Update app.py

app.py

@@ -3,96 +3,63 @@ import pandas as pd
 import numpy as np
 import faiss
 import pickle
-import html
+
 from sentence_transformers import SentenceTransformer
 
-# ===============================
-# Load data & models (ONCE)
-# ===============================
+# Load data & models ONCE
 
+# Load dataset
 df = pd.read_csv("data/hadith.csv")
+
+# Load embeddings
 hadith_embeddings = np.load("data/hadith_embeddings.npy")
 print(f"Loaded hadith embeddings: {hadith_embeddings.shape}")
 
+# Load BM25
 with open("data/bm25.pkl", "rb") as f:
     bm25 = pickle.load(f)
 
+# Load anchor FAISS index
 anchor_index = faiss.read_index("data/faiss_anchor.index")
 print(f"Anchor index dimension: {anchor_index.d}")
 
+# Load anchor mapping
 with open("data/anchor_dict.pkl", "rb") as f:
     anchor_dict = pickle.load(f)
 
 with open("data/unique_anchor_texts.pkl", "rb") as f:
     unique_anchor_texts = pickle.load(f)
 
-model = SentenceTransformer("sentence-transformers/paraphrase-multilingual-MiniLM-L12-v2")
+# Load embedding model
+model = SentenceTransformer(
+    "sentence-transformers/paraphrase-multilingual-MiniLM-L12-v2"
+)
 model.max_seq_length = 512
 
-# Quick dimension check
+# Test embedding dimension
 test_emb = model.encode("test", normalize_embeddings=True)
 print(f"Model embedding dimension: {test_emb.shape}")
+
+# Verify dimensions match
 if test_emb.shape[0] != anchor_index.d:
     raise ValueError(
         f"Dimension mismatch! Model outputs {test_emb.shape[0]}D but "
-        f"anchor_index expects {anchor_index.d}D. Rebuild your anchor_index with the same model."
+        f"anchor_index expects {anchor_index.d}D. "
+        f"Rebuild your anchor_index with the same model."
     )
 
-# Import retrieval
+# Import retrieval logic
 from retrieval import hybrid_search_fixed
 from utils import preprocess_query
-
 # ===============================
-# Helper: safe getter & short preview (robust version)
+# Search Function
 # ===============================
-def safe_get(row, col):
-    """
-    Robust getter that works if `row` is a pandas Series or a dict-like object.
-    Returns empty string if column missing or value is NaN/empty.
-    """
-    try:
-        # try Series.get first (works for pandas Series)
-        v = None
-        if hasattr(row, "get"):
-            v = row.get(col, "")
-        else:
-            # fallback for other mapping-like rows
-            v = row[col] if col in row else ""
-        if v is None:
-            return ""
-        if isinstance(v, float) and np.isnan(v):
-            return ""
-        v_str = str(v).strip()
-        return v_str
-    except Exception:
-        return ""
-
-def first_nonempty(row, cols):
-    """
-    Return the first non-empty value found in `cols` for `row`.
-    Example: first_nonempty(row, ["isnad_text","hadith_title","main_subj"])
-    """
-    for c in cols:
-        val = safe_get(row, c)
-        if val:
-            return val
-    return ""
 
-def short_preview(text, length=300):
-    text = (text or "").strip()
-    if len(text) <= length:
-        return text
-    # safe break
-    return text[:length].rsplit(' ', 1)[0] + "..."
-
-# ===============================
-# Search Function (updated)
-# ===============================
 def search_hadith(query, top_k):
-    if not query or not str(query).strip():
+    if not query.strip():
         return "<p class='empty'>Please enter a search query.</p>"
 
-    results_df, debug = hybrid_search_fixed(
+    results_df, _ = hybrid_search_fixed(
         query=query,
         df=df,
         bm25=bm25,
@@ -105,174 +72,121 @@ def search_hadith(query, top_k):
         top_k=int(top_k)
     )
 
-    final_scores = debug.get("final_scores") if isinstance(debug, dict) else None
-
-    html_parts = []
-    html_parts.append("<div class='results'>")
-
-    for rank, (_, row) in enumerate(results_df.iterrows(), start=1):
-        hadith_idx = row.name if hasattr(row, "name") else ""
-
-        # Try multiple candidate columns for the title; isnad_text is first priority
-        title_candidates = ["isnad_text", "hadith_title", "title", "main_subj"]
-        title_val = first_nonempty(row, title_candidates)
-        if not title_val:
-            # fallback to a short generated title from matn_text
-            title_val = short_preview(safe_get(row, "matn_text"), 60) or f"Hadith #{hadith_idx}"
-
-        # matn_text (main hadith body) - preserve Arabic text, escape HTML
-        matn_text = safe_get(row, "matn_text")
-        matn_text_escaped = html.escape(matn_text).replace("\n", "<br>")
+    html = "<div class='results'>"
 
-        topic = html.escape(safe_get(row, "main_subj"))
-        source_url = safe_get(row, "url")
-
-        # score (if available)
-        score_display = ""
-        if isinstance(final_scores, (list, np.ndarray)):
-            try:
-                s = float(final_scores[int(hadith_idx)])
-                score_display = f"{s:.3f}"
-            except Exception:
-                score_display = ""
-
-        preview = short_preview(matn_text, length=360)
-        preview_escaped = html.escape(preview).replace("\n", "<br>")
-
-        card_html = f"""
+    for i, row in results_df.iterrows():
+        html += f"""
         <div class="card">
             <div class="card-header">
-                <div class="left">
-                    <div class="index">#{rank}</div>
-                    <div class="title">{html.escape(title_val)}</div>
-                    <div class="topic">Topic: {topic}</div>
-                </div>
-                <div class="right-meta">
-                    <div class="meta">ID: <span class="meta-val">{hadith_idx}</span></div>
-                    {f'<div class="meta">Score: <span class="meta-val">{score_display}</span></div>' if score_display else ''}
-                </div>
+                <span class="index">#{i+1}</span>
+                <span class="topic">{row['main_subj']}</span>
             </div>
-            <div class="text-rtl">
-                <details>
-                    <summary class="summary">{preview_escaped}</summary>
-                    <div class="full-text">{matn_text_escaped}</div>
-                </details>
+            <div class="text">
+                {row['isnad_text']}
+                {row['matn_text']}
             </div>
             <div class="footer">
-                {"<a href='" + html.escape(source_url) + "' target='_blank' rel='noopener noreferrer'>📖 View source</a>" if source_url else ""}
+                <a href="{row['url']}" target="_blank">
+                    📖 View Hadith Source
+                </a>
             </div>
         </div>
         """
-        html_parts.append(card_html)
 
-    html_parts.append("</div>")
-    return "\n".join(html_parts)
+    html += "</div>"
+    return html
 
 # ===============================
-# Custom CSS (polished, professional)
+# Custom CSS
 # ===============================
-custom_css = """
-/* Page */
-body { background-color: #f5f7fb; font-family: Inter, "Segoe UI", Tahoma, Arial, sans-serif; }
-
-/* Container (results) keeps RTL for hadith text blocks but LTR for metadata */
-.results { margin-top: 10px; }
 
-/* Card */
+custom_css = """
+body {
+    background-color: #f7f9fc;
+}
+.results {
+    direction: rtl;
+    font-family: "Tahoma", "Arial", sans-serif;
+}
 .card {
-  background: #ffffff;
-  border-radius: 12px;
-  padding: 18px;
-  margin-bottom: 14px;
-  box-shadow: 0 6px 20px rgba(18, 38, 63, 0.06);
-  border: 1px solid rgba(20, 43, 74, 0.03);
+    background: white;
+    border-radius: 12px;
+    padding: 18px 20px;
+    margin-bottom: 16px;
+    box-shadow: 0 4px 14px rgba(0,0,0,0.06);
 }
-
-/* Header layout */
 .card-header {
-  display: flex;
-  justify-content: space-between;
-  align-items: flex-start;
-  gap: 12px;
-  margin-bottom: 10px;
+    display: flex;
+    align-items: center;
+    gap: 10px;
+    margin-bottom: 12px;
 }
-
-.card-header .left { display:flex; flex-direction: column; gap:6px; }
 .index {
-  display:inline-block;
-  background: linear-gradient(90deg,#2563eb,#06b6d4);
-  color: white;
-  padding: 6px 10px;
-  border-radius: 20px;
-  font-weight: 600;
-  font-size: 13px;
-  width: fit-content;
+    background: #2563eb;
+    color: white;
+    padding: 4px 10px;
+    border-radius: 20px;
+    font-size: 14px;
 }
-
-.title {
-  font-size: 18px;
-  font-weight: 700;
-  color: #0f172a;
-}
-
 .topic {
-  font-size: 13px;
-  color: #475569;
-  margin-top: 4px;
+    font-weight: bold;
+    color: #1f2937;
+    font-size: 16px;
+}
+.text {
+    line-height: 1.9;
+    font-size: 16px;
+    color: #111827;
+    margin-bottom: 14px;
 }
-
-/* Right meta */
-.right-meta { text-align: right; min-width: 120px; }
-.meta { font-size: 12px; color: #6b7280; }
-.meta-val { color: #0f172a; font-weight: 600; margin-left: 6px; }
-
-/* Hadith text: keep RTL and readable */
-.text-rtl { direction: rtl; margin-top: 6px; margin-bottom: 10px; }
-.text-rtl .summary { cursor: pointer; font-size: 16px; color: #111827; line-height: 1.9; }
-.text-rtl .full-text { margin-top: 8px; font-size: 16px; color: #111827; line-height: 1.9; }
-
-/* Footer link */
-.footer { margin-top: 12px; }
 .footer a {
-  text-decoration: none;
-  color: #2563eb; font-weight: 600;
+    text-decoration: none;
+    color: #2563eb;
+    font-weight: bold;
 }
-.footer a:hover { text-decoration: underline; }
-
-/* Empty state */
-.empty { font-size: 15px; color: #6b7280; text-align:center; padding: 12px; }
-
-/* Responsive */
-@media (max-width: 700px) {
-  .card-header { flex-direction: column; align-items: flex-start; }
-  .right-meta { text-align: left; }
+.footer a:hover {
+    text-decoration: underline;
+}
+.empty {
+    direction: ltr;
+    font-size: 16px;
+    color: #6b7280;
+    text-align: center;
 }
 """
 
 # ===============================
 # Gradio Interface
 # ===============================
+
 interface = gr.Interface(
     fn=search_hadith,
     inputs=[
         gr.Textbox(
-            label="🔍 Search Query",
-            placeholder="e.g. Importance of intention in Islam / أهمية النية",
+            label="أدخل موضوع البحث أو السؤال",
+            placeholder="مثال: أهمية النية وأثرها في قبول الأعمال",
+
             lines=2
         ),
-        gr.Slider(minimum=1, maximum=20, value=5, step=1, label="📌 Number of Results")
+        gr.Slider(
+            minimum=1,
+            maximum=20,
+            value=5,
+            step=1,
+            label="📌 Number of Results"
+        )
     ],
     outputs=gr.HTML(),
     title="📚 Intelligent Hadith Search Engine",
     description="""
     An AI-powered semantic search engine for Hadith retrieval.
-    Combines lexical (BM25), semantic embeddings, and topic-aware anchors.
+    The system combines lexical, semantic, and topic-aware retrieval
+    to return results based on meaning — not just keyword matching.
     """,
-    examples=[
+     examples=[
         ["أهمية النية وأثرها في قبول الأعمال", 5],
-        ["فضل الصدقة", 5],
-        ["حقوق الوالدين", 5]
-    ],
+        ["فضل الصلاة", 5],
+        ["حقوق الجار", 5]],
     css=custom_css,
     allow_flagging="never"
 )
@@ -280,5 +194,6 @@ interface = gr.Interface(
 # ===============================
 # Launch
 # ===============================
+
 if __name__ == "__main__":
-    interface.launch()
+    interface.launch()