Update mentions_dashboard.py

mentions_dashboard.py

@@ -1,27 +1,36 @@
-import json
 import time
-import re
-import os
 from datetime import datetime
-
+from collections import defaultdict
 from urllib.parse import quote
-from requests_oauthlib import OAuth2Session
-from textblob import TextBlob
+
 import matplotlib.pyplot as plt
+from transformers import pipeline
 
 from sessions import create_session
 
+# Load transformer-based sentiment model globally
+sentiment_pipeline = pipeline("text-classification", model="tabularisai/multilingual-sentiment-analysis")
 
 def extract_text_from_commentary(commentary):
+    import re
     return re.sub(r"{.*?}", "", commentary).strip()
 
-
-def analyze_sentiment(text):
-    return TextBlob(text).sentiment.polarity
-
+def classify_sentiment(text):
+    try:
+        result = sentiment_pipeline(text[:512])  # Limit to 512 chars for transformers
+        label = result[0]['label'].upper()
+        if label in ['POSITIVE', 'VERY POSITIVE']:
+            return 'Positive 👍'
+        elif label in ['NEGATIVE', 'VERY NEGATIVE']:
+            return 'Negative 👎'
+        elif label == 'NEUTRAL':
+            return 'Neutral 😐'
+        else:
+            return 'Unknown'
+    except Exception as e:
+        return 'Error'
 
 def generate_mentions_dashboard(comm_client_id, comm_token_dict):
-
     org_urn = "urn:li:organization:19010008"
     encoded_urn = quote(org_urn, safe='')
 
@@ -40,7 +49,6 @@ def generate_mentions_dashboard(comm_client_id, comm_token_dict):
 
     all_notifications = []
     start = 0
-
     while True:
         url = f"{base_url}&start={start}"
         resp = session.get(url)
@@ -57,7 +65,6 @@ def generate_mentions_dashboard(comm_client_id, comm_token_dict):
         start += len(elements)
         time.sleep(0.5)
 
-    # Extract mentions and their share URNs
     mention_shares = [e.get("generatedActivity") for e in all_notifications if e.get("action") == "SHARE_MENTION"]
     mention_data = []
 
@@ -77,44 +84,60 @@ def generate_mentions_dashboard(comm_client_id, comm_token_dict):
             continue
 
         commentary = extract_text_from_commentary(commentary_raw)
-        sentiment = analyze_sentiment(commentary)
+        sentiment_label = classify_sentiment(commentary)
         timestamp = post.get("createdAt", 0)
         dt = datetime.fromtimestamp(timestamp / 1000.0)
 
         mention_data.append({
             "date": dt,
             "text": commentary,
-            "sentiment": sentiment
+            "sentiment": sentiment_label
         })
 
-    # Save HTML
-    html_parts = ["<h2 style='text-align:center;'>📣 Mentions Sentiment Dashboard</h2>"]
+    # --- HTML rendering ---
+    html_parts = [
+        "<h2 style='text-align:center;'>📣 Mentions Sentiment Dashboard</h2>"
+    ]
+
     for mention in mention_data:
+        short_text = (mention["text"][:200] + "…") if len(mention["text"]) > 200 else mention["text"]
         html_parts.append(f"""
-            <div style='border:1px solid #ccc; border-radius:10px; padding:15px; margin:10px;'>
-                <p><strong>Date:</strong> {mention["date"].strftime('%Y-%m-%d')}</p>
-                <p>{mention["text"]}</p>
-                <p><strong>Sentiment:</strong> {mention["sentiment"]:.2f}</p>
+            <div style='border:1px solid #ddd; border-radius:12px; padding:15px; margin:15px; box-shadow:2px 2px 8px rgba(0,0,0,0.05); background:#fafafa;'>
+                <p><strong>📅 Date:</strong> {mention["date"].strftime('%Y-%m-%d')}</p>
+                <p style='color:#333;'>{short_text}</p>
+                <p><strong>Sentiment:</strong> {mention["sentiment"]}</p>
             </div>
         """)
 
-    html_path = "mentions_dashboard.html"
-    with open(html_path, "w", encoding="utf-8") as f:
-        f.write("\n".join(html_parts))
+    html_content = "\n".join(html_parts)
+
+    # --- Plotting ---
+    from matplotlib.figure import Figure
+    fig = Figure(figsize=(12, 6))
+    ax = fig.subplots()
+    fig.subplots_adjust(bottom=0.2)
 
-    # Plot
     if mention_data:
-        dates = [m["date"] for m in mention_data]
-        sentiments = [m["sentiment"] for m in mention_data]
-
-        plt.figure(figsize=(10, 5))
-        plt.plot(dates, sentiments, marker='o', linestyle='-', color='blue')
-        plt.title("Sentiment Over Time")
-        plt.xlabel("Date")
-        plt.ylabel("Sentiment")
-        plt.grid(True)
-        plt.tight_layout()
-        plt.savefig("mentions_sentiment_plot.png")
-        plt.close()
-
-    return html_path
+        # Sort by date
+        mention_data.sort(key=lambda x: x["date"])
+
+        date_labels = [m["date"].strftime('%Y-%m-%d') for m in mention_data]
+        sentiment_scores = [1 if m["sentiment"] == "Positive 👍" else
+                            -1 if m["sentiment"] == "Negative 👎" else
+                            0 for m in mention_data]
+
+        ax.plot(date_labels, sentiment_scores, marker='o', linestyle='-', color='#0073b1')
+        ax.set_title("📊 Mention Sentiment Over Time")
+        ax.set_xlabel("Date")
+        ax.set_ylabel("Sentiment Score (1=👍, 0=😐, -1=����)")
+        ax.tick_params(axis='x', rotation=45)
+        ax.grid(True, linestyle='--', alpha=0.6)
+        ax.set_ylim([-1.2, 1.2])
+    else:
+        ax.text(0.5, 0.5, "No mention sentiment data available.", 
+                ha='center', va='center', transform=ax.transAxes, fontsize=12, color='grey')
+        ax.set_xticks([])
+        ax.set_yticks([])
+        ax.set_title("📊 Mention Sentiment Over Time")
+
+    return html_content, fig