lionagi 0.0.314__tar.gz → 0.0.316__tar.gz

{lionagi-0.0.314/lionagi.egg-info → lionagi-0.0.316}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: lionagi
-Version: 0.0.314
+Version: 0.0.316
 Summary: Towards automated general intelligence.
 Author: HaiyangLi
 Author-email: Haiyang Li <ocean@lionagi.ai>

{lionagi-0.0.314 → lionagi-0.0.316}/lionagi/__init__.py

@@ -8,6 +8,9 @@ from dotenv import load_dotenv
 
 from .core import direct, Branch, Session, func_to_tool
 from .integrations.provider.services import Services
+from .integrations.chunker.chunk import chunk
+from .integrations.loader.load import load
+
 
 logger = logging.getLogger(__name__)
 logger.setLevel(logging.INFO)

lionagi-0.0.316/lionagi/core/direct/__init__.py

@@ -0,0 +1,10 @@
+from .predict import predict
+from .select import select
+from .score import score
+from .react import react
+from .vote import vote
+from .plan import plan
+from .cot import chain_of_thoughts, chain_of_react
+
+
+__all__ = ["predict", "select", "score", "vote", "react", "plan", "chain_of_thoughts", "chain_of_react"]

lionagi-0.0.316/lionagi/core/direct/cot.py

@@ -0,0 +1,88 @@
+from typing import Callable
+from lionagi.libs import convert
+from ..tool import func_to_tool
+from ..schema import Tool
+from .predict import predict
+from .plan import plan
+from .react import react
+
+from .utils import _process_tools
+
+
+async def chain_of_thoughts(
+    sentence=None,
+    branch=None,
+    instruction=None,
+    reason=False,
+    confidence_score=False,
+    num_steps=3,
+    directive_kwargs={},
+    return_branch=False,
+    **kwargs
+):
+    
+    out_, outs, answer, reasons, confidence_score = "", [], "", [], 0
+    if branch is not None:
+        out_ = await plan(sentence, branch=branch, instruction=instruction, num_steps=num_steps, **kwargs)
+    else:
+        out_, branch = await plan(sentence, instruction=instruction, branch=branch, num_steps=num_steps, return_branch=True, **kwargs)
+    
+    for i in range(len(out_.plan)):
+        _out = await predict(branch=branch, instruction=out_.plan[f"step_{i+1}"], reason=reason, confidence_score=confidence_score, **directive_kwargs)
+        answer += _out.answer
+        if reason:
+            reasons.append(_out.reason)
+        if confidence_score:
+            confidence_score += _out.confidence_score
+        outs.append(_out)
+    
+    setattr(out_, "chain_output", outs)
+    setattr(out_, "chain_answer", answer)
+    
+    if reason:
+        setattr(out_, "chain_reasons", reasons)
+    if confidence_score:
+        setattr(out_, "chain_confidence_score", confidence_score/len(outs))
+    
+    if return_branch:
+        return out_, branch
+    
+    return out_
+    
+
+async def chain_of_react(
+    sentence=None,
+    branch=None,
+    instruction=None,
+    num_steps=3,
+    tools=None,
+    directive_system=None,
+    directive_kwargs={},
+    return_branch=False,
+    **kwargs
+):
+    out_, outs, reasons, actions, action_responses = "", [], [], [], []
+    if branch is not None:
+        out_ = await plan(sentence, branch=branch, instruction=instruction, num_steps=num_steps, **kwargs)
+    else:
+        out_, branch = await plan(sentence, instruction=instruction, branch=branch, num_steps=num_steps, return_branch=True, **kwargs)
+    
+    _process_tools(tools, branch)
+    
+    for i in range(len(out_.plan)):
+        _out = await react(branch=branch, system=directive_system, instruction=out_.plan[f"step_{i+1}"], **directive_kwargs)
+        outs.append(_out)
+        reasons.append(_out.reason)
+        actions.append(_out.actions)
+        if _out.action_needed:
+            action_responses.append(_out.action_response)
+    
+    setattr(out_, "chain_output", convert.to_list(outs))
+    setattr(out_, "chain_reason", convert.to_list(reasons))
+    setattr(out_, "chain_actions", convert.to_list(actions))
+    setattr(out_, "chain_action_response", convert.to_list(action_responses))
+
+    if return_branch:
+        return out_, branch
+    
+    return out_

lionagi-0.0.316/lionagi/core/direct/plan.py

@@ -0,0 +1,162 @@
+# plan.py
+
+from lionagi.libs import func_call, ParseUtil
+from lionagi.integrations.bridge.pydantic_.pydantic_bridge import Field
+from ..prompt.scored_template import ScoredTemplate
+from ..branch import Branch
+
+
+class PlanTemplate(ScoredTemplate):
+    template_name: str = "default_plan"
+    sentence: str | list | dict = Field(
+        default_factory=str,
+        description="the given sentence(s) or context to generate a plan for",
+    )
+    plan: dict | str= Field(
+        default_factory=dict, description="the generated step by step plan, return as a dictionary following {step_n: {plan: ..., reason: ...}} format")
+    signature: str = "sentence -> plan"
+
+    def __init__(
+        self,
+        sentence=None,
+        instruction=None,
+        confidence_score=False,
+        reason=False,
+        num_step=3,
+        **kwargs,
+    ):
+        super().__init__(**kwargs)
+
+        self.sentence = sentence
+        self.task = f"Generate a {num_step}_step plan based on the given context. Instruction: {instruction}."
+
+        if reason:
+            self.output_fields.append("reason")
+
+        if confidence_score:
+            self.output_fields.append("confidence_score")
+
+
+async def _plan(
+    sentence,
+    *,
+    instruction=None,
+    branch=None,
+    confidence_score=False,
+    reason=False,
+    retries=2,
+    delay=0.5,
+    backoff_factor=2,
+    default_value=None,
+    timeout=None,
+    branch_name=None,
+    system=None,
+    messages=None,
+    service=None,
+    sender=None,
+    llmconfig=None,
+    tools=None,
+    datalogger=None,
+    persist_path=None,
+    tool_manager=None,
+    return_branch=False,
+    **kwargs,
+):
+    if "temperature" not in kwargs:
+        kwargs["temperature"] = 0.1
+
+    instruction = instruction or ""
+
+    branch = branch or Branch(
+        name=branch_name,
+        system=system,
+        messages=messages,
+        service=service,
+        sender=sender,
+        llmconfig=llmconfig,
+        tools=tools,
+        datalogger=datalogger,
+        persist_path=persist_path,
+        tool_manager=tool_manager,
+    )
+
+    _template = PlanTemplate(
+        sentence=sentence,
+        instruction=instruction,
+        confidence_score=confidence_score,
+        reason=reason,
+    )
+
+    await func_call.rcall(
+        branch.chat,
+        prompt_template=_template,
+        retries=retries,
+        delay=delay,
+        backoff_factor=backoff_factor,
+        default=default_value,
+        timeout=timeout,
+        **kwargs,
+    )
+    
+    _template.plan = ParseUtil.fuzzy_parse_json(_template.plan)
+
+    return (_template, branch) if return_branch else _template
+
+
+async def plan(
+    sentence,
+    *,
+    instruction=None,
+    num_instances=1,
+    branch=None,
+    confidence_score=False,
+    reason=False,
+    retries=2,
+    delay=0.5,
+    backoff_factor=2,
+    default_value=None,
+    timeout=None,
+    branch_name=None,
+    system=None,
+    messages=None,
+    service=None,
+    sender=None,
+    llmconfig=None,
+    tools=None,
+    datalogger=None,
+    persist_path=None,
+    tool_manager=None,
+    return_branch=False,
+    **kwargs,
+):
+    async def _inner(i=0):
+        return await _plan(
+            sentence=sentence,
+            instruction=instruction,
+            branch=branch,
+            confidence_score=confidence_score,
+            reason=reason,
+            retries=retries,
+            delay=delay,
+            backoff_factor=backoff_factor,
+            default_value=default_value,
+            timeout=timeout,
+            branch_name=branch_name,
+            system=system,
+            messages=messages,
+            service=service,
+            sender=sender,
+            llmconfig=llmconfig,
+            tools=tools,
+            datalogger=datalogger,
+            persist_path=persist_path,
+            tool_manager=tool_manager,
+            return_branch=return_branch,
+            **kwargs,
+        )
+
+    if num_instances == 1:
+        return await _inner()
+
+    elif num_instances > 1:
+        return await func_call.alcall(range(num_instances), _inner)

{lionagi-0.0.314 → lionagi-0.0.316}/lionagi/core/direct/predict.py

@@ -43,14 +43,15 @@ class PredictTemplate(ScoredTemplate):
         default_factory=int, description="the number of sentences to predict"
     )
     answer: str | list = Field(
-        default_factory=str, description="the predicted sentence(s)"
+        default_factory=str, description="the predicted sentence(s) or desired output"
     )
     signature: str = "sentence -> answer"
 
     def __init__(
         self,
         sentence=None,
-        num_sentences=None,
+        instruction=None,
+        num_sentences=1,
         confidence_score=False,
         reason=False,
         **kwargs,
@@ -67,9 +68,9 @@ class PredictTemplate(ScoredTemplate):
         """
         super().__init__(**kwargs)
 
-        self.sentence = sentence
+        self.sentence = sentence or ''
         self.num_sentences = num_sentences
-        self.task = f"predict the next {self.num_sentences} sentence(s)"
+        self.task = f"follow instruction to predict the next {self.num_sentences} sentence(s). Instruction: {instruction}."
 
         if reason:
             self.output_fields.append("reason")
@@ -82,6 +83,8 @@ async def predict(
     sentence=None,
     num_sentences=1,
     confidence_score=False,
+    instruction=None,
+    branch=None,
     reason=False,
     retries=2,
     delay=0.5,
@@ -128,7 +131,7 @@ async def predict(
     Returns:
         PredictTemplate: The predict template with the predicted sentence(s).
     """
-    branch = Branch(
+    branch = branch or Branch(
         name=branch_name,
         system=system,
         messages=messages,
@@ -142,6 +145,7 @@ async def predict(
     )
 
     predict_template = PredictTemplate(
+        instruction=instruction,
         sentence=sentence,
         num_sentences=num_sentences,
         confidence_score=confidence_score,

{lionagi-0.0.314 → lionagi-0.0.316}/lionagi/core/direct/react.py

@@ -3,11 +3,12 @@ from lionagi.libs import func_call, convert, AsyncUtil
 from lionagi.integrations.bridge.pydantic_.pydantic_bridge import Field
 from ..prompt.action_template import ActionedTemplate
 from ..branch import Branch
+from .utils import _process_tools
 
 
 class ReactTemplate(ActionedTemplate):
     template_name: str = "default_react"
-    sentence: str | list | dict = Field(
+    sentence: str | list | dict | None= Field(
         default_factory=str,
         description="the given sentence(s) to reason and take actions on",
     )
@@ -29,7 +30,7 @@ class ReactTemplate(ActionedTemplate):
 
 
 async def _react(
-    sentence,
+    sentence=None,
     *,
     instruction=None,
     branch=None,
@@ -58,6 +59,9 @@ async def _react(
 
     instruction = instruction or ""
 
+    if branch and tools:
+        _process_tools(tools, branch)
+        
     branch = branch or Branch(
         name=branch_name,
         system=system,
@@ -109,7 +113,7 @@ async def _react(
 
 
 async def react(
-    sentence,
+    sentence=None,
     *,
     instruction=None,
     num_instances=1,
@@ -159,7 +163,7 @@ async def react(
             return_branch=return_branch,
             **kwargs,
         )
-        
+
     if num_instances == 1:
         return await _inner()
 

{lionagi-0.0.314 → lionagi-0.0.316}/lionagi/core/direct/select.py

@@ -39,9 +39,7 @@ class SelectTemplate(ScoredTemplate):
     answer: Enum | str = Field(
         default_factory=str, description="selection from given choices"
     )
-    choices: list = Field(
-        default_factory=list, description="the given choices"
-    )
+    choices: list = Field(default_factory=list, description="the given choices")
 
     signature: str = "sentence -> answer"
 

lionagi-0.0.316/lionagi/core/direct/sentiment.py

@@ -0,0 +1 @@
+# TODO: sentiment analysis

{lionagi-0.0.314 → lionagi-0.0.316}/lionagi/core/direct/utils.py

@@ -1,3 +1,6 @@
+from typing import Callable
+from ..tool import func_to_tool
+from ..schema import Tool
 # import contextlib
 # from lionagi.libs import ParseUtil, StringMatch, convert, func_call
 
@@ -85,3 +88,20 @@
 #         return _out
 
 #     return out_ if len(out_) > 1 else out_[0]
+
+
+def _process_tools(tool_obj, branch):
+    if isinstance(tool_obj, Callable):
+        _process_tool(tool_obj, branch)
+    else:
+        for i in tool_obj:
+            _process_tool(i, branch)
+
+
+def _process_tool(tool_obj, branch):
+    if isinstance(tool_obj, Tool) and tool_obj.schema_["function"]["name"] not in branch.tool_manager.registry:
+        branch.register_tools(tool_obj)
+    if isinstance(tool_obj, Callable):
+        tool = func_to_tool(tool_obj)[0]
+        if tool.schema_["function"]["name"] not in branch.tool_manager.registry:
+            branch.register_tools(tool)

{lionagi-0.0.314 → lionagi-0.0.316}/lionagi/integrations/bridge/llamaindex_/node_parser.py

@@ -29,19 +29,18 @@ def get_llama_index_node_parser(node_parser: Any):
     import llama_index.core.node_parser
 
     if not isinstance(node_parser, str) and not issubclass(node_parser, NodeParser):
-        raise TypeError(f"node_parser must be a string or NodeParser.")
+        raise TypeError("node_parser must be a string or NodeParser.")
 
     if isinstance(node_parser, str):
         if node_parser == "CodeSplitter":
             SysUtil.check_import("tree_sitter_languages")
 
         try:
-            parser = getattr(llama_index.core.node_parser, node_parser)
-            return parser
+            return getattr(llama_index.core.node_parser, node_parser)
         except Exception as e:
             raise AttributeError(
                 f"llama_index_core has no such attribute:" f" {node_parser}, Error: {e}"
-            )
+            ) from e
 
     elif isinstance(node_parser, NodeParser):
         return node_parser
@@ -75,10 +74,8 @@ def llama_index_parse_node(
         parser = get_llama_index_node_parser(node_parser)
         try:
             parser = parser(*parser_args, **parser_kwargs)
-        except:
+        except Exception:
             parser = parser.from_defaults(*parser_args, **parser_kwargs)
-        nodes = parser.get_nodes_from_documents(documents)
-        return nodes
-
+        return parser.get_nodes_from_documents(documents)
     except Exception as e:
-        raise ValueError(f"Failed to parse. Error: {e}")
+        raise ValueError(f"Failed to parse. Error: {e}") from e

lionagi-0.0.316/lionagi/integrations/chunker/chunk.py

@@ -0,0 +1,175 @@
+from typing import Union, Callable
+
+from lionagi.libs import func_call
+from lionagi.core.schema import DataNode
+from ..bridge.langchain_.langchain_bridge import LangchainBridge
+from ..bridge.llamaindex_.llama_index_bridge import LlamaIndexBridge
+
+
+from ..loader.load_util import ChunkerType, file_to_chunks, _datanode_parser
+
+
+def datanodes_convert(documents, chunker_type):
+
+    for i in range(len(documents)):
+        if type(documents[i]) == DataNode:
+            if chunker_type == ChunkerType.LLAMAINDEX:
+                documents[i] = documents[i].to_llama_index()
+            elif chunker_type == ChunkerType.LANGCHAIN:
+                documents[i] = documents[i].to_langchain()
+    return documents
+
+
+def text_chunker(documents, args, kwargs):
+
+    def chunk_node(node):
+        chunks = file_to_chunks(node.to_dict(), *args, **kwargs)
+        func_call.lcall(chunks, lambda chunk: chunk.pop("node_id"))
+        return [DataNode.from_obj({**chunk}) for chunk in chunks]
+
+    return [chunk_node(doc) for doc in documents]
+
+
+def chunk(
+    documents,
+    chunker,
+    chunker_type=ChunkerType.PLAIN,
+    chunker_args=None,
+    chunker_kwargs=None,
+    chunking_kwargs=None,
+    documents_convert_func=None,
+    to_datanode: bool | Callable = True,
+):
+
+    if chunker_args is None:
+        chunker_args = []
+    if chunker_kwargs is None:
+        chunker_kwargs = {}
+    if chunking_kwargs is None:
+        chunking_kwargs = {}
+
+    if chunker_type == ChunkerType.PLAIN:
+        return chunk_funcs[ChunkerType.PLAIN](
+            documents, chunker, chunker_args, chunker_kwargs
+        )
+
+    elif chunker_type == ChunkerType.LANGCHAIN:
+        return chunk_funcs[ChunkerType.LANGCHAIN](
+            documents,
+            documents_convert_func,
+            chunker,
+            chunker_args,
+            chunker_kwargs,
+            to_datanode,
+        )
+
+    elif chunker_type == ChunkerType.LLAMAINDEX:
+        return chunk_funcs[ChunkerType.LLAMAINDEX](
+            documents,
+            documents_convert_func,
+            chunker,
+            chunker_args,
+            chunker_kwargs,
+            to_datanode,
+        )
+
+    elif chunker_type == ChunkerType.SELFDEFINED:
+        return chunk_funcs[ChunkerType.SELFDEFINED](
+            documents,
+            chunker,
+            chunker_args,
+            chunker_kwargs,
+            chunking_kwargs,
+            to_datanode,
+        )
+
+    else:
+        raise ValueError(
+            f"{chunker_type} is not supported. Please choose from {list(ChunkerType)}"
+        )
+
+
+def _self_defined_chunker(
+    documents,
+    chunker,
+    chunker_args,
+    chunker_kwargs,
+    chunking_kwargs,
+    to_datanode: bool | Callable,
+):
+    try:
+        splitter = chunker(*chunker_args, **chunker_kwargs)
+        nodes = splitter.split(documents, **chunking_kwargs)
+    except Exception as e:
+        raise ValueError(
+            f"Self defined chunker {chunker} is not valid. Error: {e}"
+        ) from e
+
+    if isinstance(to_datanode, bool) and to_datanode is True:
+        raise ValueError("Please define a valid parser to DataNode.")
+    elif isinstance(to_datanode, Callable):
+        nodes = _datanode_parser(nodes, to_datanode)
+    return nodes
+
+
+def _llama_index_chunker(
+    documents,
+    documents_convert_func,
+    chunker,
+    chunker_args,
+    chunker_kwargs,
+    to_datanode: bool | Callable,
+):
+    if documents_convert_func:
+        documents = documents_convert_func(documents, "llama_index")
+    nodes = LlamaIndexBridge.llama_index_parse_node(
+        documents, chunker, chunker_args, chunker_kwargs
+    )
+
+    if isinstance(to_datanode, bool) and to_datanode is True:
+        nodes = [DataNode.from_llama_index(i) for i in nodes]
+    elif isinstance(to_datanode, Callable):
+        nodes = _datanode_parser(nodes, to_datanode)
+    return nodes
+
+
+def _langchain_chunker(
+    documents,
+    documents_convert_func,
+    chunker,
+    chunker_args,
+    chunker_kwargs,
+    to_datanode: bool | Callable,
+):
+    if documents_convert_func:
+        documents = documents_convert_func(documents, "langchain")
+    nodes = LangchainBridge.langchain_text_splitter(
+        documents, chunker, chunker_args, chunker_kwargs
+    )
+    if isinstance(to_datanode, bool) and to_datanode is True:
+        if isinstance(documents, str):
+            nodes = [DataNode(content=i) for i in nodes]
+        else:
+            nodes = [DataNode.from_langchain(i) for i in nodes]
+    elif isinstance(to_datanode, Callable):
+        nodes = _datanode_parser(nodes, to_datanode)
+    return nodes
+
+
+def _plain_chunker(documents, chunker, chunker_args, chunker_kwargs):
+    try:
+        if chunker == "text_chunker":
+            chunker = text_chunker
+        return chunker(documents, chunker_args, chunker_kwargs)
+    except Exception as e:
+        raise ValueError(
+            f"Reader {chunker} is currently not supported. Error: {e}"
+        ) from e
+
+
+chunk_funcs = {
+    ChunkerType.PLAIN: _plain_chunker,
+    ChunkerType.LANGCHAIN: _langchain_chunker,
+    ChunkerType.LLAMAINDEX: _llama_index_chunker,
+    ChunkerType.SELFDEFINED: _self_defined_chunker,
+}