edsl 0.1.39.dev3__py3-none-any.whl → 0.1.39.dev5__py3-none-any.whl

edsl/data_transfer_models.py

@@ -1,73 +1,74 @@
-from typing import NamedTuple, Dict, List, Optional, Any
-from dataclasses import dataclass, fields
-import reprlib
-
-
-class ModelInputs(NamedTuple):
-    "This is what was send by the agent to the model"
-    user_prompt: str
-    system_prompt: str
-    encoded_image: Optional[str] = None
-
-
-class EDSLOutput(NamedTuple):
-    "This is the edsl dictionary that is returned by the model"
-    answer: Any
-    generated_tokens: str
-    comment: Optional[str] = None
-
-
-class ModelResponse(NamedTuple):
-    "This is the metadata that is returned by the model and includes info about the cache"
-    response: dict
-    cache_used: bool
-    cache_key: str
-    cached_response: Optional[Dict[str, Any]] = None
-    cost: Optional[float] = None
-
-
-class AgentResponseDict(NamedTuple):
-    edsl_dict: EDSLOutput
-    model_inputs: ModelInputs
-    model_outputs: ModelResponse
-
-
-class EDSLResultObjectInput(NamedTuple):
-    generated_tokens: str
-    question_name: str
-    prompts: dict
-    cached_response: str
-    raw_model_response: str
-    cache_used: bool
-    cache_key: str
-    answer: Any
-    comment: str
-    validated: bool = False
-    exception_occurred: Exception = None
-    cost: Optional[float] = None
-
-
-@dataclass
-class ImageInfo:
-    file_path: str
-    file_name: str
-    image_format: str
-    file_size: int
-    encoded_image: str
-
-    def __repr__(self):
-        reprlib_instance = reprlib.Repr()
-        reprlib_instance.maxstring = 30  # Limit the string length for the encoded image
-
-        # Get all fields except encoded_image
-        field_reprs = [
-            f"{f.name}={getattr(self, f.name)!r}"
-            for f in fields(self)
-            if f.name != "encoded_image"
-        ]
-
-        # Add the reprlib-restricted encoded_image field
-        field_reprs.append(f"encoded_image={reprlib_instance.repr(self.encoded_image)}")
-
-        # Join everything to create the repr
-        return f"{self.__class__.__name__}({', '.join(field_reprs)})"
+from typing import NamedTuple, Dict, List, Optional, Any
+from dataclasses import dataclass, fields
+
+
+class ModelInputs(NamedTuple):
+    "This is what was send by the agent to the model"
+    user_prompt: str
+    system_prompt: str
+    encoded_image: Optional[str] = None
+
+
+class EDSLOutput(NamedTuple):
+    "This is the edsl dictionary that is returned by the model"
+    answer: Any
+    generated_tokens: str
+    comment: Optional[str] = None
+
+
+class ModelResponse(NamedTuple):
+    "This is the metadata that is returned by the model and includes info about the cache"
+    response: dict
+    cache_used: bool
+    cache_key: str
+    cached_response: Optional[Dict[str, Any]] = None
+    cost: Optional[float] = None
+
+
+class AgentResponseDict(NamedTuple):
+    edsl_dict: EDSLOutput
+    model_inputs: ModelInputs
+    model_outputs: ModelResponse
+
+
+class EDSLResultObjectInput(NamedTuple):
+    generated_tokens: str
+    question_name: str
+    prompts: dict
+    cached_response: str
+    raw_model_response: str
+    cache_used: bool
+    cache_key: str
+    answer: Any
+    comment: str
+    validated: bool = False
+    exception_occurred: Exception = None
+    cost: Optional[float] = None
+
+
+@dataclass
+class ImageInfo:
+    file_path: str
+    file_name: str
+    image_format: str
+    file_size: int
+    encoded_image: str
+
+    def __repr__(self):
+        import reprlib
+
+        reprlib_instance = reprlib.Repr()
+        reprlib_instance.maxstring = 30  # Limit the string length for the encoded image
+
+        # Get all fields except encoded_image
+        field_reprs = [
+            f"{f.name}={getattr(self, f.name)!r}"
+            for f in fields(self)
+            if f.name != "encoded_image"
+        ]
+
+        # Add the reprlib-restricted encoded_image field
+        field_reprs.append(f"encoded_image={reprlib_instance.repr(self.encoded_image)}")
+
+        # Join everything to create the repr
+        return f"{self.__class__.__name__}({', '.join(field_reprs)})"

edsl/enums.py

@@ -1,175 +1,202 @@
-"""Enums for the different types of questions, language models, and inference services."""
-
-from enum import Enum
-
-
-class EnumWithChecks(Enum):
-    """Base class for all enums with checks."""
-
-    @classmethod
-    def is_value_valid(cls, value):
-        """Check if the value is valid."""
-        return any(value == item.value for item in cls)
-
-
-class QuestionType(EnumWithChecks):
-    """Enum for the question types."""
-
-    MULTIPLE_CHOICE = "multiple_choice"
-    YES_NO = "yes_no"
-    FREE_TEXT = "free_text"
-    RANK = "rank"
-    BUDGET = "budget"
-    CHECKBOX = "checkbox"
-    EXTRACT = "extract"
-    FUNCTIONAL = "functional"
-    LIST = "list"
-    NUMERICAL = "numerical"
-    TOP_K = "top_k"
-    LIKERT_FIVE = "likert_five"
-    LINEAR_SCALE = "linear_scale"
-
-
-# https://huggingface.co/meta-llama/Llama-2-70b-chat-hf
-
-
-# class LanguageModelType(EnumWithChecks):
-#     """Enum for the language model types."""
-
-#     GPT_4 = "gpt-4-1106-preview"
-#     GPT_3_5_Turbo = "gpt-3.5-turbo"
-#     LLAMA_2_70B_CHAT_HF = "llama-2-70b-chat-hf"
-#     LLAMA_2_13B_CHAT_HF = "llama-2-13b-chat-hf"
-#     GEMINI_PRO = "gemini_pro"
-#     MIXTRAL_8x7B_INSTRUCT = "mixtral-8x7B-instruct-v0.1"
-#     TEST = "test"
-#     ANTHROPIC_3_OPUS = "claude-3-opus-20240229"
-#     ANTHROPIC_3_SONNET = "claude-3-sonnet-20240229"
-#     ANTHROPIC_3_HAIKU = "claude-3-haiku-20240307"
-#     DBRX_INSTRUCT = "dbrx-instruct"
-
-
-class InferenceServiceType(EnumWithChecks):
-    """Enum for the inference service types."""
-
-    BEDROCK = "bedrock"
-    DEEP_INFRA = "deep_infra"
-    REPLICATE = "replicate"
-    OPENAI = "openai"
-    GOOGLE = "google"
-    TEST = "test"
-    ANTHROPIC = "anthropic"
-    GROQ = "groq"
-    AZURE = "azure"
-    OLLAMA = "ollama"
-    MISTRAL = "mistral"
-    TOGETHER = "together"
-    PERPLEXITY = "perplexity"
-
-
-service_to_api_keyname = {
-    InferenceServiceType.BEDROCK.value: "TBD",
-    InferenceServiceType.DEEP_INFRA.value: "DEEP_INFRA_API_KEY",
-    InferenceServiceType.REPLICATE.value: "TBD",
-    InferenceServiceType.OPENAI.value: "OPENAI_API_KEY",
-    InferenceServiceType.GOOGLE.value: "GOOGLE_API_KEY",
-    InferenceServiceType.TEST.value: "TBD",
-    InferenceServiceType.ANTHROPIC.value: "ANTHROPIC_API_KEY",
-    InferenceServiceType.GROQ.value: "GROQ_API_KEY",
-    InferenceServiceType.BEDROCK.value: ["AWS_ACCESS_KEY_ID", "AWS_SECRET_ACCESS_KEY"],
-    InferenceServiceType.MISTRAL.value: "MISTRAL_API_KEY",
-    InferenceServiceType.TOGETHER.value: "TOGETHER_API_KEY",
-    InferenceServiceType.PERPLEXITY.value: "PERPLEXITY_API_KEY",
-}
-
-
-class TokenPricing:
-    def __init__(
-        self,
-        *,
-        model_name,
-        prompt_token_price_per_k: float,
-        completion_token_price_per_k: float,
-    ):
-        self.model_name = model_name
-        self.prompt_token_price = prompt_token_price_per_k / 1_000.0
-        self.completion_token_price = completion_token_price_per_k / 1_000.0
-
-    def __eq__(self, other):
-        if not isinstance(other, TokenPricing):
-            return False
-        return (
-            self.model_name == other.model_name
-            and self.prompt_token_price == other.prompt_token_price
-            and self.completion_token_price == other.completion_token_price
-        )
-
-
-pricing = {
-    "dbrx-instruct": TokenPricing(
-        model_name="dbrx-instruct",
-        prompt_token_price_per_k=0.0,
-        completion_token_price_per_k=0.0,
-    ),
-    "claude-3-opus-20240229": TokenPricing(
-        model_name="claude-3-opus-20240229",
-        prompt_token_price_per_k=0.0,
-        completion_token_price_per_k=0.0,
-    ),
-    "claude-3-haiku-20240307": TokenPricing(
-        model_name="claude-3-haiku-20240307",
-        prompt_token_price_per_k=0.0,
-        completion_token_price_per_k=0.0,
-    ),
-    "claude-3-sonnet-20240229": TokenPricing(
-        model_name="claude-3-sonnet-20240229",
-        prompt_token_price_per_k=0.0,
-        completion_token_price_per_k=0.0,
-    ),
-    "gpt-3.5-turbo": TokenPricing(
-        model_name="gpt-3.5-turbo",
-        prompt_token_price_per_k=0.0005,
-        completion_token_price_per_k=0.0015,
-    ),
-    "gpt-4-1106-preview": TokenPricing(
-        model_name="gpt-4",
-        prompt_token_price_per_k=0.01,
-        completion_token_price_per_k=0.03,
-    ),
-    "test": TokenPricing(
-        model_name="test",
-        prompt_token_price_per_k=0.0,
-        completion_token_price_per_k=0.0,
-    ),
-    "gemini_pro": TokenPricing(
-        model_name="gemini_pro",
-        prompt_token_price_per_k=0.0,
-        completion_token_price_per_k=0.0,
-    ),
-    "llama-2-13b-chat-hf": TokenPricing(
-        model_name="llama-2-13b-chat-hf",
-        prompt_token_price_per_k=0.0,
-        completion_token_price_per_k=0.0,
-    ),
-    "llama-2-70b-chat-hf": TokenPricing(
-        model_name="llama-2-70b-chat-hf",
-        prompt_token_price_per_k=0.0,
-        completion_token_price_per_k=0.0,
-    ),
-    "mixtral-8x7B-instruct-v0.1": TokenPricing(
-        model_name="mixtral-8x7B-instruct-v0.1",
-        prompt_token_price_per_k=0.0,
-        completion_token_price_per_k=0.0,
-    ),
-}
-
-
-def get_token_pricing(model_name):
-    if model_name in pricing:
-        return pricing[model_name]
-    else:
-        return TokenPricing(
-            model_name=model_name,
-            prompt_token_price_per_k=0.0,
-            completion_token_price_per_k=0.0,
-        )
+"""Enums for the different types of questions, language models, and inference services."""
+
+from enum import Enum
+from typing import Literal
+
+
+class EnumWithChecks(Enum):
+    """Base class for all enums with checks."""
+
+    @classmethod
+    def is_value_valid(cls, value):
+        """Check if the value is valid."""
+        return any(value == item.value for item in cls)
+
+
+class QuestionType(EnumWithChecks):
+    """Enum for the question types."""
+
+    MULTIPLE_CHOICE = "multiple_choice"
+    YES_NO = "yes_no"
+    FREE_TEXT = "free_text"
+    RANK = "rank"
+    BUDGET = "budget"
+    CHECKBOX = "checkbox"
+    EXTRACT = "extract"
+    FUNCTIONAL = "functional"
+    LIST = "list"
+    NUMERICAL = "numerical"
+    TOP_K = "top_k"
+    LIKERT_FIVE = "likert_five"
+    LINEAR_SCALE = "linear_scale"
+
+
+# https://huggingface.co/meta-llama/Llama-2-70b-chat-hf
+
+
+# class LanguageModelType(EnumWithChecks):
+#     """Enum for the language model types."""
+
+#     GPT_4 = "gpt-4-1106-preview"
+#     GPT_3_5_Turbo = "gpt-3.5-turbo"
+#     LLAMA_2_70B_CHAT_HF = "llama-2-70b-chat-hf"
+#     LLAMA_2_13B_CHAT_HF = "llama-2-13b-chat-hf"
+#     GEMINI_PRO = "gemini_pro"
+#     MIXTRAL_8x7B_INSTRUCT = "mixtral-8x7B-instruct-v0.1"
+#     TEST = "test"
+#     ANTHROPIC_3_OPUS = "claude-3-opus-20240229"
+#     ANTHROPIC_3_SONNET = "claude-3-sonnet-20240229"
+#     ANTHROPIC_3_HAIKU = "claude-3-haiku-20240307"
+#     DBRX_INSTRUCT = "dbrx-instruct"
+
+
+class InferenceServiceType(EnumWithChecks):
+    """Enum for the inference service types."""
+
+    BEDROCK = "bedrock"
+    DEEP_INFRA = "deep_infra"
+    REPLICATE = "replicate"
+    OPENAI = "openai"
+    GOOGLE = "google"
+    TEST = "test"
+    ANTHROPIC = "anthropic"
+    GROQ = "groq"
+    AZURE = "azure"
+    OLLAMA = "ollama"
+    MISTRAL = "mistral"
+    TOGETHER = "together"
+    PERPLEXITY = "perplexity"
+
+
+# unavoidable violation of the DRY principle but it is necessary
+# checked w/ a unit test to make sure consistent with services in enums.py
+InferenceServiceLiteral = Literal[
+    "bedrock",
+    "deep_infra",
+    "replicate",
+    "openai",
+    "google",
+    "test",
+    "anthropic",
+    "groq",
+    "azure",
+    "ollama",
+    "mistral",
+    "together",
+    "perplexity",
+]
+
+available_models_urls = {
+    "anthropic": "https://docs.anthropic.com/en/docs/about-claude/models",
+    "openai": "https://platform.openai.com/docs/models/gp",
+    "groq": "https://console.groq.com/docs/models",
+    "google": "https://cloud.google.com/vertex-ai/generative-ai/docs/learn/models",
+}
+
+
+service_to_api_keyname = {
+    InferenceServiceType.BEDROCK.value: "TBD",
+    InferenceServiceType.DEEP_INFRA.value: "DEEP_INFRA_API_KEY",
+    InferenceServiceType.REPLICATE.value: "TBD",
+    InferenceServiceType.OPENAI.value: "OPENAI_API_KEY",
+    InferenceServiceType.GOOGLE.value: "GOOGLE_API_KEY",
+    InferenceServiceType.TEST.value: "TBD",
+    InferenceServiceType.ANTHROPIC.value: "ANTHROPIC_API_KEY",
+    InferenceServiceType.GROQ.value: "GROQ_API_KEY",
+    InferenceServiceType.BEDROCK.value: ["AWS_ACCESS_KEY_ID", "AWS_SECRET_ACCESS_KEY"],
+    InferenceServiceType.MISTRAL.value: "MISTRAL_API_KEY",
+    InferenceServiceType.TOGETHER.value: "TOGETHER_API_KEY",
+    InferenceServiceType.PERPLEXITY.value: "PERPLEXITY_API_KEY",
+}
+
+
+class TokenPricing:
+    def __init__(
+        self,
+        *,
+        model_name,
+        prompt_token_price_per_k: float,
+        completion_token_price_per_k: float,
+    ):
+        self.model_name = model_name
+        self.prompt_token_price = prompt_token_price_per_k / 1_000.0
+        self.completion_token_price = completion_token_price_per_k / 1_000.0
+
+    def __eq__(self, other):
+        if not isinstance(other, TokenPricing):
+            return False
+        return (
+            self.model_name == other.model_name
+            and self.prompt_token_price == other.prompt_token_price
+            and self.completion_token_price == other.completion_token_price
+        )
+
+
+pricing = {
+    "dbrx-instruct": TokenPricing(
+        model_name="dbrx-instruct",
+        prompt_token_price_per_k=0.0,
+        completion_token_price_per_k=0.0,
+    ),
+    "claude-3-opus-20240229": TokenPricing(
+        model_name="claude-3-opus-20240229",
+        prompt_token_price_per_k=0.0,
+        completion_token_price_per_k=0.0,
+    ),
+    "claude-3-haiku-20240307": TokenPricing(
+        model_name="claude-3-haiku-20240307",
+        prompt_token_price_per_k=0.0,
+        completion_token_price_per_k=0.0,
+    ),
+    "claude-3-sonnet-20240229": TokenPricing(
+        model_name="claude-3-sonnet-20240229",
+        prompt_token_price_per_k=0.0,
+        completion_token_price_per_k=0.0,
+    ),
+    "gpt-3.5-turbo": TokenPricing(
+        model_name="gpt-3.5-turbo",
+        prompt_token_price_per_k=0.0005,
+        completion_token_price_per_k=0.0015,
+    ),
+    "gpt-4-1106-preview": TokenPricing(
+        model_name="gpt-4",
+        prompt_token_price_per_k=0.01,
+        completion_token_price_per_k=0.03,
+    ),
+    "test": TokenPricing(
+        model_name="test",
+        prompt_token_price_per_k=0.0,
+        completion_token_price_per_k=0.0,
+    ),
+    "gemini_pro": TokenPricing(
+        model_name="gemini_pro",
+        prompt_token_price_per_k=0.0,
+        completion_token_price_per_k=0.0,
+    ),
+    "llama-2-13b-chat-hf": TokenPricing(
+        model_name="llama-2-13b-chat-hf",
+        prompt_token_price_per_k=0.0,
+        completion_token_price_per_k=0.0,
+    ),
+    "llama-2-70b-chat-hf": TokenPricing(
+        model_name="llama-2-70b-chat-hf",
+        prompt_token_price_per_k=0.0,
+        completion_token_price_per_k=0.0,
+    ),
+    "mixtral-8x7B-instruct-v0.1": TokenPricing(
+        model_name="mixtral-8x7B-instruct-v0.1",
+        prompt_token_price_per_k=0.0,
+        completion_token_price_per_k=0.0,
+    ),
+}
+
+
+def get_token_pricing(model_name):
+    if model_name in pricing:
+        return pricing[model_name]
+    else:
+        return TokenPricing(
+            model_name=model_name,
+            prompt_token_price_per_k=0.0,
+            completion_token_price_per_k=0.0,
+        )

edsl/exceptions/BaseException.py

@@ -1,21 +1,21 @@
-class BaseException(Exception):
-    relevant_doc = "https://docs.expectedparrot.com/"
-
-    def __init__(self, message, *, show_docs=True):
-        # Format main error message
-        formatted_message = [message.strip()]
-
-        # Add documentation links if requested
-        if show_docs:
-            if hasattr(self, "relevant_doc"):
-                formatted_message.append(
-                    f"\nFor more information, see:\n{self.relevant_doc}"
-                )
-            if hasattr(self, "relevant_notebook"):
-                formatted_message.append(
-                    f"\nFor a usage example, see:\n{self.relevant_notebook}"
-                )
-
-        # Join with double newlines for clear separation
-        final_message = "\n\n".join(formatted_message)
-        super().__init__(final_message)
+class BaseException(Exception):
+    relevant_doc = "https://docs.expectedparrot.com/"
+
+    def __init__(self, message, *, show_docs=True):
+        # Format main error message
+        formatted_message = [message.strip()]
+
+        # Add documentation links if requested
+        if show_docs:
+            if hasattr(self, "relevant_doc"):
+                formatted_message.append(
+                    f"\nFor more information, see:\n{self.relevant_doc}"
+                )
+            if hasattr(self, "relevant_notebook"):
+                formatted_message.append(
+                    f"\nFor a usage example, see:\n{self.relevant_notebook}"
+                )
+
+        # Join with double newlines for clear separation
+        final_message = "\n\n".join(formatted_message)
+        super().__init__(final_message)