Import and use additional HarmCategory and HarmBlockThreshold models

The update in guardrails_models.py ensures more categories of harmful content are included in our checks and result blocking schemes. This expands the capacity of our system to flag and block dangerous content, sexually explicit content, hate speech and harassment, thereby enhancing user safety and ensuring a healthier interaction environment.

guardrails_models.py

@@ -3,7 +3,7 @@ import random
 from typing import List, Optional
 
 import openai
-from google.generativeai.types import BlockedPromptException, StopCandidateException
+from google.generativeai.types import BlockedPromptException, StopCandidateException, HarmCategory, HarmBlockThreshold
 from langchain_community.chat_models import ChatAnyscale
 from langchain_core.messages import AIMessage, HumanMessage, SystemMessage
 from langchain_google_genai import ChatGoogleGenerativeAI
@@ -383,6 +383,7 @@ def gemini_pro_nemoguardrails(
             temperature=temperature,
             max_retries=6,
             metadata={"top_p": top_p, "max_output_tokens": max_output_tokens},
+            safety_settings={ HarmCategory.HARM_CATEGORY_DANGEROUS_CONTENT: HarmBlockThreshold.BLOCK_NONE, HarmCategory.HARM_CATEGORY_SEXUALLY_EXPLICIT: HarmBlockThreshold.BLOCK_NONE, HarmCategory.HARM_CATEGORY_HATE_SPEECH: HarmBlockThreshold.BLOCK_NONE, HarmCategory.HARM_CATEGORY_HARASSMENT: HarmBlockThreshold.BLOCK_NONE, },
         ),
     )
     completion = rails.generate(messages=messages)