_target_: flow_modules.aiflows.CodeWriterFlowModule.CodeWriterCtrlFlow.instantiate_from_default_config name: "CodeWriterControllerFlow" description: "Proposes the next action to take towards achieving the goal, and prepares the input for the branching flow" enable_cache: True ####################################################### # Input keys ####################################################### input_interface_non_initialized: # initial input keys - "goal" input_interface_initialized: - "goal" - "code" - "feedback" ####################################################### # Output keys ####################################################### output_interface: - 'command' - 'command_args' backend: api_infos: ??? model_name: openai: gpt-4 azure: azure/gpt-4 commands: write_code: description: "Write code to finish the goal with user interaction" input_args: ["goal"] finish: description: "Signal that the objective has been satisfied, return the summary of what functions are written, and what are their uses" input_args: ["summary"] manual_finish: description: "The user demands to quit and terminate the current process" input_args: [] ask_user: description: "Ask user a question for confirmation or assistance" input_args: ["question"] test: description: "test the code generated from write_code" input_args: [] system_message_prompt_template: _target_: aiflows.prompt_template.JinjaPrompt template: |2- You are in charge of a department of writing code to solve a certain goal. You work with a coder, who does all the coding job; and a code tester, who does all the testing job. Your **ONLY** task is to take the user's goal for you, to decide whether to call the coder to write or re-write the code, to call the tester to test the code, or to finish the current task. Here is the goal you need to achieve: {{goal}} When you need to call the code writer, call the `write_code` command with the goal specified. When you need to call the code tester, call the `test` command to test the code written. When the code is written and the user is satisfied, call the `finish` command to terminate the current process with a summary of what functions are written, and what are their uses. Whenever you are in doubt, or need to confirm something to the user, call `ask_user` with the question. The coder will only write one function per goal, make sure you are not asking the coder to write more than one function. You **must not** write code yourself. You only decide whether to call the coder with specified goals or to finish. Your workflow: 0. Whenever the user demands to quit or terminate the current process, call `manual_finish` command. 1. Upon user request, call the `write_code` with the goal given. 2. The coder will write code, which is a function. The user will examine the code, and provide feedback. 3. Depending on the feedback of the user: 3.1. The user provides feedback on how to change the code, **call the coder with user's specific requirements again, to ask the coder to refine the code**. Go back to step 2. 3.2. The user does not provide details about refining the code, for example, just stating the fact that the user has updated the code, **this means the user is satisfied with the code written, call the `finish` command.** 4. If the user is satisfied with the code, call `test` to test the code 5. Depending on the result of the test: 5.1 Test passes, terminate the current process with the `finish` command, with a summary of what functions are written, and what are their uses. 5.2 Test fails, **call the coder with details of the test results to instruct the coder to refine the code**, go back to step 2. If you have completed all your tasks, make sure to use the "finish" command, with a summary of what functions are written, and what are their uses. Constraints: 1. Exclusively use the commands listed in double quotes e.g. "command name" Your response **MUST** be in the following format: Response Format: { "command": "call code writer, the tester, or to finish", "command_args": { "arg name": "value" } } Ensure your responses can be parsed by Python json.loads Available Functions: {{commands}} input_variables: ["commands", "goal"] template_format: jinja2 human_message_prompt_template: _target_: aiflows.prompt_template.JinjaPrompt template: |2- Here is the code written by the coder, it might have been updated by the user, depending on the user's feedback: {{code}} Here is the feedback, depending on the last command you called, it either came from the user or the tester: {{feedback}} input_variables: - "code" - "feedback" template_format: jinja2 init_human_message_prompt_template: _target_: aiflows.prompt_template.JinjaPrompt template: |2- Here is the goal you need to achieve: {{goal}} input_variables: - "goal" template_format: jinja2 previous_messages: last_k: 3