brad/LLM_Engineering_OLD
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Merge pull request #921 from solisoma/solisoma-week8

Browse Source 
Solisoma week8
This commit is contained in:
Ed Donner
2025-10-30 22:30:32 -04:00
committed by GitHub
parent 2a09d50cb7 f353b61026
commit f7ee65b0ec
1 changed files with 421 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files

421
week8/community_contributions/solisoma/price_is_right_fixed.ipynb Normal file
View File

@@ -0,0 +1,421 @@
{
"cells": [
{
"cell_type": "markdown",
"metadata": {},
"source": [
"# The Price is Right - Fixed Version\n",
"\n",
"This notebook fixes the issue where existing deals disappear from the table when the system makes new calls.\n",
"\n",
"**Key Fix**: The table now continuously shows current memory during updates, so existing deals remain visible while new ones are being searched.\n"
]
},
{
"cell_type": "code",
"execution_count": 1,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"Working directory: c:\\Users\\hp\\projects\\gen-ai\\llm_engineering\\week8\n"
]
}
],
"source": [
"# Imports\n",
"import sys\n",
"import os\n",
"\n",
"# Change working directory to week8 where deal_agent_framework expects to run\n",
"# This ensures all relative paths (models, database) work correctly\n",
"notebook_dir = os.getcwd()\n",
"week8_dir = os.path.join(notebook_dir, '..', '..')\n",
"os.chdir(week8_dir)\n",
"print(f\"Working directory: {os.getcwd()}\")\n",
"\n",
"import logging\n",
"import queue\n",
"import threading\n",
"import time\n",
"import gradio as gr\n",
"from deal_agent_framework import DealAgentFramework\n",
"from agents.deals import Opportunity, Deal\n",
"from log_utils import reformat\n",
"import plotly.graph_objects as go\n"
]
},
{
"cell_type": "code",
"execution_count": 2,
"metadata": {},
"outputs": [],
"source": [
"# Helper Functions\n",
"\n",
"class QueueHandler(logging.Handler):\n",
" def __init__(self, log_queue):\n",
" super().__init__()\n",
" self.log_queue = log_queue\n",
"\n",
" def emit(self, record):\n",
" self.log_queue.put(self.format(record))\n",
"\n",
"\n",
"def html_for(log_data):\n",
" \"\"\"Convert log data to HTML format for display\"\"\"\n",
" output = '<br>'.join(log_data[-18:])\n",
" return f\"\"\"\n",
" <div id=\"scrollContent\" style=\"height: 400px; overflow-y: auto; border: 1px solid #ccc; background-color: #222229; padding: 10px;\">\n",
" {output}\n",
" </div>\n",
" \"\"\"\n",
"\n",
"\n",
"def setup_logging(log_queue):\n",
" \"\"\"Setup logging to capture messages in a queue\"\"\"\n",
" handler = QueueHandler(log_queue)\n",
" formatter = logging.Formatter(\n",
" \"[%(asctime)s] %(message)s\",\n",
" datefmt=\"%Y-%m-%d %H:%M:%S %z\",\n",
" )\n",
" handler.setFormatter(formatter)\n",
" logger = logging.getLogger()\n",
" logger.addHandler(handler)\n",
" logger.setLevel(logging.INFO)\n",
"\n",
"\n",
"def get_plot():\n",
" \"\"\"Generate 3D visualization of vector database - handles empty database gracefully\"\"\"\n",
" try:\n",
" documents, vectors, colors = DealAgentFramework.get_plot_data(max_datapoints=1000)\n",
" \n",
" # Check if we have any data\n",
" if len(vectors) == 0:\n",
" # Return placeholder plot if database is empty\n",
" fig = go.Figure()\n",
" fig.update_layout(\n",
" title='Vector Database Empty',\n",
" height=400,\n",
" annotations=[dict(\n",
" text=\"The vector database is empty.<br>Run the data loading notebook (day2.0) to populate it.\",\n",
" x=0.5,\n",
" y=0.5,\n",
" xref=\"paper\",\n",
" yref=\"paper\",\n",
" showarrow=False,\n",
" font=dict(size=14)\n",
" )]\n",
" )\n",
" return fig\n",
" \n",
" # Normal case: create 3D scatter plot\n",
" fig = go.Figure(data=[go.Scatter3d(\n",
" x=vectors[:, 0],\n",
" y=vectors[:, 1],\n",
" z=vectors[:, 2],\n",
" mode='markers',\n",
" marker=dict(size=2, color=colors, opacity=0.7),\n",
" )])\n",
" \n",
" fig.update_layout(\n",
" scene=dict(xaxis_title='x', \n",
" yaxis_title='y', \n",
" zaxis_title='z',\n",
" aspectmode='manual',\n",
" aspectratio=dict(x=2.2, y=2.2, z=1),\n",
" camera=dict(\n",
" eye=dict(x=1.6, y=1.6, z=0.8)\n",
" )),\n",
" height=400,\n",
" margin=dict(r=5, b=1, l=5, t=2)\n",
" )\n",
" return fig\n",
" except Exception as e:\n",
" # Handle any errors gracefully\n",
" fig = go.Figure()\n",
" fig.update_layout(\n",
" title='Error Loading Vector Database',\n",
" height=400,\n",
" annotations=[dict(\n",
" text=f\"Error: {str(e)}<br><br>Make sure the vector database is set up correctly.<br>Run day2.0 notebook to populate it.\",\n",
" x=0.5,\n",
" y=0.5,\n",
" xref=\"paper\",\n",
" yref=\"paper\",\n",
" showarrow=False,\n",
" font=dict(size=12)\n",
" )]\n",
" )\n",
" return fig\n",
"\n",
"\n",
"def create_opportunity_from_dict(data: dict) -> Opportunity:\n",
" \"\"\"Helper function to create Opportunity from dictionary - uses Deal and Opportunity classes\"\"\"\n",
" deal = Deal(**data['deal']) if isinstance(data['deal'], dict) else data['deal']\n",
" return Opportunity(deal=deal, estimate=data['estimate'], discount=data['discount'])\n",
"\n",
"\n",
"def validate_opportunities(opportunities) -> list:\n",
" \"\"\"Validate and ensure all items are Opportunity instances - uses Opportunity class\"\"\"\n",
" validated = []\n",
" for opp in opportunities:\n",
" if not isinstance(opp, Opportunity):\n",
" if isinstance(opp, dict):\n",
" opp = create_opportunity_from_dict(opp)\n",
" else:\n",
" continue\n",
" validated.append(opp)\n",
" return validated\n"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"# Main App Class\n",
"\n",
"class App:\n",
"\n",
" def __init__(self): \n",
" self.agent_framework = None\n",
"\n",
" def get_agent_framework(self):\n",
" \"\"\"Get or initialize the agent framework\"\"\"\n",
" if not self.agent_framework:\n",
" self.agent_framework = DealAgentFramework()\n",
" self.agent_framework.init_agents_as_needed()\n",
" return self.agent_framework\n",
"\n",
" def table_for(self, opps):\n",
" \"\"\"Convert opportunities to table format - uses Opportunity and Deal classes\"\"\"\n",
" # Validate opportunities are Opportunity instances\n",
" validated_opps = validate_opportunities(opps)\n",
" return [[opp.deal.product_description, f\"${opp.deal.price:.2f}\", f\"${opp.estimate:.2f}\", f\"${opp.discount:.2f}\", opp.deal.url] \n",
" for opp in validated_opps \n",
" if isinstance(opp, Opportunity)]\n",
"\n",
" def update_output(self, log_data, log_queue, result_queue):\n",
" \"\"\"Keep showing current memory during updates - fixes disappearing table issue\"\"\"\n",
" framework = self.get_agent_framework()\n",
" current_table = self.table_for(framework.memory)\n",
" \n",
" while True:\n",
" try:\n",
" message = log_queue.get_nowait()\n",
" log_data.append(reformat(message))\n",
" # Always refresh table from current memory during updates\n",
" current_table = self.table_for(framework.memory)\n",
" yield log_data, html_for(log_data), current_table\n",
" except queue.Empty:\n",
" try:\n",
" # When result is ready, update with final result\n",
" final_result = result_queue.get_nowait()\n",
" yield log_data, html_for(log_data), final_result\n",
" return\n",
" except queue.Empty:\n",
" # Continue showing current memory while waiting\n",
" current_table = self.table_for(framework.memory)\n",
" yield log_data, html_for(log_data), current_table\n",
" time.sleep(0.1)\n",
"\n",
" def do_run(self):\n",
" \"\"\"Run framework and return updated table\"\"\"\n",
" import datetime\n",
" framework = self.get_agent_framework()\n",
" # Log to the Gradio display (will show up in logs panel)\n",
" logging.info(f\"⏰ TIMER TRIGGERED at {datetime.datetime.now().strftime('%H:%M:%S')} - Current memory: {len(framework.memory)} deals\")\n",
" new_opportunities = framework.run()\n",
" logging.info(f\"✅ Scan complete - Total deals in memory: {len(framework.memory)}\")\n",
" return self.table_for(new_opportunities)\n",
"\n",
" def run_with_logging(self, initial_log_data):\n",
" \"\"\"Run agent framework with logging in a separate thread\"\"\"\n",
" log_queue = queue.Queue()\n",
" result_queue = queue.Queue()\n",
" setup_logging(log_queue)\n",
" \n",
" def worker():\n",
" result = self.do_run()\n",
" result_queue.put(result)\n",
" \n",
" thread = threading.Thread(target=worker)\n",
" thread.start()\n",
" \n",
" for log_data, output, final_result in self.update_output(initial_log_data, log_queue, result_queue):\n",
" yield log_data, output, final_result\n",
"\n",
" def do_select(self, selected_index: gr.SelectData):\n",
" \"\"\"Handle deal selection - send alert\"\"\"\n",
" framework = self.get_agent_framework()\n",
" opportunities = framework.memory\n",
" row = selected_index.index[0]\n",
" if row < len(opportunities):\n",
" opportunity = opportunities[row]\n",
" framework.planner.messenger.alert(opportunity)\n",
" return f\"Alert sent for: {opportunity.deal.product_description[:50]}...\"\n",
" return \"No opportunity found at that index\"\n",
"\n",
" def load_initial(self):\n",
" \"\"\"Load initial state with existing deals - uses Opportunity and Deal classes\"\"\"\n",
" framework = self.get_agent_framework()\n",
" # Ensure memory contains Opportunity instances\n",
" opportunities = validate_opportunities(framework.memory)\n",
" initial_table = self.table_for(opportunities)\n",
" return [], \"\", initial_table\n",
"\n",
" def run(self):\n",
" \"\"\"Launch the Gradio interface\"\"\"\n",
" with gr.Blocks(title=\"The Price is Right\", fill_width=True) as ui:\n",
" \n",
" log_data = gr.State([])\n",
" \n",
" with gr.Row():\n",
" gr.Markdown('<div style=\"text-align: center;font-size:24px\"><strong>The Price is Right</strong> - Autonomous Agent Framework that hunts for deals</div>')\n",
" with gr.Row():\n",
" gr.Markdown('<div style=\"text-align: center;font-size:14px\">A proprietary fine-tuned LLM deployed on Modal and a RAG pipeline with a frontier model collaborate to send push notifications with great online deals.</div>')\n",
" with gr.Row():\n",
" opportunities_dataframe = gr.Dataframe(\n",
" headers=[\"Deals found so far\", \"Price\", \"Estimate\", \"Discount\", \"URL\"],\n",
" wrap=True,\n",
" column_widths=[6, 1, 1, 1, 3],\n",
" row_count=10,\n",
" col_count=5,\n",
" max_height=400,\n",
" )\n",
" with gr.Row():\n",
" with gr.Column(scale=1):\n",
" logs = gr.HTML()\n",
" with gr.Column(scale=1):\n",
" plot = gr.Plot(value=get_plot(), show_label=False)\n",
" \n",
" # Initial load - show existing deals\n",
" ui.load(self.load_initial, inputs=[], outputs=[log_data, logs, opportunities_dataframe])\n",
"\n",
" # Timer that runs every 5 minutes (300 seconds)\n",
" timer = gr.Timer(value=10, active=True)\n",
" timer.tick(self.run_with_logging, inputs=[log_data], outputs=[log_data, logs, opportunities_dataframe])\n",
"\n",
" # Selection handler\n",
" selection_feedback = gr.Textbox(visible=False)\n",
" opportunities_dataframe.select(self.do_select, inputs=[], outputs=[selection_feedback])\n",
" \n",
" ui.launch(share=False, inbrowser=True)\n"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"* Running on local URL: http://127.0.0.1:7860\n",
"* To create a public link, set `share=True` in `launch()`.\n"
]
},
{
"data": {
"text/html": [
"<div><iframe src=\"http://127.0.0.1:7860/\" width=\"100%\" height=\"500\" allow=\"autoplay; camera; microphone; clipboard-read; clipboard-write;\" frameborder=\"0\" allowfullscreen></iframe></div>"
],
"text/plain": [
"<IPython.core.display.HTML object>"
]
},
"metadata": {},
"output_type": "display_data"
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"[2025-10-30 20:52:39 +0100] [Agents] [INFO] \u001b[44m\u001b[37m[Agent Framework] Initializing Agent Framework\u001b[0m\n",
"[2025-10-30 20:52:39 +0100] [Agents] [INFO] \u001b[40m\u001b[32m[Planning Agent] Planning Agent is initializing\u001b[0m\n",
"[2025-10-30 20:52:39 +0100] [Agents] [INFO] \u001b[40m\u001b[36m[Scanner Agent] Scanner Agent is initializing\u001b[0m\n",
"[2025-10-30 20:52:39 +0100] [Agents] [INFO] \u001b[40m\u001b[36m[Scanner Agent] Scanner Agent is ready\u001b[0m\n",
"[2025-10-30 20:52:39 +0100] [Agents] [INFO] \u001b[40m\u001b[33m[Ensemble Agent] Initializing Ensemble Agent\u001b[0m\n",
"[2025-10-30 20:52:39 +0100] [Agents] [INFO] \u001b[40m\u001b[31m[Specialist Agent] Specialist Agent is initializing - connecting to modal\u001b[0m\n",
"[2025-10-30 20:52:39 +0100] [Agents] [INFO] \u001b[40m\u001b[31m[Specialist Agent] Specialist Agent is ready\u001b[0m\n",
"[2025-10-30 20:52:39 +0100] [Agents] [INFO] \u001b[40m\u001b[34m[Frontier Agent] Initializing Frontier Agent\u001b[0m\n",
"[2025-10-30 20:52:39 +0100] [Agents] [INFO] \u001b[40m\u001b[34m[Frontier Agent] Frontier Agent is set up with DeepSeek\u001b[0m\n",
"[2025-10-30 20:52:39 +0100] [Agents] [INFO] Use pytorch device_name: cpu\n",
"[2025-10-30 20:52:39 +0100] [Agents] [INFO] Load pretrained SentenceTransformer: sentence-transformers/all-MiniLM-L6-v2\n",
"[2025-10-30 20:52:46 +0100] [Agents] [INFO] \u001b[40m\u001b[34m[Frontier Agent] Frontier Agent is ready\u001b[0m\n",
"[2025-10-30 20:52:46 +0100] [Agents] [INFO] \u001b[40m\u001b[35m[Random Forest Agent] Random Forest Agent is initializing\u001b[0m\n",
"[2025-10-30 20:52:46 +0100] [Agents] [INFO] Use pytorch device_name: cpu\n",
"[2025-10-30 20:52:46 +0100] [Agents] [INFO] Load pretrained SentenceTransformer: sentence-transformers/all-MiniLM-L6-v2\n",
"\n",
"============================================================\n",
"[20:52:49] Timer triggered - Scanning for deals...\n",
"Current memory has 2 deals\n",
"============================================================\n",
"\n",
"[2025-10-30 20:52:49 +0100] [Agents] [INFO] \u001b[44m\u001b[37m[Agent Framework] Initializing Agent Framework\u001b[0m\n",
"[2025-10-30 20:52:49 +0100] [Agents] [INFO] \u001b[40m\u001b[32m[Planning Agent] Planning Agent is initializing\u001b[0m\n",
"[2025-10-30 20:52:49 +0100] [Agents] [INFO] \u001b[40m\u001b[36m[Scanner Agent] Scanner Agent is initializing\u001b[0m\n",
"[2025-10-30 20:52:49 +0100] [Agents] [INFO] \u001b[40m\u001b[36m[Scanner Agent] Scanner Agent is ready\u001b[0m\n",
"[2025-10-30 20:52:49 +0100] [Agents] [INFO] \u001b[40m\u001b[33m[Ensemble Agent] Initializing Ensemble Agent\u001b[0m\n",
"[2025-10-30 20:52:49 +0100] [Agents] [INFO] \u001b[40m\u001b[31m[Specialist Agent] Specialist Agent is initializing - connecting to modal\u001b[0m\n",
"[2025-10-30 20:52:49 +0100] [Agents] [INFO] \u001b[40m\u001b[31m[Specialist Agent] Specialist Agent is ready\u001b[0m\n",
"[2025-10-30 20:52:49 +0100] [Agents] [INFO] \u001b[40m\u001b[34m[Frontier Agent] Initializing Frontier Agent\u001b[0m\n",
"[2025-10-30 20:52:49 +0100] [Agents] [INFO] \u001b[40m\u001b[34m[Frontier Agent] Frontier Agent is set up with DeepSeek\u001b[0m\n",
"[2025-10-30 20:52:49 +0100] [Agents] [INFO] Use pytorch device_name: cpu\n",
"[2025-10-30 20:52:49 +0100] [Agents] [INFO] Load pretrained SentenceTransformer: sentence-transformers/all-MiniLM-L6-v2\n",
"[2025-10-30 20:52:55 +0100] [Agents] [INFO] \u001b[40m\u001b[35m[Random Forest Agent] Random Forest Agent is ready\u001b[0m\n",
"[2025-10-30 20:52:55 +0100] [Agents] [INFO] \u001b[40m\u001b[33m[Ensemble Agent] Ensemble Agent is ready\u001b[0m\n",
"[2025-10-30 20:52:55 +0100] [Agents] [INFO] \u001b[40m\u001b[37m[Messaging Agent] Messaging Agent is initializing\u001b[0m\n",
"[2025-10-30 20:52:55 +0100] [Agents] [INFO] \u001b[40m\u001b[37m[Messaging Agent] Messaging Agent has initialized Pushover\u001b[0m\n",
"[2025-10-30 20:52:55 +0100] [Agents] [INFO] \u001b[40m\u001b[32m[Planning Agent] Planning Agent is ready\u001b[0m\n",
"[2025-10-30 20:52:55 +0100] [Agents] [INFO] \u001b[44m\u001b[37m[Agent Framework] Agent Framework is ready\u001b[0m\n",
"[2025-10-30 20:52:56 +0100] [Agents] [INFO] \u001b[40m\u001b[34m[Frontier Agent] Frontier Agent is ready\u001b[0m\n",
"[2025-10-30 20:52:56 +0100] [Agents] [INFO] \u001b[40m\u001b[35m[Random Forest Agent] Random Forest Agent is initializing\u001b[0m\n",
"[2025-10-30 20:52:56 +0100] [Agents] [INFO] Use pytorch device_name: cpu\n",
"[2025-10-30 20:52:56 +0100] [Agents] [INFO] Load pretrained SentenceTransformer: sentence-transformers/all-MiniLM-L6-v2\n",
"[2025-10-30 20:53:03 +0100] [Agents] [INFO] \u001b[40m\u001b[35m[Random Forest Agent] Random Forest Agent is ready\u001b[0m\n",
"[2025-10-30 20:53:03 +0100] [Agents] [INFO] \u001b[40m\u001b[33m[Ensemble Agent] Ensemble Agent is ready\u001b[0m\n",
"[2025-10-30 20:53:03 +0100] [Agents] [INFO] \u001b[40m\u001b[37m[Messaging Agent] Messaging Agent is initializing\u001b[0m\n",
"[2025-10-30 20:53:03 +0100] [Agents] [INFO] \u001b[40m\u001b[37m[Messaging Agent] Messaging Agent has initialized Pushover\u001b[0m\n",
"[2025-10-30 20:53:03 +0100] [Agents] [INFO] \u001b[40m\u001b[32m[Planning Agent] Planning Agent is ready\u001b[0m\n",
"[2025-10-30 20:53:03 +0100] [Agents] [INFO] \u001b[44m\u001b[37m[Agent Framework] Agent Framework is ready\u001b[0m\n",
"[2025-10-30 20:53:03 +0100] [Agents] [INFO] Kicking off Planning Agent\n",
"[2025-10-30 20:53:03 +0100] [Agents] [INFO] \u001b[40m\u001b[32m[Planning Agent] Planning Agent is kicking off a run\u001b[0m\n",
"[2025-10-30 20:53:03 +0100] [Agents] [INFO] \u001b[40m\u001b[36m[Scanner Agent] Scanner Agent is about to fetch deals from RSS feed\u001b[0m\n"
]
}
],
"source": [
"# Run the application\n",
"app = App()\n",
"app.run()\n"
]
}
],
"metadata": {
"kernelspec": {
"display_name": ".venv",
"language": "python",
"name": "python3"
},
"language_info": {
"codemirror_mode": {
"name": "ipython",
"version": 3
},
"file_extension": ".py",
"mimetype": "text/x-python",
"name": "python",
"nbconvert_exporter": "python",
"pygments_lexer": "ipython3",
"version": "3.12.12"
}
},
"nbformat": 4,
"nbformat_minor": 2
}