import inspect import re from pathlib import Path from typing import List, MutableMapping, Union from langchain.schema import HumanMessage, SystemMessage from termcolor import colored from gpt_engineer.core.ai import AI from gpt_engineer.core.base_execution_env import BaseExecutionEnv from gpt_engineer.core.base_memory import BaseMemory from gpt_engineer.core.chat_to_files import ( chat_to_files_dict, overwrite_code_with_edits, parse_edits, ) from gpt_engineer.core.default.constants import MAX_EDIT_REFINEMENT_STEPS from gpt_engineer.core.default.paths import ( CODE_GEN_LOG_FILE, ENTRYPOINT_FILE, ENTRYPOINT_LOG_FILE, IMPROVE_LOG_FILE, ) from gpt_engineer.core.files_dict import FilesDict from gpt_engineer.core.preprompts_holder import PrepromptsHolder def curr_fn() -> str: return inspect.stack()[1].function def setup_sys_prompt(preprompts: MutableMapping[Union[str, Path], str]) -> str: return ( preprompts["roadmap"] + preprompts["generate"].replace("FILE_FORMAT", preprompts["file_format"]) + "\nUseful to know:\n" + preprompts["philosophy"] ) def gen_code( ai: AI, prompt: str, memory: BaseMemory, preprompts_holder: PrepromptsHolder ) -> FilesDict: preprompts = preprompts_holder.get_preprompts() messages = ai.start(setup_sys_prompt(preprompts), prompt, step_name=curr_fn()) chat = messages[-1].content.strip() memory[CODE_GEN_LOG_FILE] = chat files_dict = chat_to_files_dict(chat) return files_dict def gen_entrypoint( ai: AI, files_dict: FilesDict, memory: BaseMemory, preprompts_holder: PrepromptsHolder, ) -> FilesDict: preprompts = preprompts_holder.get_preprompts() messages = ai.start( system=(preprompts["entrypoint"]), user="Information about the codebase:\n\n" + files_dict.to_chat(), step_name=curr_fn(), ) print() chat = messages[-1].content.strip() regex = r"```\S*\n(.+?)```" matches = re.finditer(regex, chat, re.DOTALL) entrypoint_code = FilesDict( {ENTRYPOINT_FILE: "\n".join(match.group(1) for match in matches)} ) memory[ENTRYPOINT_LOG_FILE] = chat return entrypoint_code def execute_entrypoint( ai: AI, execution_env: BaseExecutionEnv, files_dict: FilesDict, preprompts_holder: PrepromptsHolder = None, ) -> FilesDict: if ENTRYPOINT_FILE not in files_dict: raise FileNotFoundError( "The required entrypoint " + ENTRYPOINT_FILE + " does not exist in the code." ) command = files_dict[ENTRYPOINT_FILE] print() print( colored( "Do you want to execute this code? (Y/n)", "red", ) ) print() print(command) print() if input("").lower() not in ["", "y", "yes"]: print("Ok, not executing the code.") return files_dict print("Executing the code...") print() print( colored( "Note: If it does not work as expected, consider running the code" + " in another way than above.", "green", ) ) print() print("You can press ctrl+c *once* to stop the execution.") print() execution_env.upload(files_dict).run(f"bash {ENTRYPOINT_FILE}") return files_dict def setup_sys_prompt_existing_code( preprompts: MutableMapping[Union[str, Path], str] ) -> str: return ( preprompts["improve"].replace("FILE_FORMAT", preprompts["file_format"]) + "\nUseful to know:\n" + preprompts["philosophy"] ) def incorrect_edit(files_dict: FilesDict, chat: str) -> List[str,]: problems = [] try: edits = parse_edits(chat) except ValueError as problem: print("Not possible to parse chat to edits") problems.append(str(problem)) return problems for edit in edits: # only trigger for existing files if edit.filename in files_dict: if edit.before not in files_dict[edit.filename]: problems.append( "This section, assigned to be exchanged for an edit block, does not have an exact match in the code: " + edit.before + "\nThis is often a result of placeholders, such as ... or references to 'existing code' or 'rest of function' etc, which cannot be used the HEAD part of the edit blocks. Also, to get a match, all comments, including long doc strings may have to be reproduced in the patch HEAD" ) return problems def improve( ai: AI, prompt: str, files_dict: FilesDict, memory: BaseMemory, preprompts_holder: PrepromptsHolder, ) -> FilesDict: preprompts = preprompts_holder.get_preprompts() messages = [ SystemMessage(content=setup_sys_prompt_existing_code(preprompts)), ] # Add files as input messages.append(HumanMessage(content=f"{files_dict.to_chat()}")) messages.append(HumanMessage(content=f"Request: {prompt}")) problems = [""] # check edit correctness edit_refinements = 0 while len(problems) > 0 and edit_refinements <= MAX_EDIT_REFINEMENT_STEPS: messages = ai.next(messages, step_name=curr_fn()) chat = messages[-1].content.strip() problems = incorrect_edit(files_dict, chat) if len(problems) > 0: messages.append( HumanMessage( content="Some previously produced edits were not on the requested format, or the HEAD part was not found in the code. Details: " + "\n".join(problems) + "\n Please provide ALL the edits again, making sure that the failing ones are now on the correct format and can be found in the code. Make sure to not repeat past mistakes. \n" ) ) edit_refinements += 1 overwrite_code_with_edits(chat, files_dict) memory[IMPROVE_LOG_FILE] = chat return files_dict