Update hf_client.py
Browse files- hf_client.py +42 -30
hf_client.py
CHANGED
|
@@ -1,34 +1,56 @@
|
|
| 1 |
-
|
| 2 |
|
| 3 |
-
from huggingface_hub import InferenceClient, HfApi
|
| 4 |
-
from tavily import TavilyClient
|
| 5 |
import os
|
|
|
|
|
|
|
|
|
|
| 6 |
|
| 7 |
-
#
|
| 8 |
-
|
| 9 |
-
#
|
| 10 |
-
|
| 11 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 12 |
HF_TOKEN = os.getenv("HF_TOKEN")
|
| 13 |
if not HF_TOKEN:
|
| 14 |
-
raise RuntimeError(
|
| 15 |
-
|
| 16 |
-
|
| 17 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 18 |
|
| 19 |
def get_inference_client(model_id: str, provider: str = "auto") -> InferenceClient:
|
| 20 |
"""
|
| 21 |
-
|
| 22 |
-
|
| 23 |
-
|
| 24 |
-
|
| 25 |
-
|
|
|
|
| 26 |
"""
|
| 27 |
-
#
|
| 28 |
if model_id == "moonshotai/Kimi-K2-Instruct":
|
| 29 |
provider = "groq"
|
| 30 |
-
|
| 31 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 32 |
bill_to = provider if provider in _VALID_BILL_TO else "groq"
|
| 33 |
|
| 34 |
return InferenceClient(
|
|
@@ -36,13 +58,3 @@ def get_inference_client(model_id: str, provider: str = "auto") -> InferenceClie
|
|
| 36 |
api_key=HF_TOKEN,
|
| 37 |
bill_to=bill_to
|
| 38 |
)
|
| 39 |
-
|
| 40 |
-
# Tavily Search Client
|
| 41 |
-
TAVILY_API_KEY = os.getenv('TAVILY_API_KEY')
|
| 42 |
-
tavily_client = None
|
| 43 |
-
if TAVILY_API_KEY:
|
| 44 |
-
try:
|
| 45 |
-
tavily_client = TavilyClient(api_key=TAVILY_API_KEY)
|
| 46 |
-
except Exception as e:
|
| 47 |
-
print(f"Failed to initialize Tavily client: {e}")
|
| 48 |
-
tavily_client = None
|
|
|
|
| 1 |
+
# hf_client.py
|
| 2 |
|
|
|
|
|
|
|
| 3 |
import os
|
| 4 |
+
from typing import Optional
|
| 5 |
+
from huggingface_hub import InferenceClient
|
| 6 |
+
from tavily import TavilyClient
|
| 7 |
|
| 8 |
+
# Supported billing targets (your secrets)
|
| 9 |
+
_VALID_BILL_TO = {
|
| 10 |
+
"huggingface", # HF_TOKEN
|
| 11 |
+
"fairworksai", # if you have a special billing target
|
| 12 |
+
"groq", # GROQ_API_KEY
|
| 13 |
+
"openai", # OPENAI_API_KEY
|
| 14 |
+
"gemini", # GEMINI_API_KEY
|
| 15 |
+
"fireworks" # FIREWORKS_API_KEY
|
| 16 |
+
}
|
| 17 |
+
|
| 18 |
+
# Load your HF token
|
| 19 |
HF_TOKEN = os.getenv("HF_TOKEN")
|
| 20 |
if not HF_TOKEN:
|
| 21 |
+
raise RuntimeError("HF_TOKEN is not set. Please configure your Hugging Face token.")
|
| 22 |
+
|
| 23 |
+
# (Optional) Tavily client for web search
|
| 24 |
+
TAVILY_API_KEY = os.getenv("TAVILY_API_KEY")
|
| 25 |
+
tavily_client: Optional[TavilyClient] = None
|
| 26 |
+
if TAVILY_API_KEY:
|
| 27 |
+
try:
|
| 28 |
+
tavily_client = TavilyClient(api_key=TAVILY_API_KEY)
|
| 29 |
+
except Exception as e:
|
| 30 |
+
print(f"β Failed to init Tavily: {e}")
|
| 31 |
|
| 32 |
def get_inference_client(model_id: str, provider: str = "auto") -> InferenceClient:
|
| 33 |
"""
|
| 34 |
+
Selects the appropriate inference provider & billing target:
|
| 35 |
+
β’ moonshotai/Kimi-K2-Instruct β groq
|
| 36 |
+
β’ openai/... or GPT family β openai
|
| 37 |
+
β’ gemini/... or google/... β gemini
|
| 38 |
+
β’ fireworks/... β fireworks
|
| 39 |
+
β’ otherwise β huggingface (billing to groq)
|
| 40 |
"""
|
| 41 |
+
# override by model prefix
|
| 42 |
if model_id == "moonshotai/Kimi-K2-Instruct":
|
| 43 |
provider = "groq"
|
| 44 |
+
elif model_id.startswith("openai/") or model_id in {"gpt-4", "gpt-3.5-turbo"}:
|
| 45 |
+
provider = "openai"
|
| 46 |
+
elif model_id.startswith("gemini/") or model_id.startswith("google/"):
|
| 47 |
+
provider = "gemini"
|
| 48 |
+
elif model_id.startswith("fireworks/"):
|
| 49 |
+
provider = "fireworks"
|
| 50 |
+
else:
|
| 51 |
+
provider = "huggingface"
|
| 52 |
+
|
| 53 |
+
# ensure billing target is valid, else fallback to groq
|
| 54 |
bill_to = provider if provider in _VALID_BILL_TO else "groq"
|
| 55 |
|
| 56 |
return InferenceClient(
|
|
|
|
| 58 |
api_key=HF_TOKEN,
|
| 59 |
bill_to=bill_to
|
| 60 |
)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|