„ State management - "Change is the only constant in state management" if selected_db != st.session_state.selected_database: st.session_state.selected_database = selected_db st.session_state.selected_container = None st.session_state.selected_document_id = None st.session_state.current_index = 0 st.rerun() if st.session_state.selected_database: database = st.session_state.client.get_database_client(st.session_state.selected_database) containers = get_containers(database) selected_container = st.sidebar.selectbox("πŸ“ Select Container", containers) # πŸ”„ Container state handling - "Container changes, state arranges" if selected_container != st.session_state.selected_container: st.session_state.selected_container = selected_container st.session_state.selected_document_id = None st.session_state.current_index = 0 st.rerun() if st.session_state.selected_container: container = database.get_container_client(st.session_state.selected_container) # πŸ“¦ Export functionality - "Pack it, zip it, ship it" if st.sidebar.button("πŸ“¦ Export Container Data"): download_link = archive_current_container(st.session_state.selected_database, st.session_state.selected_container, st.session_state.client) if download_link.startswith(' num_docs_to_display: documents_to_display = documents[:num_docs_to_display] st.sidebar.info(f"Showing top {num_docs_to_display} most recent documents.") else: documents_to_display = documents st.sidebar.info(f"Showing all {len(documents_to_display)} documents.") if documents_to_display: # 🎨 View options - "Different strokes for different folks" view_options = ['Show as Markdown', 'Show as Code Editor', 'Show as Run AI', 'Clone Document', 'New Record'] selected_view = st.sidebar.selectbox("Select Viewer/Editor", view_options, index=2) if selected_view == 'Show as Markdown': Label = '#### πŸ“„ Markdown view - Mark it down, mark it up' st.markdown(Label) total_docs = len(documents) doc = documents[st.session_state.current_index] # st.markdown(f"#### Document ID: {doc.get('id', '')}") # πŸ•΅οΈ Value extraction - "Finding spaces in all the right places" values_with_space = [] def extract_values(obj): if isinstance(obj, dict): for k, v in obj.items(): extract_values(v) elif isinstance(obj, list): for item in obj: extract_values(item) elif isinstance(obj, str): if ' ' in obj: values_with_space.append(obj) extract_values(doc) st.markdown("#### πŸ”— Links for Extracted Texts") for term in values_with_space: display_glossary_entity(term) content = json.dumps(doc, indent=2) st.markdown(f"```json\n{content}\n```") # β¬…οΈβž‘οΈ Navigation - "Left and right, day and night" col_prev, col_next = st.columns([1, 1]) with col_prev: if st.button("⬅️ Previous", key='prev_markdown'): if st.session_state.current_index > 0: st.session_state.current_index -= 1 st.rerun() with col_next: if st.button("➑️ Next", key='next_markdown'): if st.session_state.current_index < total_docs - 1: st.session_state.current_index += 1 st.rerun() elif selected_view == 'Show as Code Editor': Label = '#### πŸ’» Code editor view' st.markdown(Label) total_docs = len(documents) if total_docs == 0: st.warning("No documents available.") return doc = documents[st.session_state.current_index] doc_str = st.text_area("Edit Document", value=json.dumps(doc, indent=2), height=300, key=f'code_editor_{st.session_state.current_index}') col_prev, col_next = st.columns([1, 1]) with col_prev: if st.button("⬅️ Previous", key='prev_code'): if st.session_state.current_index > 0: st.session_state.current_index -= 1 st.rerun() with col_next: if st.button("➑️ Next", key='next_code'): if st.session_state.current_index < total_docs - 1: st.session_state.current_index += 1 st.rerun() col_save, col_delete = st.columns([1, 1]) with col_save: if st.button("πŸ’Ύ Save Changes", key=f'save_button_{st.session_state.current_index}'): try: updated_doc = json.loads(doc_str) response = container.upsert_item(body=updated_doc) if response: st.success(f"Document {updated_doc['id']} saved successfully.") st.session_state.selected_document_id = updated_doc['id'] st.rerun() except json.JSONDecodeError: st.error("Invalid JSON format. Please check your edits.") except Exception as e: st.error(f"Error saving document: {str(e)}") with col_delete: if st.button("πŸ—‘οΈ Delete", key=f'delete_button_{st.session_state.current_index}'): try: current_doc = json.loads(doc_str) doc_id = current_doc.get("id") if not doc_id: st.error("Document ID not found.") return # Confirm deletion if 'confirm_delete' not in st.session_state: st.session_state.confirm_delete = False if not st.session_state.confirm_delete: if st.button("⚠️ Click to confirm deletion", key=f'confirm_delete_{st.session_state.current_index}'): st.session_state.confirm_delete = True st.rerun() else: try: # Delete the document container.delete_item(item=doc_id, partition_key=doc_id) # Update the session state st.session_state.confirm_delete = False # Update the current index if necessary if total_docs > 1: if st.session_state.current_index == total_docs - 1: st.session_state.current_index = max(0, total_docs - 2) documents.pop(st.session_state.current_index) else: st.session_state.current_index = 0 documents.clear() st.success(f"Document {doc_id} deleted successfully.") st.rerun() except Exception as e: st.error(f"Error deleting document: {str(e)}") st.session_state.confirm_delete = False except json.JSONDecodeError: st.error("Invalid JSON format. Please check the document.") except Exception as e: st.error(f"Error processing deletion: {str(e)}") elif selected_view == 'Show as Code Editor - Old': Label = '#### πŸ’» Code editor view' st.markdown(Label) total_docs = len(documents) doc = documents[st.session_state.current_index] # st.markdown(f"#### Document ID: {doc.get('id', '')}") doc_str = st.text_area("Edit Document", value=json.dumps(doc, indent=2), height=300, key=f'code_editor_{st.session_state.current_index}') col_prev, col_next = st.columns([1, 1]) with col_prev: if st.button("⬅️ Previous", key='prev_code'): if st.session_state.current_index > 0: st.session_state.current_index -= 1 st.rerun() with col_next: if st.button("➑️ Next", key='next_code'): if st.session_state.current_index < total_docs - 1: st.session_state.current_index += 1 st.rerun() col_save, col_delete = st.columns([1, 1]) with col_save: if st.button("πŸ’Ύ Save Changes", key=f'save_button_{st.session_state.current_index}'): try: updated_doc = json.loads(doc_str) response = container.upsert_item(body=updated_doc) if response: st.success(f"Document {updated_doc['id']} saved successfully.") st.session_state.selected_document_id = updated_doc['id'] st.rerun() except Exception as e: st.error(f"Error saving document: {str(e)}") with col_delete: if st.button("πŸ—‘οΈ Delete", key=f'delete_button_{st.session_state.current_index}'): try: current_doc = json.loads(doc_str) # Direct deletion using container method with id and partition key delete = container.delete_item(current_doc["id"], current_doc["id"]) if delete: st.success(f"Document {current_doc['id']} deleted successfully.") if st.session_state.current_index > 0: st.session_state.current_index -= 1 st.rerun() except Exception as e: st.error(f"Error deleting document: {str(e)}") elif selected_view == 'Show as Run AI': Label = '#### ✏️ Run AI with wisdom, save with precision' st.markdown(Label) num_cols = len(documents_to_display) cols = st.columns(num_cols) for idx, (col, doc) in enumerate(zip(cols, documents_to_display)): with col: # ID and Name fields editable_id = st.text_input("ID", value=doc.get('id', ''), key=f'edit_id_{idx}') editable_name = st.text_input("Name", value=doc.get('name', ''), key=f'edit_name_{idx}') # Create editable document copy without id and name editable_doc = doc.copy() editable_doc.pop('id', None) editable_doc.pop('name', None) doc_str = st.text_area("Document Content (in JSON format)", value=json.dumps(editable_doc, indent=2), height=300, key=f'doc_str_{idx}') # Save and AI operations columns if st.button("πŸ€– Run AI", key=f'run_with_ai_button_{idx}'): # Your existing AI processing code here values_with_space = [] def extract_values2(obj): if isinstance(obj, dict): for k, v in obj.items(): extract_values2(v) elif isinstance(obj, list): for item in obj: extract_values2(item) elif isinstance(obj, str): if ' ' in obj: values_with_space.append(obj) extract_values2(doc) for term in values_with_space: display_glossary_entity(term) search_glossary(term) if st.button("πŸ’Ύ Save Changes", key=f'save_runai_{idx}'): try: updated_doc = json.loads(doc_str) # Reinsert ID and name from editable fields updated_doc['id'] = editable_id updated_doc['name'] = editable_name response = container.upsert_item(body=updated_doc) if response: st.success(f"Document {updated_doc['id']} saved successfully.") st.session_state.selected_document_id = updated_doc['id'] st.rerun() except Exception as e: st.error(f"Error saving document: {str(e)}") elif selected_view == 'Clone Document': st.markdown("#### πŸ“„ Clone Document (Save As)") total_docs = len(documents) doc = documents[st.session_state.current_index] # Display current document info st.markdown(f"**Original Document ID:** {doc.get('id', '')}") st.markdown(f"**Original Document Name:** {doc.get('name', '')}") # Generate new unique ID and name unique_filename = gen_AI_IO_filename("Clone", doc.get('name', '')) new_id = st.text_input("New Document ID", value=unique_filename, key='new_clone_id') new_name = st.text_input("New Document Name", value=f"Clone_{unique_filename[:8]}", key='new_clone_name') # Create new document with all original content except system fields new_doc = { 'id': new_id, 'name': new_name, **{k: v for k, v in doc.items() if k not in ['id', 'name', '_rid', '_self', '_etag', '_attachments', '_ts']} } # Show editable preview of the new document doc_str = st.text_area( "Edit Document Content (in JSON format)", value=json.dumps(new_doc, indent=2), height=300, key='clone_preview' ) col1, col2 = st.columns(2) with col1: if st.button("πŸ”„ Generate New ID/Name", key='regenerate_id'): # Generate new unique filename new_unique_filename = gen_AI_IO_filename("Clone", doc.get('name', '')) st.session_state.new_clone_id = new_unique_filename st.session_state.new_clone_name = f"Clone_{new_unique_filename[:8]}" st.rerun() with col2: if st.button("πŸ’Ύ Save As New Document", key='save_clone'): try: # Parse the edited document content final_doc = json.loads(doc_str) # Ensure the new ID and name are used final_doc['id'] = new_id final_doc['name'] = new_name # Remove any system fields that might have been copied system_fields = ['_rid', '_self', '_etag', '_attachments', '_ts'] for field in system_fields: final_doc.pop(field, None) # Create the new document response = container.create_item(body=final_doc) if response: st.success(f""" βœ… New document created successfully! - ID: {final_doc['id']} - Name: {final_doc['name']} """) # Update session state to show the new document st.session_state.selected_document_id = final_doc['id'] st.rerun() else: st.error("Failed to create new document") except json.JSONDecodeError as e: st.error(f"Invalid JSON format: {str(e)}") except Exception as e: st.error(f"Error creating document: {str(e)}") # Navigation buttons for viewing other documents to clone col_prev, col_next = st.columns([1, 1]) with col_prev: if st.button("⬅️ Previous", key='prev_clone'): if st.session_state.current_index > 0: st.session_state.current_index -= 1 st.rerun() with col_next: if st.button("➑️ Next", key='next_clone'): if st.session_state.current_index < total_docs - 1: st.session_state.current_index += 1 st.rerun() elif selected_view == 'New Record': st.markdown("#### Create a new document:") if st.button("πŸ€– Insert Auto-Generated Record"): auto_doc = { "id": generate_unique_id(), "name": f"Auto-generated Record {datetime.now().strftime('%Y-%m-%d %H:%M:%S')}", "content": "This is an auto-generated record.", "timestamp": datetime.now().isoformat() } success, message = save_or_clone_to_cosmos_db(container, document=auto_doc) if success: st.success(message) st.rerun() else: st.error(message) else: new_id = st.text_input("ID", value=generate_unique_id(), key='new_id') default_doc = { "id": new_id, "name": "New Document", "content": "", "timestamp": datetime.now().isoformat() } new_doc_str = st.text_area("Document Content (in JSON format)", value=json.dumps(default_doc, indent=2), height=300) if st.button("βž• Create New Document"): try: # Preprocess the text before loading it into JSON cleaned_doc_str = preprocess_text(new_doc_str) new_doc = json.loads(cleaned_doc_str) new_doc['id'] = new_id # Ensure ID matches input field success, message = insert_record(container, new_doc) if success: st.success(f"New document created with id: {new_doc['id']} πŸŽ‰") st.session_state.selected_document_id = new_doc['id'] st.rerun() else: st.error(message) except json.JSONDecodeError as e: st.error(f"Invalid JSON: {str(e)} 🚫") st.subheader(f"πŸ“Š Container: {st.session_state.selected_container}") if st.session_state.selected_container: if documents_to_display: Label = '#### πŸ“Š Data display - Data tells tales that words cannot' st.markdown(Label) df = pd.DataFrame(documents_to_display) st.dataframe(df) else: st.info("No documents to display. 🧐") Label = '#### πŸ™ GitHub integration - Git happens' st.subheader("πŸ™ GitHub Operations") github_token = os.environ.get("GITHUB") source_repo = st.text_input("Source GitHub Repository URL", value="https://github.com/AaronCWacker/AIExamples-8-24-Streamlit") new_repo_name = st.text_input("New Repository Name (for cloning)", value=f"AIExample-Clone-{datetime.now().strftime('%Y%m%d_%H%M%S')}") col1, col2 = st.columns(2) with col1: if st.button("πŸ“₯ Clone Repository"): if github_token and source_repo: st.markdown(Label) try: local_path = f"./temp_repo_{datetime.now().strftime('%Y%m%d%H%M%S')}" download_github_repo(source_repo, local_path) zip_filename = f"{new_repo_name}.zip" create_zip_file(local_path, zip_filename[:-4]) st.markdown(get_download_link(zip_filename), unsafe_allow_html=True) st.success("Repository cloned successfully! πŸŽ‰") except Exception as e: st.error(f"An error occurred: {str(e)} 😒") finally: if os.path.exists(local_path): shutil.rmtree(local_path) if os.path.exists(zip_filename): os.remove(zip_filename) else: st.error("Please ensure GitHub token is set in environment variables and source repository URL is provided. πŸ”‘β“") with col2: if st.button("πŸ“€ Push to New Repository"): if github_token and source_repo: st.markdown(Label) try: g = Github(github_token) new_repo = create_repo(g, new_repo_name) local_path = f"./temp_repo_{datetime.now().strftime('%Y%m%d%H%M%S')}" download_github_repo(source_repo, local_path) push_to_github(local_path, new_repo, github_token) st.success(f"Repository pushed successfully to {new_repo.html_url} πŸš€") except Exception as e: st.error(f"An error occurred: {str(e)} 😒") finally: if os.path.exists(local_path): shutil.rmtree(local_path) else: st.error("Please ensure GitHub token is set in environment variables and source repository URL is provided. πŸ”‘β“") st.subheader("πŸ’¬ Chat with Claude") user_input = st.text_area("Message πŸ“¨:", height=100) if st.button("Send πŸ“¨"): Label = '#### πŸ’¬ Chat functionality - Every chat is a chance to learn' st.markdown(Label) if user_input: response = anthropicclient.messages.create( model="claude-3-sonnet-20240229", max_tokens=1000, messages=[ {"role": "user", "content": user_input} ] ) st.write("Claude's reply 🧠:") st.write(response.content[0].text) filename = generate_filename(user_input, "md") create_file(filename, user_input, response.content[0].text) st.session_state.chat_history.append({"user": user_input, "claude": response.content[0].text}) # Save to Cosmos DB save_to_cosmos_db(container, user_input, response.content[0].text, "") # πŸ“œ Chat history display - "History repeats itself, first as chat, then as wisdom" st.subheader("Past Conversations πŸ“œ") for chat in st.session_state.chat_history: st.text_area("You said πŸ’¬:", chat["user"], height=100, disabled=True) st.text_area("Claude replied πŸ€–:", chat["claude"], height=200, disabled=True) st.markdown("---") # πŸ“ File editor - "Edit with care, save with flair" if hasattr(st.session_state, 'current_file'): st.subheader(f"Editing: {st.session_state.current_file} πŸ› ") new_content = st.text_area("File Content ✏️:", st.session_state.file_content, height=300) # Preprocess the text before loading it into JSON - Added to protect copy paste into JSON to keep format. cleaned_doc_str = preprocess_text(new_content) new_doc = json.loads(cleaned_doc_str) new_content = cleaned_doc_str if st.button("Save Changes πŸ’Ύ"): with open(st.session_state.current_file, 'w', encoding='utf-8') as file: file.write(new_content) st.success("File updated successfully! πŸŽ‰") # πŸ“‚ File management - "Manage many, maintain order" update_file_management_section() except exceptions.CosmosHttpResponseError as e: st.error(f"Failed to connect to Cosmos DB. HTTP error: {str(e)} 🚨") except Exception as e: st.error(f"An unexpected error occurred: {str(e)} 😱") if st.session_state.logged_in and st.sidebar.button("πŸšͺ Logout"): Label = '#### πŸšͺ Logout - All good things must come to an end' st.markdown(Label) st.session_state.logged_in = False st.session_state.selected_records.clear() st.session_state.client = None st.session_state.selected_database = None st.session_state.selected_container = None st.session_state.selected_document_id = None st.session_state.current_index = 0 st.rerun() if __name__ == "__main__": main()