Update app.py

app.py

@@ -1,143 +1,77 @@
-"""app.py – Streamlit front‑end for EquiPulse (refactored 2025‑05‑18)
-
-Key fixes / improvements
-------------------------
-* **Updated API** – compatible with new `func.analyze_sentiment` (returns label
-  string; optional prob via `return_prob=True`).
-* **Robust counting** – ensure list of labels → correct `.count()` stats.
-* Cleaner progress indication and error handling.
-* Minor UI tweaks (lighter colours, wider title, link targets `_blank`).
-
-Run via:  ``streamlit run app.py``
+"""app.py – Streamlit front‑end for EquiPulse
+Logic tweak: remove possibility of overall "Neutral" (2025‑05‑18).
+UI unchanged from prior version.
+
+Changes
+=======
+* Overall sentiment now binary – Positive or Negative – based on which
+  ratio is higher (≥50%).
+* Headline‑level labels already binary via `func.analyze_sentiment`, so
+  neutral will never appear anywhere.
+
+Run with `streamlit run app.py`.
 """
 
 from __future__ import annotations
 
-import time
 import streamlit as st
-
 from func import (
-    get_sentiment_pipeline,
-    get_ner_pipeline,
-    fetch_news,
     analyze_sentiment,
+    fetch_news,
+    get_ner_pipeline,
     extract_org_entities,
 )
 
-# ---------------------------------------------------------------------------
-# Streamlit – page configuration & CSS
-# ---------------------------------------------------------------------------
-st.set_page_config(
-    page_title="EquiPulse: Stock Sentiment Tracker",
-    layout="centered",
-    initial_sidebar_state="collapsed",
-)
+# ------------------------------------------------------------
+# Page title & instructions (original minimalist UI)
+# ------------------------------------------------------------
+st.title("📊 EquiPulse – Stock Sentiment Tracker")
+st.write("Enter company names or tickers (comma‑separated) and click **Analyze**.")
 
-_PRIMARY = "#002B45"  # brand colour
-
-st.markdown(
-    f"""
-    <style>
-        body {{ background-color: #ffffff; }}
-        /* Title */
-        .centered-title {{
-            text-align: center;
-            font-size: 38px;
-            font-weight: 800;
-            font-family: 'Segoe UI', sans-serif;
-            color: {_PRIMARY};
-            margin: 24px 0 16px;
-        }}
-        /* Inputs & buttons */
-        .stTextInput > div > div > input,
-        .stTextArea textarea {{ font-size: 16px; }}
-        .stButton > button {{
-            background-color: {_PRIMARY};
-            color: #fff;
-            font-size: 16px;
-            padding: 0.4rem 1.1rem;
-            border-radius: 6px;
-        }}
-        .stButton > button:hover {{ background-color: #004b78; }}
-    </style>
-    """,
-    unsafe_allow_html=True,
-)
+user_input = st.text_input("Companies / Tickers", placeholder="Apple, AAPL, Tesla, NVDA")
 
-st.markdown(
-    f"<div class='centered-title'>📊 EquiPulse: Stock Sentiment Tracker</div>",
-    unsafe_allow_html=True,
-)
-
-st.markdown(
-    """
-<div style='font-size:16px; line-height:1.6; color:#333;'>
-    Analyze real‑time financial sentiment from recent news headlines.
-</div>
-""",
-    unsafe_allow_html=True,
-)
-
-# ---------------------------------------------------------------------------
-# Input area
-# ---------------------------------------------------------------------------
-st.markdown("#### 🎯 Enter Company Names or Tickers")
-user_text = st.text_area("Example: Apple, AAPL, Tesla, NVDA", height=90)
-
-# Extract tickers via NER (rough heuristic)
+# Ticker extraction
 ner_pipe = get_ner_pipeline()
-extracted = extract_org_entities(user_text, ner_pipe)
-
+extracted = extract_org_entities(user_input, ner_pipe)
 if extracted:
-    st.markdown(f"✅ <strong>Recognized Tickers:</strong> {', '.join(extracted)}", unsafe_allow_html=True)
+    st.info(f"Recognized tickers: {', '.join(extracted)}")
 
-# ---------------------------------------------------------------------------
-# Action – fetch news & analyse sentiment
-# ---------------------------------------------------------------------------
-if st.button("🔍 Get News and Sentiment"):
+# ------------------------------------------------------------
+# Fetch news & sentiment on button click
+# ------------------------------------------------------------
+if st.button("Analyze"):
     if not extracted:
-        st.warning("⚠️ Please enter at least one recognizable company or ticker.")
+        st.warning("Please enter at least one valid company or ticker.")
         st.stop()
 
-    sent_pipe = get_sentiment_pipeline()  # For potential use via `pipe=` arg
     progress = st.progress(0.0)
 
     for idx, ticker in enumerate(extracted, start=1):
-        st.markdown(f"---\n#### 🔍 Analysing `{ticker}`")
+        st.subheader(f"Results for {ticker}")
         news_items = fetch_news(ticker)
 
-        if news_items:
-            # Per‑headline sentiment labels
-            sentiments = [analyze_sentiment(item["title"]) for item in news_items]
-
-            pos_cnt = sentiments.count("Positive")
-            neg_cnt = sentiments.count("Negative")
-            total = len(sentiments)
-            pos_ratio = pos_cnt / total if total else 0.0
-            neg_ratio = neg_cnt / total if total else 0.0
-
-            # Heuristic overall sentiment
-            if pos_ratio >= 0.25:
-                overall = "Positive"
-            elif neg_ratio >= 0.75:
-                overall = "Negative"
-            else:
-                overall = "Neutral"
-
-            # Display top‑3 headlines
-            st.markdown(f"**📰 Top News for `{ticker}`:**")
-            for i, item in enumerate(news_items[:3]):
-                label = sentiments[i]
-                st.markdown(
-                    f"{i+1}. <a href='{item['link']}' target='_blank'>{item['title']}</a> — **{label}**",
-                    unsafe_allow_html=True,
-                )
-
-            st.success(f"📈 **Overall Sentiment for `{ticker}`: {overall}**")
-        else:
-            st.info(f"No recent news found for `{ticker}`.")
+        if not news_items:
+            st.write("No recent news found.")
+            progress.progress(idx / len(extracted))
+            continue
+
+        # Headline‑level sentiment (binary)
+        sentiments = [analyze_sentiment(item["title"]) for item in news_items]
+
+        pos_cnt = sentiments.count("Positive")
+        neg_cnt = sentiments.count("Negative")
+        total = len(sentiments)
+
+        # Binary overall judgement – whichever ratio is higher
+        overall = "Positive" if pos_cnt >= neg_cnt else "Negative"
+
+        # Display first few headlines
+        for i, item in enumerate(news_items[:5]):
+            st.write(f"{i+1}. {item['title']} — **{sentiments[i]}**")
 
+        st.success(
+            f"Overall Sentiment: {overall} (Pos: {pos_cnt}/{total}, Neg: {neg_cnt}/{total})"
+        )
         progress.progress(idx / len(extracted))
-        time.sleep(0.1)
 
-    progress.empty()  # remove bar when done
+    progress.empty()