Spaces:
Sleeping
Sleeping
Update app.py
Browse files
app.py
CHANGED
|
@@ -51,8 +51,7 @@ def semantic_search(query, top_k=3):
|
|
| 51 |
results.append({
|
| 52 |
'text': data[idx]['text'],
|
| 53 |
'similarity_score': 1 - distances[0][i] / 2,
|
| 54 |
-
'entity_groups': get_entity_groups(data[idx]['entities'])
|
| 55 |
-
'entities': data[idx]['entities']
|
| 56 |
})
|
| 57 |
return results
|
| 58 |
|
|
@@ -60,42 +59,10 @@ def search_and_format(query):
|
|
| 60 |
results = semantic_search(query)
|
| 61 |
formatted_results = ""
|
| 62 |
for i, result in enumerate(results, 1):
|
| 63 |
-
formatted_results += f"
|
| 64 |
-
formatted_results += "
|
| 65 |
-
|
| 66 |
-
# Split the text into words
|
| 67 |
-
words = result['text'].split()
|
| 68 |
-
|
| 69 |
-
# Create a dictionary to map entity groups to colors
|
| 70 |
-
color_map = {
|
| 71 |
-
'PERSON': 'lightpink',
|
| 72 |
-
'ORG': 'lightblue',
|
| 73 |
-
'PLACE': 'lightyellow',
|
| 74 |
-
'TECHNOLOGY': 'lightgreen',
|
| 75 |
-
'MALWARE': 'plum',
|
| 76 |
-
'ATTACK': 'peachpuff'
|
| 77 |
-
}
|
| 78 |
-
|
| 79 |
-
# Format each word based on its entity group
|
| 80 |
-
formatted_text = []
|
| 81 |
-
for word in words:
|
| 82 |
-
found = False
|
| 83 |
-
for entity in result['entities']:
|
| 84 |
-
if word in entity['word']:
|
| 85 |
-
color = color_map.get(entity['entity_group'], 'lightgray')
|
| 86 |
-
formatted_word = f'<span style="background-color: {color};">{word} <sup>{entity["entity_group"]}</sup></span>'
|
| 87 |
-
formatted_text.append(formatted_word)
|
| 88 |
-
found = True
|
| 89 |
-
break
|
| 90 |
-
if not found:
|
| 91 |
-
formatted_text.append(word)
|
| 92 |
-
|
| 93 |
-
# Join the formatted words back into a sentence
|
| 94 |
-
formatted_results += ' '.join(formatted_text) + "<br><br>"
|
| 95 |
-
formatted_results += f"<strong>相似度分數:</strong> {result['similarity_score']:.4f}<br><br>"
|
| 96 |
-
|
| 97 |
return formatted_results
|
| 98 |
-
|
| 99 |
# 示例問題
|
| 100 |
example_queries = [
|
| 101 |
"Tell me about recent cyber attacks from Russia",
|
|
@@ -112,35 +79,38 @@ example_queries = [
|
|
| 112 |
|
| 113 |
# 自定義 CSS
|
| 114 |
custom_css = """
|
| 115 |
-
|
| 116 |
-
.
|
| 117 |
-
.
|
| 118 |
-
.
|
| 119 |
-
.
|
| 120 |
-
.examples-grid {display: grid; grid-template-columns: repeat(auto-fill, minmax(200px, 1fr)); gap: 10px; margin-top: 20px;}
|
| 121 |
-
.example-button {width: 100%;}
|
| 122 |
-
span sup {font-size: 0.7em; font-weight: bold;}
|
| 123 |
"""
|
| 124 |
|
| 125 |
# 創建Gradio界面
|
| 126 |
with gr.Blocks(css=custom_css) as iface:
|
| 127 |
gr.Markdown("# AskCTI")
|
| 128 |
-
gr.Markdown("輸入查詢以搜索相關威脅情報,將顯示前3
|
| 129 |
-
|
| 130 |
-
with gr.Row(class_name="input-row"):
|
| 131 |
-
query_input = gr.Textbox(lines=3, label="", placeholder="輸入你的查詢...", class_name="query-input")
|
| 132 |
-
submit_btn = gr.Button("查詢", variant="primary")
|
| 133 |
|
| 134 |
-
|
| 135 |
-
|
| 136 |
-
|
| 137 |
-
|
| 138 |
-
|
| 139 |
-
|
| 140 |
-
|
| 141 |
-
)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 142 |
|
| 143 |
submit_btn.click(search_and_format, inputs=[query_input], outputs=[output])
|
|
|
|
| 144 |
|
| 145 |
# 啟動Gradio界面
|
| 146 |
iface.launch()
|
|
|
|
| 51 |
results.append({
|
| 52 |
'text': data[idx]['text'],
|
| 53 |
'similarity_score': 1 - distances[0][i] / 2,
|
| 54 |
+
'entity_groups': get_entity_groups(data[idx]['entities'])
|
|
|
|
| 55 |
})
|
| 56 |
return results
|
| 57 |
|
|
|
|
| 59 |
results = semantic_search(query)
|
| 60 |
formatted_results = ""
|
| 61 |
for i, result in enumerate(results, 1):
|
| 62 |
+
formatted_results += f"{i}. 相似度分數: {result['similarity_score']:.4f}\n"
|
| 63 |
+
formatted_results += f" 情資: {result['text']}\n"
|
| 64 |
+
formatted_results += f" 命名實體: {', '.join(result['entity_groups'])}\n\n"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 65 |
return formatted_results
|
|
|
|
| 66 |
# 示例問題
|
| 67 |
example_queries = [
|
| 68 |
"Tell me about recent cyber attacks from Russia",
|
|
|
|
| 79 |
|
| 80 |
# 自定義 CSS
|
| 81 |
custom_css = """
|
| 82 |
+
.container {display: flex; flex-direction: row;}
|
| 83 |
+
.input-column {flex: 1; padding-right: 20px;}
|
| 84 |
+
.output-column {flex: 2;}
|
| 85 |
+
.examples-list {display: flex; flex-wrap: wrap; gap: 10px;}
|
| 86 |
+
.examples-list > * {flex-basis: calc(50% - 5px);}
|
|
|
|
|
|
|
|
|
|
| 87 |
"""
|
| 88 |
|
| 89 |
# 創建Gradio界面
|
| 90 |
with gr.Blocks(css=custom_css) as iface:
|
| 91 |
gr.Markdown("# AskCTI")
|
| 92 |
+
gr.Markdown("輸入查詢以搜索相關威脅情報,將顯示前3個最相關的結果,包括實體組。")
|
|
|
|
|
|
|
|
|
|
|
|
|
| 93 |
|
| 94 |
+
with gr.Row(equal_height=True):
|
| 95 |
+
with gr.Column(scale=1, min_width=300):
|
| 96 |
+
query_input = gr.Textbox(lines=3, label="")
|
| 97 |
+
submit_btn = gr.Button("查詢")
|
| 98 |
+
#clear_btn = gr.Button("清除")
|
| 99 |
+
|
| 100 |
+
gr.Markdown("### 範例查詢")
|
| 101 |
+
for i in range(0, len(example_queries), 2):
|
| 102 |
+
with gr.Row():
|
| 103 |
+
for j in range(2):
|
| 104 |
+
if i + j < len(example_queries):
|
| 105 |
+
gr.Button(example_queries[i+j]).click(
|
| 106 |
+
lambda x: x, inputs=[gr.Textbox(value=example_queries[i+j], visible=False)], outputs=[query_input]
|
| 107 |
+
)
|
| 108 |
+
|
| 109 |
+
with gr.Column(scale=2):
|
| 110 |
+
output = gr.Textbox(lines=20, label="")
|
| 111 |
|
| 112 |
submit_btn.click(search_and_format, inputs=[query_input], outputs=[output])
|
| 113 |
+
#clear_btn.click(lambda: "", outputs=[query_input])
|
| 114 |
|
| 115 |
# 啟動Gradio界面
|
| 116 |
iface.launch()
|