Spaces:
Running
Running
talexm
commited on
Commit
·
c3c1187
1
Parent(s):
e9a8c67
update
Browse files
rag_sec/document_search_system.py
CHANGED
@@ -1,42 +1,79 @@
|
|
1 |
-
|
2 |
-
from
|
3 |
-
|
4 |
-
from
|
|
|
|
|
|
|
|
|
5 |
|
6 |
class DocumentSearchSystem:
|
7 |
def __init__(self):
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
8 |
self.detector = BadQueryDetector()
|
9 |
self.transformer = QueryTransformer()
|
10 |
self.retriever = DocumentRetriever()
|
11 |
self.response_generator = SemanticResponseGenerator()
|
12 |
|
13 |
def process_query(self, query):
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
14 |
if self.detector.is_bad_query(query):
|
15 |
return {"status": "rejected", "message": "Query blocked due to detected malicious intent."}
|
16 |
|
|
|
17 |
transformed_query = self.transformer.transform_query(query)
|
18 |
-
|
19 |
|
|
|
|
|
20 |
if not retrieved_docs:
|
21 |
return {"status": "no_results", "message": "No relevant documents found for your query."}
|
22 |
|
|
|
23 |
response = self.response_generator.generate_response(retrieved_docs)
|
24 |
return {"status": "success", "response": response}
|
25 |
|
26 |
|
27 |
def test_system():
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
28 |
system = DocumentSearchSystem()
|
29 |
-
system.retriever.load_documents(
|
30 |
|
31 |
-
#
|
32 |
normal_query = "Tell me about great acting performances."
|
33 |
print("\nNormal Query Result:")
|
34 |
print(system.process_query(normal_query))
|
35 |
|
36 |
-
#
|
37 |
malicious_query = "DROP TABLE users; SELECT * FROM sensitive_data;"
|
38 |
print("\nMalicious Query Result:")
|
39 |
print(system.process_query(malicious_query))
|
40 |
|
|
|
41 |
if __name__ == "__main__":
|
42 |
test_system()
|
|
|
1 |
+
import os
|
2 |
+
from pathlib import Path
|
3 |
+
|
4 |
+
from .bad_query_detector import BadQueryDetector
|
5 |
+
from .query_transformer import QueryTransformer
|
6 |
+
from .document_retriver import DocumentRetriever
|
7 |
+
from .senamtic_response_generator import SemanticResponseGenerator
|
8 |
+
|
9 |
|
10 |
class DocumentSearchSystem:
|
11 |
def __init__(self):
|
12 |
+
"""
|
13 |
+
Initializes the DocumentSearchSystem with:
|
14 |
+
- BadQueryDetector for identifying malicious or inappropriate queries.
|
15 |
+
- QueryTransformer for improving or rephrasing queries.
|
16 |
+
- DocumentRetriever for semantic document retrieval.
|
17 |
+
- SemanticResponseGenerator for generating context-aware responses.
|
18 |
+
"""
|
19 |
self.detector = BadQueryDetector()
|
20 |
self.transformer = QueryTransformer()
|
21 |
self.retriever = DocumentRetriever()
|
22 |
self.response_generator = SemanticResponseGenerator()
|
23 |
|
24 |
def process_query(self, query):
|
25 |
+
"""
|
26 |
+
Processes a user query through the following steps:
|
27 |
+
1. Detect if the query is malicious.
|
28 |
+
2. Transform the query if needed.
|
29 |
+
3. Retrieve relevant documents based on the query.
|
30 |
+
4. Generate a response using the retrieved documents.
|
31 |
+
|
32 |
+
:param query: The user query as a string.
|
33 |
+
:return: A dictionary with the status and response or error message.
|
34 |
+
"""
|
35 |
if self.detector.is_bad_query(query):
|
36 |
return {"status": "rejected", "message": "Query blocked due to detected malicious intent."}
|
37 |
|
38 |
+
# Transform the query
|
39 |
transformed_query = self.transformer.transform_query(query)
|
40 |
+
print(f"Transformed Query: {transformed_query}")
|
41 |
|
42 |
+
# Retrieve relevant documents
|
43 |
+
retrieved_docs = self.retriever.retrieve(transformed_query)
|
44 |
if not retrieved_docs:
|
45 |
return {"status": "no_results", "message": "No relevant documents found for your query."}
|
46 |
|
47 |
+
# Generate a response based on the retrieved documents
|
48 |
response = self.response_generator.generate_response(retrieved_docs)
|
49 |
return {"status": "success", "response": response}
|
50 |
|
51 |
|
52 |
def test_system():
|
53 |
+
"""
|
54 |
+
Test the DocumentSearchSystem with normal and malicious queries.
|
55 |
+
- Load documents from a dataset directory.
|
56 |
+
- Perform a normal query and display results.
|
57 |
+
- Perform a malicious query to ensure proper blocking.
|
58 |
+
"""
|
59 |
+
# Define the path to the dataset directory
|
60 |
+
home_dir = Path(os.getenv("HOME", "/"))
|
61 |
+
data_dir = home_dir / "data-sets/aclImdb/train"
|
62 |
+
|
63 |
+
# Initialize the system
|
64 |
system = DocumentSearchSystem()
|
65 |
+
system.retriever.load_documents(data_dir)
|
66 |
|
67 |
+
# Perform a normal query
|
68 |
normal_query = "Tell me about great acting performances."
|
69 |
print("\nNormal Query Result:")
|
70 |
print(system.process_query(normal_query))
|
71 |
|
72 |
+
# Perform a malicious query
|
73 |
malicious_query = "DROP TABLE users; SELECT * FROM sensitive_data;"
|
74 |
print("\nMalicious Query Result:")
|
75 |
print(system.process_query(malicious_query))
|
76 |
|
77 |
+
|
78 |
if __name__ == "__main__":
|
79 |
test_system()
|
rag_sec/senamtic_response_generator.py
CHANGED
@@ -1,10 +1,21 @@
|
|
1 |
-
from transformers import
|
|
|
2 |
|
3 |
class SemanticResponseGenerator:
|
4 |
-
def __init__(self):
|
5 |
-
self.
|
|
|
|
|
|
|
6 |
|
7 |
def generate_response(self, retrieved_docs):
|
8 |
-
combined_docs = " ".join(retrieved_docs[:2])
|
9 |
-
|
10 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
from transformers import AutoModelForSeq2SeqLM, AutoTokenizer
|
2 |
+
|
3 |
|
4 |
class SemanticResponseGenerator:
|
5 |
+
def __init__(self, model_name="google/flan-t5-small", max_input_length=512, max_new_tokens=50):
|
6 |
+
self.tokenizer = AutoTokenizer.from_pretrained(model_name)
|
7 |
+
self.model = AutoModelForSeq2SeqLM.from_pretrained(model_name)
|
8 |
+
self.max_input_length = max_input_length
|
9 |
+
self.max_new_tokens = max_new_tokens
|
10 |
|
11 |
def generate_response(self, retrieved_docs):
|
12 |
+
combined_docs = " ".join(retrieved_docs[:2])
|
13 |
+
truncated_docs = combined_docs[:self.max_input_length - 50]
|
14 |
+
input_text = f"Based on the following information: {truncated_docs}"
|
15 |
+
inputs = self.tokenizer(input_text, return_tensors="pt", truncation=True, max_length=self.max_input_length)
|
16 |
+
outputs = self.model.generate(
|
17 |
+
**inputs,
|
18 |
+
max_new_tokens=self.max_new_tokens,
|
19 |
+
pad_token_id=self.tokenizer.eos_token_id
|
20 |
+
)
|
21 |
+
return self.tokenizer.decode(outputs[0], skip_special_tokens=True)
|