lionagi 0.7.5__tar.gz → 0.7.7__tar.gz

{lionagi-0.7.5 → lionagi-0.7.7}/.env.example

@@ -4,3 +4,4 @@
 # PERPLEXITY_API_KEY=
 # GROQ_API_KEY=
 # OPENROUTER_API_KEY=
+# EXA_API_KEY=

{lionagi-0.7.5 → lionagi-0.7.7}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: lionagi
-Version: 0.7.5
+Version: 0.7.7
 Summary: An Intelligence Operating System.
 Author-email: HaiyangLi <quantocean.li@gmail.com>
 License:            Apache License
@@ -214,6 +214,7 @@ Classifier: License :: OSI Approved :: Apache Software License
 Classifier: Operating System :: OS Independent
 Classifier: Programming Language :: Python :: 3
 Classifier: Programming Language :: Python :: 3 :: Only
+Classifier: Programming Language :: Python :: 3.10
 Classifier: Programming Language :: Python :: 3.11
 Classifier: Programming Language :: Python :: 3.12
 Classifier: Programming Language :: Python :: 3.13

{lionagi-0.7.5 → lionagi-0.7.7}/lionagi/__init__.py

@@ -2,6 +2,8 @@
 #
 # SPDX-License-Identifier: Apache-2.0
 
+from pydantic import BaseModel, Field
+
 from . import _types as types
 from .operations import types as op
 from .operatives import types as ops_types  # deprecated
@@ -20,4 +22,6 @@ __all__ = (
     "ops_types",
     "op",
     "__version__",
+    "BaseModel",
+    "Field",
 )

{lionagi-0.7.5 → lionagi-0.7.7}/lionagi/libs/token_transform/synthlang.py

@@ -387,10 +387,16 @@ async def translate_to_synthlang(
     calculator = TokenCalculator()
 
     len_tokens = calculator.tokenize(text, return_tokens=False)
-    out = await branch.communicate(
+
+    kwargs["guidance"] = (
+        "Following SynthLang, translate the provided text into SynthLang syntax. "
+        "Shrink the token size by 60-85%. Return only the translated text.\n\n"
+        + kwargs.get("guidance", "")
+    )
+
+    out = await branch.chat(
         instruction=f"Converts the given text into SynthLang's hyper-efficient format.",
         context="Text to convert:\n\n" + text,
-        guidance=f"Following SynthLang, translate the provided text into SynthLang syntax. Shrink the token size by 60-85%. Return only the translated text.",
         **kwargs,
     )
     if sys1:

{lionagi-0.7.5 → lionagi-0.7.7}/lionagi/operations/interpret/interpret.py

@@ -13,6 +13,7 @@ async def interpret(
     text: str,
     domain: str | None = None,
     style: str | None = None,
+    sample_writing: str | None = None,
     **kwargs,
 ) -> str:
     instruction = (
@@ -25,14 +26,17 @@ async def interpret(
         f"Desired style: {style or 'concise'}. "
         "You can add or clarify context if needed."
     )
+    if sample_writing:
+        guidance += f" Sample writing: {sample_writing}"
+
     context = [f"User input: {text}"]
 
     # Default temperature if none provided
     kwargs["temperature"] = kwargs.get("temperature", 0.1)
+    kwargs["guidance"] = guidance + "\n" + kwargs.get("guidance", "")
 
     refined_prompt = await branch.chat(
         instruction=instruction,
-        guidance=guidance,
         context=context,
         **kwargs,
     )

lionagi-0.7.7/lionagi/operations/manager.py

@@ -0,0 +1,22 @@
+from typing import Callable
+
+from lionagi.protocols._concepts import Manager
+
+"""
+experimental
+"""
+
+
+class OperationManager(Manager):
+
+    def __init__(self, *args, **kwargs):
+        super().__init__()
+        self.registry: dict[str, Callable] = {}
+        self.register_operations(*args, **kwargs)
+
+    def register_operations(self, *args, **kwargs) -> None:
+        operations = {}
+        if args:
+            operations = {i.__name__ for i in args if hasattr(i, "__name__")}
+        operations.update(kwargs)
+        self.registry.update(operations)

{lionagi-0.7.5 → lionagi-0.7.7}/lionagi/operations/operate/operate.py

@@ -47,6 +47,9 @@ async def operate(
     actions: bool = False,
     reason: bool = False,
     action_kwargs: dict = None,
+    action_strategy: Literal[
+        "sequential", "concurrent", "batch"
+    ] = "concurrent",
     field_models: list[FieldModel] = None,
     exclude_fields: list | dict | None = None,
     request_params: ModelParams = None,
@@ -97,6 +100,8 @@ async def operate(
         instruct.reason = True
     if actions:
         instruct.actions = True
+        if action_strategy:
+            instruct.action_strategy = action_strategy
 
     # 1) Create or update the Operative
     operative = Step.request_operative(
@@ -178,9 +183,15 @@ async def operate(
         getattr(response_model, "action_required", None) is True
         and getattr(response_model, "action_requests", None) is not None
     ):
+        action_kwargs = action_kwargs or {}
+        action_kwargs["strategy"] = (
+            instruct.action_strategy
+            if instruct.action_strategy
+            else action_kwargs.get("strategy", "concurrent")
+        )
+
         action_response_models = await branch.act(
-            response_model.action_requests,
-            **(action_kwargs or {}),
+            response_model.action_requests, **action_kwargs
         )
         # Possibly refine the operative with the tool outputs
         operative = Step.respond_operative(

{lionagi-0.7.5 → lionagi-0.7.7}/lionagi/operatives/instruct/base.py

@@ -2,6 +2,8 @@
 #
 # SPDX-License-Identifier: Apache-2.0
 
+"""deprecated"""
+
 from pydantic import JsonValue
 
 from lionagi.libs.validate.common_field_validators import (

lionagi-0.7.7/lionagi/operatives/instruct/instruct.py

@@ -0,0 +1,153 @@
+# Copyright (c) 2023 - 2024, HaiyangLi <quantocean.li at gmail dot com>
+#
+# SPDX-License-Identifier: Apache-2.0
+
+from typing import Any, ClassVar, Literal
+
+from pydantic import Field, JsonValue, field_validator
+
+from lionagi.libs.validate.common_field_validators import (
+    validate_boolean_field,
+    validate_nullable_jsonvalue_field,
+)
+from lionagi.utils import HashableModel, to_num
+
+from ..models.field_model import FieldModel
+
+__all__ = (
+    "Instruct",
+    "InstructResponse",
+    "INSTRUCT_FIELD",
+    "LIST_INSTRUCT_FIELD",
+)
+
+
+class Instruct(HashableModel):
+    """Model for defining instruction parameters and execution requirements.
+
+    Attributes:
+        instruction (JsonValue | None): The primary instruction.
+        guidance (JsonValue | None): Execution guidance.
+        context (JsonValue | None): Task context.
+    """
+
+    reserved_kwargs: ClassVar[list[str]] = [
+        "operative_model",
+        "field_models",
+        "operative",
+        "reason",
+        "actions",
+        "action_strategy",
+        "batch_size",
+        "request_params",
+        "response_params",
+    ]
+    instruction: JsonValue | None = Field(
+        None,
+        title="Primary Instruction",
+        description=(
+            "A clear, actionable task definition. Specify:\n"
+            "1) The primary goal or objective\n"
+            "2) Key success criteria or constraints\n"
+            "\n"
+            "Guidelines:\n"
+            "- Start with a direct action verb (e.g., 'Analyze', 'Generate', 'Create')\n"
+            "- Include scope, boundaries, or constraints\n"
+            "- Provide success criteria if relevant\n"
+            "- For complex tasks, break them into logical steps"
+        ),
+    )
+    guidance: JsonValue | None = Field(
+        None,
+        title="Guidance",
+        description=(
+            "Strategic direction and constraints for executing the task. "
+            "Include:\n"
+            "1) Preferred methods or frameworks\n"
+            "2) Quality benchmarks (e.g., speed, clarity)\n"
+            "3) Resource or environmental constraints\n"
+            "4) Relevant compliance or standards\n"
+            "Use None if no special guidance."
+        ),
+    )
+    context: JsonValue | None = Field(
+        None,
+        description=(
+            "Background information and current-state data needed for the task. "
+            "Should be:\n"
+            "1) Directly relevant\n"
+            "2) Sufficient to perform the task\n"
+            "3) Free of extraneous detail\n"
+            "Include environment, prior outcomes, system states, or dependencies. "
+            "Use None if no additional context is needed."
+        ),
+    )
+    reason: bool | None = Field(
+        None,
+        description=(
+            "Include a thoughtful explanation of decisions, trade-offs, "
+            "and insights. Encourage deeper introspection on why certain "
+            "choices were made, potential alternatives, and how confidence "
+            "was shaped. If not needed, set to None."
+        ),
+    )
+    actions: bool | None = Field(
+        None,
+        description=(
+            "Controls execution mode. "
+            "True: Execute specified actions. "
+            "False: Analysis/recommendations only. "
+            "None: Contextual execution."
+        ),
+    )
+    action_strategy: Literal["batch", "sequential", "concurrent"] | None = (
+        Field(
+            None,
+            description="Action strategy to use for executing actions. Default "
+            "is 'concurrent'. Only provide for if actions are enabled.",
+        )
+    )
+    batch_size: int | None = Field(
+        None,
+        description="Batch size for executing actions. Only provide for 'batch' strategy.",
+    )
+
+    @field_validator("instruction", "guidance", "context", mode="before")
+    def _validate_instruction(cls, v):
+        return validate_nullable_jsonvalue_field(cls, v)
+
+    @field_validator("reason", "actions", mode="before")
+    def _validate_reason(cls, v):
+        return validate_boolean_field(cls, v)
+
+    @field_validator("action_strategy", mode="before")
+    def _validate_action_strategy(cls, v):
+        if v not in ["batch", "sequential", "concurrent"]:
+            return "concurrent"
+        return v
+
+    @field_validator("batch_size", mode="before")
+    def _validate_batch_size(cls, v):
+        try:
+            return to_num(v, num_type=int)
+        except Exception:
+            return None
+
+
+INSTRUCT_FIELD = FieldModel(
+    name="instruct_model",
+    annotation=Instruct | None,
+    default=None,
+)
+
+
+class InstructResponse(HashableModel):
+    instruct: Instruct
+    response: Any | None = None
+
+
+LIST_INSTRUCT_FIELD_MODEL = FieldModel(
+    name="instruct_models",
+    annotation=list[Instruct] | None,
+    default=None,
+)

{lionagi-0.7.5 → lionagi-0.7.7}/lionagi/operatives/instruct/prompts.py

@@ -2,6 +2,8 @@
 #
 # SPDX-License-Identifier: Apache-2.0
 
+"""deprecated"""
+
 instruction_field_description = (
     "A clear, actionable task definition. Specify:\n"
     "1) The primary goal or objective\n"

{lionagi-0.7.5 → lionagi-0.7.7}/lionagi/operatives/instruct/reason.py

@@ -2,13 +2,19 @@
 #
 # SPDX-License-Identifier: Apache-2.0
 
-from pydantic import BaseModel, field_validator
+from pydantic import BaseModel, Field, field_validator
 
 from lionagi.utils import to_num
 
 from ..models.field_model import FieldModel
 
+__all__ = (
+    "Reason",
+    "REASON_FIELD",
+)
+
 
+# deprecated
 def validate_confidence_score(cls, value) -> float:
     try:
         return to_num(
@@ -22,6 +28,7 @@ def validate_confidence_score(cls, value) -> float:
         return -1
 
 
+# deprecated
 confidence_description = (
     "Numeric confidence score (0.0 to 1.0, up to three decimals) indicating "
     "how well you've met user expectations. Use this guide:\n"
@@ -31,6 +38,7 @@ confidence_description = (
     "  • 0.0-0.5: Off track"
 )
 
+# deprecated
 CONFIDENCE_SCORE_FIELD = FieldModel(
     name="confidence_score",
     annotation=float | None,
@@ -47,13 +55,32 @@ class Reason(BaseModel):
 
     title: str | None = None
     content: str | None = None
-    confidence_score: float | None = CONFIDENCE_SCORE_FIELD.field_info
-
-    @field_validator(
-        "confidence_score", **CONFIDENCE_SCORE_FIELD.validator_kwargs
+    confidence_score: float | None = Field(
+        None,
+        title="Confidence Score",
+        description=(
+            "Numeric confidence score (0.0 to 1.0, up to three decimals) indicating "
+            "how well you've met user expectations. Use this guide:\n"
+            "  • 1.0: Highly confident\n"
+            "  • 0.8-1.0: Reasonably sure\n"
+            "  • 0.5-0.8: Re-check or refine\n"
+            "  • 0.0-0.5: Off track"
+        ),
+        examples=[0.821, 0.257, 0.923, 0.439],
     )
+
+    @field_validator("confidence_score", mode="before")
     def _validate_confidence(cls, v):
-        return CONFIDENCE_SCORE_FIELD.validator(cls, v)
+        try:
+            return to_num(
+                v,
+                upper_bound=1,
+                lower_bound=0,
+                num_type=float,
+                precision=3,
+            )
+        except Exception:
+            return -1
 
 
 REASON_FIELD = FieldModel(
@@ -62,6 +89,3 @@ REASON_FIELD = FieldModel(
     title="Reason",
     description="**Provide a concise reason for the decision made.**",
 )
-
-
-__all__ = ["Reason"]

lionagi-0.7.7/lionagi/service/endpoints/match_endpoint.py

@@ -0,0 +1,65 @@
+# Copyright (c) 2023 - 2024, HaiyangLi <quantocean.li at gmail dot com>
+#
+# SPDX-License-Identifier: Apache-2.0
+
+from .chat_completion import EndPoint
+
+
+def match_endpoint(
+    provider: str,
+    base_url: str,
+    endpoint: str,
+    endpoint_params: list[str] | None = None,
+) -> EndPoint:
+
+    if endpoint in ["chat/completions", "chat", "messages"]:
+        from ..providers.openai_.chat_completions import (
+            OpenAIChatCompletionEndPoint,
+        )
+
+        if provider == "openai":
+            return OpenAIChatCompletionEndPoint()
+
+        if provider == "anthropic":
+            from ..providers.anthropic_.messages import (
+                AnthropicChatCompletionEndPoint,
+            )
+
+            return AnthropicChatCompletionEndPoint()
+
+        if provider == "groq":
+            from ..providers.groq_.chat_completions import (
+                GroqChatCompletionEndPoint,
+            )
+
+            return GroqChatCompletionEndPoint()
+
+        if provider == "perplexity":
+            from ..providers.perplexity_.chat_completions import (
+                PerplexityChatCompletionEndPoint,
+            )
+
+            return PerplexityChatCompletionEndPoint()
+
+        if provider == "openrouter":
+            from ..providers.openrouter_.chat_completions import (
+                OpenRouterChatCompletionEndPoint,
+            )
+
+            return OpenRouterChatCompletionEndPoint()
+
+        return OpenAIChatCompletionEndPoint(
+            config={
+                "provider": provider,
+                "base_url": base_url,
+                "endpoint": endpoint,
+                "endpoint_params": endpoint_params,
+            }
+        )
+
+    if provider == "exa" and endpoint == "search":
+        from ..providers.exa_.search import ExaSearchEndPoint
+
+        return ExaSearchEndPoint()
+
+    raise ValueError(f"Unsupported endpoint: {endpoint}")

{lionagi-0.7.5 → lionagi-0.7.7}/lionagi/service/imodel.py

@@ -102,6 +102,8 @@ class iModel:
                     api_key = "PERPLEXITY_API_KEY"
                 case "groq":
                     api_key = "GROQ_API_KEY"
+                case "exa":
+                    api_key = "EXA_API_KEY"
 
         if os.getenv(api_key, None) is not None:
             self.api_key_scheme = api_key
@@ -198,7 +200,9 @@ class iModel:
         except Exception as e:
             raise ValueError(f"Failed to stream API call: {e}")
 
-    async def invoke(self, **kwargs) -> APICalling | None:
+    async def invoke(
+        self, api_call: APICalling = None, **kwargs
+    ) -> APICalling | None:
         """Invokes a rate-limited API call with the given arguments.
 
         Args:
@@ -215,8 +219,9 @@ class iModel:
                 If the call fails or if an error occurs during invocation.
         """
         try:
-            kwargs.pop("stream", None)
-            api_call = self.create_api_calling(**kwargs)
+            if api_call is None:
+                kwargs.pop("stream", None)
+                api_call = self.create_api_calling(**kwargs)
             if (
                 self.executor.processor is None
                 or self.executor.processor.is_stopped()
@@ -252,6 +257,15 @@ class iModel:
         """
         return self.endpoint.sequential_exchange
 
+    @property
+    def model_name(self) -> str:
+        """str: The name of the model used by the endpoint.
+
+        Returns:
+            The model name if available; otherwise, an empty string.
+        """
+        return self.kwargs.get("model", "")
+
     def to_dict(self):
         kwargs = self.kwargs
         if "kwargs" in self.kwargs:

lionagi-0.7.7/lionagi/service/providers/exa_/models.py

@@ -0,0 +1,160 @@
+from enum import Enum
+
+from pydantic import BaseModel, Field
+
+
+class CategoryEnum(str, Enum):
+    company = "company"
+    research_paper = "research paper"
+    news = "news"
+    pdf = "pdf"
+    github = "github"
+    tweet = "tweet"
+    personal_site = "personal site"
+    linkedin_profile = "linkedin profile"
+    financial_report = "financial report"
+
+
+class LivecrawlEnum(str, Enum):
+    never = "never"
+    fallback = "fallback"
+    always = "always"
+
+
+class SearchTypeEnum(str, Enum):
+    keyword = "keyword"
+    neural = "neural"
+    auto = "auto"
+
+
+class ContentsText(BaseModel):
+    includeHtmlTags: bool | None = Field(
+        default=False,
+        description="Whether to include HTML tags in the text. Set to True if you want to retain HTML structure for the LLM to interpret.",
+    )
+    maxCharacters: int | None = Field(
+        default=None,
+        description="The maximum number of characters to return from the webpage text.",
+    )
+
+
+class ContentsHighlights(BaseModel):
+    highlightsPerUrl: int | None = Field(
+        default=1,
+        description="The number of highlight snippets you want per page.",
+    )
+    numSentences: int | None = Field(
+        default=5,
+        description="Number of sentences to return in each highlight snippet.",
+    )
+    query: None | str = Field(
+        default=None,
+        description="A specific query used to generate the highlight snippets.",
+    )
+
+
+class ContentsSummary(BaseModel):
+    query: None | str = Field(
+        default=None,
+        description="A specific query used to generate a summary of the webpage.",
+    )
+
+
+class ContentsExtras(BaseModel):
+    links: int | None = Field(
+        default=None, description="Number of links to return from each page."
+    )
+    imageLinks: int | None = Field(
+        default=None, description="Number of images to return for each result."
+    )
+
+
+class Contents(BaseModel):
+    text: None | ContentsText = Field(
+        default=None,
+        description="Return full or partial text for each page, with optional HTML structure or size limit.",
+    )
+    highlights: None | ContentsHighlights = Field(
+        default=None, description="Return snippet highlights for each page."
+    )
+    summary: None | ContentsSummary = Field(
+        default=None, description="Return a short summary of each page."
+    )
+    livecrawl: None | LivecrawlEnum = Field(
+        default=LivecrawlEnum.never,
+        description="Livecrawling setting for each page. Options: never, fallback, always.",
+    )
+    livecrawlTimeout: int | None = Field(
+        default=10000,
+        description="Timeout in milliseconds for livecrawling. Default 10000.",
+    )
+    subpages: int | None = Field(
+        default=None,
+        description="Number of subpages to crawl within each URL.",
+    )
+    subpageTarget: None | str | list[str] = Field(
+        default=None,
+        description="A target subpage or multiple subpages (list) to crawl, e.g. 'cited papers'.",
+    )
+    extras: None | ContentsExtras = Field(
+        default=None,
+        description="Additional extras like links or images to return for each page.",
+    )
+
+
+class ExaSearchRequest(BaseModel):
+    query: str = Field(
+        ...,
+        description="The main query string describing what you're looking for.",
+    )
+    category: None | CategoryEnum = Field(
+        default=None,
+        description="A data category to focus on, such as 'company', 'research paper', 'news', etc.",
+    )
+    type: None | SearchTypeEnum = Field(
+        default=None,
+        description="The type of search to run. Can be 'auto', 'keyword', or 'neural'.",
+    )
+    useAutoprompt: None | bool = Field(
+        default=False,
+        description="If True, Exa auto-optimizes your query for best results (neural or auto search only).",
+    )
+    numResults: int | None = Field(
+        default=10, description="Number of results to return. Default is 10."
+    )
+    includeDomains: None | list[str] = Field(
+        default=None,
+        description="List of domains you want to include exclusively.",
+    )
+    excludeDomains: None | list[str] = Field(
+        default=None,
+        description="List of domains you do NOT want to see in the results.",
+    )
+    startCrawlDate: None | str = Field(
+        default=None,
+        description="Include results crawled after this ISO date (e.g., '2023-01-01T00:00:00.000Z').",
+    )
+    endCrawlDate: None | str = Field(
+        default=None,
+        description="Include results crawled before this ISO date.",
+    )
+    startPublishedDate: None | str = Field(
+        default=None,
+        description="Only return results published after this ISO date.",
+    )
+    endPublishedDate: None | str = Field(
+        default=None,
+        description="Only return results published before this ISO date.",
+    )
+    includeText: None | list[str] = Field(
+        default=None,
+        description="Strings that must appear in the webpage text. Only a single string up to 5 words is currently supported.",
+    )
+    excludeText: None | list[str] = Field(
+        default=None,
+        description="Strings that must NOT appear in the webpage text. Only a single string up to 5 words is currently supported.",
+    )
+    contents: None | Contents = Field(
+        default=None,
+        description="Dict defining the different ways you want to retrieve webpage contents, including text, highlights, or summaries.",
+    )