diff options
Diffstat (limited to 'autogpts/autogpt/autogpt/agent_factory/profile_generator.py')
| -rw-r--r-- | autogpts/autogpt/autogpt/agent_factory/profile_generator.py | 255 |
1 files changed, 255 insertions, 0 deletions
diff --git a/autogpts/autogpt/autogpt/agent_factory/profile_generator.py b/autogpts/autogpt/autogpt/agent_factory/profile_generator.py new file mode 100644 index 000000000..8d9ff399e --- /dev/null +++ b/autogpts/autogpt/autogpt/agent_factory/profile_generator.py @@ -0,0 +1,255 @@ +import json +import logging + +from autogpt.config import AIDirectives, AIProfile, Config +from autogpt.core.configuration import SystemConfiguration, UserConfigurable +from autogpt.core.prompting import ( + ChatPrompt, + LanguageModelClassification, + PromptStrategy, +) +from autogpt.core.prompting.utils import json_loads +from autogpt.core.resource.model_providers.schema import ( + AssistantChatMessage, + ChatMessage, + ChatModelProvider, + CompletionModelFunction, +) +from autogpt.core.utils.json_schema import JSONSchema + +logger = logging.getLogger(__name__) + + +class AgentProfileGeneratorConfiguration(SystemConfiguration): + model_classification: LanguageModelClassification = UserConfigurable( + default=LanguageModelClassification.SMART_MODEL + ) + _example_call: object = [ + { + "type": "function", + "function": { + "name": "create_agent", + "arguments": { + "name": "CMOGPT", + "description": ( + "a professional digital marketer AI that assists Solopreneurs " + "in growing their businesses by providing " + "world-class expertise in solving marketing problems " + "for SaaS, content products, agencies, and more." + ), + "directives": { + "best_practices": [ + ( + "Engage in effective problem-solving, prioritization, " + "planning, and supporting execution to address your " + "marketing needs as your virtual " + "Chief Marketing Officer." + ), + ( + "Provide specific, actionable, and concise advice to " + "help you make informed decisions without the use of " + "platitudes or overly wordy explanations." + ), + ( + "Identify and prioritize quick wins and cost-effective " + "campaigns that maximize results with minimal time and " + "budget investment." + ), + ( + "Proactively take the lead in guiding you and offering " + "suggestions when faced with unclear information or " + "uncertainty to ensure your marketing strategy remains " + "on track." + ), + ], + "constraints": [ + "Do not suggest illegal or unethical plans or strategies.", + "Take reasonable budgetary limits into account.", + ], + }, + }, + }, + } + ] + system_prompt: str = UserConfigurable( + default=( + "Your job is to respond to a user-defined task, given in triple quotes, by " + "invoking the `create_agent` function to generate an autonomous agent to " + "complete the task. " + "You should supply a role-based name for the agent (_GPT), " + "an informative description for what the agent does, and 1 to 5 directives " + "in each of the categories Best Practices and Constraints, " + "that are optimally aligned with the successful completion " + "of its assigned task.\n" + "\n" + "Example Input:\n" + '"""Help me with marketing my business"""\n\n' + "Example Call:\n" + "```\n" + f"{json.dumps(_example_call, indent=4)}" + "\n```" + ) + ) + user_prompt_template: str = UserConfigurable(default='"""{user_objective}"""') + create_agent_function: dict = UserConfigurable( + default=CompletionModelFunction( + name="create_agent", + description="Create a new autonomous AI agent to complete a given task.", + parameters={ + "name": JSONSchema( + type=JSONSchema.Type.STRING, + description="A short role-based name for an autonomous agent.", + required=True, + ), + "description": JSONSchema( + type=JSONSchema.Type.STRING, + description=( + "An informative one sentence description " + "of what the AI agent does" + ), + required=True, + ), + "directives": JSONSchema( + type=JSONSchema.Type.OBJECT, + properties={ + "best_practices": JSONSchema( + type=JSONSchema.Type.ARRAY, + minItems=1, + maxItems=5, + items=JSONSchema( + type=JSONSchema.Type.STRING, + ), + description=( + "One to five highly effective best practices " + "that are optimally aligned with the completion " + "of the given task" + ), + required=True, + ), + "constraints": JSONSchema( + type=JSONSchema.Type.ARRAY, + minItems=1, + maxItems=5, + items=JSONSchema( + type=JSONSchema.Type.STRING, + ), + description=( + "One to five reasonable and efficacious constraints " + "that are optimally aligned with the completion " + "of the given task" + ), + required=True, + ), + }, + required=True, + ), + }, + ).schema + ) + + +class AgentProfileGenerator(PromptStrategy): + default_configuration: AgentProfileGeneratorConfiguration = ( + AgentProfileGeneratorConfiguration() + ) + + def __init__( + self, + model_classification: LanguageModelClassification, + system_prompt: str, + user_prompt_template: str, + create_agent_function: dict, + ): + self._model_classification = model_classification + self._system_prompt_message = system_prompt + self._user_prompt_template = user_prompt_template + self._create_agent_function = CompletionModelFunction.parse( + create_agent_function + ) + + @property + def model_classification(self) -> LanguageModelClassification: + return self._model_classification + + def build_prompt(self, user_objective: str = "", **kwargs) -> ChatPrompt: + system_message = ChatMessage.system(self._system_prompt_message) + user_message = ChatMessage.user( + self._user_prompt_template.format( + user_objective=user_objective, + ) + ) + prompt = ChatPrompt( + messages=[system_message, user_message], + functions=[self._create_agent_function], + ) + return prompt + + def parse_response_content( + self, + response_content: AssistantChatMessage, + ) -> tuple[AIProfile, AIDirectives]: + """Parse the actual text response from the objective model. + + Args: + response_content: The raw response content from the objective model. + + Returns: + The parsed response. + + """ + try: + if not response_content.tool_calls: + raise ValueError( + f"LLM did not call {self._create_agent_function.name} function; " + "agent profile creation failed" + ) + arguments: object = json_loads( + response_content.tool_calls[0].function.arguments + ) + ai_profile = AIProfile( + ai_name=arguments.get("name"), + ai_role=arguments.get("description"), + ) + ai_directives = AIDirectives( + best_practices=arguments.get("directives", {}).get("best_practices"), + constraints=arguments.get("directives", {}).get("constraints"), + resources=[], + ) + except KeyError: + logger.debug(f"Failed to parse this response content: {response_content}") + raise + return ai_profile, ai_directives + + +async def generate_agent_profile_for_task( + task: str, + app_config: Config, + llm_provider: ChatModelProvider, +) -> tuple[AIProfile, AIDirectives]: + """Generates an AIConfig object from the given string. + + Returns: + AIConfig: The AIConfig object tailored to the user's input + """ + agent_profile_generator = AgentProfileGenerator( + **AgentProfileGenerator.default_configuration.dict() # HACK + ) + + prompt = agent_profile_generator.build_prompt(task) + + # Call LLM with the string as user input + output = ( + await llm_provider.create_chat_completion( + prompt.messages, + model_name=app_config.smart_llm, + functions=prompt.functions, + ) + ).response + + # Debug LLM Output + logger.debug(f"AI Config Generator Raw Output: {output}") + + # Parse the output + ai_profile, ai_directives = agent_profile_generator.parse_response_content(output) + + return ai_profile, ai_directives |