feat: Add OpenTelemetry dependencies

[christinaexyou]

Add OpenTelemetry dependencies to enable tracing via OpenShift

Example actions.py script

from typing import Optional
from nemoguardrails.actions import action

# OpenTelemetry imports and initialization
from opentelemetry import trace
from opentelemetry.sdk.trace import TracerProvider
from opentelemetry.sdk.trace.export import BatchSpanProcessor
from opentelemetry.exporter.otlp.proto.grpc.trace_exporter import OTLPSpanExporter
from opentelemetry.sdk.resources import Resource
import os

resource = Resource.create({"service.name": "nemo-guardrails"})
tracer_provider = TracerProvider(resource=resource)
trace.set_tracer_provider(tracer_provider)
tracer = trace.get_tracer(__name__)

otlp_exporter = OTLPSpanExporter(endpoint=os.getenv("OTEL_EXPORTER_OTLP_ENDPOINT"), insecure=True)
tracer_provider.add_span_processor(BatchSpanProcessor(otlp_exporter))

@action(is_system_action=True)
async def check_message_length(context: Optional[dict] = None) -> str:
    with tracer.start_as_current_span("check_message_length"):
        """Check if user message is within acceptable length limits."""
        user_message = context.get("user_message", "")
        word_count = len(user_message.split())
        MAX_WORDS = 100
        if word_count > MAX_WORDS:
            return "blocked_too_long"
        elif word_count > MAX_WORDS * 0.8:
            return "warning_long"
        return "allowed"

@action(is_system_action=True)
async def check_forbidden_words(context: Optional[dict] = None) -> str:
    with tracer.start_as_current_span("check_forbidden_words"):
        """Check for forbidden words or topics."""
        user_message = context.get("user_message", "").lower()
        forbidden_topics = {
            "security": ["password", "hack", "exploit", "vulnerability"],
            "inappropriate": ["violence", "illegal", "harmful"],
            "competitors": ["chatgpt", "openai", "claude", "anthropic"],
        }
        for category, words in forbidden_topics.items():
            for word in words:
                if word in user_message:
                    return f"blocked_{category}_{word}"
        return "allowed"