kyber-runtime 0.1.0__py3-none-any.whl → 0.1.2__py3-none-any.whl

kyber_runtime/__init__.py

@@ -0,0 +1,4 @@
+from .llm import call_llm, LLMCallContext
+
+__version__ = "0.1.2"
+__all__ = ["llm_call", "LLMCallContext"]

kyber_runtime/config.py

@@ -0,0 +1,4 @@
+LLM_pricing = {
+      "gpt-4o":{"input":5.00,"output":15.00},
+      "gpt-4o-mini":{"input":0.15,"output":0.6}
+}

kyber_runtime/llm.py

@@ -0,0 +1,110 @@
+import time 
+import uuid 
+import json
+import logging
+from typing import Any, Dict, Optional
+from openai import OpenAI
+from pydantic import BaseModel,Field
+
+from .config import LLM_pricing
+
+logging.basicConfig(level=logging.INFO)
+logger = logging.getLogger("kyber.telemetry")
+
+class LLMCallContect(BaseModel):
+      execution_id: str
+      agent_id: str
+      tenent_id: str
+      circuit_id: Optional[str] = None
+      
+def _calculate_cost(model: str, prompt_tokens: int, completion_tokens: int) -> float:
+      if model not in LLM_pricing:
+            return 0.0
+      princing = LLM_pricing[model]
+      input_cost = (prompt_tokens / 1000) * princing["input"]
+      output_cost = (completion_tokens / 1000) * princing["output"]
+      return round(input_cost + output_cost, 6)
+
+def llm_call(
+      *,
+      model: str,
+      messages: list[Dict[str, str]],
+      purpose: str,
+      context: LLMCallContxet,
+      temperature: float = 0.7,
+      max_retries: int = 3,
+      client: Optional[OpenAI] = None
+) -> Dict[str, Any]:
+      
+      if not purpose:
+            raise ValueError("Purpose must be provided for LLM call telemetry.")
+      
+      call_id = str(uuid.uuid4())
+      start_time = time.time()
+      retry_count = 0
+      client = client or OpenAI()
+      
+      while retry_count <= max_retries:
+            try:
+                  response = client.chat.completions.create(
+                        model=model,
+                        messages=messages,
+                        temperature=temperature,
+                        timeout=30.0
+                  )
+                  
+                  
+                  usage = response.usage
+                  latency_ms = int((time.time() - start_time) * 1000)
+                  cost_usd = _calculate_cost(model,usage.prompt_tokens, usage.completion_tokens)
+                  
+                  telemetry = {
+                        "call_id": call_id,
+                        "execution_id": context.execution_id,
+                        "agent_id": context.agent_id,
+                        "tenent_id": context.tenent_id,
+                        "circuit_id": context.circuit_id,
+                        "model": model,
+                        "purpose": purpose,
+                        "latency_ms": latency_ms,
+                        "prompt_tokens": usage.prompt_tokens,
+                        "completion_tokens": usage.completion_tokens,
+                        "total_tokens": usage.total_tokens,
+                        "llm_cost_usd": cost_usd,
+                        "success": True,
+                        "error_message": None,
+                        "temperature": temperature,
+                        "retry_count": retry_count,
+                        "timestamp": time.time()
+                  }
+                  
+                  logger.info("LLM Call Telemetry:", extra={"telemetry": telemetry})
+                  
+                  return json.loads(response.model_dump_json())
+            except Exception as e:
+                  retry_count += 1
+                  if retry_count > max_retries:
+                        latency_ms = int((time.time() - start_time) * 1000)
+                        telemetry = {
+                              "call_id": call_id,
+                              "execution_id": context.execution_id,
+                              "agent_id": context.agent_id,
+                              "tenent_id": context.tenent_id,
+                              "circuit_id": context.circuit_id,
+                              "model": model,
+                              "purpose": purpose,
+                              "latency_ms": latency_ms,
+                              "prompt_tokens": 0,
+                              "completion_tokens": 0,
+                              "total_tokens": 0,
+                              "llm_cost_usd": 0.0,
+                              "success": False,
+                              "error_message": str(e),
+                              "temperature": temperature,
+                              "retry_count": retry_count,
+                              "timestamp": time.time()
+                        }
+                        
+                        logger.error("LLM Call Telemetry:", extra={"telemetry": telemetry})
+                        raise e
+                  time.sleep(0.5 *(2 ** (retry_count - 1)))  # Exponential backoff

{kyber_runtime-0.1.0.dist-info → kyber_runtime-0.1.2.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
-Name: kyber-runtime
-Version: 0.1.0
+Name: kyber_runtime
+Version: 0.1.2
 Summary: Kyber Agent Runtime SDK for instrumented LLM calls
 Author: Your Team
 Requires-Python: >=3.9

kyber_runtime-0.1.2.dist-info/RECORD

@@ -0,0 +1,7 @@
+kyber_runtime/__init__.py,sha256=pRzznOBdKJ9EDxRSBj72taVABtMoHyyddzDiY73cgF4,105
+kyber_runtime/config.py,sha256=L_9YtY_q6i7NtyN8JVdxkqSgZ3ahCx18Q_xs8NeSCEM,111
+kyber_runtime/llm.py,sha256=DKRYF-HyEf_zAJHRRTtEtlQl6ay64WTVcn6vrAYZwGc,4395
+kyber_runtime-0.1.2.dist-info/METADATA,sha256=fIPFA0_W9nRcGhqGbVADx0kPHgtaUw86vR5W99aCy2k,373
+kyber_runtime-0.1.2.dist-info/WHEEL,sha256=wUyA8OaulRlbfwMtmQsvNngGrxQHAvkKcvRmdizlJi0,92
+kyber_runtime-0.1.2.dist-info/top_level.txt,sha256=wLlG1BFH_QzUNPebY6Xww6TE4m9Er26DP2a36iwhpkY,14
+kyber_runtime-0.1.2.dist-info/RECORD,,

kyber_runtime-0.1.2.dist-info/top_level.txt

@@ -0,0 +1 @@
+kyber_runtime

kyber_runtime-0.1.0.dist-info/RECORD

@@ -1,4 +0,0 @@
-kyber_runtime-0.1.0.dist-info/METADATA,sha256=RkFtls6Cefl7oqJgdtFEyrMqcKpiEIZzLptDOhLBstk,373
-kyber_runtime-0.1.0.dist-info/WHEEL,sha256=wUyA8OaulRlbfwMtmQsvNngGrxQHAvkKcvRmdizlJi0,92
-kyber_runtime-0.1.0.dist-info/top_level.txt,sha256=AbpHGcgLb-kRsJGnwFEktk7uzpZOCcBY74-YBdrKVGs,1
-kyber_runtime-0.1.0.dist-info/RECORD,,

kyber_runtime-0.1.0.dist-info/top_level.txt

@@ -1 +0,0 @@
-