refactor: clean up deprecated warnings and improve code formatting

lion_service/__init__.py

@@ -2,7 +2,6 @@
 from .rate_limiter import RateLimiter, RateLimitError
 from .service import Service, register_service
 from .service_util import invoke_retry
-from .version import __version__
 
 __all__ = [
     "Service",

lion_service/complete_request_info.py

@@ -6,4 +6,6 @@ class CompleteRequestInfo(BaseModel):
 
 
 class CompleteRequestTokenInfo(CompleteRequestInfo):
-    token_usage: int = Field(description="Number of tokens used in the request")
+    token_usage: int = Field(
+        description="Number of tokens used in the request"
+    )

lion_service/imodel.py

@@ -1,7 +1,7 @@
 import warnings
 
-from lion_service.service import Service
-from lion_service.service_match_util import (
+from .service import Service
+from .service_match_util import (
     match_parameters,
     match_service,
     match_task_method,
@@ -25,74 +25,6 @@ def __init__(
         elif api_key_schema is not None:
             api_key = api_key_schema
 
-        # Deprecation
-        if "config" in kwargs:
-            warnings.warn(
-                "'config' is deprecated in 'iModel'. Please pass configurations directly as keyword arguments. "
-                "If 'config' is a parameter name for further request settings, it will still be passed along.",
-                DeprecationWarning,
-                stacklevel=2,
-            )
-        if "provider_schema" in kwargs:
-            warnings.warn(
-                "'provider_schema' is deprecated in 'iModel'. Please refer to the corresponding service's "
-                "Pydantic models for the schema structure. "
-                "If 'provider_schema' is a parameter name for further request settings, it will still be passed along.",
-                DeprecationWarning,
-                stacklevel=2,
-            )
-        if "endpoint" in kwargs:
-            warnings.warn(
-                "'endpoint' is deprecated in 'iModel'. Please specify 'task' instead. "
-                "If 'endpoint' is a parameter name for further request settings, it will still be passed along.",
-                DeprecationWarning,
-                stacklevel=2,
-            )
-            task = kwargs["endpoint"]
-        if "token_encoding_name" in kwargs:
-            warnings.warn(
-                "'token_encoding_name' is deprecated in 'iModel' as it is now automatically detected. "
-                "To explicitly set the encoding method, please refer to TokenCalculator.",
-                DeprecationWarning,
-                stacklevel=2,
-            )
-        if "interval" in kwargs:
-            warnings.warn(
-                "'interval' is deprecated in 'iModel' as the unit is now set per minute. "
-                "If 'interval' is a parameter name for further request settings, it will still be passed along.",
-                DeprecationWarning,
-                stacklevel=2,
-            )
-        if "service" in kwargs:
-            warnings.warn(
-                "'service' parameter is deprecated in 'iModel'. Please use 'provider' instead. For more "
-                "details about service configurations, refer to 'lion_service' or the corresponding service's package. "
-                "If 'service' is a parameter name for further request settings, it will still be passed along.",
-                DeprecationWarning,
-                stacklevel=2,
-            )
-        if "allowed_parameters" in kwargs:
-            warnings.warn(
-                "'allowed_parameters' is deprecated in 'iModel'. If 'allowed_parameters' is a parameter "
-                "name for further request settings, it will still be passed along.",
-                DeprecationWarning,
-                stacklevel=2,
-            )
-        if "device" in kwargs:
-            warnings.warn(
-                "'device' is deprecated in 'iModel'. "
-                "If 'device' is a parameter name for further request settings, it will still be passed along.",
-                DeprecationWarning,
-                stacklevel=2,
-            )
-        if "costs" in kwargs:
-            warnings.warn(
-                "'cost' is deprecated in 'iModel'. "
-                "If 'costs' is a parameter name for further request settings, it will still be passed along.",
-                DeprecationWarning,
-                stacklevel=2,
-            )
-
         if isinstance(provider, str):
             self.service = match_service(provider, api_key=api_key, **kwargs)
         elif isinstance(provider, Service):

lion_service/rate_limiter.py

@@ -3,7 +3,10 @@
 
 from pydantic import BaseModel, Field
 
-from .complete_request_info import CompleteRequestInfo, CompleteRequestTokenInfo
+from .complete_request_info import (
+    CompleteRequestInfo,
+    CompleteRequestTokenInfo,
+)
 
 
 class RateLimiter(BaseModel):
@@ -48,7 +51,8 @@ def release_tokens(self):
         self.last_check_timestamp = datetime.now(UTC).timestamp()
         while self.unreleased_requests:
             if (
-                datetime.now(UTC).timestamp() - self.unreleased_requests[0].timestamp
+                datetime.now(UTC).timestamp()
+                - self.unreleased_requests[0].timestamp
                 > 60
             ):
                 release_info = self.unreleased_requests.popleft()
@@ -62,7 +66,9 @@ def release_tokens(self):
             else:
                 break
 
-    def update_rate_limit(self, request_datetime_header, total_token_usage: int = None):
+    def update_rate_limit(
+        self, request_datetime_header, total_token_usage: int = None
+    ):
         # rate limiter tokens only tracks if there are token usage info
         # otherwise, tracks requests num
         date_format = "%a, %d %b %Y %H:%M:%S GMT"  # the format of the date string according to RFC 1123
@@ -76,14 +82,19 @@ def update_rate_limit(self, request_datetime_header, total_token_usage: int = No
                 timestamp=request_timestamp, token_usage=total_token_usage
             )
         else:
-            complete_request_info = CompleteRequestInfo(timestamp=request_timestamp)
+            complete_request_info = CompleteRequestInfo(
+                timestamp=request_timestamp
+            )
         self.append_complete_request_token_info(complete_request_info)
 
     def check_availability(
         self, request_token_len: int = 0, estimated_output_len: int = 0
     ):
         if self.remaining_tokens is not None:
-            if request_token_len + estimated_output_len > self.remaining_tokens:
+            if (
+                request_token_len + estimated_output_len
+                > self.remaining_tokens
+            ):
                 return False
         if self.remaining_requests is not None:
             if self.remaining_requests <= 0:

lion_service/service_util.py

@@ -4,7 +4,9 @@
 from .rate_limiter import RateLimitError
 
 
-def invoke_retry(max_retries: int = 3, base_delay: int = 1, max_delay: int = 60):
+def invoke_retry(
+    max_retries: int = 3, base_delay: int = 1, max_delay: int = 60
+):
     def decorator(func):
         @wraps(func)
         async def wrapper(request_model, *args, **kwargs):
@@ -24,7 +26,10 @@ async def wrapper(request_model, *args, **kwargs):
 
                     # RateLimitError for Model only
                     if isinstance(e, RateLimitError):
-                        if e.requested_tokens > request_model.rate_limiter.limit_tokens:
+                        if (
+                            e.requested_tokens
+                            > request_model.rate_limiter.limit_tokens
+                        ):
                             raise ValueError(
                                 "Requested tokens exceed the model's token limit. "
                                 "Please modify the input, adjust the expected output tokens, or increase the token limit. "
@@ -42,11 +47,14 @@ async def wrapper(request_model, *args, **kwargs):
                     elif error_code := getattr(
                         e, "status", None
                     ):  # http request errors
-                        if error_code == 429 and "exceeded your current quota" in str(
-                            e
+                        if (
+                            error_code == 429
+                            and "exceeded your current quota" in str(e)
                         ):  # RateLimitError (account quota reached)
                             raise e
-                        if error_code == 429 or error_code >= 500:  # ServerError
+                        if (
+                            error_code == 429 or error_code >= 500
+                        ):  # ServerError
                             if retry_after := getattr(e, "headers", {}).get(
                                 "Retry-After"
                             ):

lion_service/token_calculator.py

@@ -43,7 +43,9 @@ def tokenize(
     ):
         enc = tiktoken.get_encoding(self.encoding_name)
         enc_text = self.encode(text)
-        tokens = [enc.decode_single_token_bytes(token_id) for token_id in enc_text]
+        tokens = [
+            enc.decode_single_token_bytes(token_id) for token_id in enc_text
+        ]
         if decode_byte_str:
             tokens = [token.decode(decoder) for token in tokens]
         return tokens

lion_service/version.py

@@ -1 +1 @@
-__version__ = "1.1.0"
+__version__ = "1.1.1"