-
-
Notifications
You must be signed in to change notification settings - Fork 286
Commit
This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.
Merge pull request #37 from jannikstdl/prompt_injection_filter
- Loading branch information
Showing
1 changed file
with
81 additions
and
0 deletions.
There are no files selected for viewing
81 changes: 81 additions & 0 deletions
81
examples/filters/llmguard_prompt_injection_filter_pipeline.py
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Original file line number | Diff line number | Diff line change |
---|---|---|
@@ -0,0 +1,81 @@ | ||
""" | ||
title: LLM Guard Filter Pipeline | ||
author: jannikstdl | ||
date: 2024-05-30 | ||
version: 1.0 | ||
license: MIT | ||
description: A pipeline for filtering out potential prompt injections using the LLM Guard library. | ||
requirements: llm-guard | ||
""" | ||
|
||
from typing import List, Optional | ||
from schemas import OpenAIChatMessage | ||
from pydantic import BaseModel | ||
from llm_guard.input_scanners import PromptInjection | ||
from llm_guard.input_scanners.prompt_injection import MatchType | ||
import os | ||
|
||
class Pipeline: | ||
def __init__(self): | ||
# Pipeline filters are only compatible with Open WebUI | ||
# You can think of filter pipeline as a middleware that can be used to edit the form data before it is sent to the OpenAI API. | ||
self.type = "filter" | ||
|
||
# Optionally, you can set the id and name of the pipeline. | ||
# Assign a unique identifier to the pipeline. | ||
# The identifier must be unique across all pipelines. | ||
# The identifier must be an alphanumeric string that can include underscores or hyphens. It cannot contain spaces, special characters, slashes, or backslashes. | ||
self.id = "llmguard_prompt_injection_filter_pipeline" | ||
self.name = "LLMGuard Prompt Injection Filter" | ||
|
||
class Valves(BaseModel): | ||
# List target pipeline ids (models) that this filter will be connected to. | ||
# If you want to connect this filter to all pipelines, you can set pipelines to ["*"] | ||
# e.g. ["llama3:latest", "gpt-3.5-turbo"] | ||
pipelines: List[str] = [] | ||
|
||
# Assign a priority level to the filter pipeline. | ||
# The priority level determines the order in which the filter pipelines are executed. | ||
# The lower the number, the higher the priority. | ||
priority: int = 0 | ||
|
||
# Initialize | ||
self.valves = Valves( | ||
**{ | ||
"pipelines": ["*"], # Connect to all pipelines | ||
} | ||
) | ||
|
||
self.model = None | ||
|
||
pass | ||
|
||
async def on_startup(self): | ||
# This function is called when the server is started. | ||
print(f"on_startup:{__name__}") | ||
|
||
self.model = PromptInjection(threshold=0.8, match_type=MatchType.FULL) | ||
pass | ||
|
||
async def on_shutdown(self): | ||
# This function is called when the server is stopped. | ||
print(f"on_shutdown:{__name__}") | ||
pass | ||
|
||
async def on_valves_updated(self): | ||
# This function is called when the valves are updated. | ||
pass | ||
|
||
async def inlet(self, body: dict, user: Optional[dict] = None) -> dict: | ||
# This filter is applied to the form data before it is sent to the OpenAI API. | ||
print(f"inlet:{__name__}") | ||
|
||
user_message = body["messages"][-1]["content"] | ||
|
||
# Filter out prompt injection messages | ||
sanitized_prompt, is_valid, risk_score = self.model.scan(user_message) | ||
|
||
if risk_score > 0.8: | ||
raise Exception("Prompt injection detected") | ||
|
||
return body |