Update app.py

app.py

@@ -3,7 +3,6 @@ from typing import List, Dict, Any
 import numpy as np
 import yaml
 from smolagents import CodeAgent, HfApiModel, load_tool, tool
-from textblob import TextBlob
 from tools.final_answer import FinalAnswerTool
 from Gradio_UI import GradioUI
 
@@ -71,30 +70,7 @@ def fetch_btc_news() -> List[Dict[str, str]]:
     return [{"title": post["title"], "url": post["url"]} for post in data.get("results", [])[:5]]
 
 
-### 🔹 2. Analyze News Sentiment ###
-@tool
-def analyze_btc_news(articles: List[Dict[str, str]]) -> Dict[str, Any]:
-    """
-    Analyzes the sentiment of Bitcoin news articles.
-
-    :param articles: A list of dictionaries, each with keys 'title' and 'url' representing a news article.
-    :return: A dictionary with the sentiment analysis results, including:
-             - 'sentiment_score': The average sentiment polarity.
-             - 'sentiment_label': A label ("Positive 😊", "Negative 😡", or "Neutral 😐") based on the score.
-             - 'keywords': A list of keywords extracted from the titles.
-    """
-    sentiments = [TextBlob(article["title"]).sentiment.polarity for article in articles]
-    avg_sentiment = np.mean(sentiments) if sentiments else 0
-    sentiment_label = "Positive 😊" if avg_sentiment > 0 else "Negative 😡" if avg_sentiment < 0 else "Neutral 😐"
-    keywords = list(set(word for article in articles for word in article["title"].split()[:5]))
-    return {
-        "sentiment_score": avg_sentiment,
-        "sentiment_label": sentiment_label,
-        "keywords": keywords
-    }
-
-
-### 🔹 3. Get BTC Market Insights ###
+### 🔹 2. Get BTC Market Insights ###
 @tool
 def btc_market_insights() -> Dict[str, Any]:
     """
@@ -124,11 +100,11 @@ def btc_market_insights() -> Dict[str, Any]:
     prices = [entry[1] for entry in hist_data.get("prices", [])]
 
     # Compute moving averages for different timeframes
-    ma_5m = compute_moving_average(prices, 1)   # 5-minute
-    ma_1h = compute_moving_average(prices, 12)  # 1-hour
-    ma_1d = compute_moving_average(prices, 24)  # 1-day
-    ma_7d = compute_moving_average(prices, 168) # 7-day
-    ma_1m = compute_moving_average(prices, 720) # 1-month
+    ma_5m = compute_moving_average(prices, 1)    # 5-minute
+    ma_1h = compute_moving_average(prices, 12)   # 1-hour
+    ma_1d = compute_moving_average(prices, 24)   # 1-day
+    ma_7d = compute_moving_average(prices, 168)  # 7-day
+    ma_1m = compute_moving_average(prices, 720)  # 1-month
     ma_1y = compute_moving_average(prices, 8760) # 1-year
 
     # Fetch Fear & Greed Index
@@ -151,19 +127,17 @@ def btc_market_insights() -> Dict[str, Any]:
     }
 
 
-### 🔹 4. Make Buy/Sell/HODL Decision ###
+### 🔹 3. Make Buy/Sell/HODL Decision ###
 @tool
-def btc_decision(news_analysis: Dict[str, Any], market_insights: Dict[str, Any]) -> Dict[str, str]:
+def btc_decision(market_insights: Dict[str, Any]) -> Dict[str, str]:
     """
-    Aggregates news sentiment and market insights to suggest a Buy, Sell, or HODL decision.
+    Aggregates market insights to suggest a Buy, Sell, or HODL decision.
 
-    :param news_analysis: Output from the news sentiment analysis tool.
     :param market_insights: BTC market data from the market insights tool.
     :return: A dictionary mapping timeframes to a decision ("BUY 🚀", "SELL 📉", or "HODL 🤔").
     """
-    sentiment_score = news_analysis["sentiment_score"]
-    fgi = int(market_insights["fear_greed"])
     btc_price = market_insights["price"]
+    fgi = int(market_insights["fear_greed"])
 
     # Moving averages for different timeframes
     ma_5m = market_insights["ma_5m"]
@@ -182,20 +156,18 @@ def btc_decision(news_analysis: Dict[str, Any], market_insights: Dict[str, Any])
         "1Y": "HODL 🤔"
     }
 
-    def make_decision(timeframe: str, ma: float, price: float, sentiment: float, fear_greed: int) -> str:
-        """
-        Helper function to decide action based on moving average, current price, sentiment, and fear & greed index.
-        """
-        if price > ma and fear_greed < 40 and sentiment > 0:
+    # Helper function to update decision based on parameters
+    def make_decision(timeframe: str, ma: float) -> str:
+        if btc_price > ma and fgi < 40:
             decision[timeframe] = "BUY 🚀"
-        elif price < ma and fear_greed > 60 and sentiment < 0:
+        elif btc_price < ma and fgi > 60:
             decision[timeframe] = "SELL 📉"
         return decision[timeframe]
 
-    decision["5m-1H"] = make_decision("5m-1H", ma_5m, btc_price, sentiment_score, fgi)
-    decision["1D"] = make_decision("1D", ma_1d, btc_price, sentiment_score, fgi)
-    decision["7D"] = make_decision("7D", ma_7d, btc_price, sentiment_score, fgi)
-    decision["1M"] = make_decision("1M", ma_1m, btc_price, sentiment_score, fgi)
+    decision["5m-1H"] = make_decision("5m-1H", ma_5m)
+    decision["1D"] = make_decision("1D", ma_1d)
+    decision["7D"] = make_decision("7D", ma_7d)
+    decision["1M"] = make_decision("1M", ma_1m)
 
     return decision
 
@@ -203,7 +175,7 @@ def btc_decision(news_analysis: Dict[str, Any], market_insights: Dict[str, Any])
 # --- Agent Setup ---
 agent = CodeAgent(
     model=model,
-    tools=[final_answer, fetch_btc_news, analyze_btc_news, btc_market_insights, btc_decision],
+    tools=[final_answer, fetch_btc_news, btc_market_insights, btc_decision],
     max_steps=6,
     verbosity_level=1,
     prompt_templates=prompt_templates