semantic-router 0.1.0.dev2__tar.gz → 0.1.0.dev3__tar.gz

{semantic_router-0.1.0.dev2 → semantic_router-0.1.0.dev3}/LICENSE

@@ -1,6 +1,6 @@
 MIT License
 
-Copyright (c) 2023 Aurelio AI
+Copyright (c) 2024 Aurelio AI
 
 Permission is hereby granted, free of charge, to any person obtaining a copy
 of this software and associated documentation files (the "Software"), to deal

{semantic_router-0.1.0.dev2 → semantic_router-0.1.0.dev3}/PKG-INFO

@@ -1,16 +1,18 @@
 Metadata-Version: 2.1
 Name: semantic-router
-Version: 0.1.0.dev2
+Version: 0.1.0.dev3
 Summary: Super fast semantic router for AI decision making
 License: MIT
 Author: Aurelio AI
 Author-email: hello@aurelio.ai
-Requires-Python: >=3.9,<3.13
+Requires-Python: >=3.9,<3.14
 Classifier: License :: OSI Approved :: MIT License
 Classifier: Programming Language :: Python :: 3
 Classifier: Programming Language :: Python :: 3.9
 Classifier: Programming Language :: Python :: 3.10
 Classifier: Programming Language :: Python :: 3.11
+Classifier: Programming Language :: Python :: 3.12
+Classifier: Programming Language :: Python :: 3.13
 Provides-Extra: bedrock
 Provides-Extra: cohere
 Provides-Extra: docs
@@ -30,17 +32,17 @@ Requires-Dist: botocore (>=1.34.110,<2.0.0) ; extra == "bedrock"
 Requires-Dist: cohere (>=5.9.4,<6.00) ; extra == "cohere"
 Requires-Dist: colorama (>=0.4.6,<0.5.0)
 Requires-Dist: colorlog (>=6.8.0,<7.0.0)
-Requires-Dist: fastembed (>=0.3.0,<0.4.0) ; extra == "fastembed"
+Requires-Dist: fastembed (>=0.3.0,<0.4.0) ; (python_version < "3.13") and (extra == "fastembed")
 Requires-Dist: google-cloud-aiplatform (>=1.45.0,<2.0.0) ; extra == "google"
-Requires-Dist: llama-cpp-python (>=0.2.28,<0.2.86) ; extra == "local"
+Requires-Dist: llama-cpp-python (>=0.2.28,<0.2.86) ; (python_version < "3.13") and (extra == "local")
 Requires-Dist: mistralai (>=0.0.12,<0.1.0) ; extra == "mistralai"
 Requires-Dist: numpy (>=1.25.2,<2.0.0)
 Requires-Dist: openai (>=1.10.0,<2.0.0)
-Requires-Dist: pillow (>=10.2.0,<11.0.0) ; extra == "vision"
+Requires-Dist: pillow (>=10.2.0,<11.0.0) ; (python_version < "3.13") and (extra == "vision")
 Requires-Dist: pinecone (>=5.0.0) ; extra == "pinecone"
 Requires-Dist: pinecone-text (>=0.7.1,<0.10.0) ; extra == "hybrid"
 Requires-Dist: psycopg2-binary (>=2.9.9,<3.0.0)
-Requires-Dist: pydantic (>=2.5.3,<3.0.0)
+Requires-Dist: pydantic (>=2.10.2,<3.0.0)
 Requires-Dist: pyyaml (>=6.0.1,<7.0.0)
 Requires-Dist: qdrant-client (>=1.11.1,<2.0.0) ; extra == "qdrant"
 Requires-Dist: regex (>=2023.12.25)
@@ -48,11 +50,11 @@ Requires-Dist: requests-mock (>=1.12.1,<2.0.0)
 Requires-Dist: sphinx (>=7.0.0,<8.0.0) ; extra == "docs"
 Requires-Dist: sphinxawesome-theme (>=5.2.0,<6.0.0) ; extra == "docs"
 Requires-Dist: tiktoken (>=0.6.0,<1.0.0)
-Requires-Dist: tokenizers (>=0.19) ; extra == "local"
-Requires-Dist: torch (>=2.1.0,<2.6.0) ; extra == "local" or extra == "vision"
-Requires-Dist: torchvision (>=0.17.0,<0.18.0) ; extra == "vision"
+Requires-Dist: tokenizers (>=0.19) ; (python_version < "3.13") and (extra == "local")
+Requires-Dist: torch (>=2.1.0,<2.6.0) ; (python_version < "3.13") and (extra == "local" or extra == "vision")
+Requires-Dist: torchvision (>=0.17.0,<0.18.0) ; (python_version < "3.13") and (extra == "vision")
 Requires-Dist: tornado (>=6.4.2,<7.0.0)
-Requires-Dist: transformers (>=4.36.2) ; extra == "local" or extra == "vision"
+Requires-Dist: transformers (>=4.36.2) ; (python_version < "3.13") and (extra == "local" or extra == "vision")
 Description-Content-Type: text/markdown
 
 [![Semantic Router](https://i.ibb.co.com/g423grt/semantic-router-banner.png)](https://aurelio.ai)

{semantic_router-0.1.0.dev2 → semantic_router-0.1.0.dev3}/pyproject.toml

@@ -1,6 +1,6 @@
 [tool.poetry]
 name = "semantic-router"
-version = "0.1.0.dev2"
+version = "0.1.0.dev3"
 description = "Super fast semantic router for AI decision making"
 authors = ["Aurelio AI <hello@aurelio.ai>"]
 readme = "README.md"
@@ -8,8 +8,8 @@ packages = [{include = "semantic_router"}]
 license = "MIT"
 
 [tool.poetry.dependencies]
-python = ">=3.9,<3.13"
-pydantic = "^2.5.3"
+python = ">=3.9,<3.14"
+pydantic = "^2.10.2"
 openai = ">=1.10.0,<2.0.0"
 cohere = {version = ">=5.9.4,<6.00", optional = true}
 mistralai= {version = ">=0.0.12,<0.1.0", optional = true}
@@ -18,15 +18,15 @@ colorlog = "^6.8.0"
 pyyaml = "^6.0.1"
 aurelio-sdk = {version = "^0.0.16"}
 pinecone-text = {version = ">=0.7.1,<0.10.0", optional = true}
-torch = {version = ">=2.1.0,<2.6.0", optional = true}
-transformers = {version = ">=4.36.2", optional = true}
-tokenizers = {version = ">=0.19", optional = true}
-llama-cpp-python = {version = ">=0.2.28,<0.2.86", optional = true}
+torch = {version = ">=2.1.0,<2.6.0", optional = true, python = "<3.13" }
+transformers = {version = ">=4.36.2", optional = true, python = "<3.13" }
+tokenizers = {version = ">=0.19", optional = true, python = "<3.13" }
+llama-cpp-python = {version = ">=0.2.28,<0.2.86", optional = true, python = "<3.13" }
 colorama = "^0.4.6"
 pinecone = {version=">=5.0.0", optional = true}
 regex = ">=2023.12.25"
-torchvision = { version = ">=0.17.0,<0.18.0", optional = true}
-pillow = { version = ">=10.2.0,<11.0.0", optional = true}
+torchvision = { version = ">=0.17.0,<0.18.0", optional = true, python = "<3.13" }
+pillow = { version = ">=10.2.0,<11.0.0", optional = true, python = "<3.13" }
 tiktoken = ">=0.6.0,<1.0.0"
 qdrant-client = {version = "^1.11.1", optional = true}
 google-cloud-aiplatform = {version = "^1.45.0", optional = true}
@@ -34,7 +34,7 @@ requests-mock = "^1.12.1"
 boto3 = { version = "^1.34.98", optional = true }
 botocore = {version = "^1.34.110", optional = true}
 aiohttp = "^3.10.11"
-fastembed = {version = "^0.3.0", optional = true}
+fastembed = {version = "^0.3.0", optional = true, python = "<3.13" }
 psycopg2-binary = {version = "^2.9.9", optional = true}
 sphinx = {version = "^7.0.0", optional = true}
 sphinxawesome-theme = {version = "^5.2.0", optional = true}

{semantic_router-0.1.0.dev2 → semantic_router-0.1.0.dev3}/semantic_router/__init__.py

@@ -3,4 +3,4 @@ from semantic_router.route import Route
 
 __all__ = ["SemanticRouter", "HybridRouter", "Route", "RouterConfig"]
 
-__version__ = "0.1.0.dev2"
+__version__ = "0.1.0.dev3"

{semantic_router-0.1.0.dev2 → semantic_router-0.1.0.dev3}/semantic_router/encoders/aurelio.py

@@ -1,6 +1,6 @@
 import os
 from typing import Any, List, Optional
-from pydantic.v1 import Field
+from pydantic import Field
 
 from aurelio_sdk import AurelioClient, AsyncAurelioClient, EmbeddingResponse
 

{semantic_router-0.1.0.dev2 → semantic_router-0.1.0.dev3}/semantic_router/encoders/base.py

@@ -1,6 +1,6 @@
 from typing import Any, Coroutine, List, Optional
 
-from pydantic.v1 import BaseModel, Field, validator
+from pydantic import BaseModel, Field, field_validator
 import numpy as np
 
 from semantic_router.schema import SparseEmbedding
@@ -14,7 +14,7 @@ class DenseEncoder(BaseModel):
     class Config:
         arbitrary_types_allowed = True
 
-    @validator("score_threshold", pre=True, always=True)
+    @field_validator("score_threshold")
     def set_score_threshold(cls, v):
         return float(v) if v is not None else None
 

{semantic_router-0.1.0.dev2 → semantic_router-0.1.0.dev3}/semantic_router/encoders/clip.py

@@ -1,7 +1,7 @@
 from typing import Any, List, Optional
 
 import numpy as np
-from pydantic.v1 import PrivateAttr
+from pydantic import PrivateAttr
 from typing import Dict
 from semantic_router.encoders import DenseEncoder
 
@@ -9,7 +9,6 @@ from semantic_router.encoders import DenseEncoder
 class CLIPEncoder(DenseEncoder):
     name: str = "openai/clip-vit-base-patch16"
     type: str = "huggingface"
-    score_threshold: float = 0.2
     tokenizer_kwargs: Dict = {}
     processor_kwargs: Dict = {}
     model_kwargs: Dict = {}
@@ -21,6 +20,8 @@ class CLIPEncoder(DenseEncoder):
     _Image: Any = PrivateAttr()
 
     def __init__(self, **data):
+        if data.get("score_threshold") is None:
+            data["score_threshold"] = 0.2
         super().__init__(**data)
         self._tokenizer, self._processor, self._model = self._initialize_hf_model()
 

{semantic_router-0.1.0.dev2 → semantic_router-0.1.0.dev3}/semantic_router/encoders/cohere.py

@@ -1,7 +1,7 @@
 import os
 from typing import Any, List, Optional
 
-from pydantic.v1 import PrivateAttr
+from pydantic import PrivateAttr
 
 from semantic_router.encoders import DenseEncoder
 from semantic_router.utils.defaults import EncoderDefault

{semantic_router-0.1.0.dev2 → semantic_router-0.1.0.dev3}/semantic_router/encoders/fastembed.py

@@ -1,7 +1,7 @@
 from typing import Any, List, Optional
 
 import numpy as np
-from pydantic.v1 import PrivateAttr
+from pydantic import PrivateAttr
 
 from semantic_router.encoders import DenseEncoder
 

{semantic_router-0.1.0.dev2 → semantic_router-0.1.0.dev3}/semantic_router/encoders/huggingface.py

@@ -25,7 +25,7 @@ import time
 import os
 from typing import Any, List, Optional, Dict
 
-from pydantic.v1 import PrivateAttr
+from pydantic import PrivateAttr
 
 from semantic_router.encoders import DenseEncoder
 from semantic_router.utils.logger import logger
@@ -34,7 +34,6 @@ from semantic_router.utils.logger import logger
 class HuggingFaceEncoder(DenseEncoder):
     name: str = "sentence-transformers/all-MiniLM-L6-v2"
     type: str = "huggingface"
-    score_threshold: float = 0.5
     tokenizer_kwargs: Dict = {}
     model_kwargs: Dict = {}
     device: Optional[str] = None
@@ -43,6 +42,8 @@ class HuggingFaceEncoder(DenseEncoder):
     _torch: Any = PrivateAttr()
 
     def __init__(self, **data):
+        if data.get("score_threshold") is None:
+            data["score_threshold"] = 0.5
         super().__init__(**data)
         self._tokenizer, self._model = self._initialize_hf_model()
 
@@ -153,7 +154,6 @@ class HFEndpointEncoder(DenseEncoder):
     name: str = "hugging_face_custom_endpoint"
     huggingface_url: Optional[str] = None
     huggingface_api_key: Optional[str] = None
-    score_threshold: float = 0.8
 
     def __init__(
         self,
@@ -180,6 +180,8 @@ class HFEndpointEncoder(DenseEncoder):
         """
         huggingface_url = huggingface_url or os.getenv("HF_API_URL")
         huggingface_api_key = huggingface_api_key or os.getenv("HF_API_KEY")
+        if score_threshold is None:
+            score_threshold = 0.8
 
         super().__init__(name=name, score_threshold=score_threshold)  # type: ignore
 

{semantic_router-0.1.0.dev2 → semantic_router-0.1.0.dev3}/semantic_router/encoders/mistral.py

@@ -4,7 +4,7 @@ import os
 from time import sleep
 from typing import Any, List, Optional
 
-from pydantic.v1 import PrivateAttr
+from pydantic import PrivateAttr
 
 from semantic_router.encoders import DenseEncoder
 from semantic_router.utils.defaults import EncoderDefault

{semantic_router-0.1.0.dev2 → semantic_router-0.1.0.dev3}/semantic_router/encoders/openai.py

@@ -2,7 +2,7 @@ from asyncio import sleep as asleep
 import os
 from time import sleep
 from typing import Any, List, Optional, Union
-from pydantic.v1 import PrivateAttr
+from pydantic import PrivateAttr
 
 import openai
 from openai import OpenAIError
@@ -36,8 +36,8 @@ model_configs = {
 
 
 class OpenAIEncoder(DenseEncoder):
-    client: Optional[openai.Client]
-    async_client: Optional[openai.AsyncClient]
+    _client: Optional[openai.Client] = PrivateAttr(default=None)
+    _async_client: Optional[openai.AsyncClient] = PrivateAttr(default=None)
     dimensions: Union[int, NotGiven] = NotGiven()
     token_limit: int = 8192  # default value, should be replaced by config
     _token_encoder: Any = PrivateAttr()
@@ -77,10 +77,10 @@ class OpenAIEncoder(DenseEncoder):
         if max_retries is not None:
             self.max_retries = max_retries
         try:
-            self.client = openai.Client(
+            self._client = openai.Client(
                 base_url=base_url, api_key=api_key, organization=openai_org_id
             )
-            self.async_client = openai.AsyncClient(
+            self._async_client = openai.AsyncClient(
                 base_url=base_url, api_key=api_key, organization=openai_org_id
             )
         except Exception as e:
@@ -103,7 +103,7 @@ class OpenAIEncoder(DenseEncoder):
             False and a document exceeds the token limit, an error will be
             raised.
         :return: List of embeddings for each document."""
-        if self.client is None:
+        if self._client is None:
             raise ValueError("OpenAI client is not initialized.")
         embeds = None
 
@@ -114,7 +114,7 @@ class OpenAIEncoder(DenseEncoder):
         # Exponential backoff
         for j in range(self.max_retries + 1):
             try:
-                embeds = self.client.embeddings.create(
+                embeds = self._client.embeddings.create(
                     input=docs,
                     model=self.name,
                     dimensions=self.dimensions,
@@ -160,7 +160,7 @@ class OpenAIEncoder(DenseEncoder):
         return text
 
     async def acall(self, docs: List[str], truncate: bool = True) -> List[List[float]]:
-        if self.async_client is None:
+        if self._async_client is None:
             raise ValueError("OpenAI async client is not initialized.")
         embeds = None
 
@@ -171,7 +171,7 @@ class OpenAIEncoder(DenseEncoder):
         # Exponential backoff
         for j in range(self.max_retries + 1):
             try:
-                embeds = await self.async_client.embeddings.create(
+                embeds = await self._async_client.embeddings.create(
                     input=docs,
                     model=self.name,
                     dimensions=self.dimensions,

{semantic_router-0.1.0.dev2 → semantic_router-0.1.0.dev3}/semantic_router/encoders/vit.py

@@ -1,6 +1,6 @@
-from typing import Any, List, Optional, Dict
+from typing import Any, Dict, List, Optional
 
-from pydantic.v1 import PrivateAttr
+from pydantic import PrivateAttr
 
 from semantic_router.encoders import DenseEncoder
 
@@ -8,7 +8,6 @@ from semantic_router.encoders import DenseEncoder
 class VitEncoder(DenseEncoder):
     name: str = "google/vit-base-patch16-224"
     type: str = "huggingface"
-    score_threshold: float = 0.5
     processor_kwargs: Dict = {}
     model_kwargs: Dict = {}
     device: Optional[str] = None
@@ -19,6 +18,8 @@ class VitEncoder(DenseEncoder):
     _Image: Any = PrivateAttr()
 
     def __init__(self, **data):
+        if data.get("score_threshold") is None:
+            data["score_threshold"] = 0.5
         super().__init__(**data)
         self._processor, self._model = self._initialize_hf_model()
 

{semantic_router-0.1.0.dev2 → semantic_router-0.1.0.dev3}/semantic_router/index/base.py

@@ -2,7 +2,7 @@ from typing import Any, List, Optional, Tuple, Union, Dict
 import json
 
 import numpy as np
-from pydantic.v1 import BaseModel
+from pydantic import BaseModel
 
 from semantic_router.schema import ConfigParameter, SparseEmbedding, Utterance
 from semantic_router.route import Route

{semantic_router-0.1.0.dev2 → semantic_router-0.1.0.dev3}/semantic_router/index/pinecone.py

@@ -8,7 +8,7 @@ import json
 from typing import Any, Dict, List, Optional, Union, Tuple
 
 import numpy as np
-from pydantic.v1 import BaseModel, Field
+from pydantic import BaseModel, Field
 
 from semantic_router.index.base import BaseIndex
 from semantic_router.schema import ConfigParameter, SparseEmbedding

{semantic_router-0.1.0.dev2 → semantic_router-0.1.0.dev3}/semantic_router/index/postgres.py

@@ -103,7 +103,6 @@ class PostgresIndex(BaseIndex):
     connection_string: Optional[str] = None
     index_prefix: str = "semantic_router_"
     index_name: str = "index"
-    dimensions: int = 1536
     metric: Metric = Metric.COSINE
     namespace: Optional[str] = ""
     conn: Optional["psycopg2.extensions.connection"] = None
@@ -115,9 +114,9 @@ class PostgresIndex(BaseIndex):
         connection_string: Optional[str] = None,
         index_prefix: str = "semantic_router_",
         index_name: str = "index",
-        dimensions: int = 1536,
         metric: Metric = Metric.COSINE,
         namespace: Optional[str] = "",
+        dimensions: int | None = None,
     ):
         """
         Initializes the Postgres index with the specified parameters.
@@ -135,6 +134,8 @@ class PostgresIndex(BaseIndex):
         :param namespace: An optional namespace for the index.
         :type namespace: Optional[str]
         """
+        if dimensions is None:
+            dimensions = 1536
         super().__init__()
         # try and import psycopg2
         try:

{semantic_router-0.1.0.dev2 → semantic_router-0.1.0.dev3}/semantic_router/index/qdrant.py

@@ -1,7 +1,7 @@
 from typing import Any, Dict, List, Optional, Tuple, Union
 
 import numpy as np
-from pydantic.v1 import Field
+from pydantic import Field
 
 from semantic_router.index.base import BaseIndex
 from semantic_router.schema import ConfigParameter, Metric, SparseEmbedding, Utterance
@@ -40,7 +40,7 @@ class QdrantIndex(BaseIndex):
         default=6334,
         description="Port of the gRPC interface.",
     )
-    prefer_grpc: bool = Field(
+    prefer_grpc: Optional[bool] = Field(
         default=None,
         description="Whether to use gPRC interface whenever possible in methods",
     )

{semantic_router-0.1.0.dev2 → semantic_router-0.1.0.dev3}/semantic_router/llms/base.py

@@ -1,7 +1,7 @@
 import json
 from typing import Any, List, Optional, Dict
 
-from pydantic.v1 import BaseModel
+from pydantic import BaseModel
 
 from semantic_router.schema import Message
 from semantic_router.utils.logger import logger
@@ -9,6 +9,8 @@ from semantic_router.utils.logger import logger
 
 class BaseLLM(BaseModel):
     name: str
+    temperature: Optional[float] = 0.0
+    max_tokens: Optional[int] = None
 
     class Config:
         arbitrary_types_allowed = True

{semantic_router-0.1.0.dev2 → semantic_router-0.1.0.dev3}/semantic_router/llms/cohere.py

@@ -1,7 +1,7 @@
 import os
 from typing import Any, List, Optional
 
-from pydantic.v1 import PrivateAttr
+from pydantic import PrivateAttr
 
 from semantic_router.llms import BaseLLM
 from semantic_router.schema import Message

{semantic_router-0.1.0.dev2 → semantic_router-0.1.0.dev3}/semantic_router/llms/llamacpp.py

@@ -2,7 +2,7 @@ from contextlib import contextmanager
 from pathlib import Path
 from typing import Any, Optional, List, Dict
 
-from pydantic.v1 import PrivateAttr
+from pydantic import PrivateAttr
 
 from semantic_router.llms.base import BaseLLM
 from semantic_router.schema import Message
@@ -11,8 +11,6 @@ from semantic_router.utils.logger import logger
 
 class LlamaCppLLM(BaseLLM):
     llm: Any
-    temperature: float
-    max_tokens: Optional[int] = 200
     grammar: Optional[Any] = None
     _llama_cpp: Any = PrivateAttr()
 

{semantic_router-0.1.0.dev2 → semantic_router-0.1.0.dev3}/semantic_router/llms/mistral.py

@@ -1,7 +1,7 @@
 import os
 from typing import Any, List, Optional
 
-from pydantic.v1 import PrivateAttr
+from pydantic import PrivateAttr
 
 from semantic_router.llms import BaseLLM
 from semantic_router.schema import Message
@@ -11,8 +11,6 @@ from semantic_router.utils.logger import logger
 
 class MistralAILLM(BaseLLM):
     _client: Any = PrivateAttr()
-    temperature: Optional[float]
-    max_tokens: Optional[int]
     _mistralai: Any = PrivateAttr()
 
     def __init__(

{semantic_router-0.1.0.dev2 → semantic_router-0.1.0.dev3}/semantic_router/llms/ollama.py

@@ -8,22 +8,17 @@ from semantic_router.utils.logger import logger
 
 
 class OllamaLLM(BaseLLM):
-    temperature: Optional[float]
-    llm_name: Optional[str]
-    max_tokens: Optional[int]
-    stream: Optional[bool]
+    stream: bool = False
 
     def __init__(
         self,
-        name: str = "ollama",
+        name: str = "openhermes",
         temperature: float = 0.2,
-        llm_name: str = "openhermes",
         max_tokens: Optional[int] = 200,
         stream: bool = False,
     ):
         super().__init__(name=name)
         self.temperature = temperature
-        self.llm_name = llm_name
         self.max_tokens = max_tokens
         self.stream = stream
 
@@ -31,19 +26,19 @@ class OllamaLLM(BaseLLM):
         self,
         messages: List[Message],
         temperature: Optional[float] = None,
-        llm_name: Optional[str] = None,
+        name: Optional[str] = None,
         max_tokens: Optional[int] = None,
         stream: Optional[bool] = None,
     ) -> str:
         # Use instance defaults if not overridden
         temperature = temperature if temperature is not None else self.temperature
-        llm_name = llm_name if llm_name is not None else self.llm_name
+        name = name if name is not None else self.name
         max_tokens = max_tokens if max_tokens is not None else self.max_tokens
         stream = stream if stream is not None else self.stream
 
         try:
             payload = {
-                "model": llm_name,
+                "model": name,
                 "messages": [m.to_openai() for m in messages],
                 "options": {"temperature": temperature, "num_predict": max_tokens},
                 "format": "json",

{semantic_router-0.1.0.dev2 → semantic_router-0.1.0.dev3}/semantic_router/llms/openai.py

@@ -1,5 +1,6 @@
 import os
 from typing import List, Optional, Any, Callable, Dict, Union
+from pydantic import PrivateAttr
 
 import openai
 from openai._types import NotGiven, NOT_GIVEN
@@ -21,10 +22,8 @@ from openai.types.chat.chat_completion_message_tool_call import (
 
 
 class OpenAILLM(BaseLLM):
-    client: Optional[openai.OpenAI]
-    async_client: Optional[openai.AsyncOpenAI]
-    temperature: Optional[float]
-    max_tokens: Optional[int]
+    _client: Optional[openai.OpenAI] = PrivateAttr(default=None)
+    _async_client: Optional[openai.AsyncOpenAI] = PrivateAttr(default=None)
 
     def __init__(
         self,
@@ -40,8 +39,8 @@ class OpenAILLM(BaseLLM):
         if api_key is None:
             raise ValueError("OpenAI API key cannot be 'None'.")
         try:
-            self.async_client = openai.AsyncOpenAI(api_key=api_key)
-            self.client = openai.OpenAI(api_key=api_key)
+            self._async_client = openai.AsyncOpenAI(api_key=api_key)
+            self._client = openai.OpenAI(api_key=api_key)
         except Exception as e:
             raise ValueError(
                 f"OpenAI API client failed to initialize. Error: {e}"
@@ -88,14 +87,14 @@ class OpenAILLM(BaseLLM):
         messages: List[Message],
         function_schemas: Optional[List[Dict[str, Any]]] = None,
     ) -> str:
-        if self.client is None:
+        if self._client is None:
             raise ValueError("OpenAI client is not initialized.")
         try:
             tools: Union[List[Dict[str, Any]], NotGiven] = (
                 function_schemas if function_schemas else NOT_GIVEN
             )
 
-            completion = self.client.chat.completions.create(
+            completion = self._client.chat.completions.create(
                 model=self.name,
                 messages=[m.to_openai() for m in messages],
                 temperature=self.temperature,
@@ -132,14 +131,14 @@ class OpenAILLM(BaseLLM):
         messages: List[Message],
         function_schemas: Optional[List[Dict[str, Any]]] = None,
     ) -> str:
-        if self.async_client is None:
+        if self._async_client is None:
             raise ValueError("OpenAI async_client is not initialized.")
         try:
             tools: Union[List[Dict[str, Any]], NotGiven] = (
                 function_schemas if function_schemas is not None else NOT_GIVEN
             )
 
-            completion = await self.async_client.chat.completions.create(
+            completion = await self._async_client.chat.completions.create(
                 model=self.name,
                 messages=[m.to_openai() for m in messages],
                 temperature=self.temperature,

{semantic_router-0.1.0.dev2 → semantic_router-0.1.0.dev3}/semantic_router/llms/openrouter.py

@@ -1,6 +1,8 @@
 import os
 from typing import List, Optional
 
+from pydantic import PrivateAttr
+
 import openai
 
 from semantic_router.llms import BaseLLM
@@ -9,10 +11,8 @@ from semantic_router.utils.logger import logger
 
 
 class OpenRouterLLM(BaseLLM):
-    client: Optional[openai.OpenAI]
-    base_url: Optional[str]
-    temperature: Optional[float]
-    max_tokens: Optional[int]
+    _client: Optional[openai.OpenAI] = PrivateAttr(default=None)
+    _base_url: str = PrivateAttr(default="https://openrouter.ai/api/v1")
 
     def __init__(
         self,
@@ -27,12 +27,12 @@ class OpenRouterLLM(BaseLLM):
                 "OPENROUTER_CHAT_MODEL_NAME", "mistralai/mistral-7b-instruct"
             )
         super().__init__(name=name)
-        self.base_url = base_url
+        self._base_url = base_url
         api_key = openrouter_api_key or os.getenv("OPENROUTER_API_KEY")
         if api_key is None:
             raise ValueError("OpenRouter API key cannot be 'None'.")
         try:
-            self.client = openai.OpenAI(api_key=api_key, base_url=self.base_url)
+            self._client = openai.OpenAI(api_key=api_key, base_url=self._base_url)
         except Exception as e:
             raise ValueError(
                 f"OpenRouter API client failed to initialize. Error: {e}"
@@ -41,10 +41,10 @@ class OpenRouterLLM(BaseLLM):
         self.max_tokens = max_tokens
 
     def __call__(self, messages: List[Message]) -> str:
-        if self.client is None:
+        if self._client is None:
             raise ValueError("OpenRouter client is not initialized.")
         try:
-            completion = self.client.chat.completions.create(
+            completion = self._client.chat.completions.create(
                 model=self.name,
                 messages=[m.to_openai() for m in messages],
                 temperature=self.temperature,

{semantic_router-0.1.0.dev2 → semantic_router-0.1.0.dev3}/semantic_router/llms/zure.py

@@ -1,5 +1,6 @@
 import os
 from typing import List, Optional
+from pydantic import PrivateAttr
 
 import openai
 
@@ -10,9 +11,7 @@ from semantic_router.utils.logger import logger
 
 
 class AzureOpenAILLM(BaseLLM):
-    client: Optional[openai.AzureOpenAI]
-    temperature: Optional[float]
-    max_tokens: Optional[int]
+    _client: Optional[openai.AzureOpenAI] = PrivateAttr(default=None)
 
     def __init__(
         self,
@@ -33,7 +32,7 @@ class AzureOpenAILLM(BaseLLM):
         if azure_endpoint is None:
             raise ValueError("Azure endpoint API key cannot be 'None'.")
         try:
-            self.client = openai.AzureOpenAI(
+            self._client = openai.AzureOpenAI(
                 api_key=api_key, azure_endpoint=azure_endpoint, api_version=api_version
             )
         except Exception as e:
@@ -42,10 +41,10 @@ class AzureOpenAILLM(BaseLLM):
         self.max_tokens = max_tokens
 
     def __call__(self, messages: List[Message]) -> str:
-        if self.client is None:
+        if self._client is None:
             raise ValueError("AzureOpenAI client is not initialized.")
         try:
-            completion = self.client.chat.completions.create(
+            completion = self._client.chat.completions.create(
                 model=self.name,
                 messages=[m.to_openai() for m in messages],
                 temperature=self.temperature,

{semantic_router-0.1.0.dev2 → semantic_router-0.1.0.dev3}/semantic_router/route.py

@@ -2,18 +2,13 @@ import json
 import re
 from typing import Any, Callable, Dict, List, Optional, Union
 
-from pydantic.v1 import BaseModel
+from pydantic import BaseModel
 
 from semantic_router.llms import BaseLLM
 from semantic_router.schema import Message, RouteChoice
 from semantic_router.utils import function_call
 from semantic_router.utils.logger import logger
 
-try:
-    from PIL.Image import Image
-except ImportError:
-    pass
-
 
 def is_valid(route_config: str) -> bool:
     try:
@@ -45,7 +40,7 @@ def is_valid(route_config: str) -> bool:
 
 class Route(BaseModel):
     name: str
-    utterances: Union[List[str], List[Union[Any, "Image"]]]
+    utterances: Union[List[str], List[Any]]
     description: Optional[str] = None
     function_schemas: Optional[List[Dict[str, Any]]] = None
     llm: Optional[BaseLLM] = None

{semantic_router-0.1.0.dev2 → semantic_router-0.1.0.dev3}/semantic_router/routers/base.py

@@ -4,7 +4,7 @@ import os
 import random
 import hashlib
 from typing import Any, Callable, Dict, List, Optional, Tuple, Union
-from pydantic.v1 import BaseModel, Field
+from pydantic import BaseModel, Field
 
 import numpy as np
 import yaml  # type: ignore

{semantic_router-0.1.0.dev2 → semantic_router-0.1.0.dev3}/semantic_router/routers/hybrid.py

@@ -1,6 +1,6 @@
 from typing import Dict, List, Optional
 import asyncio
-from pydantic.v1 import Field
+from pydantic import Field
 
 import numpy as np
 

{semantic_router-0.1.0.dev2 → semantic_router-0.1.0.dev3}/semantic_router/schema.py

@@ -1,9 +1,10 @@
 from datetime import datetime
 from difflib import Differ
 from enum import Enum
+import json
 import numpy as np
 from typing import List, Optional, Union, Any, Dict, Tuple
-from pydantic.v1 import BaseModel, Field
+from pydantic import BaseModel, Field
 from semantic_router.utils.logger import logger
 from aurelio_sdk.schema import BM25Embedding
 
@@ -126,7 +127,18 @@ class Utterance(BaseModel):
 
     def to_str(self, include_metadata: bool = False):
         if include_metadata:
-            return f"{self.route}: {self.utterance} | {self.function_schemas} | {self.metadata}"
+            # we sort the dicts to ensure consistent order as we need this to compare
+            # stringified function schemas accurately
+            if self.function_schemas is not None:
+                function_schemas_sorted: List[str] | None = [
+                    json.dumps(schema, sort_keys=True)
+                    for schema in self.function_schemas
+                ]
+            else:
+                function_schemas_sorted = None
+            # we must do the same for metadata
+            metadata_sorted = json.dumps(self.metadata, sort_keys=True)
+            return f"{self.route}: {self.utterance} | {function_schemas_sorted} | {metadata_sorted}"
         return f"{self.route}: {self.utterance}"
 
     def to_diff_str(self, include_metadata: bool = False):

{semantic_router-0.1.0.dev2 → semantic_router-0.1.0.dev3}/semantic_router/utils/function_call.py

@@ -1,7 +1,7 @@
 import inspect
 from typing import Any, Callable, Dict, List, Optional, Union
 
-from pydantic.v1 import BaseModel
+from pydantic import BaseModel
 
 from semantic_router.llms import BaseLLM
 from semantic_router.schema import Message, RouteChoice