cc / api_usage.py
superdup95's picture
Update api_usage.py
0360007
raw
history blame
3.62 kB
import requests
import os
import openai
import anthropic
BASE_URL = 'https://api.openai.com/v1'
GPT_TYPES = ["gpt-3.5-turbo", "gpt-4", "gpt-4-32k"]
RATE_LIMIT_PER_MODEL = {
"gpt-3.5-turbo": 2000, # new pay turbo will have 2000 RPM for the first 48 hours then become 3500
"gpt-4": 200,
"gpt-4-32k": 1000
}
def get_headers(key):
headers = {'Authorization': f'Bearer {key}'}
return headers
def get_subscription(key, available_models):
headers = get_headers(key)
rpm = "0"
tpm = "0"
tpm_left = "0"
org = ""
quota = ""
key_highest_model = ""
has_gpt4_32k = False
has_gpt4 = False
if check_gpt4_32k_availability(available_models):
key_highest_model = GPT_TYPES[2]
has_gpt4_32k = True
has_gpt4 = True
elif check_gpt4_availability(available_models):
key_highest_model = GPT_TYPES[1]
has_gpt4 = True
else:
key_highest_model = GPT_TYPES[0]
req_body = {"model": key_highest_model, "messages": [{'role':'user', 'content': ''}], "max_tokens": 1}
r = requests.post(f"{BASE_URL}/chat/completions", headers=headers, json=req_body)
result = r.json()
if "id" in result:
rpm = r.headers.get("x-ratelimit-limit-requests", "0")
tpm = r.headers.get("x-ratelimit-limit-tokens", "0")
tpm_left = r.headers.get("x-ratelimit-remaining-tokens", "0")
org = r.headers.get('openai-organization', "")
quota = check_key_type(key_highest_model, int(rpm))
else:
e = result.get("error", {}).get("code", "")
quota = f"Error: {e}"
org = get_org_name(key)
return {"has_gpt4_32k": has_gpt4_32k,
"has_gpt4": has_gpt4,
"organization": org,
"rpm": f"{rpm} ({key_highest_model})",
"tpm": f"{tpm} ({tpm_left} left)",
"quota": quota}
def get_org_name(key):
headers=get_headers(key)
r = requests.post(f"{BASE_URL}/images/generations", headers=headers)
return r.headers['openai-organization']
def check_key_type(model, rpm):
if rpm < RATE_LIMIT_PER_MODEL[model]:
return "yes | trial"
else:
return "yes | pay"
def check_gpt4_availability(available_models):
if 'gpt-4' in available_models:
return True
else:
return False
def check_gpt4_32k_availability(available_models):
if 'gpt-4-32k' in available_models:
return True
else:
return False
def check_key_availability():
try:
avai_models = openai.Model.list()
return [model["root"] for model in avai_models["data"] if model["root"] in GPT_TYPES]
except:
return False
def check_key_ant_availability(ant):
try:
ant.completions.create(
prompt=f"{anthropic.HUMAN_PROMPT}Hi{anthropic.AI_PROMPT}",
max_tokens_to_sample=1,
model="claude-instant-v1",
)
return True, "Working"
except anthropic.APIConnectionError as e:
print(e.__cause__) # an underlying Exception, likely raised within httpx.
return False, "Error: The server could not be reached"
except anthropic.RateLimitError as e:
return True, "Error: 429, rate limited; we should back off a bit."
except anthropic.APIStatusError as e:
err_msg = e.body.get('error', {}).get('message', '')
return False, f"Error: {e.status_code}, {err_msg}"
if __name__ == "__main__":
key = os.getenv("OPENAI_API_KEY")
key_ant = os.getenv("ANTHROPIC_API_KEY")
results = get_subscription(key)