Vritika's branch- Added compare companies dashboard updates

app/copilot_engine.py

@@ -0,0 +1,90 @@
+import pandas as pd
+
+
+def answer_query(query, latest_snapshot, high_risk, high_credibility, signal_screener, master_features):
+    q = query.lower().strip()
+
+    if not q:
+        return "Please type a question."
+
+
+    # Top risky companies
+
+    if "top risky" in q or "high risk" in q:
+        top = high_risk.head(5)[["ticker", "risk_score", "ret_20d", "credibility_score"]]
+        return "Top high-risk companies:\n\n" + top.to_string(index=False)
+
+
+    # Top credibility companies
+
+    if "high credibility" in q or "top credibility" in q:
+        top = high_credibility.head(5)[["ticker", "credibility_score", "ret_20d", "risk_score"]]
+        return "Top high-credibility companies:\n\n" + top.to_string(index=False)
+
+
+    # Compare two companies
+
+    if "compare" in q:
+        tickers = [t for t in latest_snapshot["ticker"].dropna().unique() if t.lower() in q]
+        if len(tickers) >= 2:
+            t1, t2 = tickers[0], tickers[1]
+
+            row1 = latest_snapshot[latest_snapshot["ticker"] == t1].tail(1)
+            row2 = latest_snapshot[latest_snapshot["ticker"] == t2].tail(1)
+
+            if not row1.empty and not row2.empty:
+                r1 = row1.iloc[0]
+                r2 = row2.iloc[0]
+
+                return (
+                    f"Comparison: {t1} vs {t2}\n\n"
+                    f"{t1}: risk={r1['risk_score']:.4f}, credibility={r1['credibility_score']:.4f}, ret_20d={r1['ret_20d']:.4f}\n"
+                    f"{t2}: risk={r2['risk_score']:.4f}, credibility={r2['credibility_score']:.4f}, ret_20d={r2['ret_20d']:.4f}"
+                )
+
+        return "I could not identify two valid tickers to compare."
+
+    # Summarize one company
+
+    possible_tickers = [t for t in latest_snapshot["ticker"].dropna().unique() if t.lower() in q]
+
+    if possible_tickers:
+        ticker = possible_tickers[0]
+        company_df = master_features[master_features["ticker"] == ticker].sort_values("date")
+
+        if company_df.empty:
+            return f"No data found for {ticker}."
+
+        latest = company_df.tail(1).iloc[0]
+
+        lines = [f"Summary for {ticker}:"]
+        lines.append(f"- Latest close: {latest['dlyclose']:.2f}")
+
+        if pd.notna(latest.get("ret_20d")):
+            lines.append(f"- 20-day return: {latest['ret_20d']:.4f}")
+
+        if pd.notna(latest.get("risk_score")):
+            lines.append(f"- Risk score: {latest['risk_score']:.4f}")
+
+        if pd.notna(latest.get("credibility_score")):
+            lines.append(f"- Credibility score: {latest['credibility_score']:.4f}")
+
+        if pd.notna(latest.get("misalignment_score")):
+            lines.append(f"- Misalignment score: {latest['misalignment_score']:.4f}")
+
+        if pd.notna(latest.get("net_positivity")):
+            lines.append(f"- Net positivity: {latest['net_positivity']:.4f}")
+
+        if pd.notna(latest.get("numeric_transparency")):
+            lines.append(f"- Numeric transparency: {latest['numeric_transparency']:.4f}")
+
+        return "\n".join(lines)
+
+    # Fallback
+    return (
+        "I can help with:\n"
+        "- show top risky companies\n"
+        "- show high credibility companies\n"
+        "- compare AAPL and MSFT\n"
+        "- summarize NVDA"
+    )