Initial upload: local files to Hugging Face repo

.env

@@ -0,0 +1,3 @@
+FINNHUB_API_KEY=d3ql0apr01quv7kbbqe0d3ql0apr01quv7kbbqeg 
+ALPHAVANTAGE_API_KEY=GSA9BWLJRJI0E4GN 
+GNEWS_API_KEY=27e5b9ee0a57864a86e2b6f13390ad35 

README.md

@@ -1,14 +1,14 @@
 ---
-title: NLP GenAI Final
-emoji: 🌖
-colorFrom: purple
-colorTo: pink
+title: Asset News Notifcation System
+emoji: 👀
+colorFrom: blue
+colorTo: indigo
 sdk: gradio
 sdk_version: 5.49.1
 app_file: app.py
 pinned: false
 license: mit
-short_description: agentic_agent_finance
+short_description: Agentic AI system to notify a user via of market events
 ---
 
 Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference

agents/alertcoordinator.py

@@ -0,0 +1,22 @@
+from datetime import datetime
+
+class AlertCoordinatorAgent:
+    def __init__(self, config, database):
+        self.config = config
+        self.db = database
+
+    def process_alerts(self, analyzed):
+        for article in analyzed:
+            keywords = [k for k in self.config.IMPACT_KEYWORDS if k in article.get('headline','').lower()]
+            urgency = len(keywords) * 0.2 + abs(article.get('sentiment',0)) * 0.3
+            if urgency > 0.5:  # Threshold, adjust as needed
+                alert = {
+                    'timestamp': datetime.now().isoformat(),
+                    'ticker': article.get('symbol',''),
+                    'headline': article.get('headline',''),
+                    'summary': article.get('summary',''),
+                    'sentiment': article['sentiment'],
+                    'impact_keywords': ','.join(keywords),
+                    'urgency_score': round(min(urgency, 1.0),2)
+                }
+                self.db.save_alert(alert)

agents/datacollector.py

@@ -0,0 +1,11 @@
+import requests
+from config.config import Config
+
+class DataCollectionAgent:
+    def __init__(self, config):
+        self.config = config
+
+    def collect_news(self, ticker):
+        url = f"https://finnhub.io/api/v1/company-news?symbol={ticker}&from=2023-01-01&to=2025-12-31&token={self.config.FINNHUB_API_KEY}"
+        resp = requests.get(url)
+        return resp.json() if resp.status_code == 200 else []

agents/filterclassifier.py

@@ -0,0 +1,15 @@
+from config.config import Config
+
+class FilterClassificationAgent:
+    def __init__(self, config):
+        self.config = config
+
+    def filter_articles(self, articles):
+        impact_kw = self.config.IMPACT_KEYWORDS
+        filtered = []
+        for a in articles:
+            in_head = any(k in a.get('headline', '').lower() for k in impact_kw)
+            in_sum = any(k in a.get('summary', '').lower() for k in impact_kw)
+            if in_head or in_sum:
+                filtered.append(a)
+        return filtered

agents/learningagent.py

@@ -0,0 +1,8 @@
+class LearningAgent:
+    def __init__(self, config, database):
+        self.config = config
+        self.db = database
+
+    def learn_and_optimize(self):
+        # Placeholder: you can add logic to update thresholds, expand keywords, or adjust urgency rules based on data
+        print("Learning/optimization step placeholder.")

agents/orchestrator.py

@@ -0,0 +1,14 @@
+class OrchestratorAgent:
+    def __init__(self, datacol, filterer, sentiment, alert, learner):
+        self.datacollector = datacol
+        self.filterclassifier = filterer
+        self.sentimentanalyzer = sentiment
+        self.alertcoordinator = alert
+        self.learningagent = learner
+
+    def process(self, ticker):
+        news = self.datacollector.collect_news(ticker)
+        filtered = self.filterclassifier.filter_articles(news)
+        analyzed = self.sentimentanalyzer.batch_analyze(filtered)
+        self.alertcoordinator.process_alerts(analyzed)
+        return analyzed

agents/sentimentanalyzer.py

@@ -0,0 +1,17 @@
+from vaderSentiment.vaderSentiment import SentimentIntensityAnalyzer
+from textblob import TextBlob
+
+class SentimentAnalysisAgent:
+    def __init__(self):
+        self.analyzer = SentimentIntensityAnalyzer()
+
+    def analyze(self, news_item):
+        text = news_item.get("headline","") + " " + news_item.get("summary","")
+        vs = self.analyzer.polarity_scores(text)
+        blob = TextBlob(text)
+        final_score = 0.7*vs['compound'] + 0.3*blob.sentiment.polarity
+        news_item['sentiment'] = final_score
+        return news_item
+
+    def batch_analyze(self, articles):
+        return [self.analyze(a) for a in articles]

app.py

@@ -0,0 +1,54 @@
+import gradio as gr
+from config.config import Config
+from utils.database import NewsDatabase
+from agents.datacollector import DataCollectionAgent
+from agents.filterclassifier import FilterClassificationAgent
+from agents.sentimentanalyzer import SentimentAnalysisAgent
+from agents.alertcoordinator import AlertCoordinatorAgent
+from agents.learningagent import LearningAgent
+from agents.orchestrator import OrchestratorAgent
+from apscheduler.schedulers.background import BackgroundScheduler
+
+config = Config()
+db = NewsDatabase(config.DATABASE_PATH)
+datacol = DataCollectionAgent(config)
+filterer = FilterClassificationAgent(config)
+sentiment = SentimentAnalysisAgent()
+alert = AlertCoordinatorAgent(config, db)
+learner = LearningAgent(config, db)
+orchestrator = OrchestratorAgent(datacol, filterer, sentiment, alert, learner)
+
+scheduler = BackgroundScheduler()
+scheduler.add_job(lambda: orchestrator.process("AAPL"), 'interval', minutes=config.CHECK_INTERVAL_MINUTES)
+scheduler.add_job(learner.learn_and_optimize, 'cron', hour=0)
+scheduler.start()
+
+
+def user_request(user_input):
+    # Simple ticker extraction (can expand this for more NLP):
+    words = user_input.lower().split()
+    ticker = None
+    for word in words:
+        if word.isalpha() and len(word) <= 5:
+            ticker = word.upper()
+            break
+    if not ticker:
+        return "Please specify a stock ticker (e.g. AAPL)."
+    # Run through orchestrator:
+    results = orchestrator.process(ticker)
+    if not results:
+        return f"No recent news found for {ticker}."
+    output = []
+    for r in results:
+        output.append(f"Headline: {r['headline']}\nSentiment: {r['sentiment']:.2f}\nSummary: {r['summary']}\n")
+    return "\n".join(output)
+
+
+#def user_request(ticker):
+ #   result = orchestrator.process(ticker.upper())
+  #  return str(result)
+
+#iface = gr.Interface(fn=user_request, inputs=gr.Textbox(label="Stock Symbol"), outputs=gr.Textbox(label="News/Alerts"), title="Agentic Financial News Monitor")
+iface = gr.Interface(fn=user_request, inputs=gr.Textbox(label="What should I track?"), outputs=gr.Textbox(label="Latest News/Sentiment"))
+iface.launch()
+

config.py

@@ -0,0 +1,17 @@
+import os
+from dotenv import load_dotenv
+load_dotenv()
+
+class Config:
+    FINNHUB_API_KEY = os.getenv("FINNHUB_API_KEY")
+    ALPHAVANTAGE_API_KEY = os.getenv("ALPHAVANTAGE_API_KEY")
+    GNEWS_API_KEY = os.getenv("GNEWS_API_KEY")
+<<<<<<< HEAD
+=======
+    DATABASE_PATH = "data/newsalerts.db"
+    IMPACT_KEYWORDS = ["pandemic", "war", "conflict", "earnings", "bankruptcy", "inflation", "recession", "federal reserve", "interest rate"]
+>>>>>>> 8273ac1405391cea658ce600557f91e4f7b7219b
+    CHECK_INTERVAL_MINUTES = 15
+    IMPACT_KEYWORDS = ["pandemic","war","conflict","earnings","bankruptcy","inflation","recession","federal reserve","interest rate"]
+    DATABASE_PATH = "data/newsalerts.db"
+

dashboard.py

@@ -0,0 +1,17 @@
+import sqlite3
+
+class NewsDatabase:
+    def __init__(self, db_path):
+        self.conn = sqlite3.connect(db_path)
+        self.init_db()
+    def init_db(self):
+        c = self.conn.cursor()
+        c.execute('''CREATE TABLE IF NOT EXISTS alerts (timestamp TEXT, ticker TEXT, headline TEXT, summary TEXT, sentiment REAL, impact_keywords TEXT, urgency REAL)''')
+        self.conn.commit()
+    def save_alert(self, alert):
+        c = self.conn.cursor()
+        c.execute('''INSERT INTO alerts VALUES (?,?,?,?,?,?,?)''', (
+            alert['timestamp'], alert['ticker'], alert['headline'], alert['summary'],
+            alert['sentiment'], alert['impact_keywords'], alert['urgency_score']
+        ))
+        self.conn.commit()

requirements.txt

@@ -0,0 +1,14 @@
+requests
+aiohttp
+pandas
+nltk
+vaderSentiment
+textblob
+apscheduler
+flask
+python-dotenv
+gradio
+transformers       # Optional, for advanced NLP
+torch              # Optional, if using transformers
+finnhub-python
+alpha-vantage