Initial commit

.gitignore

@@ -0,0 +1,268 @@
+# Warning!!! Generated by Copilot, haven't checked it fully yet.
+
+# Visual Studio Code
+.vscode/
+.settings/
+
+# Gradio
+flagged/
+
+# Python
+__pycache__/
+*.pyc
+*.pyo
+*.pyd
+*.pyz
+*.pyw
+*.pyi
+*.egg-info/
+dist/
+build/
+*.egg
+*.whl
+*.swp
+*.swo
+*.swn
+*.swm
+*.swl
+*.swk
+*.swj
+*.swi
+*.swg
+*.swf
+*.swe
+*.swd
+*.swc
+*.swb
+*.swa
+*.sw9
+*.sw8
+*.sw7
+*.sw6
+*.sw5
+*.sw4
+*.sw3
+*.sw2
+*.sw1
+*.sw0
+*.s9
+*.s8
+*.s7
+*.s6
+*.s5
+*.s4
+*.s3
+*.s2
+*.s1
+*.s0
+*.s
+*.r9
+*.r8
+*.r7
+*.r6
+*.r5
+*.r4
+*.r3
+*.r2
+*.r1
+*.r0
+*.r
+*.q9
+*.q8
+*.q7
+*.q6
+*.q5
+*.q4
+*.q3
+*.q2
+*.q1
+*.q0
+*.q
+*.p9
+*.p8
+*.p7
+*.p6
+*.p5
+*.p4
+*.p3
+*.p2
+*.p1
+*.p0
+*.p
+*.o9
+*.o8
+*.o7
+*.o6
+*.o5
+*.o4
+*.o3
+*.o2
+*.o1
+*.o0
+*.o
+*.n9
+*.n8
+*.n7
+*.n6
+*.n5
+*.n4
+*.n3
+*.n2
+*.n1
+*.n0
+*.n
+*.m9
+*.m8
+*.m7
+*.m6
+*.m5
+*.m4
+*.m3
+*.m2
+*.m1
+*.m0
+*.m
+*.l9
+*.l8
+*.l7
+*.l6
+*.l5
+*.l4
+*.l3
+*.l2
+*.l1
+*.l0
+*.l
+*.k9
+*.k8
+*.k7
+*.k6
+*.k5
+*.k4
+*.k3
+*.k2
+*.k1
+*.k0
+*.k
+*.j9
+*.j8
+*.j7
+*.j6
+*.j5
+*.j4
+*.j3
+*.j2
+*.j1
+*.j0
+*.j
+*.i9
+*.i8
+*.i7
+*.i6
+*.i5
+*.i4
+*.i3
+*.i2
+*.i1
+*.i0
+*.i
+*.h9
+*.h8
+*.h7
+*.h6
+*.h5
+*.h4
+*.h3
+*.h2
+*.h1
+*.h0
+*.h
+*.g9
+*.g8
+*.g7
+*.g6
+*.g5
+*.g4
+*.g3
+*.g2
+*.g1
+*.g0
+*.g
+*.f9
+*.f8
+*.f7
+*.f6
+*.f5
+*.f4
+*.f3
+*.f2
+*.f1
+*.f0
+*.f
+*.e9
+*.e8
+*.e7
+*.e6
+*.e5
+*.e4
+*.e3
+*.e2
+*.e1
+*.e0
+*.e
+*.d9
+*.d8
+*.d7
+*.d6
+*.d5
+*.d4
+*.d3
+*.d2
+*.d1
+*.d0
+*.d
+*.c9
+*.c8
+*.c7
+*.c6
+*.c5
+*.c4
+*.c3
+*.c2
+*.c1
+*.c0
+*.c
+*.b9
+*.b8
+*.b7
+*.b6
+*.b5
+*.b4
+*.b3
+*.b2
+*.b1
+*.b0
+*.b
+*.a9
+*.a8
+*.a7
+*.a6
+*.a5
+*.a4
+*.a3
+*.a2
+*.a1
+*.a0
+*.a
+*.9
+*.8
+*.7
+*.6
+*.5
+*.4
+*.3
+*.2
+*.1
+*.0
+*.~
+*~

LLM/LLMGuard/GuardProcessor.py

@@ -0,0 +1,33 @@
+from llm_guard.input_scanners import Anonymize
+from llm_guard.vault import Vault
+from llm_guard.input_scanners.anonymize_helpers import BERT_LARGE_NER_CONF
+
+
+def process_output_with_llmguard(output):
+    print(f"Processed with LLMGuard: {output}")
+
+    # Need to not disclose student id or other sensitive information using LLMGuard
+    
+    # Use the LLMGuard to scan the output
+    vault = Vault()
+    scanner = Anonymize(vault, preamble="Insert before prompt", recognizer_conf=BERT_LARGE_NER_CONF, language="en")
+
+    # Anonymize the output
+    anonymized_output, is_valid, risk_score = scanner.scan(output)
+
+    return anonymized_output
+
+
+def process_input_with_llmguard(input):
+    print(f"Processed with LLMGuard: {input}")
+
+    # Need to not disclose student id or other sensitive information using LLMGuard
+    
+    # Use the LLMGuard to scan the output
+    vault = Vault()
+    scanner = Anonymize(vault, preamble="Insert before prompt", recognizer_conf=BERT_LARGE_NER_CONF, language="en")
+
+    # Anonymize the output
+    anonymized_output, is_valid, risk_score = scanner.scan(input)
+
+    return anonymized_output

LLM/MockLLM.py

@@ -0,0 +1,3 @@
+def generate_mock_output(prompt):
+    # Simple mock output for testing
+    return f"Mock output for: {prompt}"

README.md

@@ -0,0 +1,21 @@
+LLMGuard is a project that utilizes Gradio and LLMGuard to protect users from leaking personal information in a university environment.
+
+## Installation
+
+To install LLMGuard, follow these steps:
+
+1. Clone the repository: `git clone https://github.com/your-username/LLMGuard.git`
+2. Install the required dependencies: `pip install -r requirements.txt`
+
+## Usage
+
+To run this project, follow these steps:
+
+1. Run `main.py`: `python main.py`
+
+## TODO
+
+- [] Implement scanners to detect personal information leakage. :mag:
+- [] Consider using regex scanner for more accurate detection. :dart:
+- [] Implement real LLM (Leakage Limiting Mechanism) instead of the mock version. :gear:
+- [] Generate tests to automate the testing process. :test_tube:

main.py

@@ -0,0 +1,24 @@
+import gradio as gr
+from LLM.MockLLM import generate_mock_output
+from LLM.LLMGuard.GuardProcessor import process_output_with_llmguard, process_input_with_llmguard
+
+def run_llm_guard(prompt):
+    # Generate mock LLM output
+    mock_output = generate_mock_output(process_input_with_llmguard(prompt))
+    
+    # Process the output with LLMGuard (placeholder)
+    processed_output = process_output_with_llmguard(mock_output)
+    
+    return processed_output
+
+# Gradio Interface
+iface = gr.Interface(
+    fn=run_llm_guard,
+    inputs="text",
+    outputs="text",
+    title="LLMGuard Tester",
+    description="Enter a prompt to generate mock LLM output and process it with LLMGuard."
+)
+
+if __name__ == "__main__":
+    iface.launch()