update sentiment_tool.py

tools/sentiment_tool.py

@@ -1,3 +1,5 @@
+# tools/sentiment_tool.py
+
 import os
 import requests
 from crewai.tools import BaseTool
@@ -5,6 +7,9 @@ from openai import OpenAI
 from typing import Type
 from pydantic import BaseModel, Field
 
+# -------------------------
+# Environment Variables
+# -------------------------
 SERPER_API_KEY = os.getenv("SERPER_API_KEY")
 OPENAI_API_KEY = os.getenv("OPENAI_API_KEY")
 
@@ -14,8 +19,7 @@ client = OpenAI(api_key=OPENAI_API_KEY)
 # Input Schema
 # -------------------------
 class SentimentInput(BaseModel):
-    query: str = Field(default="bitcoin", description="Cryptocurrency name to evaluate sentiment for.")
-
+    query: str = Field(default="bitcoin", description="Cryptocurrency name to analyze.")
 
 # -------------------------
 # Sentiment Tool
@@ -23,110 +27,134 @@ class SentimentInput(BaseModel):
 class SentimentTool(BaseTool):
     name: str = "get_crypto_sentiment"
     description: str = (
-        "Fetches cryptocurrency sentiment from Google News and Reddit "
-        "via Serper.dev, then classifies sentiment as bullish, bearish, or neutral "
-        "and returns structured JSON."
+        "Fetches cryptocurrency sentiment using Serper (news + Reddit), "
+        "then uses GPT-4.1 to classify sentiment as bullish, bearish, or neutral."
     )
     args_schema: Type[BaseModel] = SentimentInput
 
-    # -------------------------
-    # Internal execution
-    # -------------------------
     def _run(self, query: str = "bitcoin") -> dict:
+        headers = {
+            "X-API-KEY": SERPER_API_KEY,
+            "Content-Type": "application/json"
+        }
+
+        # ----------------------------------------
+        # 1) FETCH NEWS USING SERPER
+        # ----------------------------------------
         news_headlines = []
-        reddit_titles = []
         news_error = None
-        reddit_error = None
-
-        # -------------------------
-        # 1) Fetch Google News (Serper News API)
-        # -------------------------
         try:
-            news_url = "https://google.serper.dev/news"
-            headers = {"X-API-KEY": SERPER_API_KEY, "Content-Type": "application/json"}
-            payload = {"q": f"{query} crypto", "num": 5}
+            news_payload = {
+                "q": f"{query} cryptocurrency",
+                "num": 5
+            }
 
-            resp = requests.post(news_url, headers=headers, json=payload, timeout=10)
-            resp.raise_for_status()
+            news_res = requests.post(
+                "https://google.serper.dev/news",
+                headers=headers,
+                json=news_payload,
+                timeout=10
+            )
+            news_res.raise_for_status()
 
-            news_items = resp.json().get("news", [])
-            news_headlines = [item["title"] for item in news_items[:5]]
+            news_json = news_res.json()
+            news_items = news_json.get("news", [])
+            news_headlines = [item.get("title") for item in news_items if "title" in item]
 
         except Exception as e:
             news_error = str(e)
 
-        # -------------------------
-        # 2) Fetch Reddit Discussions (Serper Search API)
-        # -------------------------
+        # ----------------------------------------
+        # 2) FETCH REDDIT POSTS USING SERPER (CORRECT ENDPOINT)
+        # ----------------------------------------
+        reddit_titles = []
+        reddit_error = None
+
         try:
-            reddit_url = "https://google.serper.dev/search"
-            headers = {"X-API-KEY": SERPER_API_KEY, "Content-Type": "application/json"}
-
-            # OPTION C — hybrid search (most accurate)
-            payload = {
-                "q": (
-                    f'"{query}" '
-                    f'(site:reddit.com/r/CryptoCurrency OR '
-                    f'site:reddit.com/r/{query} OR '
-                    f'site:reddit.com/r/Bitcoin OR '
-                    f'site:reddit.com)'
-                ),
-                "num": 5
+            reddit_payload = {
+                "q": query,
+                "num": 5,
+                "reddit": True  # <-- THIS IS THE CORRECT WAY TO FILTER REDDIT RESULTS
             }
 
-            resp = requests.post(reddit_url, headers=headers, json=payload, timeout=10)
-            resp.raise_for_status()
+            reddit_res = requests.post(
+                "https://google.serper.dev/search",
+                headers=headers,
+                json=reddit_payload,
+                timeout=10
+            )
+            reddit_res.raise_for_status()
 
-            organic = resp.json().get("organic", [])
-            reddit_titles = [item["title"] for item in organic[:5]]
+            reddit_json = reddit_res.json()
+            reddit_items = reddit_json.get("reddit", [])
+
+            reddit_titles = [
+                item.get("title")
+                for item in reddit_items
+                if "title" in item
+            ]
 
         except Exception as e:
             reddit_error = str(e)
 
-        # -------------------------
-        # 3) Sentiment Analysis (OpenAI)
-        # -------------------------
+        # ----------------------------------------
+        # 3) COMBINE TEXT FOR SENTIMENT CLASSIFICATION
+        # ----------------------------------------
+        combined_text = (
+            "News Headlines:\n" + "\n".join(news_headlines) +
+            "\n\nReddit Posts:\n" + "\n".join(reddit_titles)
+        )
+
+        # ----------------------------------------
+        # 4) GPT-4.1 SENTIMENT INTERPRETATION
+        # ----------------------------------------
         sentiment_prompt = f"""
 You are a cryptocurrency sentiment analyst.
 
-Based on the following data:
-News Headlines: {news_headlines}
-Reddit Titles: {reddit_titles}
-
-Classify the overall sentiment toward "{query}" as bullish, bearish, or neutral.
-Write a concise reasoning sentence.
+Classify sentiment toward **{query}** based on NEWS and REDDIT content.
 
-Return ONLY valid JSON exactly in this format:
+Return ONLY valid JSON using this format:
 
 {{
-  "sentiment": "<bullish/bearish/neutral>",
-  "reasoning": "<short explanation>",
+  "sentiment": "bullish" | "bearish" | "neutral",
+  "reasoning": "short explanation",
   "news_headlines": [...],
   "reddit_titles": [...],
-  "news_error": "<error or null>",
-  "reddit_error": "<error or null>"
+  "news_error": "... or null",
+  "reddit_error": "... or null"
 }}
+
+Content to analyze:
+{combined_text}
 """
 
         try:
             completion = client.chat.completions.create(
                 model="gpt-4.1",
                 messages=[
-                    {"role": "system", "content": "You return ONLY valid JSON, no prose."},
+                    {"role": "system", "content": "You extract structured sentiment from crypto news and Reddit posts."},
                     {"role": "user", "content": sentiment_prompt}
-                ]
+                ],
+                temperature=0.2
             )
 
-            raw_output = completion.choices[0].message.content
+            # GPT returns JSON string → CrewAI needs dict
+            result_text = completion.choices[0].message.content
+            import json
+            parsed = json.loads(result_text)
+
+            # Ensure errors included even if GPT forgets them
+            parsed["news_error"] = news_error
+            parsed["reddit_error"] = reddit_error
 
-            # The model returns correct JSON; just forward it:
-            return raw_output
+            return parsed
 
         except Exception as e:
             return {
-                "error": f"Sentiment classification failed: {str(e)}",
+                "sentiment": "unknown",
+                "reasoning": "GPT sentiment classification failed.",
                 "news_headlines": news_headlines,
                 "reddit_titles": reddit_titles,
                 "news_error": news_error,
-                "reddit_error": reddit_error
+                "reddit_error": str(e)
             }