agent
Agent(config=AgentConfig())
¶
Bases: ABC
An Agent is an abstraction that encapsulates mainly two components:
- a language model (LLM)
- a vector store (vecdb)
plus associated components such as a parser, and variables that hold information about any tool/function-calling messages that have been defined.
Source code in langroid/agent/base.py
indent: str
property
writable
¶
Indentation to print before any responses from the agent's entities.
entity_responders()
¶
Sequence of (entity, response_method) pairs. This sequence is used
in a Task
to respond to the current pending message.
See Task.step()
for details.
Returns:
Sequence of (entity, response_method) pairs.
Source code in langroid/agent/base.py
entity_responders_async()
¶
Async version of entity_responders
. See there for details.
Source code in langroid/agent/base.py
enable_message_handling(message_class=None)
¶
Enable an agent to RESPOND (i.e. handle) a "tool" message of a specific type
from LLM. Also "registers" (i.e. adds) the message_class
to the
self.llm_tools_map
dict.
Parameters:
Name | Type | Description | Default |
---|---|---|---|
message_class |
Optional[Type[ToolMessage]]
|
The message class to enable; Optional; if None, all known message classes are enabled for handling. |
None
|
Source code in langroid/agent/base.py
disable_message_handling(message_class=None)
¶
Disable a message class from being handled by this Agent.
Parameters:
Name | Type | Description | Default |
---|---|---|---|
message_class |
Optional[Type[ToolMessage]]
|
The message class to disable. If None, all message classes are disabled. |
None
|
Source code in langroid/agent/base.py
sample_multi_round_dialog()
¶
Generate a sample multi-round dialog based on enabled message classes. Returns: str: The sample dialog string.
Source code in langroid/agent/base.py
create_agent_response(content=None)
¶
agent_response(msg=None)
¶
Response from the "agent itself", typically (but not only)
used to handle LLM's "tool message" or function_call
(e.g. OpenAI function_call
).
Args:
msg (str|ChatDocument): the input to respond to: if msg is a string,
and it contains a valid JSON-structured "tool message", or
if msg is a ChatDocument, and it contains a function_call
.
Returns:
Optional[ChatDocument]: the response, packaged as a ChatDocument
Source code in langroid/agent/base.py
create_user_response(content=None)
¶
user_response(msg=None)
¶
Get user response to current message. Could allow (human) user to intervene with an actual answer, or quit using "q" or "x"
Parameters:
Name | Type | Description | Default |
---|---|---|---|
msg |
str | ChatDocument
|
the string to respond to. |
None
|
Returns:
Type | Description |
---|---|
Optional[ChatDocument]
|
(str) User response, packaged as a ChatDocument |
Source code in langroid/agent/base.py
llm_can_respond(message=None)
¶
Whether the LLM can respond to a message. Args: message (str|ChatDocument): message or ChatDocument object to respond to.
Returns:
Source code in langroid/agent/base.py
create_llm_response(content=None)
¶
llm_response_async(msg=None)
async
¶
Asynch version of llm_response
. See there for details.
Source code in langroid/agent/base.py
llm_response(msg=None)
¶
LLM response to a prompt. Args: msg (str|ChatDocument): prompt string, or ChatDocument object
Returns:
Type | Description |
---|---|
Optional[ChatDocument]
|
Response from LLM, packaged as a ChatDocument |
Source code in langroid/agent/base.py
has_tool_message_attempt(msg)
¶
Check whether msg contains a Tool/fn-call attempt (by the LLM)
Source code in langroid/agent/base.py
get_json_tool_messages(input_str)
¶
Returns ToolMessage objects (tools) corresponding to JSON substrings, if any.
Parameters:
Name | Type | Description | Default |
---|---|---|---|
input_str |
str
|
input string, typically a message sent by an LLM |
required |
Returns:
Type | Description |
---|---|
List[ToolMessage]
|
List[ToolMessage]: list of ToolMessage objects |
Source code in langroid/agent/base.py
tool_validation_error(ve)
¶
Handle a validation error raised when parsing a tool message, when there is a legit tool name used, but it has missing/bad fields. Args: tool (ToolMessage): The tool message that failed validation ve (ValidationError): The exception raised
Returns:
Name | Type | Description |
---|---|---|
str |
str
|
The error message to send back to the LLM |
Source code in langroid/agent/base.py
handle_message(msg)
¶
Handle a "tool" message either a string containing one or more
valid "tool" JSON substrings, or a
ChatDocument containing a function_call
attribute.
Handle with the corresponding handler method, and return
the results as a combined string.
Parameters:
Name | Type | Description | Default |
---|---|---|---|
msg |
str | ChatDocument
|
The string or ChatDocument to handle |
required |
Returns:
Type | Description |
---|---|
None | str | ChatDocument
|
Optional[Str]: The result of the handler method in string form so it can |
None | str | ChatDocument
|
be sent back to the LLM, or None if |
None | str | ChatDocument
|
handled by a method. |
Source code in langroid/agent/base.py
handle_message_fallback(msg)
¶
Fallback method to handle possible "tool" msg if no other method applies or if an error is thrown. This method can be overridden by subclasses.
Parameters:
Name | Type | Description | Default |
---|---|---|---|
msg |
str | ChatDocument
|
The input msg to handle |
required |
Returns: str: The result of the handler method in string form so it can be sent back to the LLM.
Source code in langroid/agent/base.py
handle_tool_message(tool)
¶
Respond to a tool request from the LLM, in the form of an ToolMessage object. Args: tool: ToolMessage object representing the tool request.
Returns:
Source code in langroid/agent/base.py
update_token_usage(response, prompt, stream, chat=True, print_response_stats=True)
¶
Updates response.usage
obj (token usage and cost fields).the usage memebr
It updates the cost after checking the cache and updates the
tokens (prompts and completion) if the response stream is True, because OpenAI
doesn't returns these fields.
Parameters:
Name | Type | Description | Default |
---|---|---|---|
response |
LLMResponse
|
LLMResponse object |
required |
prompt |
str | List[LLMMessage]
|
prompt or list of LLMMessage objects |
required |
stream |
bool
|
whether to update the usage in the response object if the response is not cached. |
required |
chat |
bool
|
whether this is a chat model or a completion model |
True
|
print_response_stats |
bool
|
whether to print the response stats |
True
|
Source code in langroid/agent/base.py
ask_agent(agent, request, no_answer=NO_ANSWER, user_confirm=True)
¶
Send a request to another agent, possibly after confirming with the user.
This is not currently used, since we rely on the task loop and
RecipientTool
to address requests to other agents. It is generally best to
avoid using this method.
Parameters:
Name | Type | Description | Default |
---|---|---|---|
agent |
Agent
|
agent to ask |
required |
request |
str
|
request to send |
required |
no_answer |
str
|
expected response when agent does not know the answer |
NO_ANSWER
|
user_confirm |
bool
|
whether to gate the request with a human confirmation |
True
|
Returns:
Name | Type | Description |
---|---|---|
str |
Optional[str]
|
response from agent |
Source code in langroid/agent/base.py
AgentConfig
¶
Bases: BaseSettings
General config settings for an LLM agent. This is nested, combining configs of various components.
ChatDocument(**data)
¶
Bases: Document
Source code in langroid/agent/chat_document.py
delete_id(id)
staticmethod
¶
Remove ChatDocument with given id from ObjectRegistry, and all its descendants.
Source code in langroid/agent/chat_document.py
get_json_tools()
¶
Get names of attempted JSON tool usages in the content of the message. Returns: List[str]: list of JSON tool names
Source code in langroid/agent/chat_document.py
log_fields()
¶
Fields for logging in csv/tsv logger Returns: List[str]: list of fields
Source code in langroid/agent/chat_document.py
pop_tool_ids()
¶
from_LLMResponse(response, displayed=False)
staticmethod
¶
Convert LLMResponse to ChatDocument. Args: response (LLMResponse): LLMResponse to convert. displayed (bool): Whether this response was displayed to the user. Returns: ChatDocument: ChatDocument representation of this LLMResponse.
Source code in langroid/agent/chat_document.py
to_LLMMessage(message)
staticmethod
¶
Convert to LLMMessage for use with LLM.
Parameters:
Name | Type | Description | Default |
---|---|---|---|
message |
str | ChatDocument
|
Message to convert. |
required |
Returns: LLMMessage: LLMMessage representation of this str or ChatDocument.
Source code in langroid/agent/chat_document.py
ChatAgentConfig
¶
Bases: AgentConfig
Configuration for ChatAgent
Attributes:
system_message: system message to include in message sequence
(typically defines role and task of agent).
Used only if task
is not specified in the constructor.
user_message: user message to include in message sequence.
Used only if task
is not specified in the constructor.
use_tools: whether to use our own ToolMessages mechanism
use_functions_api: whether to use functions native to the LLM API
(e.g. OpenAI's function_call
mechanism)
ChatAgent(config=ChatAgentConfig(), task=None)
¶
Bases: Agent
Chat Agent interacting with external env
(could be human, or external tools).
The agent (the LLM actually) is provided with an optional "Task Spec",
which is a sequence of LLMMessage
s. These are used to initialize
the task_messages
of the agent.
In most applications we will use a ChatAgent
rather than a bare Agent
.
The Agent
class mainly exists to hold various common methods and attributes.
One difference between ChatAgent
and Agent
is that ChatAgent
's
llm_response
method uses "chat mode" API (i.e. one that takes a
message sequence rather than a single message),
whereas the same method in the Agent
class uses "completion mode" API (i.e. one
that takes a single message).
config: settings for the agent
Source code in langroid/agent/chat_agent.py
task_messages: List[LLMMessage]
property
¶
The task messages are the initial messages that define the task of the agent. There will be at least a system message plus possibly a user msg. Returns: List[LLMMessage]: the task messages
from_id(id)
staticmethod
¶
Get an agent from its ID Args: agent_id (str): ID of the agent Returns: ChatAgent: The agent with the given ID
clone(i=0)
¶
Create i'th clone of this agent, ensuring tool use/handling is cloned. Important: We assume all member variables are in the init method here and in the Agent class. TODO: We are attempting to clone an agent after its state has been changed in possibly many ways. Below is an imperfect solution. Caution advised. Revisit later.
Source code in langroid/agent/chat_agent.py
clear_history(start=-2)
¶
Clear the message history, starting at the index start
Parameters:
Name | Type | Description | Default |
---|---|---|---|
start |
int
|
index of first message to delete; default = -2 (i.e. delete last 2 messages, typically these are the last user and assistant messages) |
-2
|
Source code in langroid/agent/chat_agent.py
update_history(message, response)
¶
Update the message history with the latest user message and LLM response. Args: message (str): user message response: (str): LLM response
Source code in langroid/agent/chat_agent.py
json_format_rules()
¶
Specification of JSON formatting rules, based on the currently enabled
usable ToolMessage
s
Returns:
Name | Type | Description |
---|---|---|
str |
str
|
formatting rules |
Source code in langroid/agent/chat_agent.py
tool_instructions()
¶
Instructions for tools or function-calls, for enabled and usable Tools. These are inserted into system prompt regardless of whether we are using our own ToolMessage mechanism or the LLM's function-call mechanism.
Returns:
Name | Type | Description |
---|---|---|
str |
str
|
concatenation of instructions for all usable tools |
Source code in langroid/agent/chat_agent.py
augment_system_message(message)
¶
Augment the system message with the given message. Args: message (str): system message
last_message_with_role(role)
¶
from message_history
, return the last message with role role
Source code in langroid/agent/chat_agent.py
nth_message_idx_with_role(role, n)
¶
Index of n
th message in message_history, with specified role.
(n is assumed to be 1-based, i.e. 1 is the first message with that role).
Return -1 if not found. Index = 0 is the first message in the history.
Source code in langroid/agent/chat_agent.py
update_last_message(message, role=Role.USER)
¶
Update the last message that has role role
in the message history.
Useful when we want to replace a long user prompt, that may contain context
documents plus a question, with just the question.
Args:
message (str): new message to replace with
role (str): role of message to replace
Source code in langroid/agent/chat_agent.py
enable_message(message_class, use=True, handle=True, force=False, require_recipient=False, include_defaults=True)
¶
Add the tool (message class) to the agent, and enable either - tool USE (i.e. the LLM can generate JSON to use this tool), - tool HANDLING (i.e. the agent can handle JSON from this tool),
Parameters:
Name | Type | Description | Default |
---|---|---|---|
message_class |
Optional[Type[ToolMessage]]
|
The ToolMessage class to enable, for USE, or HANDLING, or both. Optional; if None, then apply the enabling to all tools in the agent's toolset that have been enabled so far. |
required |
use |
bool
|
IF True, allow the agent (LLM) to use this tool (or all tools), else disallow |
True
|
handle |
bool
|
if True, allow the agent (LLM) to handle (i.e. respond to) this tool (or all tools) |
True
|
force |
bool
|
whether to FORCE the agent (LLM) to USE the specific
tool represented by |
False
|
require_recipient |
bool
|
whether to require that recipient be specified
when using the tool message (only applies if |
False
|
require_defaults |
whether to include fields that have default values, in the "properties" section of the JSON format instructions. (Normally the OpenAI completion API ignores these fields, but the Assistant fn-calling seems to pay attn to these, and if we don't want this, we should set this to False.) |
required |
Source code in langroid/agent/chat_agent.py
disable_message_handling(message_class=None)
¶
Disable this agent from RESPONDING to a message_class
(Tool). If
message_class
is None, then disable this agent from responding to ALL.
Args:
message_class: The ToolMessage class to disable; Optional.
Source code in langroid/agent/chat_agent.py
disable_message_use(message_class)
¶
Disable this agent from USING a message class (Tool).
If message_class
is None, then disable this agent from USING ALL tools.
Args:
message_class: The ToolMessage class to disable.
If None, disable all.
Source code in langroid/agent/chat_agent.py
disable_message_use_except(message_class)
¶
Disable this agent from USING ALL messages EXCEPT a message class (Tool) Args: message_class: The only ToolMessage class to allow
Source code in langroid/agent/chat_agent.py
llm_response(message=None)
¶
Respond to a single user message, appended to the message history, in "chat" mode Args: message (str|ChatDocument): message or ChatDocument object to respond to. If None, use the self.task_messages Returns: LLM response as a ChatDocument object
Source code in langroid/agent/chat_agent.py
llm_response_async(message=None)
async
¶
Async version of llm_response
. See there for details.
Source code in langroid/agent/chat_agent.py
init_message_history()
¶
Initialize the message history with the system message and user message
Source code in langroid/agent/chat_agent.py
llm_response_messages(messages, output_len=None)
¶
Respond to a series of messages, e.g. with OpenAI ChatCompletion Args: messages: seq of messages (with role, content fields) sent to LLM output_len: max number of tokens expected in response. If None, use the LLM's default max_output_tokens. Returns: Document (i.e. with fields "content", "metadata")
Source code in langroid/agent/chat_agent.py
llm_response_messages_async(messages, output_len=None)
async
¶
Async version of llm_response_messages
. See there for details.
Source code in langroid/agent/chat_agent.py
llm_response_forget(message)
¶
LLM Response to single message, and restore message_history. In effect a "one-off" message & response that leaves agent message history state intact.
Parameters:
Name | Type | Description | Default |
---|---|---|---|
message |
str
|
user message |
required |
Returns:
Type | Description |
---|---|
ChatDocument
|
A Document object with the response. |
Source code in langroid/agent/chat_agent.py
llm_response_forget_async(message)
async
¶
Async version of llm_response_forget
. See there for details.
Source code in langroid/agent/chat_agent.py
chat_num_tokens(messages=None)
¶
Total number of tokens in the message history so far.
Parameters:
Name | Type | Description | Default |
---|---|---|---|
messages |
Optional[List[LLMMessage]]
|
if provided, compute the number of tokens in this list of messages, rather than the current message history. |
None
|
Returns: int: number of tokens in message history
Source code in langroid/agent/chat_agent.py
message_history_str(i=None)
¶
Return a string representation of the message history Args: i: if provided, return only the i-th message when i is postive, or last k messages when i = -k. Returns:
Source code in langroid/agent/chat_agent.py
ToolMessage
¶
Bases: ABC
, BaseModel
Abstract Class for a class that defines the structure of a "Tool" message from an LLM. Depending on context, "tools" are also referred to as "plugins", or "function calls" (in the context of OpenAI LLMs). Essentially, they are a way for the LLM to express its intent to run a special function or method. Currently these "tools" are handled by methods of the agent.
Attributes:
Name | Type | Description |
---|---|---|
request |
str
|
name of agent method to map to. |
purpose |
str
|
purpose of agent method, expressed in general terms. (This is used when auto-generating the tool instruction to the LLM) |
examples()
classmethod
¶
Examples to use in few-shot demos with JSON formatting instructions. Each example can be either: - just a ToolMessage instance, e.g. MyTool(param1=1, param2="hello"), or - a tuple (description, ToolMessage instance), where the description is a natural language "thought" that leads to the tool usage, e.g. ("I want to find the square of 5", SquareTool(num=5)) In some scenarios, including such a description can significantly enhance reliability of tool use. Returns:
Source code in langroid/agent/tool_message.py
usage_examples(random=False)
classmethod
¶
Instruction to the LLM showing examples of how to use the tool-message.
Parameters:
Name | Type | Description | Default |
---|---|---|---|
random |
bool
|
whether to pick a random example from the list of examples.
Set to |
False
|
Returns: str: examples of how to use the tool/function-call
Source code in langroid/agent/tool_message.py
default_value(f)
classmethod
¶
Returns the default value of the given field, for the message-class Args: f (str): field name
Returns:
Name | Type | Description |
---|---|---|
Any |
Any
|
default value of the field, or None if not set or if the field does not exist. |
Source code in langroid/agent/tool_message.py
json_instructions(tool=False)
classmethod
¶
Default Instructions to the LLM showing how to use the tool/function-call. Works for GPT4 but override this for weaker LLMs if needed.
Parameters:
Name | Type | Description | Default |
---|---|---|---|
tool |
bool
|
instructions for Langroid-native tool use? (e.g. for non-OpenAI LLM) (or else it would be for OpenAI Function calls) |
False
|
Returns: str: instructions on how to use the message
Source code in langroid/agent/tool_message.py
json_group_instructions()
staticmethod
¶
Template for instructions for a group of tools. Works with GPT4 but override this for weaker LLMs if needed.
Source code in langroid/agent/tool_message.py
llm_function_schema(request=False, defaults=True)
classmethod
¶
Clean up the schema of the Pydantic class (which can recursively contain other Pydantic classes), to create a version compatible with OpenAI Function-call API.
Adapted from this excellent library: https://github.com/jxnl/instructor/blob/main/instructor/function_calls.py
Parameters:
Name | Type | Description | Default |
---|---|---|---|
request |
bool
|
whether to include the "request" field in the schema. (we set this to True when using Langroid-native TOOLs as opposed to OpenAI Function calls) |
False
|
defaults |
bool
|
whether to include fields with default values in the schema, in the "properties" section. |
True
|
Returns:
Name | Type | Description |
---|---|---|
LLMFunctionSpec |
LLMFunctionSpec
|
the schema as an LLMFunctionSpec |
Source code in langroid/agent/tool_message.py
simple_schema()
classmethod
¶
Return a simplified schema for the message, with only the request and required fields. Returns: Dict[str, Any]: simplified schema
Source code in langroid/agent/tool_message.py
Task(agent=None, name='', llm_delegate=False, single_round=False, system_message='', user_message='', restart=True, default_human_response=None, interactive=True, only_user_quits_root=True, erase_substeps=False, allow_null_result=False, max_stalled_steps=5, done_if_no_response=[], done_if_response=[], config=TaskConfig(), **kwargs)
¶
A Task
wraps an Agent
object, and sets up the Agent
's goals and instructions.
A Task
maintains two key variables:
self.pending_message
, which is the message awaiting a response, andself.pending_sender
, which is the entity that sent the pending message.
The possible responders to self.pending_message
are the Agent
's own "native"
responders (agent_response
, llm_response
, and user_response
), and
the run()
methods of any sub-tasks. All responders have the same type-signature
(somewhat simplified):
The main top-level method in the Task
class is run()
, which repeatedly calls
step()
until done()
returns true. The step()
represents a "turn" in the
conversation: this method sequentially (in round-robin fashion) calls the responders
until it finds one that generates a valid response to the pending_message
(as determined by the valid()
method). Once a valid response is found,
step()
updates the pending_message
and pending_sender
variables,
and on the next iteration, step()
re-starts its search for a valid response
from the beginning of the list of responders (the exception being that the
human user always gets a chance to respond after each non-human valid response).
This process repeats until done()
returns true, at which point run()
returns
the value of result()
, which is the final result of the task.
Parameters:
Name | Type | Description | Default |
---|---|---|---|
agent |
Agent
|
agent associated with the task |
None
|
name |
str
|
name of the task |
''
|
llm_delegate |
bool
|
Whether to delegate "control" to LLM; conceptually,
the "controlling entity" is the one "seeking" responses to its queries,
and has a goal it is aiming to achieve, and decides when a task is done.
The "controlling entity" is either the LLM or the USER.
(Note within a Task there is just one
LLM, and all other entities are proxies of the "User" entity).
See also: |
False
|
single_round |
bool
|
If true, task runs until one message by "controller"
(i.e. LLM if |
False
|
system_message |
str
|
if not empty, overrides agent's system_message |
''
|
user_message |
str
|
if not empty, overrides agent's user_message |
''
|
restart |
bool
|
if true, resets the agent's message history at every run. |
True
|
default_human_response |
str | None
|
default response from user; useful for
testing, to avoid interactive input from user.
[Instead of this, setting |
None
|
interactive |
bool
|
if true, wait for human input after each non-human
response (prevents infinite loop of non-human responses).
Default is true. If false, then |
True
|
only_user_quits_root |
bool
|
if true, when interactive=True, only user can quit the root task (Ignored when interactive=False). |
True
|
erase_substeps |
bool
|
if true, when task completes, erase intermediate
conversation with subtasks from this agent's |
False
|
allow_null_result |
bool
|
If true, create dummy NO_ANSWER response when no valid response is found
in a step.
Optional, default is False.
Note: In non-interactive mode, when this is set to True,
you can have a situation where an LLM generates (non-tool) text,
and no other responders have valid responses, and a "Null result"
is inserted as a dummy response from the User entity, so the LLM
will now respond to this Null result, and this will continue
until the LLM emits a DONE signal (if instructed to do so),
otherwise langroid detects a potential infinite loop after
a certain number of such steps (= |
False
|
max_stalled_steps |
int
|
task considered done after this many consecutive steps with no progress. Default is 3. |
5
|
done_if_no_response |
List[Responder]
|
consider task done if NULL response from any of these responders. Default is empty list. |
[]
|
done_if_response |
List[Responder]
|
consider task done if NON-NULL response from any of these responders. Default is empty list. |
[]
|
Source code in langroid/agent/task.py
136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 |
|
clone(i)
¶
Returns a copy of this task, with a new agent.
Source code in langroid/agent/task.py
kill_session(session_id='')
classmethod
¶
Kill the session with the given session_id.
kill()
¶
add_sub_task(task)
¶
Add a sub-task (or list of subtasks) that this task can delegate (or fail-over) to. Note that the sequence of sub-tasks is important, since these are tried in order, as the parent task searches for a valid response (unless a sub-task is explicitly addressed).
Parameters:
Name | Type | Description | Default |
---|---|---|---|
task |
Task | List[Task] | Tuple[Task, TaskConfig] | List[Tuple[Task, TaskConfig]]
|
A task, or list of tasks, or a tuple of task and task config, or a list of tuples of task and task config. These tasks are added as sub-tasks of the current task. The task configs (if any) dictate how the tasks are run when invoked as sub-tasks of other tasks. This allows users to specify behavior applicable only in the context of a particular task-subtask combination. |
required |
Source code in langroid/agent/task.py
init(msg=None)
¶
Initialize the task, with an optional message to start the conversation.
Initializes self.pending_message
and self.pending_sender
.
Args:
msg (str|ChatDocument): optional message to start the conversation.
Returns:
Type | Description |
---|---|
ChatDocument | None
|
the initialized |
ChatDocument | None
|
Currently not used in the code, but provided for convenience. |
Source code in langroid/agent/task.py
reset_all_sub_tasks()
¶
Recursively reset message history of own agent and all sub-tasks
run(msg=None, turns=-1, caller=None, max_cost=0, max_tokens=0, session_id='')
¶
Synchronous version of run_async()
.
See run_async()
for details.
Source code in langroid/agent/task.py
579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 |
|
run_async(msg=None, turns=-1, caller=None, max_cost=0, max_tokens=0, session_id='')
async
¶
Loop over step()
until task is considered done or turns
is reached.
Runs asynchronously.
Parameters:
Name | Type | Description | Default |
---|---|---|---|
msg |
str | ChatDocument
|
initial user-role message to process; if None,
the LLM will respond to its initial |
None
|
turns |
int
|
number of turns to run the task for; default is -1, which means run until task is done. |
-1
|
caller |
Task | None
|
the calling task, if any |
None
|
max_cost |
float
|
max cost allowed for the task (default 0 -> no limit) |
0
|
max_tokens |
int
|
max tokens allowed for the task (default 0 -> no limit) |
0
|
session_id |
str
|
session id for the task |
''
|
Returns:
Type | Description |
---|---|
Optional[ChatDocument]
|
Optional[ChatDocument]: valid result of the task. |
Source code in langroid/agent/task.py
674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 |
|
step(turns=-1)
¶
Synchronous version of step_async()
. See step_async()
for details.
TODO: Except for the self.response() calls, this fn should be identical to
step_async()
. Consider refactoring to avoid duplication.
Source code in langroid/agent/task.py
873 874 875 876 877 878 879 880 881 882 883 884 885 886 887 888 889 890 891 892 893 894 895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924 925 926 927 928 929 930 931 932 933 934 935 936 937 938 939 940 941 942 943 944 945 946 947 948 949 950 951 952 953 954 955 956 957 958 959 960 961 962 963 |
|
step_async(turns=-1)
async
¶
A single "turn" in the task conversation: The "allowed" responders in this
turn (which can be either the 3 "entities", or one of the sub-tasks) are
tried in sequence, until a valid response is obtained; a valid
response is one that contributes to the task, either by ending it,
or producing a response to be further acted on.
Update self.pending_message
to the latest valid response (or NO_ANSWER
if no valid response was obtained from any responder).
Parameters:
Name | Type | Description | Default |
---|---|---|---|
turns |
int
|
number of turns to process. Typically used in testing where there is no human to "quit out" of current level, or in cases where we want to limit the number of turns of a delegated agent. |
-1
|
Returns (ChatDocument|None):
Updated self.pending_message
. Currently the return value is not used
by the task.run()
method, but we return this as a convenience for
other use-cases, e.g. where we want to run a task step by step in a
different context.
Source code in langroid/agent/task.py
965 966 967 968 969 970 971 972 973 974 975 976 977 978 979 980 981 982 983 984 985 986 987 988 989 990 991 992 993 994 995 996 997 998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055 1056 1057 1058 1059 1060 1061 1062 1063 1064 1065 1066 1067 |
|
response(e, turns=-1)
¶
Sync version of response_async()
. See response_async()
for details.
Source code in langroid/agent/task.py
response_async(e, turns=-1)
async
¶
Get response to self.pending_message
from a responder.
If response is valid (i.e. it ends the current turn of seeking
responses):
-then return the response as a ChatDocument object,
-otherwise return None.
Args:
e (Responder): responder to get response from.
turns (int): number of turns to run the task for.
Default is -1, which means run until task is done.
Returns:
Type | Description |
---|---|
Optional[ChatDocument]
|
Optional[ChatDocument]: response to |
Optional[ChatDocument]
|
valid, None otherwise |
Source code in langroid/agent/task.py
result(status=None)
¶
Get result of task. This is the default behavior. Derived classes can override this.
Note the result of a task is returned as if it is from the User entity.
Parameters:
Name | Type | Description | Default |
---|---|---|---|
status |
StatusCode
|
status of the task when it ended |
None
|
Returns: ChatDocument: result of task
Source code in langroid/agent/task.py
done(result=None, r=None)
¶
Check if task is done. This is the default behavior. Derived classes can override this. Args: result (ChatDocument|None): result from a responder r (Responder|None): responder that produced the result Not used here, but could be used by derived classes. Returns: bool: True if task is done, False otherwise StatusCode: status code indicating why task is done
Source code in langroid/agent/task.py
1455 1456 1457 1458 1459 1460 1461 1462 1463 1464 1465 1466 1467 1468 1469 1470 1471 1472 1473 1474 1475 1476 1477 1478 1479 1480 1481 1482 1483 1484 1485 1486 1487 1488 1489 1490 1491 1492 1493 1494 1495 1496 1497 1498 1499 1500 1501 1502 1503 1504 1505 1506 1507 1508 1509 1510 1511 1512 1513 1514 1515 1516 1517 1518 1519 1520 1521 1522 |
|
valid(result, r)
¶
Is the result from a Responder (i.e. an entity or sub-task) such that we can stop searching for responses in this step?
Source code in langroid/agent/task.py
log_message(resp, msg=None, mark=False)
¶
Log current pending message, and related state, for lineage/debugging purposes.
Parameters:
Name | Type | Description | Default |
---|---|---|---|
resp |
Responder
|
Responder that generated the |
required |
msg |
ChatDocument
|
Message to log. Defaults to None. |
None
|
mark |
bool
|
Whether to mark the message as the final result of
a |
False
|
Source code in langroid/agent/task.py
set_color_log(enable=True)
¶
Flag to enable/disable color logging using rich.console.
In some contexts, such as Colab notebooks, we may want to disable color logging
using rich.console, since those logs show up in the cell output rather than
in the log file. Turning off this feature will still create logs, but without
the color formatting from rich.console
Args:
enable (bool): value of self.color_log
to set to,
which will enable/diable rich logging