sij/khoj
1
0
Fork 0
mirror of https://github.com/khoj-ai/khoj.git synced 2024-12-18 18:47:11 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions 23

Track usage costs from DeepInfra OpenAI compatible API

Browse source
This commit is contained in:
Debanjum 2024-12-12 14:08:27 -08:00
parent b0abec39d5
commit 12c976dcb2
2 changed files with 9 additions and 5 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

6
src/khoj/processor/conversation/openai/utils.py
View file

@ -97,7 +97,8 @@ def completion_with_backoff(
# Calculate cost of chat
input_tokens = chunk.usage.prompt_tokens if hasattr(chunk, "usage") and chunk.usage else 0
output_tokens = chunk.usage.completion_tokens if hasattr(chunk, "usage") and chunk.usage else 0
tracer["usage"] = get_chat_usage_metrics(model_name, input_tokens, output_tokens, tracer.get("usage"))
cost = chunk.usage.model_extra.get("estimated_cost") or 0 # Estimated costs returned by DeepInfra API
tracer["usage"] = get_chat_usage_metrics(model_name, input_tokens, output_tokens, tracer.get("usage"), cost)
# Save conversation trace
tracer["chat_model"] = model_name
@ -208,7 +209,8 @@ def llm_thread(
# Calculate cost of chat
input_tokens = chunk.usage.prompt_tokens if hasattr(chunk, "usage") and chunk.usage else 0
output_tokens = chunk.usage.completion_tokens if hasattr(chunk, "usage") and chunk.usage else 0
tracer["usage"] = get_chat_usage_metrics(model_name, input_tokens, output_tokens, tracer.get("usage"))
cost = chunk.usage.model_extra.get("estimated_cost") or 0 # Estimated costs returned by DeepInfra API
tracer["usage"] = get_chat_usage_metrics(model_name, input_tokens, output_tokens, tracer.get("usage"), cost)
# Save conversation trace
tracer["chat_model"] = model_name

8
src/khoj/utils/helpers.py
View file

@ -584,13 +584,15 @@ def get_cost_of_chat_message(model_name: str, input_tokens: int = 0, output_toke
return input_cost + output_cost + prev_cost
def get_chat_usage_metrics(model_name: str, input_tokens: int = 0, output_tokens: int = 0, usage: dict = {}):
def get_chat_usage_metrics(
model_name: str, input_tokens: int = 0, output_tokens: int = 0, usage: dict = {}, cost: float = None
):
"""
Get usage metrics for chat message based on input and output tokens
Get usage metrics for chat message based on input and output tokens and cost
"""
prev_usage = usage or {"input_tokens": 0, "output_tokens": 0, "cost": 0.0}
return {
"input_tokens": prev_usage["input_tokens"] + input_tokens,
"output_tokens": prev_usage["output_tokens"] + output_tokens,
"cost": get_cost_of_chat_message(model_name, input_tokens, output_tokens, prev_cost=prev_usage["cost"]),
"cost": cost or get_cost_of_chat_message(model_name, input_tokens, output_tokens, prev_cost=prev_usage["cost"]),
}