Spaces:
Sleeping
Sleeping
Update app.py
Browse files
app.py
CHANGED
|
@@ -51,73 +51,153 @@ if test_emb.shape[0] != anchor_index.d:
|
|
| 51 |
# Import retrieval logic
|
| 52 |
from retrieval import hybrid_search_fixed
|
| 53 |
from utils import preprocess_query
|
| 54 |
-
|
| 55 |
-
#
|
| 56 |
-
#
|
| 57 |
-
# -----------------------------
|
| 58 |
|
| 59 |
def search_hadith(query, top_k):
|
| 60 |
-
if query.strip()
|
| 61 |
-
return
|
| 62 |
-
|
| 63 |
-
|
| 64 |
-
|
| 65 |
-
|
| 66 |
-
|
| 67 |
-
|
| 68 |
-
|
| 69 |
-
|
| 70 |
-
|
| 71 |
-
|
| 72 |
-
|
| 73 |
-
|
| 74 |
-
top_k=int(top_k)
|
| 75 |
-
)
|
| 76 |
-
|
| 77 |
-
return results_df[["main_subj", "matn_text", "url"]] \
|
| 78 |
-
.rename(columns={
|
| 79 |
-
"main_subj": "الموضوع",
|
| 80 |
-
"matn_text": "نص الحديث",
|
| 81 |
-
"url": "hadith page on Islamweb.net"
|
| 82 |
-
})
|
| 83 |
-
except Exception as e:
|
| 84 |
-
print(f"Error in search: {e}")
|
| 85 |
-
return pd.DataFrame({
|
| 86 |
-
"الموضوع": ["Error"],
|
| 87 |
-
"نص الحديث": [str(e)],
|
| 88 |
-
"hadith page on Islamweb.net": [""]
|
| 89 |
-
})
|
| 90 |
-
|
| 91 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 92 |
# Gradio Interface
|
|
|
|
|
|
|
| 93 |
interface = gr.Interface(
|
| 94 |
fn=search_hadith,
|
| 95 |
inputs=[
|
| 96 |
gr.Textbox(
|
| 97 |
-
label="
|
| 98 |
-
placeholder="
|
|
|
|
| 99 |
),
|
| 100 |
gr.Slider(
|
| 101 |
minimum=1,
|
| 102 |
maximum=20,
|
| 103 |
value=5,
|
| 104 |
step=1,
|
| 105 |
-
label="
|
| 106 |
)
|
| 107 |
],
|
| 108 |
-
outputs=gr.
|
| 109 |
-
|
| 110 |
-
|
| 111 |
-
|
| 112 |
-
|
| 113 |
-
|
| 114 |
-
|
| 115 |
-
|
| 116 |
-
|
| 117 |
-
["حقوق الجار", 5],
|
| 118 |
-
]
|
| 119 |
)
|
| 120 |
|
| 121 |
-
#
|
|
|
|
|
|
|
|
|
|
| 122 |
if __name__ == "__main__":
|
| 123 |
interface.launch()
|
|
|
|
| 51 |
# Import retrieval logic
|
| 52 |
from retrieval import hybrid_search_fixed
|
| 53 |
from utils import preprocess_query
|
| 54 |
+
# ===============================
|
| 55 |
+
# Search Function
|
| 56 |
+
# ===============================
|
|
|
|
| 57 |
|
| 58 |
def search_hadith(query, top_k):
|
| 59 |
+
if not query.strip():
|
| 60 |
+
return "<p class='empty'>Please enter a search query.</p>"
|
| 61 |
+
|
| 62 |
+
results_df, _ = hybrid_search_fixed(
|
| 63 |
+
query=query,
|
| 64 |
+
df=df,
|
| 65 |
+
bm25=bm25,
|
| 66 |
+
model=model,
|
| 67 |
+
hadith_embeddings=hadith_embeddings,
|
| 68 |
+
anchor_index=anchor_index,
|
| 69 |
+
anchor_dict=anchor_dict,
|
| 70 |
+
unique_anchor_texts=unique_anchor_texts,
|
| 71 |
+
top_k=int(top_k)
|
| 72 |
+
)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 73 |
|
| 74 |
+
html = "<div class='results'>"
|
| 75 |
+
|
| 76 |
+
for i, row in results_df.iterrows():
|
| 77 |
+
html += f"""
|
| 78 |
+
<div class="card">
|
| 79 |
+
<div class="card-header">
|
| 80 |
+
<span class="index">#{i+1}</span>
|
| 81 |
+
<span class="topic">{row['main_subj']}</span>
|
| 82 |
+
</div>
|
| 83 |
+
|
| 84 |
+
<div class="text">
|
| 85 |
+
{row['clean_text']}
|
| 86 |
+
</div>
|
| 87 |
+
|
| 88 |
+
<div class="footer">
|
| 89 |
+
<a href="{row['url']}" target="_blank">
|
| 90 |
+
📖 View Hadith Source
|
| 91 |
+
</a>
|
| 92 |
+
</div>
|
| 93 |
+
</div>
|
| 94 |
+
"""
|
| 95 |
+
|
| 96 |
+
html += "</div>"
|
| 97 |
+
return html
|
| 98 |
+
|
| 99 |
+
# ===============================
|
| 100 |
+
# Custom CSS
|
| 101 |
+
# ===============================
|
| 102 |
+
|
| 103 |
+
custom_css = """
|
| 104 |
+
body {
|
| 105 |
+
background-color: #f7f9fc;
|
| 106 |
+
}
|
| 107 |
+
|
| 108 |
+
.results {
|
| 109 |
+
direction: rtl;
|
| 110 |
+
font-family: "Tahoma", "Arial", sans-serif;
|
| 111 |
+
}
|
| 112 |
+
|
| 113 |
+
.card {
|
| 114 |
+
background: white;
|
| 115 |
+
border-radius: 12px;
|
| 116 |
+
padding: 18px 20px;
|
| 117 |
+
margin-bottom: 16px;
|
| 118 |
+
box-shadow: 0 4px 14px rgba(0,0,0,0.06);
|
| 119 |
+
}
|
| 120 |
+
|
| 121 |
+
.card-header {
|
| 122 |
+
display: flex;
|
| 123 |
+
align-items: center;
|
| 124 |
+
gap: 10px;
|
| 125 |
+
margin-bottom: 12px;
|
| 126 |
+
}
|
| 127 |
+
|
| 128 |
+
.index {
|
| 129 |
+
background: #2563eb;
|
| 130 |
+
color: white;
|
| 131 |
+
padding: 4px 10px;
|
| 132 |
+
border-radius: 20px;
|
| 133 |
+
font-size: 14px;
|
| 134 |
+
}
|
| 135 |
+
|
| 136 |
+
.topic {
|
| 137 |
+
font-weight: bold;
|
| 138 |
+
color: #1f2937;
|
| 139 |
+
font-size: 16px;
|
| 140 |
+
}
|
| 141 |
+
|
| 142 |
+
.text {
|
| 143 |
+
line-height: 1.9;
|
| 144 |
+
font-size: 16px;
|
| 145 |
+
color: #111827;
|
| 146 |
+
margin-bottom: 14px;
|
| 147 |
+
}
|
| 148 |
+
|
| 149 |
+
.footer a {
|
| 150 |
+
text-decoration: none;
|
| 151 |
+
color: #2563eb;
|
| 152 |
+
font-weight: bold;
|
| 153 |
+
}
|
| 154 |
+
|
| 155 |
+
.footer a:hover {
|
| 156 |
+
text-decoration: underline;
|
| 157 |
+
}
|
| 158 |
+
|
| 159 |
+
.empty {
|
| 160 |
+
direction: ltr;
|
| 161 |
+
font-size: 16px;
|
| 162 |
+
color: #6b7280;
|
| 163 |
+
text-align: center;
|
| 164 |
+
}
|
| 165 |
+
"""
|
| 166 |
+
|
| 167 |
+
# ===============================
|
| 168 |
# Gradio Interface
|
| 169 |
+
# ===============================
|
| 170 |
+
|
| 171 |
interface = gr.Interface(
|
| 172 |
fn=search_hadith,
|
| 173 |
inputs=[
|
| 174 |
gr.Textbox(
|
| 175 |
+
label="🔍 Search Query",
|
| 176 |
+
placeholder="e.g. Importance of intention in Islam",
|
| 177 |
+
lines=2
|
| 178 |
),
|
| 179 |
gr.Slider(
|
| 180 |
minimum=1,
|
| 181 |
maximum=20,
|
| 182 |
value=5,
|
| 183 |
step=1,
|
| 184 |
+
label="📌 Number of Results"
|
| 185 |
)
|
| 186 |
],
|
| 187 |
+
outputs=gr.HTML(),
|
| 188 |
+
title="📚 Intelligent Hadith Search Engine",
|
| 189 |
+
description="""
|
| 190 |
+
An AI-powered semantic search engine for Hadith retrieval.
|
| 191 |
+
The system combines lexical, semantic, and topic-aware retrieval
|
| 192 |
+
to return results based on meaning — not just keyword matching.
|
| 193 |
+
""",
|
| 194 |
+
css=custom_css,
|
| 195 |
+
allow_flagging="never"
|
|
|
|
|
|
|
| 196 |
)
|
| 197 |
|
| 198 |
+
# ===============================
|
| 199 |
+
# Launch
|
| 200 |
+
# ===============================
|
| 201 |
+
|
| 202 |
if __name__ == "__main__":
|
| 203 |
interface.launch()
|