diff --git a/.pre-commit-config.yaml b/.pre-commit-config.yaml index 95b449726..9222c3e25 100644 --- a/.pre-commit-config.yaml +++ b/.pre-commit-config.yaml @@ -1,5 +1,5 @@ repos: - - repo: https://github.com/pycqa/isort + - repo: https://github.com/PyCQA/isort rev: 5.12.0 hooks: - id: isort diff --git a/HISTORY.md b/HISTORY.md index 1284773f3..70f86560a 100644 --- a/HISTORY.md +++ b/HISTORY.md @@ -1,9 +1,24 @@ ---- -nav_order: 999 ---- # Release history +### v0.39.0 + +- Use `--sonnet` for Claude 3.5 Sonnet, which is the top model on [aider's LLM code editing leaderboard](https://aider.chat/docs/leaderboards/#claude-35-sonnet-takes-the-top-spot). +- All `AIDER_xxx` environment variables can now be set in `.env` (by @jpshack-at-palomar). +- Use `--llm-history-file` to log raw messages sent to the LLM (by @daniel-vainsencher). +- Commit messages are no longer prefixed with "aider:". Instead the git author and committer names have "(aider)" added. + +### v0.38.0 + +- Use `--vim` for [vim keybindings](https://aider.chat/docs/commands.html#vi) in the chat. +- [Add LLM metadata](https://aider.chat/docs/llms/warnings.html#specifying-context-window-size-and-token-costs) via `.aider.models.json` file (by @caseymcc). +- More detailed [error messages on token limit errors](https://aider.chat/docs/troubleshooting/token-limits.html). +- Single line commit messages, without the recent chat messages. +- Ensure `--commit --dry-run` does nothing. +- Have playwright wait for idle network to better scrape js sites. +- Documentation updates, moved into website/ subdir. +- Moved tests/ into aider/tests/. + ### v0.37.0 - Repo map is now optimized based on text of chat history as well as files added to chat. diff --git a/README.md b/README.md index 9f932f331..3bdbc20a4 100644 --- a/README.md +++ b/README.md @@ -1,11 +1,13 @@ + + # Aider is AI pair programming in your terminal Aider lets you pair program with LLMs, to edit code in your local git repository. Start a new project or work with an existing git repo. -Aider works best with GPT-4o and Claude 3 Opus -and can [connect to almost any LLM](https://aider.chat/docs/llms.html). +Aider can [connect to almost any LLM](https://aider.chat/docs/llms.html). +and works best with GPT-4o, Claude 3.5 Sonnet, Claude 3 Opus and DeepSeek Coder V2.

## Getting started + +You can get started quickly like this: -{% include get-started.md %} +``` +$ pip install aider-chat -**See the +# Change directory into a git repo +$ cd /to/your/git/repo + +# Work with GPT-4o on your repo +$ export OPENAI_API_KEY=your-key-goes-here +$ aider + +# Or, work with Anthropic's models +$ export ANTHROPIC_API_KEY=your-key-goes-here + +# Claude 3 Opus +$ aider --opus + +# Claude 3.5 Sonnet +$ aider --sonnet +``` + + +See the [installation instructions](https://aider.chat/docs/install.html) and other [documentation](https://aider.chat/docs/usage.html) -for more details.** +for more details. ## Features -- Chat with aider about your code: `aider ...` +- Run aider with the files you want to edit: `aider ...` - Ask for changes: - - New features, test cases, improvements. - - Bug fixes, updated docs or code refactors. - - Paste in a GitHub issue that needs to be solved. + - Add new features or test cases. + - Describe a bug. + - Paste in an error message or or GitHub issue URL. + - Refactor code. + - Update docs. - Aider will edit your files to complete your request. - Aider [automatically git commits](https://aider.chat/docs/git.html) changes with a sensible commit message. - Aider works with [most popular languages](https://aider.chat/docs/languages.html): python, javascript, typescript, php, html, css, and more... - Aider works best with GPT-4o and Claude 3 Opus and can [connect to almost any LLM](https://aider.chat/docs/llms.html). -- Aider can make coordinated changes across multiple files at once. +- Aider can edit multiple files at once for complex requests. - Aider uses a [map of your entire git repo](https://aider.chat/docs/repomap.html), which helps it work well in larger codebases. -- You can also edit files in your editor while chatting with aider. -Aider will notice and always use the latest version. -So you can bounce back and forth between aider and your editor, to collaboratively code with AI. -- Images can be added to the chat (GPT-4o, GPT-4 Turbo, etc). -- URLs can be added to the chat and aider will read their content. -- [Code with your voice](https://aider.chat/docs/voice.html) using speech recognition. +- Edit files in your editor while chatting with aider, +and it will always use the latest version. +Pair program with AI. +- Add images to the chat (GPT-4o, GPT-4 Turbo, etc). +- Add URLs to the chat and aider will read their content. +- [Code with your voice](https://aider.chat/docs/voice.html). -## State of the art +## Top tier performance -Aider has the -[top score on SWE Bench](https://aider.chat/2024/06/02/main-swe-bench.html). +[Aider has the one of the top scores on SWE Bench](https://aider.chat/2024/06/02/main-swe-bench.html). SWE Bench is a challenging software engineering benchmark where aider solved *real* GitHub issues from popular open source projects like django, scikitlearn, matplotlib, etc. -

- - aider swe bench - -

- -## Documentation +## More info +- [Documentation](https://aider.chat/) - [Installation](https://aider.chat/docs/install.html) - [Usage](https://aider.chat/docs/usage.html) - [Tutorial videos](https://aider.chat/docs/tutorials.html) - [Connecting to LLMs](https://aider.chat/docs/llms.html) +- [Configuration](https://aider.chat/docs/config.html) +- [Troubleshooting](https://aider.chat/docs/troubleshooting.html) - [LLM Leaderboards](https://aider.chat/docs/leaderboards/) -- [FAQ](https://aider.chat/docs/faq.html) +- [GitHub](https://github.com/paul-gauthier/aider) - [Discord](https://discord.gg/Tv2uQnR88V) - [Blog](https://aider.chat/blog/) diff --git a/aider/__init__.py b/aider/__init__.py index 316f3d226..57bacca7f 100644 --- a/aider/__init__.py +++ b/aider/__init__.py @@ -1 +1 @@ -__version__ = "0.37.1-dev" +__version__ = "0.39.1-dev" diff --git a/aider/__main__.py b/aider/__main__.py new file mode 100644 index 000000000..40e2b013f --- /dev/null +++ b/aider/__main__.py @@ -0,0 +1,4 @@ +from .main import main + +if __name__ == "__main__": + main() diff --git a/aider/args.py b/aider/args.py index b67756db2..47efeb7ed 100644 --- a/aider/args.py +++ b/aider/args.py @@ -7,30 +7,47 @@ import sys import configargparse from aider import __version__, models -from aider.args_formatter import MarkdownHelpFormatter, YamlHelpFormatter +from aider.args_formatter import ( + DotEnvFormatter, + MarkdownHelpFormatter, + YamlHelpFormatter, +) from .dump import dump # noqa: F401 +def default_env_file(git_root): + return os.path.join(git_root, ".env") if git_root else ".env" + + +def get_preparser(git_root): + parser = configargparse.ArgumentParser(add_help=False) + parser.add_argument( + "--env-file", + metavar="ENV_FILE", + default=default_env_file(git_root), + help="Specify the .env file to load (default: .env in git root)", + ) + return parser + + def get_parser(default_config_files, git_root): parser = configargparse.ArgumentParser( description="aider is GPT powered coding in your terminal", add_config_file_help=True, default_config_files=default_config_files, + config_file_parser_class=configargparse.YAMLConfigFileParser, auto_env_var_prefix="AIDER_", ) group = parser.add_argument_group("Main") group.add_argument( - "--vim", - action="store_true", - help="Use VI editing mode in the terminal (default: False)", - default=False, + "--llm-history-file", + metavar="LLM_HISTORY_FILE", + default=None, + help="Log the conversation with the LLM to this file (for example, .aider.llm.history)", ) group.add_argument( - "files", - metavar="FILE", - nargs="*", - help="files to edit with an LLM (optional)", + "files", metavar="FILE", nargs="*", help="files to edit with an LLM (optional)" ) group.add_argument( "--openai-api-key", @@ -59,7 +76,7 @@ def get_parser(default_config_files, git_root): const=opus_model, help=f"Use {opus_model} model for the main chat", ) - sonnet_model = "claude-3-sonnet-20240229" + sonnet_model = "claude-3-5-sonnet-20240620" group.add_argument( "--sonnet", action="store_const", @@ -146,13 +163,18 @@ def get_parser(default_config_files, git_root): metavar="MODEL_SETTINGS_FILE", default=None, help="Specify a file with aider model settings for unknown models", - ) group.add_argument( "--model-metadata-file", metavar="MODEL_METADATA_FILE", default=None, help="Specify a file with context window and costs for unknown models", ) + group.add_argument( + "--verify-ssl", + action=argparse.BooleanOptionalAction, + default=True, + help="Verify the SSL cert when connecting to models (default: True)", + ) group.add_argument( "--edit-format", metavar="EDIT_FORMAT", @@ -189,11 +211,12 @@ def get_parser(default_config_files, git_root): " max_chat_history_tokens." ), ) - default_env_file = os.path.join(git_root, ".env") if git_root else ".env" + # This is a duplicate of the argument in the preparser and is a no-op by this time of + # argument parsing, but it's here so that the help is displayed as expected. group.add_argument( "--env-file", metavar="ENV_FILE", - default=default_env_file, + default=default_env_file(git_root), help="Specify the .env file to load (default: .env in git root)", ) @@ -375,6 +398,12 @@ def get_parser(default_config_files, git_root): ########## group = parser.add_argument_group("Other Settings") + group.add_argument( + "--vim", + action="store_true", + help="Use VI editing mode in the terminal (default: False)", + default=False, + ) group.add_argument( "--voice-language", metavar="VOICE_LANGUAGE", @@ -500,11 +529,27 @@ def get_sample_yaml(): return parser.format_help() +def get_sample_dotenv(): + os.environ["COLUMNS"] = "120" + sys.argv = ["aider"] + parser = get_parser([], None) + + # This instantiates all the action.env_var values + parser.parse_known_args() + + parser.formatter_class = DotEnvFormatter + + return argparse.ArgumentParser.format_help(parser) + return parser.format_help() + + def main(): arg = sys.argv[1] if len(sys.argv[1:]) else None if arg == "md": print(get_md_help()) + elif arg == "dotenv": + print(get_sample_dotenv()) else: print(get_sample_yaml()) diff --git a/aider/args_formatter.py b/aider/args_formatter.py index 9ad00ff66..cc7e3e73d 100644 --- a/aider/args_formatter.py +++ b/aider/args_formatter.py @@ -1,8 +1,83 @@ import argparse +from aider import urls + from .dump import dump # noqa: F401 +class DotEnvFormatter(argparse.HelpFormatter): + def start_section(self, heading): + res = "\n\n" + res += "#" * (len(heading) + 3) + res += f"\n# {heading}" + super().start_section(res) + + def _format_usage(self, usage, actions, groups, prefix): + return "" + + def _format_text(self, text): + return f""" +########################################################## +# Sample aider .env file. +# Place at the root of your git repo. +# Or use `aider --env ` to specify. +########################################################## + +################# +# LLM parameters: +# +# Include xxx_API_KEY parameters and other params needed for your LLMs. +# See {urls.llms} for details. + +## OpenAI +#OPENAI_API_KEY= + +## Anthropic +#ANTHROPIC_API_KEY= + +##... +""" + + def _format_action(self, action): + if not action.option_strings: + return "" + + if not action.env_var: + return + + parts = [""] + + default = action.default + if default == argparse.SUPPRESS: + default = "" + elif isinstance(default, str): + pass + elif isinstance(default, list) and not default: + default = "" + elif action.default is not None: + default = "true" if default else "false" + else: + default = "" + + if action.help: + parts.append(f"## {action.help}") + + if action.env_var: + env_var = action.env_var + if default: + parts.append(f"#{env_var}={default}\n") + else: + parts.append(f"#{env_var}=\n") + + return "\n".join(parts) + "\n" + + def _format_action_invocation(self, action): + return "" + + def _format_args(self, action, default_metavar): + return "" + + class YamlHelpFormatter(argparse.HelpFormatter): def start_section(self, heading): res = "\n\n" @@ -17,6 +92,7 @@ class YamlHelpFormatter(argparse.HelpFormatter): return """ ########################################################## # Sample .aider.conf.yaml +# This file lists *all* the valid configuration entries. # Place in your home dir, or at the root of your git repo. ########################################################## diff --git a/aider/coders/base_coder.py b/aider/coders/base_coder.py index 16dc55c78..76df78985 100755 --- a/aider/coders/base_coder.py +++ b/aider/coders/base_coder.py @@ -27,7 +27,7 @@ from aider.mdstream import MarkdownStream from aider.repo import GitRepo from aider.repomap import RepoMap from aider.sendchat import send_with_retries -from aider.utils import is_image_file +from aider.utils import format_content, format_messages, is_image_file from ..dump import dump # noqa: F401 @@ -783,6 +783,8 @@ class Coder: messages = self.format_messages() + self.io.log_llm_history("TO LLM", format_messages(messages)) + if self.verbose: utils.show_messages(messages, functions=self.functions) @@ -795,21 +797,23 @@ class Coder: except ExhaustedContextWindow: exhausted = True except litellm.exceptions.BadRequestError as err: - self.io.tool_error(f"BadRequestError: {err}") - return + if "ContextWindowExceededError" in err.message: + exhausted = True + else: + self.io.tool_error(f"BadRequestError: {err}") + return except openai.BadRequestError as err: if "maximum context length" in str(err): exhausted = True else: raise err + except Exception as err: + self.io.tool_error(f"Unexpected error: {err}") + return if exhausted: + self.show_exhausted_error() self.num_exhausted_context_windows += 1 - self.io.tool_error("The chat session is larger than the context window!\n") - self.commands.cmd_tokens("") - self.io.tool_error("\nTo reduce token usage:") - self.io.tool_error(" - Use /drop to remove unneeded files from the chat session.") - self.io.tool_error(" - Use /clear to clear chat history.") return if self.partial_response_function_call: @@ -825,6 +829,8 @@ class Coder: self.io.tool_output() + self.io.log_llm_history("LLM RESPONSE", format_content("ASSISTANT", content)) + if interrupted: content += "\n^C KeyboardInterrupt" self.cur_messages += [dict(role="assistant", content=content)] @@ -878,6 +884,63 @@ class Coder: else: self.reflected_message = add_rel_files_message + def show_exhausted_error(self): + output_tokens = 0 + if self.partial_response_content: + output_tokens = self.main_model.token_count(self.partial_response_content) + max_output_tokens = self.main_model.info.get("max_output_tokens", 0) + + input_tokens = self.main_model.token_count(self.format_messages()) + max_input_tokens = self.main_model.info.get("max_input_tokens", 0) + + total_tokens = input_tokens + output_tokens + + if output_tokens >= max_output_tokens: + out_err = " -- exceeded output limit!" + else: + out_err = "" + + if input_tokens >= max_input_tokens: + inp_err = " -- context window exhausted!" + else: + inp_err = "" + + if total_tokens >= max_input_tokens: + tot_err = " -- context window exhausted!" + else: + tot_err = "" + + res = ["", ""] + res.append(f"Model {self.main_model.name} has hit a token limit!") + res.append("") + res.append(f"Input tokens: {input_tokens:,} of {max_input_tokens:,}{inp_err}") + res.append(f"Output tokens: {output_tokens:,} of {max_output_tokens:,}{out_err}") + res.append(f"Total tokens: {total_tokens:,} of {max_input_tokens:,}{tot_err}") + + if output_tokens >= max_output_tokens: + res.append("") + res.append("To reduce output tokens:") + res.append("- Ask for smaller changes in each request.") + res.append("- Break your code into smaller source files.") + if "diff" not in self.main_model.edit_format: + res.append( + "- Try using a stronger model like gpt-4o or opus that can return diffs." + ) + + if input_tokens >= max_input_tokens or total_tokens >= max_input_tokens: + res.append("") + res.append("To reduce input tokens:") + res.append("- Use /tokens to see token usage.") + res.append("- Use /drop to remove unneeded files from the chat session.") + res.append("- Use /clear to clear the chat history.") + res.append("- Break your code into smaller source files.") + + res.append("") + res.append(f"For more info: {urls.token_limits}") + + res = "".join([line + "\n" for line in res]) + self.io.tool_error(res) + def lint_edited(self, fnames): res = "" for fname in fnames: @@ -1321,7 +1384,7 @@ class Coder: def auto_commit(self, edited): # context = self.get_context_from_history(self.cur_messages) - res = self.repo.commit(fnames=edited, prefix="aider: ") + res = self.repo.commit(fnames=edited, aider_edits=True) if res: commit_hash, commit_message = res self.last_aider_commit_hash = commit_hash diff --git a/aider/coders/editblock_coder.py b/aider/coders/editblock_coder.py index 017e40408..557e692b6 100644 --- a/aider/coders/editblock_coder.py +++ b/aider/coders/editblock_coder.py @@ -414,16 +414,8 @@ def find_original_update_blocks(content, fence=DEFAULT_FENCE): processed.append(cur) # original_marker - filename = strip_filename(processed[-2].splitlines()[-1], fence) - try: - if not filename: - filename = strip_filename(processed[-2].splitlines()[-2], fence) - if not filename: - if current_filename: - filename = current_filename - else: - raise ValueError(missing_filename_err.format(fence=fence)) - except IndexError: + filename = find_filename(processed[-2].splitlines(), fence) + if not filename: if current_filename: filename = current_filename else: @@ -460,6 +452,35 @@ def find_original_update_blocks(content, fence=DEFAULT_FENCE): raise ValueError(f"{processed}\n^^^ Error parsing SEARCH/REPLACE block.") +def find_filename(lines, fence): + """ + Deepseek Coder v2 has been doing this: + + + ```python + word_count.py + ``` + ```python + <<<<<<< SEARCH + ... + + This is a more flexible search back for filenames. + """ + # Go back through the 3 preceding lines + lines.reverse() + lines = lines[:3] + + for line in lines: + # If we find a filename, done + filename = strip_filename(line, fence) + if filename: + return filename + + # Only continue as long as we keep seeing fences + if not line.startswith(fence[0]): + return + + if __name__ == "__main__": edit = """ Here's the change: diff --git a/aider/commands.py b/aider/commands.py index f6b0c3cba..00189fecd 100644 --- a/aider/commands.py +++ b/aider/commands.py @@ -332,7 +332,7 @@ class Commands: last_commit = self.coder.repo.repo.head.commit if ( - not last_commit.message.startswith("aider:") + not last_commit.author.name.endswith(" (aider)") or last_commit.hexsha[:7] != self.coder.last_aider_commit_hash ): self.io.tool_error("The last commit was not made by aider in this chat session.") diff --git a/aider/gui.py b/aider/gui.py index 16a13a004..b2239b162 100755 --- a/aider/gui.py +++ b/aider/gui.py @@ -110,9 +110,6 @@ class GUI: show_undo = False res = "" if commit_hash: - prefix = "aider: " - if commit_message.startswith(prefix): - commit_message = commit_message[len(prefix) :] res += f"Commit `{commit_hash}`: {commit_message} \n" if commit_hash == self.coder.last_aider_commit_hash: show_undo = True diff --git a/aider/io.py b/aider/io.py index 65093332c..1cf09f41f 100644 --- a/aider/io.py +++ b/aider/io.py @@ -107,6 +107,7 @@ class InputOutput: tool_error_color="red", encoding="utf-8", dry_run=False, + llm_history_file=None, editingmode=EditingMode.EMACS, ): self.editingmode = editingmode @@ -128,6 +129,7 @@ class InputOutput: self.yes = yes self.input_history_file = input_history_file + self.llm_history_file = llm_history_file if chat_history_file is not None: self.chat_history_file = Path(chat_history_file) else: @@ -209,10 +211,11 @@ class InputOutput: else: style = None + completer_instance = AutoCompleter( + root, rel_fnames, addable_rel_fnames, commands, self.encoding + ) + while True: - completer_instance = AutoCompleter( - root, rel_fnames, addable_rel_fnames, commands, self.encoding - ) if multiline_input: show = ". " @@ -271,6 +274,14 @@ class InputOutput: fh = FileHistory(self.input_history_file) return fh.load_history_strings() + def log_llm_history(self, role, content): + if not self.llm_history_file: + return + timestamp = datetime.now().isoformat(timespec='seconds') + with open(self.llm_history_file, 'a', encoding=self.encoding) as log_file: + log_file.write(f"{role.upper()} {timestamp}\n") + log_file.write(content + "\n") + def user_input(self, inp, log_only=True): if not log_only: style = dict(style=self.user_input_color) if self.user_input_color else dict() diff --git a/aider/main.py b/aider/main.py index 4fe21aba1..3acaf5afc 100644 --- a/aider/main.py +++ b/aider/main.py @@ -5,12 +5,13 @@ import sys from pathlib import Path import git +import httpx from dotenv import load_dotenv from prompt_toolkit.enums import EditingMode from streamlit.web import cli from aider import __version__, models, utils -from aider.args import get_parser +from aider.args import get_parser, get_preparser from aider.coders import Coder from aider.commands import SwitchModel from aider.io import InputOutput @@ -124,12 +125,18 @@ def check_gitignore(git_root, io, ask=True): def format_settings(parser, args): show = scrub_sensitive_info(args, parser.format_values()) + # clean up the headings for consistency w/ new lines + heading_env = "Environment Variables:" + heading_defaults = "Defaults:" + if heading_env in show: + show = show.replace(heading_env, "\n" + heading_env) + show = show.replace(heading_defaults, "\n" + heading_defaults) show += "\n" show += "Option settings:\n" for arg, val in sorted(vars(args).items()): if val: val = scrub_sensitive_info(args, str(val)) - show += f" - {arg}: {val}\n" + show += f" - {arg}: {val}\n" # noqa: E221 return show @@ -266,9 +273,18 @@ def main(argv=None, input=None, output=None, force_git_root=None, return_coder=F default_config_files.append(Path.home() / conf_fname) # homedir default_config_files = list(map(str, default_config_files)) + preparser = get_preparser(git_root) + pre_args, _ = preparser.parse_known_args(argv) + + # Load the .env file specified in the arguments + load_dotenv(pre_args.env_file) + parser = get_parser(default_config_files, git_root) args = parser.parse_args(argv) + if not args.verify_ssl: + litellm.client_session = httpx.Client(verify=False) + if args.gui and not return_coder: launch_gui(argv) return @@ -302,6 +318,7 @@ def main(argv=None, input=None, output=None, force_git_root=None, return_coder=F tool_error_color=args.tool_error_color, dry_run=args.dry_run, encoding=args.encoding, + llm_history_file=args.llm_history_file, editingmode=editing_mode, ) @@ -360,9 +377,6 @@ def main(argv=None, input=None, output=None, force_git_root=None, return_coder=F cmd_line = scrub_sensitive_info(args, cmd_line) io.tool_output(cmd_line, log_only=True) - if args.env_file: - load_dotenv(args.env_file) - if args.anthropic_api_key: os.environ["ANTHROPIC_API_KEY"] = args.anthropic_api_key diff --git a/aider/models.py b/aider/models.py index 19e04167a..bed59ea7e 100644 --- a/aider/models.py +++ b/aider/models.py @@ -179,6 +179,43 @@ MODEL_SETTINGS = [ "whole", weak_model_name="claude-3-haiku-20240307", ), + ModelSettings( + "claude-3-5-sonnet-20240620", + "diff", + weak_model_name="claude-3-haiku-20240307", + use_repo_map=True, + ), + ModelSettings( + "anthropic/claude-3-5-sonnet-20240620", + "diff", + weak_model_name="claude-3-haiku-20240307", + use_repo_map=True, + ), + ModelSettings( + "openrouter/anthropic/claude-3.5-sonnet", + "diff", + weak_model_name="openrouter/anthropic/claude-3-haiku-20240307", + use_repo_map=True, + ), + # Vertex AI Claude models + ModelSettings( + "vertex_ai/claude-3-5-sonnet@20240620", + "diff", + weak_model_name="vertex_ai/claude-3-haiku@20240307", + use_repo_map=True, + ), + ModelSettings( + "vertex_ai/claude-3-opus@20240229", + "diff", + weak_model_name="vertex_ai/claude-3-haiku@20240307", + use_repo_map=True, + send_undo_reply=True, + ), + ModelSettings( + "vertex_ai/claude-3-sonnet@20240229", + "whole", + weak_model_name="vertex_ai/claude-3-haiku@20240307", + ), # Cohere ModelSettings( "command-r-plus", @@ -219,7 +256,7 @@ MODEL_SETTINGS = [ send_undo_reply=True, ), ModelSettings( - "openai/deepseek-chat", + "deepseek/deepseek-chat", "diff", use_repo_map=True, send_undo_reply=True, @@ -227,7 +264,15 @@ MODEL_SETTINGS = [ reminder_as_sys_msg=True, ), ModelSettings( - "deepseek/deepseek-chat", + "deepseek/deepseek-coder", + "diff", + use_repo_map=True, + send_undo_reply=True, + examples_as_sys_msg=True, + reminder_as_sys_msg=True, + ), + ModelSettings( + "openrouter/deepseek/deepseek-coder", "diff", use_repo_map=True, send_undo_reply=True, diff --git a/aider/repo.py b/aider/repo.py index 666a8f084..38f5ef8f5 100644 --- a/aider/repo.py +++ b/aider/repo.py @@ -59,7 +59,7 @@ class GitRepo: if aider_ignore_file: self.aider_ignore_file = Path(aider_ignore_file) - def commit(self, fnames=None, context=None, prefix=None, message=None): + def commit(self, fnames=None, context=None, message=None, aider_edits=False): if not fnames and not self.repo.is_dirty(): return @@ -75,9 +75,6 @@ class GitRepo: if not commit_message: commit_message = "(no commit message provided)" - if prefix: - commit_message = prefix + commit_message - full_commit_message = commit_message if context: full_commit_message += "\n\n# Aider chat conversation:\n\n" + context @@ -91,10 +88,32 @@ class GitRepo: else: cmd += ["-a"] + original_user_name = self.repo.config_reader().get_value("user", "name") + original_committer_name_env = os.environ.get("GIT_COMMITTER_NAME") + + committer_name = f"{original_user_name} (aider)" + os.environ["GIT_COMMITTER_NAME"] = committer_name + + if aider_edits: + original_auther_name_env = os.environ.get("GIT_AUTHOR_NAME") + os.environ["GIT_AUTHOR_NAME"] = committer_name + self.repo.git.commit(cmd) commit_hash = self.repo.head.commit.hexsha[:7] self.io.tool_output(f"Commit {commit_hash} {commit_message}") + # Restore the original GIT_COMMITTER_NAME + if aider_edits: + if original_auther_name_env is not None: + os.environ["GIT_AUTHOR_NAME"] = original_auther_name_env + else: + del os.environ["GIT_AUTHOR_NAME"] + + if original_committer_name_env is not None: + os.environ["GIT_COMMITTER_NAME"] = original_committer_name_env + else: + del os.environ["GIT_COMMITTER_NAME"] + return commit_hash, commit_message def get_rel_repo_dir(self): diff --git a/aider/scrape.py b/aider/scrape.py index 147203c48..2ac29b6a2 100755 --- a/aider/scrape.py +++ b/aider/scrape.py @@ -10,7 +10,7 @@ from bs4 import BeautifulSoup from playwright.sync_api import sync_playwright from aider import __version__, urls -from aider.dump import dump +from aider.dump import dump # noqa: F401 aider_user_agent = f"Aider/{__version__} +{urls.website}" @@ -21,7 +21,7 @@ For better web scraping, install Playwright chromium with this command in your t playwright install --with-deps chromium -See {urls.enable_playwrite} for more info. +See {urls.enable_playwright} for more info. """ @@ -53,7 +53,6 @@ class Scraper: else: content = self.scrape_with_httpx(url) - dump(content) if not content: return diff --git a/aider/tests/test_commands.py b/aider/tests/test_commands.py index 8cc67b6d8..f6629f481 100644 --- a/aider/tests/test_commands.py +++ b/aider/tests/test_commands.py @@ -523,16 +523,20 @@ class TestCommands(TestCase): other_path.write_text("other content") repo.git.add(str(other_path)) + os.environ["GIT_AUTHOR_NAME"] = "Foo (aider)" + # Create and commit a file filename = "test_file.txt" file_path = Path(repo_dir) / filename file_path.write_text("first content") repo.git.add(filename) - repo.git.commit("-m", "aider: first commit") + repo.git.commit("-m", "first commit") file_path.write_text("second content") repo.git.add(filename) - repo.git.commit("-m", "aider: second commit") + repo.git.commit("-m", "second commit") + + del os.environ["GIT_AUTHOR_NAME"] # Store the commit hash last_commit_hash = repo.head.commit.hexsha[:7] diff --git a/aider/tests/test_editblock.py b/aider/tests/test_editblock.py index 0c1143232..40a0d4572 100644 --- a/aider/tests/test_editblock.py +++ b/aider/tests/test_editblock.py @@ -398,6 +398,32 @@ Hope you like it! ], ) + def test_deepseek_coder_v2_filename_mangling(self): + edit = """ +Here's the change: + + ```python +foo.txt +``` +```python +<<<<<<< SEARCH +one +======= +two +>>>>>>> REPLACE +``` + +Hope you like it! +""" + + edits = list(eb.find_original_update_blocks(edit)) + self.assertEqual( + edits, + [ + ("foo.txt", "one\n", "two\n"), + ], + ) + if __name__ == "__main__": unittest.main() diff --git a/aider/tests/test_main.py b/aider/tests/test_main.py index d319a78dd..b576d3754 100644 --- a/aider/tests/test_main.py +++ b/aider/tests/test_main.py @@ -1,7 +1,7 @@ import os -import shutil import subprocess import tempfile +from io import StringIO from pathlib import Path from unittest import TestCase from unittest.mock import MagicMock, patch @@ -13,24 +13,28 @@ from prompt_toolkit.output import DummyOutput from aider.dump import dump # noqa: F401 from aider.io import InputOutput from aider.main import check_gitignore, main, setup_git -from aider.utils import GitTemporaryDirectory, make_repo +from aider.utils import GitTemporaryDirectory, IgnorantTemporaryDirectory, make_repo class TestMain(TestCase): def setUp(self): + self.original_env = os.environ.copy() os.environ["OPENAI_API_KEY"] = "deadbeef" self.original_cwd = os.getcwd() - self.tempdir = tempfile.mkdtemp() + self.tempdir_obj = IgnorantTemporaryDirectory() + self.tempdir = self.tempdir_obj.name os.chdir(self.tempdir) def tearDown(self): os.chdir(self.original_cwd) - shutil.rmtree(self.tempdir, ignore_errors=True) + self.tempdir_obj.cleanup() + os.environ.clear() + os.environ.update(self.original_env) def test_main_with_empty_dir_no_files_on_command(self): main(["--no-git"], input=DummyInput(), output=DummyOutput()) - def test_main_with_empty_dir_new_file(self): + def test_main_with_emptqy_dir_new_file(self): main(["foo.txt", "--yes", "--no-git"], input=DummyInput(), output=DummyOutput()) self.assertTrue(os.path.exists("foo.txt")) @@ -237,3 +241,82 @@ class TestMain(TestCase): main(["--message", test_message]) args, kwargs = MockInputOutput.call_args self.assertEqual(args[1], None) + + def test_dark_mode_sets_code_theme(self): + # Mock Coder.create to capture the configuration + with patch("aider.coders.Coder.create") as MockCoder: + main(["--dark-mode", "--no-git"], input=DummyInput(), output=DummyOutput()) + # Ensure Coder.create was called + MockCoder.assert_called_once() + # Check if the code_theme setting is for dark mode + _, kwargs = MockCoder.call_args + self.assertEqual(kwargs["code_theme"], "monokai") + + def test_light_mode_sets_code_theme(self): + # Mock Coder.create to capture the configuration + with patch("aider.coders.Coder.create") as MockCoder: + main(["--light-mode", "--no-git"], input=DummyInput(), output=DummyOutput()) + # Ensure Coder.create was called + MockCoder.assert_called_once() + # Check if the code_theme setting is for light mode + _, kwargs = MockCoder.call_args + self.assertEqual(kwargs["code_theme"], "default") + + def create_env_file(self, file_name, content): + env_file_path = Path(self.tempdir) / file_name + env_file_path.write_text(content) + return env_file_path + + def test_env_file_flag_sets_automatic_variable(self): + env_file_path = self.create_env_file(".env.test", "AIDER_DARK_MODE=True") + with patch("aider.coders.Coder.create") as MockCoder: + main( + ["--env-file", str(env_file_path), "--no-git"], + input=DummyInput(), + output=DummyOutput(), + ) + MockCoder.assert_called_once() + # Check if the color settings are for dark mode + _, kwargs = MockCoder.call_args + self.assertEqual(kwargs["code_theme"], "monokai") + + def test_default_env_file_sets_automatic_variable(self): + self.create_env_file(".env", "AIDER_DARK_MODE=True") + with patch("aider.coders.Coder.create") as MockCoder: + main(["--no-git"], input=DummyInput(), output=DummyOutput()) + # Ensure Coder.create was called + MockCoder.assert_called_once() + # Check if the color settings are for dark mode + _, kwargs = MockCoder.call_args + self.assertEqual(kwargs["code_theme"], "monokai") + + def test_false_vals_in_env_file(self): + self.create_env_file(".env", "AIDER_SHOW_DIFFS=off") + with patch("aider.coders.Coder.create") as MockCoder: + main(["--no-git"], input=DummyInput(), output=DummyOutput()) + MockCoder.assert_called_once() + _, kwargs = MockCoder.call_args + self.assertEqual(kwargs["show_diffs"], False) + + def test_true_vals_in_env_file(self): + self.create_env_file(".env", "AIDER_SHOW_DIFFS=on") + with patch("aider.coders.Coder.create") as MockCoder: + main(["--no-git"], input=DummyInput(), output=DummyOutput()) + MockCoder.assert_called_once() + _, kwargs = MockCoder.call_args + self.assertEqual(kwargs["show_diffs"], True) + + def test_verbose_mode_lists_env_vars(self): + self.create_env_file(".env", "AIDER_DARK_MODE=on") + with patch("sys.stdout", new_callable=StringIO) as mock_stdout: + main(["--no-git", "--verbose"], input=DummyInput(), output=DummyOutput()) + output = mock_stdout.getvalue() + relevant_output = "\n".join( + line + for line in output.splitlines() + if "AIDER_DARK_MODE" in line or "dark_mode" in line + ) # this bit just helps failing assertions to be easier to read + self.assertIn("AIDER_DARK_MODE", relevant_output) + self.assertIn("dark_mode", relevant_output) + self.assertRegex(relevant_output, r"AIDER_DARK_MODE:\s+on") + self.assertRegex(relevant_output, r"dark_mode:\s+True") diff --git a/aider/tests/test_repo.py b/aider/tests/test_repo.py index bdee89c0c..92d074b9a 100644 --- a/aider/tests/test_repo.py +++ b/aider/tests/test_repo.py @@ -1,4 +1,5 @@ import os +import platform import tempfile import unittest from pathlib import Path @@ -137,6 +138,52 @@ class TestRepo(unittest.TestCase): # Assert that the returned message is the expected one self.assertEqual(result, 'a good "commit message"') + @patch("aider.repo.GitRepo.get_commit_message") + def test_commit_with_custom_committer_name(self, mock_send): + mock_send.return_value = '"a good commit message"' + + # Cleanup of the git temp dir explodes on windows + if platform.system() == "Windows": + return + + with GitTemporaryDirectory(): + # new repo + raw_repo = git.Repo() + raw_repo.config_writer().set_value("user", "name", "Test User").release() + + # add a file and commit it + fname = Path("file.txt") + fname.touch() + raw_repo.git.add(str(fname)) + raw_repo.git.commit("-m", "initial commit") + + io = InputOutput() + git_repo = GitRepo(io, None, None) + + # commit a change + fname.write_text("new content") + git_repo.commit(fnames=[str(fname)], aider_edits=True) + + # check the committer name + commit = raw_repo.head.commit + self.assertEqual(commit.author.name, "Test User (aider)") + self.assertEqual(commit.committer.name, "Test User (aider)") + + # commit a change without aider_edits + fname.write_text("new content again!") + git_repo.commit(fnames=[str(fname)], aider_edits=False) + + # check the committer name + commit = raw_repo.head.commit + self.assertEqual(commit.author.name, "Test User") + self.assertEqual(commit.committer.name, "Test User (aider)") + + # check that the original committer name is restored + original_committer_name = os.environ.get("GIT_COMMITTER_NAME") + self.assertIsNone(original_committer_name) + original_author_name = os.environ.get("GIT_AUTHOR_NAME") + self.assertIsNone(original_author_name) + def test_get_tracked_files(self): # Create a temporary directory tempdir = Path(tempfile.mkdtemp()) diff --git a/aider/urls.py b/aider/urls.py index 62f95503d..8e21742d3 100644 --- a/aider/urls.py +++ b/aider/urls.py @@ -2,6 +2,8 @@ website = "https://aider.chat/" add_all_files = "https://aider.chat/docs/faq.html#how-can-i-add-all-the-files-to-the-chat" edit_errors = "https://aider.chat/docs/troubleshooting/edit-errors.html" git = "https://aider.chat/docs/git.html" -enable_playwrite = "https://aider.chat/docs/install/optional.html#enable-playwright" +enable_playwright = "https://aider.chat/docs/install/optional.html#enable-playwright" favicon = "https://aider.chat/assets/icons/favicon-32x32.png" model_warnings = "https://aider.chat/docs/llms/warnings.html" +token_limits = "https://aider.chat/docs/troubleshooting/token-limits.html" +llms = "https://aider.chat/docs/llms.html" diff --git a/aider/utils.py b/aider/utils.py index 31767aec6..6d097fbe7 100644 --- a/aider/utils.py +++ b/aider/utils.py @@ -17,11 +17,17 @@ class IgnorantTemporaryDirectory: return self.temp_dir.__enter__() def __exit__(self, exc_type, exc_val, exc_tb): + self.cleanup() + + def cleanup(self): try: - self.temp_dir.__exit__(exc_type, exc_val, exc_tb) + self.temp_dir.cleanup() except (OSError, PermissionError): pass # Ignore errors (Windows) + def __getattr__(self, item): + return getattr(self.temp_dir, item) + class ChdirTemporaryDirectory(IgnorantTemporaryDirectory): def __init__(self): @@ -84,24 +90,38 @@ def safe_abs_path(res): return str(res) -def show_messages(messages, title=None, functions=None): +def format_content(role, content): + formatted_lines = [] + for line in content.splitlines(): + formatted_lines.append(f"{role} {line}") + return "\n".join(formatted_lines) + + +def format_messages(messages, title=None): + output = [] if title: - print(title.upper(), "*" * 50) + output.append(f"{title.upper()} {'*' * 50}") for msg in messages: - print() + output.append("") role = msg["role"].upper() content = msg.get("content") if isinstance(content, list): # Handle list content (e.g., image messages) for item in content: if isinstance(item, dict) and "image_url" in item: - print(role, "Image URL:", item["image_url"]["url"]) + output.append(f"{role} Image URL: {item['image_url']['url']}") elif isinstance(content, str): # Handle string content - for line in content.splitlines(): - print(role, line) + output.append(format_content(role, content)) content = msg.get("function_call") if content: - print(role, content) + output.append(f"{role} {content}") + + return "\n".join(output) + + +def show_messages(messages, title=None, functions=None): + formatted_output = format_messages(messages, title) + print(formatted_output) if functions: dump(functions) diff --git a/requirements.txt b/requirements.txt index 844b4e6b5..f99a01346 100644 --- a/requirements.txt +++ b/requirements.txt @@ -10,9 +10,9 @@ aiosignal==1.3.1 # via aiohttp altair==5.3.0 # via streamlit -annotated-types==0.6.0 +annotated-types==0.7.0 # via pydantic -anyio==4.3.0 +anyio==4.4.0 # via # httpx # openai @@ -31,7 +31,7 @@ cachetools==5.3.3 # via # google-auth # streamlit -certifi==2024.2.2 +certifi==2024.6.2 # via # httpcore # httpx @@ -54,15 +54,15 @@ diskcache==5.6.3 # via -r requirements.in distro==1.9.0 # via openai -filelock==3.14.0 +filelock==3.15.3 # via huggingface-hub -flake8==7.0.0 +flake8==7.1.0 # via -r requirements.in frozenlist==1.4.1 # via # aiohttp # aiosignal -fsspec==2024.5.0 +fsspec==2024.6.0 # via huggingface-hub gitdb==4.0.11 # via gitpython @@ -70,16 +70,16 @@ gitpython==3.1.43 # via # -r requirements.in # streamlit -google-ai-generativelanguage==0.6.4 +google-ai-generativelanguage==0.6.5 # via google-generativeai google-api-core[grpc]==2.19.0 # via # google-ai-generativelanguage # google-api-python-client # google-generativeai -google-api-python-client==2.129.0 +google-api-python-client==2.134.0 # via google-generativeai -google-auth==2.29.0 +google-auth==2.30.0 # via # google-ai-generativelanguage # google-api-core @@ -88,9 +88,9 @@ google-auth==2.29.0 # google-generativeai google-auth-httplib2==0.2.0 # via google-api-python-client -google-generativeai==0.5.4 +google-generativeai==0.7.0 # via -r requirements.in -googleapis-common-protos==1.63.0 +googleapis-common-protos==1.63.1 # via # google-api-core # grpcio-status @@ -98,7 +98,7 @@ greenlet==3.0.3 # via playwright grep-ast==0.3.2 # via -r requirements.in -grpcio==1.63.0 +grpcio==1.64.1 # via # google-api-core # grpcio-status @@ -114,7 +114,7 @@ httplib2==0.22.0 # google-auth-httplib2 httpx==0.27.0 # via openai -huggingface-hub==0.23.0 +huggingface-hub==0.23.4 # via tokenizers idna==3.7 # via @@ -122,7 +122,9 @@ idna==3.7 # httpx # requests # yarl -importlib-metadata==7.1.0 +ijson==3.3.0 + # via litellm +importlib-metadata==7.2.0 # via litellm jinja2==3.1.4 # via @@ -135,7 +137,7 @@ jsonschema==4.22.0 # altair jsonschema-specifications==2023.12.1 # via jsonschema -litellm==1.37.16 +litellm==1.40.21 # via -r requirements.in markdown-it-py==3.0.0 # via rich @@ -151,7 +153,7 @@ multidict==6.0.5 # yarl networkx==3.2.1 # via -r requirements.in -numpy==1.26.4 +numpy==2.0.0 # via # -r requirements.in # altair @@ -160,11 +162,11 @@ numpy==1.26.4 # pydeck # scipy # streamlit -openai==1.30.1 +openai==1.35.3 # via # -r requirements.in # litellm -packaging==24.0 +packaging==24.1 # via # -r requirements.in # altair @@ -184,9 +186,9 @@ pillow==10.3.0 # streamlit playwright==1.44.0 # via -r requirements.in -prompt-toolkit==3.0.43 +prompt-toolkit==3.0.47 # via -r requirements.in -proto-plus==1.23.0 +proto-plus==1.24.0 # via # google-ai-generativelanguage # google-api-core @@ -207,15 +209,16 @@ pyasn1==0.6.0 # rsa pyasn1-modules==0.4.0 # via google-auth -pycodestyle==2.11.1 +pycodestyle==2.12.0 # via flake8 pycparser==2.22 # via cffi -pydantic==2.7.1 +pydantic==2.7.4 # via # google-generativeai + # litellm # openai -pydantic-core==2.18.2 +pydantic-core==2.18.4 # via pydantic pydeck==0.9.1 # via streamlit @@ -245,7 +248,7 @@ referencing==0.35.1 # jsonschema-specifications regex==2024.5.15 # via tiktoken -requests==2.31.0 +requests==2.32.3 # via # google-api-core # huggingface-hub @@ -262,7 +265,7 @@ rpds-py==0.18.1 # referencing rsa==4.9 # via google-auth -scipy==1.13.0 +scipy==1.13.1 # via -r requirements.in six==1.16.0 # via python-dateutil @@ -273,15 +276,15 @@ sniffio==1.3.1 # anyio # httpx # openai -sounddevice==0.4.6 +sounddevice==0.4.7 # via -r requirements.in soundfile==0.12.1 # via -r requirements.in soupsieve==2.5 # via beautifulsoup4 -streamlit==1.34.0 +streamlit==1.36.0 # via -r requirements.in -tenacity==8.3.0 +tenacity==8.4.1 # via streamlit tiktoken==0.7.0 # via @@ -293,7 +296,7 @@ toml==0.10.2 # via streamlit toolz==0.12.1 # via altair -tornado==6.4 +tornado==6.4.1 # via streamlit tqdm==4.66.4 # via @@ -306,7 +309,7 @@ tree-sitter==0.21.3 # tree-sitter-languages tree-sitter-languages==1.10.2 # via grep-ast -typing-extensions==4.11.0 +typing-extensions==4.12.2 # via # google-generativeai # huggingface-hub @@ -319,7 +322,7 @@ tzdata==2024.1 # via pandas uritemplate==4.1.1 # via google-api-python-client -urllib3==2.2.1 +urllib3==2.2.2 # via requests watchdog==4.0.1 # via -r requirements.in @@ -327,5 +330,5 @@ wcwidth==0.2.13 # via prompt-toolkit yarl==1.9.4 # via aiohttp -zipp==3.18.2 +zipp==3.19.2 # via importlib-metadata diff --git a/scripts/update-docs.sh b/scripts/update-docs.sh index 756a61c69..fc16c4c66 100755 --- a/scripts/update-docs.sh +++ b/scripts/update-docs.sh @@ -9,8 +9,12 @@ else ARG=$1 fi +# README.md before index.md, because index.md uses cog to include README.md cog $ARG \ + README.md \ website/index.md \ + website/HISTORY.md \ + website/docs/dotenv.md \ website/docs/commands.md \ website/docs/languages.md \ website/docs/options.md \ diff --git a/website/HISTORY.md b/website/HISTORY.md new file mode 100644 index 000000000..4e84bf086 --- /dev/null +++ b/website/HISTORY.md @@ -0,0 +1,425 @@ +--- +title: Release history +parent: More info +nav_order: 999 +--- + + + +# Release history + +### v0.39.0 + +- Use `--sonnet` for Claude 3.5 Sonnet, which is the top model on [aider's LLM code editing leaderboard](https://aider.chat/docs/leaderboards/#claude-35-sonnet-takes-the-top-spot). +- All `AIDER_xxx` environment variables can now be set in `.env` (by @jpshack-at-palomar). +- Use `--llm-history-file` to log raw messages sent to the LLM (by @daniel-vainsencher). +- Commit messages are no longer prefixed with "aider:". Instead the git author and committer names have "(aider)" added. + +### v0.38.0 + +- Use `--vim` for [vim keybindings](https://aider.chat/docs/commands.html#vi) in the chat. +- [Add LLM metadata](https://aider.chat/docs/llms/warnings.html#specifying-context-window-size-and-token-costs) via `.aider.models.json` file (by @caseymcc). +- More detailed [error messages on token limit errors](https://aider.chat/docs/troubleshooting/token-limits.html). +- Single line commit messages, without the recent chat messages. +- Ensure `--commit --dry-run` does nothing. +- Have playwright wait for idle network to better scrape js sites. +- Documentation updates, moved into website/ subdir. +- Moved tests/ into aider/tests/. + +### v0.37.0 + +- Repo map is now optimized based on text of chat history as well as files added to chat. +- Improved prompts when no files have been added to chat to solicit LLM file suggestions. +- Aider will notice if you paste a URL into the chat, and offer to scrape it. +- Performance improvements the repo map, especially in large repos. +- Aider will not offer to add bare filenames like `make` or `run` which may just be words. +- Properly override `GIT_EDITOR` env for commits if it is already set. +- Detect supported audio sample rates for `/voice`. +- Other small bug fixes. + +### v0.36.0 + +- [Aider can now lint your code and fix any errors](https://aider.chat/2024/05/22/linting.html). + - Aider automatically lints and fixes after every LLM edit. + - You can manually lint-and-fix files with `/lint` in the chat or `--lint` on the command line. + - Aider includes built in basic linters for all supported tree-sitter languages. + - You can also configure aider to use your preferred linter with `--lint-cmd`. +- Aider has additional support for running tests and fixing problems. + - Configure your testing command with `--test-cmd`. + - Run tests with `/test` or from the command line with `--test`. + - Aider will automatically attempt to fix any test failures. + + +### v0.35.0 + +- Aider now uses GPT-4o by default. + - GPT-4o tops the [aider LLM code editing leaderboard](https://aider.chat/docs/leaderboards/) at 72.9%, versus 68.4% for Opus. + - GPT-4o takes second on [aider's refactoring leaderboard](https://aider.chat/docs/leaderboards/#code-refactoring-leaderboard) with 62.9%, versus Opus at 72.3%. +- Added `--restore-chat-history` to restore prior chat history on launch, so you can continue the last conversation. +- Improved reflection feedback to LLMs using the diff edit format. +- Improved retries on `httpx` errors. + +### v0.34.0 + +- Updated prompting to use more natural phrasing about files, the git repo, etc. Removed reliance on read-write/read-only terminology. +- Refactored prompting to unify some phrasing across edit formats. +- Enhanced the canned assistant responses used in prompts. +- Added explicit model settings for `openrouter/anthropic/claude-3-opus`, `gpt-3.5-turbo` +- Added `--show-prompts` debug switch. +- Bugfix: catch and retry on all litellm exceptions. + + +### v0.33.0 + +- Added native support for [Deepseek models](https://aider.chat/docs/llms.html#deepseek) using `DEEPSEEK_API_KEY` and `deepseek/deepseek-chat`, etc rather than as a generic OpenAI compatible API. + +### v0.32.0 + +- [Aider LLM code editing leaderboards](https://aider.chat/docs/leaderboards/) that rank popular models according to their ability to edit code. + - Leaderboards include GPT-3.5/4 Turbo, Opus, Sonnet, Gemini 1.5 Pro, Llama 3, Deepseek Coder & Command-R+. +- Gemini 1.5 Pro now defaults to a new diff-style edit format (diff-fenced), enabling it to work better with larger code bases. +- Support for Deepseek-V2, via more a flexible config of system messages in the diff edit format. +- Improved retry handling on errors from model APIs. +- Benchmark outputs results in YAML, compatible with leaderboard. + +### v0.31.0 + +- [Aider is now also AI pair programming in your browser!](https://aider.chat/2024/05/02/browser.html) Use the `--browser` switch to launch an experimental browser based version of aider. +- Switch models during the chat with `/model ` and search the list of available models with `/models `. + +### v0.30.1 + +- Adding missing `google-generativeai` dependency + +### v0.30.0 + +- Added [Gemini 1.5 Pro](https://aider.chat/docs/llms.html#free-models) as a recommended free model. +- Allow repo map for "whole" edit format. +- Added `--models ` to search the available models. +- Added `--no-show-model-warnings` to silence model warnings. + +### v0.29.2 + +- Improved [model warnings](https://aider.chat/docs/llms.html#model-warnings) for unknown or unfamiliar models + +### v0.29.1 + +- Added better support for groq/llama3-70b-8192 + +### v0.29.0 + +- Added support for [directly connecting to Anthropic, Cohere, Gemini and many other LLM providers](https://aider.chat/docs/llms.html). +- Added `--weak-model ` which allows you to specify which model to use for commit messages and chat history summarization. +- New command line switches for working with popular models: + - `--4-turbo-vision` + - `--opus` + - `--sonnet` + - `--anthropic-api-key` +- Improved "whole" and "diff" backends to better support [Cohere's free to use Command-R+ model](https://aider.chat/docs/llms.html#cohere). +- Allow `/add` of images from anywhere in the filesystem. +- Fixed crash when operating in a repo in a detached HEAD state. +- Fix: Use the same default model in CLI and python scripting. + +### v0.28.0 + +- Added support for new `gpt-4-turbo-2024-04-09` and `gpt-4-turbo` models. + - Benchmarked at 61.7% on Exercism benchmark, comparable to `gpt-4-0613` and worse than the `gpt-4-preview-XXXX` models. See [recent Exercism benchmark results](https://aider.chat/2024/03/08/claude-3.html). + - Benchmarked at 34.1% on the refactoring/laziness benchmark, significantly worse than the `gpt-4-preview-XXXX` models. See [recent refactor bencmark results](https://aider.chat/2024/01/25/benchmarks-0125.html). + - Aider continues to default to `gpt-4-1106-preview` as it performs best on both benchmarks, and significantly better on the refactoring/laziness benchmark. + +### v0.27.0 + +- Improved repomap support for typescript, by @ryanfreckleton. +- Bugfix: Only /undo the files which were part of the last commit, don't stomp other dirty files +- Bugfix: Show clear error message when OpenAI API key is not set. +- Bugfix: Catch error for obscure languages without tags.scm file. + +### v0.26.1 + +- Fixed bug affecting parsing of git config in some environments. + +### v0.26.0 + +- Use GPT-4 Turbo by default. +- Added `-3` and `-4` switches to use GPT 3.5 or GPT-4 (non-Turbo). +- Bug fix to avoid reflecting local git errors back to GPT. +- Improved logic for opening git repo on launch. + +### v0.25.0 + +- Issue a warning if user adds too much code to the chat. + - https://aider.chat/docs/faq.html#how-can-i-add-all-the-files-to-the-chat +- Vocally refuse to add files to the chat that match `.aiderignore` + - Prevents bug where subsequent git commit of those files will fail. +- Added `--openai-organization-id` argument. +- Show the user a FAQ link if edits fail to apply. +- Made past articles part of https://aider.chat/blog/ + +### v0.24.1 + +- Fixed bug with cost computations when --no-steam in effect + +### v0.24.0 + +- New `/web ` command which scrapes the url, turns it into fairly clean markdown and adds it to the chat. +- Updated all OpenAI model names, pricing info +- Default GPT 3.5 model is now `gpt-3.5-turbo-0125`. +- Bugfix to the `!` alias for `/run`. + +### v0.23.0 + +- Added support for `--model gpt-4-0125-preview` and OpenAI's alias `--model gpt-4-turbo-preview`. The `--4turbo` switch remains an alias for `--model gpt-4-1106-preview` at this time. +- New `/test` command that runs a command and adds the output to the chat on non-zero exit status. +- Improved streaming of markdown to the terminal. +- Added `/quit` as alias for `/exit`. +- Added `--skip-check-update` to skip checking for the update on launch. +- Added `--openrouter` as a shortcut for `--openai-api-base https://openrouter.ai/api/v1` +- Fixed bug preventing use of env vars `OPENAI_API_BASE, OPENAI_API_TYPE, OPENAI_API_VERSION, OPENAI_API_DEPLOYMENT_ID`. + +### v0.22.0 + +- Improvements for unified diff editing format. +- Added ! as an alias for /run. +- Autocomplete for /add and /drop now properly quotes filenames with spaces. +- The /undo command asks GPT not to just retry reverted edit. + +### v0.21.1 + +- Bugfix for unified diff editing format. +- Added --4turbo and --4 aliases for --4-turbo. + +### v0.21.0 + +- Support for python 3.12. +- Improvements to unified diff editing format. +- New `--check-update` arg to check if updates are available and exit with status code. + +### v0.20.0 + +- Add images to the chat to automatically use GPT-4 Vision, by @joshuavial + +- Bugfixes: + - Improved unicode encoding for `/run` command output, by @ctoth + - Prevent false auto-commits on Windows, by @ctoth + +### v0.19.1 + +- Removed stray debug output. + +### v0.19.0 + +- [Significantly reduced "lazy" coding from GPT-4 Turbo due to new unified diff edit format](https://aider.chat/docs/unified-diffs.html) + - Score improves from 20% to 61% on new "laziness benchmark". + - Aider now uses unified diffs by default for `gpt-4-1106-preview`. +- New `--4-turbo` command line switch as a shortcut for `--model gpt-4-1106-preview`. + +### v0.18.1 + +- Upgraded to new openai python client v1.3.7. + +### v0.18.0 + +- Improved prompting for both GPT-4 and GPT-4 Turbo. + - Far fewer edit errors from GPT-4 Turbo (`gpt-4-1106-preview`). + - Significantly better benchmark results from the June GPT-4 (`gpt-4-0613`). Performance leaps from 47%/64% up to 51%/71%. +- Fixed bug where in-chat files were marked as both read-only and ready-write, sometimes confusing GPT. +- Fixed bug to properly handle repos with submodules. + +### v0.17.0 + +- Support for OpenAI's new 11/06 models: + - gpt-4-1106-preview with 128k context window + - gpt-3.5-turbo-1106 with 16k context window +- [Benchmarks for OpenAI's new 11/06 models](https://aider.chat/docs/benchmarks-1106.html) +- Streamlined [API for scripting aider, added docs](https://aider.chat/docs/faq.html#can-i-script-aider) +- Ask for more concise SEARCH/REPLACE blocks. [Benchmarked](https://aider.chat/docs/benchmarks.html) at 63.9%, no regression. +- Improved repo-map support for elisp. +- Fixed crash bug when `/add` used on file matching `.gitignore` +- Fixed misc bugs to catch and handle unicode decoding errors. + +### v0.16.3 + +- Fixed repo-map support for C#. + +### v0.16.2 + +- Fixed docker image. + +### v0.16.1 + +- Updated tree-sitter dependencies to streamline the pip install process + +### v0.16.0 + +- [Improved repository map using tree-sitter](https://aider.chat/docs/repomap.html) +- Switched from "edit block" to "search/replace block", which reduced malformed edit blocks. [Benchmarked](https://aider.chat/docs/benchmarks.html) at 66.2%, no regression. +- Improved handling of malformed edit blocks targeting multiple edits to the same file. [Benchmarked](https://aider.chat/docs/benchmarks.html) at 65.4%, no regression. +- Bugfix to properly handle malformed `/add` wildcards. + + +### v0.15.0 + +- Added support for `.aiderignore` file, which instructs aider to ignore parts of the git repo. +- New `--commit` cmd line arg, which just commits all pending changes with a sensible commit message generated by gpt-3.5. +- Added universal ctags and multiple architectures to the [aider docker image](https://aider.chat/docs/docker.html) +- `/run` and `/git` now accept full shell commands, like: `/run (cd subdir; ls)` +- Restored missing `--encoding` cmd line switch. + +### v0.14.2 + +- Easily [run aider from a docker image](https://aider.chat/docs/docker.html) +- Fixed bug with chat history summarization. +- Fixed bug if `soundfile` package not available. + +### v0.14.1 + +- /add and /drop handle absolute filenames and quoted filenames +- /add checks to be sure files are within the git repo (or root) +- If needed, warn users that in-chat file paths are all relative to the git repo +- Fixed /add bug in when aider launched in repo subdir +- Show models supported by api/key if requested model isn't available + +### v0.14.0 + +- [Support for Claude2 and other LLMs via OpenRouter](https://aider.chat/docs/faq.html#accessing-other-llms-with-openrouter) by @joshuavial +- Documentation for [running the aider benchmarking suite](https://github.com/paul-gauthier/aider/tree/main/benchmark) +- Aider now requires Python >= 3.9 + + +### v0.13.0 + +- [Only git commit dirty files that GPT tries to edit](https://aider.chat/docs/faq.html#how-did-v0130-change-git-usage) +- Send chat history as prompt/context for Whisper voice transcription +- Added `--voice-language` switch to constrain `/voice` to transcribe to a specific language +- Late-bind importing `sounddevice`, as it was slowing down aider startup +- Improved --foo/--no-foo switch handling for command line and yml config settings + +### v0.12.0 + +- [Voice-to-code](https://aider.chat/docs/voice.html) support, which allows you to code with your voice. +- Fixed bug where /diff was causing crash. +- Improved prompting for gpt-4, refactor of editblock coder. +- [Benchmarked](https://aider.chat/docs/benchmarks.html) at 63.2% for gpt-4/diff, no regression. + +### v0.11.1 + +- Added a progress bar when initially creating a repo map. +- Fixed bad commit message when adding new file to empty repo. +- Fixed corner case of pending chat history summarization when dirty committing. +- Fixed corner case of undefined `text` when using `--no-pretty`. +- Fixed /commit bug from repo refactor, added test coverage. +- [Benchmarked](https://aider.chat/docs/benchmarks.html) at 53.4% for gpt-3.5/whole (no regression). + +### v0.11.0 + +- Automatically summarize chat history to avoid exhausting context window. +- More detail on dollar costs when running with `--no-stream` +- Stronger GPT-3.5 prompt against skipping/eliding code in replies (51.9% [benchmark](https://aider.chat/docs/benchmarks.html), no regression) +- Defend against GPT-3.5 or non-OpenAI models suggesting filenames surrounded by asterisks. +- Refactored GitRepo code out of the Coder class. + +### v0.10.1 + +- /add and /drop always use paths relative to the git root +- Encourage GPT to use language like "add files to the chat" to ask users for permission to edit them. + +### v0.10.0 + +- Added `/git` command to run git from inside aider chats. +- Use Meta-ENTER (Esc+ENTER in some environments) to enter multiline chat messages. +- Create a `.gitignore` with `.aider*` to prevent users from accidentaly adding aider files to git. +- Check pypi for newer versions and notify user. +- Updated keyboard interrupt logic so that 2 ^C in 2 seconds always forces aider to exit. +- Provide GPT with detailed error if it makes a bad edit block, ask for a retry. +- Force `--no-pretty` if aider detects it is running inside a VSCode terminal. +- [Benchmarked](https://aider.chat/docs/benchmarks.html) at 64.7% for gpt-4/diff (no regression) + + +### v0.9.0 + +- Support for the OpenAI models in [Azure](https://aider.chat/docs/faq.html#azure) +- Added `--show-repo-map` +- Improved output when retrying connections to the OpenAI API +- Redacted api key from `--verbose` output +- Bugfix: recognize and add files in subdirectories mentioned by user or GPT +- [Benchmarked](https://aider.chat/docs/benchmarks.html) at 53.8% for gpt-3.5-turbo/whole (no regression) + +### v0.8.3 + +- Added `--dark-mode` and `--light-mode` to select colors optimized for terminal background +- Install docs link to [NeoVim plugin](https://github.com/joshuavial/aider.nvim) by @joshuavial +- Reorganized the `--help` output +- Bugfix/improvement to whole edit format, may improve coding editing for GPT-3.5 +- Bugfix and tests around git filenames with unicode characters +- Bugfix so that aider throws an exception when OpenAI returns InvalidRequest +- Bugfix/improvement to /add and /drop to recurse selected directories +- Bugfix for live diff output when using "whole" edit format + +### v0.8.2 + +- Disabled general availability of gpt-4 (it's rolling out, not 100% available yet) + +### v0.8.1 + +- Ask to create a git repo if none found, to better track GPT's code changes +- Glob wildcards are now supported in `/add` and `/drop` commands +- Pass `--encoding` into ctags, require it to return `utf-8` +- More robust handling of filepaths, to avoid 8.3 windows filenames +- Added [FAQ](https://aider.chat/docs/faq.html) +- Marked GPT-4 as generally available +- Bugfix for live diffs of whole coder with missing filenames +- Bugfix for chats with multiple files +- Bugfix in editblock coder prompt + +### v0.8.0 + +- [Benchmark comparing code editing in GPT-3.5 and GPT-4](https://aider.chat/docs/benchmarks.html) +- Improved Windows support: + - Fixed bugs related to path separators in Windows + - Added a CI step to run all tests on Windows +- Improved handling of Unicode encoding/decoding + - Explicitly read/write text files with utf-8 encoding by default (mainly benefits Windows) + - Added `--encoding` switch to specify another encoding + - Gracefully handle decoding errors +- Added `--code-theme` switch to control the pygments styling of code blocks (by @kwmiebach) +- Better status messages explaining the reason when ctags is disabled + +### v0.7.2: + +- Fixed a bug to allow aider to edit files that contain triple backtick fences. + +### v0.7.1: + +- Fixed a bug in the display of streaming diffs in GPT-3.5 chats + +### v0.7.0: + +- Graceful handling of context window exhaustion, including helpful tips. +- Added `--message` to give GPT that one instruction and then exit after it replies and any edits are performed. +- Added `--no-stream` to disable streaming GPT responses. + - Non-streaming responses include token usage info. + - Enables display of cost info based on OpenAI advertised pricing. +- Coding competence benchmarking tool against suite of programming tasks based on Execism's python repo. + - https://github.com/exercism/python +- Major refactor in preparation for supporting new function calls api. +- Initial implementation of a function based code editing backend for 3.5. + - Initial experiments show that using functions makes 3.5 less competent at coding. +- Limit automatic retries when GPT returns a malformed edit response. + +### v0.6.2 + +* Support for `gpt-3.5-turbo-16k`, and all OpenAI chat models +* Improved ability to correct when gpt-4 omits leading whitespace in code edits +* Added `--openai-api-base` to support API proxies, etc. + +### v0.5.0 + +- Added support for `gpt-3.5-turbo` and `gpt-4-32k`. +- Added `--map-tokens` to set a token budget for the repo map, along with a PageRank based algorithm for prioritizing which files and identifiers to include in the map. +- Added in-chat command `/tokens` to report on context window token usage. +- Added in-chat command `/clear` to clear the conversation history. + diff --git a/website/_config.yml b/website/_config.yml index 74021c3fe..15e28ec7f 100644 --- a/website/_config.yml +++ b/website/_config.yml @@ -4,6 +4,7 @@ url: "https://aider.chat" plugins: - jekyll-redirect-from - jekyll-sitemap + - jekyll-feed defaults: - scope: @@ -19,6 +20,7 @@ exclude: - "**/OLD/**" - "OLD/**" - vendor + - feed.xml aux_links: "GitHub": diff --git a/website/_data/edit_leaderboard.yml b/website/_data/edit_leaderboard.yml index 5a4802184..c7c4a0efc 100644 --- a/website/_data/edit_leaderboard.yml +++ b/website/_data/edit_leaderboard.yml @@ -44,29 +44,6 @@ seconds_per_case: 23.1 total_cost: 0.0000 -- dirname: 2024-04-29-19-17-28--deepseek-coder-whole - test_cases: 132 - model: deepseek-coder - released: 2024-01-25 - edit_format: whole - commit_hash: c07f793-dirty - pass_rate_1: 47.0 - pass_rate_2: 54.5 - percent_cases_well_formed: 100.0 - error_outputs: 0 - num_malformed_responses: 0 - user_asks: 0 - lazy_comments: 2 - syntax_errors: 13 - indentation_errors: 0 - exhausted_context_windows: 0 - test_timeouts: 2 - command: aider --model deepseek/deepseek-coder - date: 2024-04-29 - versions: 0.30.2-dev - seconds_per_case: 26.7 - total_cost: 0.0000 - - dirname: 2024-05-03-20-47-24--gemini-1.5-pro-diff-fenced test_cases: 133 model: gemini-1.5-pro-latest @@ -611,4 +588,97 @@ date: 2024-06-08 versions: 0.37.1-dev seconds_per_case: 280.6 - total_cost: 0.0000 \ No newline at end of file + total_cost: 0.0000 + +- dirname: 2024-06-20-15-09-26--claude-3.5-sonnet-whole + test_cases: 133 + model: claude-3.5-sonnet (whole) + edit_format: whole + commit_hash: 068609e + pass_rate_1: 61.7 + pass_rate_2: 78.2 + percent_cases_well_formed: 100.0 + error_outputs: 4 + num_malformed_responses: 0 + num_with_malformed_responses: 0 + user_asks: 2 + lazy_comments: 0 + syntax_errors: 0 + indentation_errors: 0 + exhausted_context_windows: 0 + test_timeouts: 0 + command: aider --model openrouter/anthropic/claude-3.5-sonnet --edit-format whole + date: 2024-06-20 + versions: 0.38.1-dev + seconds_per_case: 15.4 + total_cost: 0.0000 + +- dirname: 2024-06-20-15-16-41--claude-3.5-sonnet-diff + test_cases: 133 + model: claude-3.5-sonnet (diff) + edit_format: diff + commit_hash: 068609e-dirty + pass_rate_1: 57.9 + pass_rate_2: 74.4 + percent_cases_well_formed: 97.0 + error_outputs: 48 + num_malformed_responses: 11 + num_with_malformed_responses: 4 + user_asks: 0 + lazy_comments: 0 + syntax_errors: 0 + indentation_errors: 0 + exhausted_context_windows: 0 + test_timeouts: 1 + command: aider --model openrouter/anthropic/claude-3.5-sonnet + date: 2024-06-20 + versions: 0.38.1-dev + seconds_per_case: 21.6 + total_cost: 0.0000 + +- dirname: 2024-06-17-14-45-54--deepseek-coder2-whole + test_cases: 133 + model: DeepSeek Coder V2 (whole) + edit_format: whole + commit_hash: ca8672b + pass_rate_1: 63.9 + pass_rate_2: 75.2 + percent_cases_well_formed: 100.0 + error_outputs: 1 + num_malformed_responses: 0 + num_with_malformed_responses: 0 + user_asks: 1 + lazy_comments: 0 + syntax_errors: 1 + indentation_errors: 0 + exhausted_context_windows: 0 + test_timeouts: 7 + command: aider --model deepseek/deepseek-coder + date: 2024-06-17 + versions: 0.38.1-dev + seconds_per_case: 21.1 + total_cost: 0.0537 + +- dirname: 2024-06-21-15-29-08--deepseek-coder2-diff-again3 + test_cases: 133 + model: DeepSeek Coder V2 (diff) + edit_format: diff + commit_hash: 515ab3e + pass_rate_1: 58.6 + pass_rate_2: 66.2 + percent_cases_well_formed: 98.5 + error_outputs: 23 + num_malformed_responses: 5 + num_with_malformed_responses: 2 + user_asks: 2 + lazy_comments: 0 + syntax_errors: 0 + indentation_errors: 1 + exhausted_context_windows: 0 + test_timeouts: 2 + command: aider --model deepseek/deepseek-coder + date: 2024-06-21 + versions: 0.39.1-dev + seconds_per_case: 30.2 + total_cost: 0.0857 + \ No newline at end of file diff --git a/website/_data/refactor_leaderboard.yml b/website/_data/refactor_leaderboard.yml index db4d3483f..11773ac39 100644 --- a/website/_data/refactor_leaderboard.yml +++ b/website/_data/refactor_leaderboard.yml @@ -143,4 +143,25 @@ seconds_per_case: 67.8 total_cost: 20.4889 - \ No newline at end of file + +- dirname: 2024-06-20-16-39-18--refac-claude-3.5-sonnet-diff + test_cases: 89 + model: claude-3.5-sonnet (diff) + edit_format: diff + commit_hash: e5e07f9 + pass_rate_1: 55.1 + percent_cases_well_formed: 70.8 + error_outputs: 240 + num_malformed_responses: 54 + num_with_malformed_responses: 26 + user_asks: 10 + lazy_comments: 2 + syntax_errors: 0 + indentation_errors: 3 + exhausted_context_windows: 0 + test_timeouts: 0 + command: aider --model openrouter/anthropic/claude-3.5-sonnet + date: 2024-06-20 + versions: 0.38.1-dev + seconds_per_case: 51.9 + total_cost: 0.0000 \ No newline at end of file diff --git a/website/_includes/get-started.md b/website/_includes/get-started.md index 498d310dc..e31d2ca3a 100644 --- a/website/_includes/get-started.md +++ b/website/_includes/get-started.md @@ -10,7 +10,12 @@ $ cd /to/your/git/repo $ export OPENAI_API_KEY=your-key-goes-here $ aider -# Or, work with Claude 3 Opus on your repo +# Or, work with Anthropic's models $ export ANTHROPIC_API_KEY=your-key-goes-here + +# Claude 3 Opus $ aider --opus + +# Claude 3.5 Sonnet +$ aider --sonnet ``` diff --git a/website/_includes/head_custom.html b/website/_includes/head_custom.html index a6d4cca22..95a83cf10 100644 --- a/website/_includes/head_custom.html +++ b/website/_includes/head_custom.html @@ -5,6 +5,7 @@ {% endif %} + diff --git a/website/_includes/special-keys.md b/website/_includes/special-keys.md new file mode 100644 index 000000000..ec5d59f9c --- /dev/null +++ b/website/_includes/special-keys.md @@ -0,0 +1,9 @@ +Aider has special support for providing +OpenAI and Anthropic API keys +via command line switches and yaml config settings. +*All other LLM providers* must +have their keys and settings +specified in environment variables. +This can be done in your shell, +or by using a +[`.env` file](/docs/dotenv.html). diff --git a/website/assets/sample.aider.conf.yml b/website/assets/sample.aider.conf.yml index a4ecb0928..c7873aecd 100644 --- a/website/assets/sample.aider.conf.yml +++ b/website/assets/sample.aider.conf.yml @@ -1,5 +1,6 @@ ########################################################## # Sample .aider.conf.yaml +# This file lists *all* the valid configuration entries. # Place in your home dir, or at the root of your git repo. ########################################################## @@ -12,8 +13,8 @@ ####### # Main: -## Use VI editing mode in the terminal (default: False) -#vim: false +## Log the conversation with the LLM to this file (for example, .aider.llm.history) +#llm-history-file: ## Specify the OpenAI API key #openai-api-key: @@ -27,7 +28,7 @@ ## Use claude-3-opus-20240229 model for the main chat #opus: false -## Use claude-3-sonnet-20240229 model for the main chat +## Use claude-3-5-sonnet-20240620 model for the main chat #sonnet: false ## Use gpt-4-0613 model for the main chat @@ -63,6 +64,9 @@ ## Specify the OpenAI organization ID #openai-organization-id: +## Verify the SSL cert when connecting to models (default: True) +#verify-ssl: true + ## Specify a file with context window and costs for unknown models #model-metadata-file: @@ -177,6 +181,9 @@ ################# # Other Settings: +## Use VI editing mode in the terminal (default: False) +#vim: false + ## Specify the language for voice using ISO 639-1 code (default: auto) #voice-language: en diff --git a/website/assets/sample.env b/website/assets/sample.env new file mode 100644 index 000000000..af76039dc --- /dev/null +++ b/website/assets/sample.env @@ -0,0 +1,229 @@ +########################################################## +# Sample aider .env file. +# Place at the root of your git repo. +# Or use `aider --env ` to specify. +########################################################## + +################# +# LLM parameters: +# +# Include xxx_API_KEY parameters and other params needed for your LLMs. +# See https://aider.chat/docs/llms.html for details. + +## OpenAI +#OPENAI_API_KEY= + +## Anthropic +#ANTHROPIC_API_KEY= + +##... + +####### +# Main: + +## Log the conversation with the LLM to this file (for example, .aider.llm.history) +#AIDER_LLM_HISTORY_FILE= + +## Specify the OpenAI API key +#OPENAI_API_KEY= + +## Specify the Anthropic API key +#ANTHROPIC_API_KEY= + +## Specify the model to use for the main chat (default: gpt-4o) +#AIDER_MODEL=gpt-4o + +## Use claude-3-opus-20240229 model for the main chat +#AIDER_OPUS= + +## Use claude-3-5-sonnet-20240620 model for the main chat +#AIDER_SONNET= + +## Use gpt-4-0613 model for the main chat +#AIDER_4= + +## Use gpt-4o model for the main chat +#AIDER_4O= + +## Use gpt-4-1106-preview model for the main chat +#AIDER_4_TURBO= + +## Use gpt-3.5-turbo model for the main chat +#AIDER_35TURBO= + +################# +# Model Settings: + +## List known models which match the (partial) MODEL name +#AIDER_MODELS= + +## Specify the api base url +#OPENAI_API_BASE= + +## Specify the api_type +#OPENAI_API_TYPE= + +## Specify the api_version +#OPENAI_API_VERSION= + +## Specify the deployment_id +#OPENAI_API_DEPLOYMENT_ID= + +## Specify the OpenAI organization ID +#OPENAI_ORGANIZATION_ID= + +## Verify the SSL cert when connecting to models (default: True) +#AIDER_VERIFY_SSL=true + +## Specify a file with context window and costs for unknown models +#AIDER_MODEL_METADATA_FILE= + +## Specify what edit format the LLM should use (default depends on model) +#AIDER_EDIT_FORMAT= + +## Specify the model to use for commit messages and chat history summarization (default depends on --model) +#AIDER_WEAK_MODEL= + +## Only work with models that have meta-data available (default: True) +#AIDER_SHOW_MODEL_WARNINGS=true + +## Max number of tokens to use for repo map, use 0 to disable (default: 1024) +#AIDER_MAP_TOKENS=true + +## Maximum number of tokens to use for chat history. If not specified, uses the model's max_chat_history_tokens. +#AIDER_MAX_CHAT_HISTORY_TOKENS= + +## Specify the .env file to load (default: .env in git root) +#AIDER_ENV_FILE=.env + +################ +# History Files: + +## Specify the chat input history file (default: .aider.input.history) +#AIDER_INPUT_HISTORY_FILE=.aider.input.history + +## Specify the chat history file (default: .aider.chat.history.md) +#AIDER_CHAT_HISTORY_FILE=.aider.chat.history.md + +## Restore the previous chat history messages (default: False) +#AIDER_RESTORE_CHAT_HISTORY=false + +################## +# Output Settings: + +## Use colors suitable for a dark terminal background (default: False) +#AIDER_DARK_MODE=false + +## Use colors suitable for a light terminal background (default: False) +#AIDER_LIGHT_MODE=false + +## Enable/disable pretty, colorized output (default: True) +#AIDER_PRETTY=true + +## Enable/disable streaming responses (default: True) +#AIDER_STREAM=true + +## Set the color for user input (default: #00cc00) +#AIDER_USER_INPUT_COLOR=#00cc00 + +## Set the color for tool output (default: None) +#AIDER_TOOL_OUTPUT_COLOR= + +## Set the color for tool error messages (default: red) +#AIDER_TOOL_ERROR_COLOR=#FF2222 + +## Set the color for assistant output (default: #0088ff) +#AIDER_ASSISTANT_OUTPUT_COLOR=#0088ff + +## Set the markdown code theme (default: default, other options include monokai, solarized-dark, solarized-light) +#AIDER_CODE_THEME=default + +## Show diffs when committing changes (default: False) +#AIDER_SHOW_DIFFS=false + +############### +# Git Settings: + +## Enable/disable looking for a git repo (default: True) +#AIDER_GIT=true + +## Enable/disable adding .aider* to .gitignore (default: True) +#AIDER_GITIGNORE=true + +## Specify the aider ignore file (default: .aiderignore in git root) +#AIDER_AIDERIGNORE=.aiderignore + +## Enable/disable auto commit of LLM changes (default: True) +#AIDER_AUTO_COMMITS=true + +## Enable/disable commits when repo is found dirty (default: True) +#AIDER_DIRTY_COMMITS=true + +## Perform a dry run without modifying files (default: False) +#AIDER_DRY_RUN=false + +######################## +# Fixing and committing: + +## Commit all pending changes with a suitable commit message, then exit +#AIDER_COMMIT=false + +## Lint and fix provided files, or dirty files if none provided +#AIDER_LINT=false + +## Specify lint commands to run for different languages, eg: "python: flake8 --select=..." (can be used multiple times) +#AIDER_LINT_CMD= + +## Enable/disable automatic linting after changes (default: True) +#AIDER_AUTO_LINT=true + +## Specify command to run tests +#AIDER_TEST_CMD= + +## Enable/disable automatic testing after changes (default: False) +#AIDER_AUTO_TEST=false + +## Run tests and fix problems found +#AIDER_TEST=false + +################# +# Other Settings: + +## Use VI editing mode in the terminal (default: False) +#AIDER_VIM=false + +## Specify the language for voice using ISO 639-1 code (default: auto) +#AIDER_VOICE_LANGUAGE=en + +## Check for updates and return status in the exit code +#AIDER_CHECK_UPDATE=false + +## Skips checking for the update when the program runs +#AIDER_SKIP_CHECK_UPDATE=false + +## Apply the changes from the given file instead of running the chat (debug) +#AIDER_APPLY= + +## Always say yes to every confirmation +#AIDER_YES= + +## Enable verbose output +#AIDER_VERBOSE=false + +## Print the repo map and exit (debug) +#AIDER_SHOW_REPO_MAP=false + +## Print the system prompts and exit (debug) +#AIDER_SHOW_PROMPTS=false + +## Specify a single message to send the LLM, process reply then exit (disables chat mode) +#AIDER_MESSAGE= + +## Specify a file containing the message to send the LLM, process reply, then exit (disables chat mode) +#AIDER_MESSAGE_FILE= + +## Specify the encoding for input and output (default: utf-8) +#AIDER_ENCODING=utf-8 + +## Run aider in your browser +#AIDER_GUI=false diff --git a/website/docs/aider_conf.md b/website/docs/aider_conf.md index 344ebc534..e6e89fc25 100644 --- a/website/docs/aider_conf.md +++ b/website/docs/aider_conf.md @@ -10,6 +10,8 @@ Most of aider's options can be set in an `.aider.conf.yml` file, which can be placed in your home directory or at the root of your git repo. +{% include special-keys.md %} + Below is a sample of the file, which you can also [download from GitHub](https://github.com/paul-gauthier/aider/blob/main/website/assets/sample.aider.conf.yml). @@ -26,6 +28,7 @@ cog.outl("```") ``` ########################################################## # Sample .aider.conf.yaml +# This file lists *all* the valid configuration entries. # Place in your home dir, or at the root of your git repo. ########################################################## @@ -38,8 +41,8 @@ cog.outl("```") ####### # Main: -## Use VI editing mode in the terminal (default: False) -#vim: false +## Log the conversation with the LLM to this file (for example, .aider.llm.history) +#llm-history-file: ## Specify the OpenAI API key #openai-api-key: @@ -53,7 +56,7 @@ cog.outl("```") ## Use claude-3-opus-20240229 model for the main chat #opus: false -## Use claude-3-sonnet-20240229 model for the main chat +## Use claude-3-5-sonnet-20240620 model for the main chat #sonnet: false ## Use gpt-4-0613 model for the main chat @@ -89,6 +92,9 @@ cog.outl("```") ## Specify the OpenAI organization ID #openai-organization-id: +## Verify the SSL cert when connecting to models (default: True) +#verify-ssl: true + ## Specify a file with context window and costs for unknown models #model-metadata-file: @@ -203,6 +209,9 @@ cog.outl("```") ################# # Other Settings: +## Use VI editing mode in the terminal (default: False) +#vim: false + ## Specify the language for voice using ISO 639-1 code (default: auto) #voice-language: en diff --git a/website/docs/dotenv.md b/website/docs/dotenv.md index 6620e2254..e1eb3a425 100644 --- a/website/docs/dotenv.md +++ b/website/docs/dotenv.md @@ -4,28 +4,263 @@ nav_order: 900 description: Using a .env file to store LLM API keys for aider. --- -# Storing LLM params in .env +# Config with .env You can use a `.env` file to store API keys and other settings for the models you use with aider. -You currently can not set general aider options -in the `.env` file, only LLM environment variables. +You can also set many general aider options +in the `.env` file. + +{% include special-keys.md %} Aider will look for a `.env` file in the root of your git repo or in the current directory. You can give it an explicit file to load with the `--env-file ` parameter. -Here is an example `.env` file: +Below is a sample `.env` file, which you +can also +[download from GitHub](https://github.com/paul-gauthier/aider/blob/main/website/assets/sample.env). + ``` -OPENAI_API_KEY= -ANTHROPIC_API_KEY= -GROQ_API_KEY= -OPENROUTER_API_KEY= +########################################################## +# Sample aider .env file. +# Place at the root of your git repo. +# Or use `aider --env ` to specify. +########################################################## -AZURE_API_KEY= -AZURE_API_VERSION=2023-05-15 -AZURE_API_BASE=https://example-endpoint.openai.azure.com +################# +# LLM parameters: +# +# Include xxx_API_KEY parameters and other params needed for your LLMs. +# See https://aider.chat/docs/llms.html for details. -OLLAMA_API_BASE=http://127.0.0.1:11434 +## OpenAI +#OPENAI_API_KEY= + +## Anthropic +#ANTHROPIC_API_KEY= + +##... + +####### +# Main: + +## Log the conversation with the LLM to this file (for example, .aider.llm.history) +#AIDER_LLM_HISTORY_FILE= + +## Specify the OpenAI API key +#OPENAI_API_KEY= + +## Specify the Anthropic API key +#ANTHROPIC_API_KEY= + +## Specify the model to use for the main chat (default: gpt-4o) +#AIDER_MODEL=gpt-4o + +## Use claude-3-opus-20240229 model for the main chat +#AIDER_OPUS= + +## Use claude-3-5-sonnet-20240620 model for the main chat +#AIDER_SONNET= + +## Use gpt-4-0613 model for the main chat +#AIDER_4= + +## Use gpt-4o model for the main chat +#AIDER_4O= + +## Use gpt-4-1106-preview model for the main chat +#AIDER_4_TURBO= + +## Use gpt-3.5-turbo model for the main chat +#AIDER_35TURBO= + +################# +# Model Settings: + +## List known models which match the (partial) MODEL name +#AIDER_MODELS= + +## Specify the api base url +#OPENAI_API_BASE= + +## Specify the api_type +#OPENAI_API_TYPE= + +## Specify the api_version +#OPENAI_API_VERSION= + +## Specify the deployment_id +#OPENAI_API_DEPLOYMENT_ID= + +## Specify the OpenAI organization ID +#OPENAI_ORGANIZATION_ID= + +## Verify the SSL cert when connecting to models (default: True) +#AIDER_VERIFY_SSL=true + +## Specify a file with context window and costs for unknown models +#AIDER_MODEL_METADATA_FILE= + +## Specify what edit format the LLM should use (default depends on model) +#AIDER_EDIT_FORMAT= + +## Specify the model to use for commit messages and chat history summarization (default depends on --model) +#AIDER_WEAK_MODEL= + +## Only work with models that have meta-data available (default: True) +#AIDER_SHOW_MODEL_WARNINGS=true + +## Max number of tokens to use for repo map, use 0 to disable (default: 1024) +#AIDER_MAP_TOKENS=true + +## Maximum number of tokens to use for chat history. If not specified, uses the model's max_chat_history_tokens. +#AIDER_MAX_CHAT_HISTORY_TOKENS= + +## Specify the .env file to load (default: .env in git root) +#AIDER_ENV_FILE=.env + +################ +# History Files: + +## Specify the chat input history file (default: .aider.input.history) +#AIDER_INPUT_HISTORY_FILE=.aider.input.history + +## Specify the chat history file (default: .aider.chat.history.md) +#AIDER_CHAT_HISTORY_FILE=.aider.chat.history.md + +## Restore the previous chat history messages (default: False) +#AIDER_RESTORE_CHAT_HISTORY=false + +################## +# Output Settings: + +## Use colors suitable for a dark terminal background (default: False) +#AIDER_DARK_MODE=false + +## Use colors suitable for a light terminal background (default: False) +#AIDER_LIGHT_MODE=false + +## Enable/disable pretty, colorized output (default: True) +#AIDER_PRETTY=true + +## Enable/disable streaming responses (default: True) +#AIDER_STREAM=true + +## Set the color for user input (default: #00cc00) +#AIDER_USER_INPUT_COLOR=#00cc00 + +## Set the color for tool output (default: None) +#AIDER_TOOL_OUTPUT_COLOR= + +## Set the color for tool error messages (default: red) +#AIDER_TOOL_ERROR_COLOR=#FF2222 + +## Set the color for assistant output (default: #0088ff) +#AIDER_ASSISTANT_OUTPUT_COLOR=#0088ff + +## Set the markdown code theme (default: default, other options include monokai, solarized-dark, solarized-light) +#AIDER_CODE_THEME=default + +## Show diffs when committing changes (default: False) +#AIDER_SHOW_DIFFS=false + +############### +# Git Settings: + +## Enable/disable looking for a git repo (default: True) +#AIDER_GIT=true + +## Enable/disable adding .aider* to .gitignore (default: True) +#AIDER_GITIGNORE=true + +## Specify the aider ignore file (default: .aiderignore in git root) +#AIDER_AIDERIGNORE=.aiderignore + +## Enable/disable auto commit of LLM changes (default: True) +#AIDER_AUTO_COMMITS=true + +## Enable/disable commits when repo is found dirty (default: True) +#AIDER_DIRTY_COMMITS=true + +## Perform a dry run without modifying files (default: False) +#AIDER_DRY_RUN=false + +######################## +# Fixing and committing: + +## Commit all pending changes with a suitable commit message, then exit +#AIDER_COMMIT=false + +## Lint and fix provided files, or dirty files if none provided +#AIDER_LINT=false + +## Specify lint commands to run for different languages, eg: "python: flake8 --select=..." (can be used multiple times) +#AIDER_LINT_CMD= + +## Enable/disable automatic linting after changes (default: True) +#AIDER_AUTO_LINT=true + +## Specify command to run tests +#AIDER_TEST_CMD= + +## Enable/disable automatic testing after changes (default: False) +#AIDER_AUTO_TEST=false + +## Run tests and fix problems found +#AIDER_TEST=false + +################# +# Other Settings: + +## Use VI editing mode in the terminal (default: False) +#AIDER_VIM=false + +## Specify the language for voice using ISO 639-1 code (default: auto) +#AIDER_VOICE_LANGUAGE=en + +## Check for updates and return status in the exit code +#AIDER_CHECK_UPDATE=false + +## Skips checking for the update when the program runs +#AIDER_SKIP_CHECK_UPDATE=false + +## Apply the changes from the given file instead of running the chat (debug) +#AIDER_APPLY= + +## Always say yes to every confirmation +#AIDER_YES= + +## Enable verbose output +#AIDER_VERBOSE=false + +## Print the repo map and exit (debug) +#AIDER_SHOW_REPO_MAP=false + +## Print the system prompts and exit (debug) +#AIDER_SHOW_PROMPTS=false + +## Specify a single message to send the LLM, process reply then exit (disables chat mode) +#AIDER_MESSAGE= + +## Specify a file containing the message to send the LLM, process reply, then exit (disables chat mode) +#AIDER_MESSAGE_FILE= + +## Specify the encoding for input and output (default: utf-8) +#AIDER_ENCODING=utf-8 + +## Run aider in your browser +#AIDER_GUI=false ``` + + + diff --git a/website/docs/faq.md b/website/docs/faq.md index a1443b1e8..f32872828 100644 --- a/website/docs/faq.md +++ b/website/docs/faq.md @@ -1,9 +1,9 @@ --- -nav_order: 85 +nav_order: 90 description: Frequently asked questions about aider. --- -# Frequently asked questions +# FAQ {: .no_toc } - TOC diff --git a/website/docs/git.md b/website/docs/git.md index b25418b93..ea2b145fe 100644 --- a/website/docs/git.md +++ b/website/docs/git.md @@ -1,4 +1,5 @@ --- +parent: More info nav_order: 800 description: Aider is tightly integrated with git. --- @@ -14,14 +15,21 @@ Aider is tightly integrated with git, which makes it easy to: Aider specifically uses git in these ways: - - It asks to create a git repo if you launch it in a directory without one. - - Whenever aider edits a file, it commits those changes with a descriptive commit message. This makes it easy to undo or review aider's changes. - - Aider takes special care before editing files that already have uncommitted changes (dirty files). Aider will first commit any preexisting changes with a descriptive commit message. This keeps your edits separate from aider's edits, and makes sure you never lose your work if aider makes an inappropriate change. +- It asks to create a git repo if you launch it in a directory without one. +- Whenever aider edits a file, it commits those changes with a descriptive commit message. This makes it easy to undo or review aider's changes. +These commits will have "(aider)" appended to their git author and git committer metadata. +- Aider takes special care before editing files that already have uncommitted changes (dirty files). Aider will first commit any preexisting changes with a descriptive commit message. +This keeps your edits separate from aider's edits, and makes sure you never lose your work if aider makes an inappropriate change. +These commits will have "(aider)" appended to their git committer metadata. + +## In-chat commands Aider also allows you to use in-chat commands to `/diff` or `/undo` the last change. To do more complex management of your git history, you cat use raw `git` commands, either by using `/git` within the chat, or with standard git tools outside of aider. +## Disabling git integration + While it is not recommended, you can disable aider's use of git in a few ways: - `--no-auto-commits` will stop aider from git committing each of its changes. diff --git a/website/docs/languages.md b/website/docs/languages.md index 889e4caa9..d74f748c9 100644 --- a/website/docs/languages.md +++ b/website/docs/languages.md @@ -1,4 +1,5 @@ --- +parent: More info nav_order: 900 description: Aider supports pretty much all popular coding languages. --- diff --git a/website/docs/leaderboards/index.md b/website/docs/leaderboards/index.md index 8d0f189e7..78d2abafc 100644 --- a/website/docs/leaderboards/index.md +++ b/website/docs/leaderboards/index.md @@ -16,6 +16,20 @@ While [aider can connect to almost any LLM](/docs/llms.html), it works best with models that score well on the benchmarks. +## Claude 3.5 Sonnet takes the top spot + +Claude 3.5 Sonnet is now the top ranked model on aider's code editing leaderboard. +DeepSeek Coder V2 only spent 4 days in the top spot. + +The new Sonnet came in 3rd on aider's refactoring leaderboard, behind GPT-4o and Opus. + +Sonnet ranked #1 when using the "whole" editing format, +but it also scored very well with +aider's "diff" editing format. +This format allows it to return code changes as diffs -- saving time and token costs, +and making it practical to work with larger source files. +As such, aider uses "diff" by default with this new Sonnet model. + ## Code editing leaderboard [Aider's code editing benchmark](/docs/benchmarks.html#the-benchmark) asks the LLM to edit python source files to complete 133 small coding exercises. This benchmark measures the LLM's coding ability, but also whether it can consistently emit code edits in the format specified in the system prompt. diff --git a/website/docs/llms.md b/website/docs/llms.md index e2cdf7ba6..37cd82f7a 100644 --- a/website/docs/llms.md +++ b/website/docs/llms.md @@ -10,22 +10,27 @@ description: Aider can connect to most LLMs for AI pair programming. [![connecting to many LLMs](/assets/llms.jpg)](https://aider.chat/assets/llms.jpg) + ## Best models {: .no_toc } -**Aider works best with [GPT-4o](/docs/llms/openai.html) and -[Claude 3 Opus](/docs/llms/anthropic.html),** -as they are the very best models for editing code. +Aider works best with these models, which are skilled at editing code: + +- [GPT-4o](/docs/llms/openai.html) +- [Claude 3.5 Sonnet](/docs/llms/anthropic.html) +- [Claude 3 Opus](/docs/llms/anthropic.html) +- [DeepSeek Coder V2](/docs/llms/deepseek.html) + ## Free models {: .no_toc } Aider works with a number of **free** API providers: -- Google's [Gemini 1.5 Pro](/docs/llms/gemini.html) is the most capable free model to use with aider, with +- The [DeepSeek Coder V2](/docs/llms/deepseek.html) model gets the top score on aider's code editing benchmark. DeepSeek has been offering 5M free tokens of API usage. +- Google's [Gemini 1.5 Pro](/docs/llms/gemini.html) works with aider, with code editing capabilities similar to GPT-3.5. - You can use [Llama 3 70B on Groq](/docs/llms/groq.html) which is comparable to GPT-3.5 in code editing performance. -- The [Deepseek Chat v2](/docs/llms/deepseek.html) model work well with aider, better than GPT-3.5. Deepseek currently offers 5M free tokens of API usage. - Cohere also offers free API access to their [Command-R+ model](/docs/llms/cohere.html), which works with aider as a *very basic* coding assistant. ## Local models diff --git a/website/docs/llms/anthropic.md b/website/docs/llms/anthropic.md index c2d2ff78f..027feef44 100644 --- a/website/docs/llms/anthropic.md +++ b/website/docs/llms/anthropic.md @@ -22,7 +22,7 @@ setx ANTHROPIC_API_KEY # Windows # Claude 3 Opus aider --opus -# Claude 3 Sonnet +# Claude 3.5 Sonnet aider --sonnet # List models available from Anthropic diff --git a/website/docs/llms/azure.md b/website/docs/llms/azure.md index ada2beb8d..55aa791f4 100644 --- a/website/docs/llms/azure.md +++ b/website/docs/llms/azure.md @@ -25,3 +25,6 @@ aider --model azure/ # List models available from Azure aider --models azure/ ``` + +Note that aider will also use environment variables +like `AZURE_OPENAI_API_xxx`. diff --git a/website/docs/llms/deepseek.md b/website/docs/llms/deepseek.md index af086a99d..7ff93ca88 100644 --- a/website/docs/llms/deepseek.md +++ b/website/docs/llms/deepseek.md @@ -3,10 +3,11 @@ parent: Connecting to LLMs nav_order: 500 --- -# Deepseek +# DeepSeek -Aider can connect to the Deepseek.com API. -Deepseek appears to grant 5M tokens of free API usage to new accounts. +Aider can connect to the DeepSeek.com API. +The DeepSeek Coder V2 model gets the top score on aider's code editing benchmark. +DeepSeek appears to grant 5M tokens of free API usage to new accounts. ``` pip install aider-chat @@ -14,8 +15,8 @@ pip install aider-chat export DEEPSEEK_API_KEY= # Mac/Linux setx DEEPSEEK_API_KEY # Windows -# Use Deepseek Chat v2 -aider --model deepseek/deepseek-chat +# Use DeepSeek Coder V2 +aider --model deepseek/deepseek-coder ``` See the [model warnings](warnings.html) diff --git a/website/docs/more-info.md b/website/docs/more-info.md new file mode 100644 index 000000000..3b40cb9e1 --- /dev/null +++ b/website/docs/more-info.md @@ -0,0 +1,8 @@ +--- +has_children: true +nav_order: 85 +--- + +# More info + +See below for more info about aider, including some advanced topics. diff --git a/website/docs/options.md b/website/docs/options.md index 69f4d7e7d..204c68337 100644 --- a/website/docs/options.md +++ b/website/docs/options.md @@ -21,6 +21,15 @@ usage: aider [-h] [--vim] [--openai-api-key] [--anthropic-api-key] [--openai-api-deployment-id] [--openai-organization-id] [--model-settings-file] [--model-metadata-file] [--edit-format] [--weak-model] +======= +usage: aider [-h] [--llm-history-file] [--openai-api-key] + [--anthropic-api-key] [--model] [--opus] [--sonnet] + [--4] [--4o] [--4-turbo] [--35turbo] [--models] + [--openai-api-base] [--openai-api-type] + [--openai-api-version] [--openai-api-deployment-id] + [--openai-organization-id] + [--verify-ssl | --no-verify-ssl] + [--model-metadata-file] [--edit-format] [--weak-model] [--show-model-warnings | --no-show-model-warnings] [--map-tokens] [--max-chat-history-tokens] [--env-file] [--input-history-file] [--chat-history-file] @@ -36,7 +45,7 @@ usage: aider [-h] [--vim] [--openai-api-key] [--anthropic-api-key] [--dry-run | --no-dry-run] [--commit] [--lint] [--lint-cmd] [--auto-lint | --no-auto-lint] [--test-cmd] [--auto-test | --no-auto-test] [--test] - [--voice-language] [--version] [--check-update] + [--vim] [--voice-language] [--version] [--check-update] [--skip-check-update] [--apply] [--yes] [-v] [--show-repo-map] [--show-prompts] [--message] [--message-file] [--encoding] [-c] [--gui] @@ -53,10 +62,9 @@ Aliases: ## Main: -### `--vim` -Use VI editing mode in the terminal (default: False) -Default: False -Environment variable: `AIDER_VIM` +### `--llm-history-file LLM_HISTORY_FILE` +Log the conversation with the LLM to this file (for example, .aider.llm.history) +Environment variable: `AIDER_LLM_HISTORY_FILE` ### `--openai-api-key OPENAI_API_KEY` Specify the OpenAI API key @@ -76,7 +84,7 @@ Use claude-3-opus-20240229 model for the main chat Environment variable: `AIDER_OPUS` ### `--sonnet` -Use claude-3-sonnet-20240229 model for the main chat +Use claude-3-5-sonnet-20240620 model for the main chat Environment variable: `AIDER_SONNET` ### `--4` @@ -132,6 +140,14 @@ Environment variable: `OPENAI_ORGANIZATION_ID` ### `--model-settings-file MODEL_FILE` Specify a file with aider model settings for unknown models Environment variable: `AIDER_MODEL_SETTINGS_FILE` +======= +### `--verify-ssl` +Verify the SSL cert when connecting to models (default: True) +Default: True +Environment variable: `AIDER_VERIFY_SSL` +Aliases: + - `--verify-ssl` + - `--no-verify-ssl` ### `--model-metadata-file MODEL_FILE` Specify a file with context window and costs for unknown models @@ -336,6 +352,11 @@ Environment variable: `AIDER_TEST` ## Other Settings: +### `--vim` +Use VI editing mode in the terminal (default: False) +Default: False +Environment variable: `AIDER_VIM` + ### `--voice-language VOICE_LANGUAGE` Specify the language for voice using ISO 639-1 code (default: auto) Default: en diff --git a/website/docs/repomap.md b/website/docs/repomap.md index 57a0cc415..359243103 100644 --- a/website/docs/repomap.md +++ b/website/docs/repomap.md @@ -1,4 +1,5 @@ --- +parent: More info highlight_image: /assets/robot-ast.png nav_order: 900 description: Aider uses a map of your git repository to provide code context to LLMs. @@ -12,23 +13,22 @@ Aider uses a **concise map of your whole git repository** that includes the most important classes and functions along with their types and call signatures. -This helps the LLM understand the code it needs to change, +This helps aider understand the code it's editing and how it relates to the other parts of the codebase. -The repo map also helps the LLM write new code +The repo map also helps aider write new code that respects and utilizes existing libraries, modules and abstractions found elsewhere in the codebase. ## Using a repo map to provide context Aider sends a **repo map** to the LLM along with -each request from the user to make a code change. -The map contains a list of the files in the +each change request from the user. +The repo map contains a list of the files in the repo, along with the key symbols which are defined in each file. -It shows how each of these symbols are defined in the -source code, by including the critical lines of code for each definition. +It shows how each of these symbols are defined, by including the critical lines of code for each definition. -Here's a -sample of the map of the aider repo, just showing the maps of +Here's a part of +the repo map of aider's repo, for [base_coder.py](https://github.com/paul-gauthier/aider/blob/main/aider/coders/base_coder.py) and [commands.py](https://github.com/paul-gauthier/aider/blob/main/aider/commands.py) @@ -71,7 +71,7 @@ aider/commands.py: Mapping out the repo like this provides some key benefits: - The LLM can see classes, methods and function signatures from everywhere in the repo. This alone may give it enough context to solve many tasks. For example, it can probably figure out how to use the API exported from a module just based on the details shown in the map. - - If it needs to see more code, the LLM can use the map to figure out by itself which files it needs to look at in more detail. The LLM will then ask to see these specific files, and aider will automatically add them to the chat context. + - If it needs to see more code, the LLM can use the map to figure out which files it needs to look at. The LLM can ask to see these specific files, and aider will offer to add them to the chat context. ## Optimizing the map diff --git a/website/docs/scripting.md b/website/docs/scripting.md index 2d6149d31..984ecb864 100644 --- a/website/docs/scripting.md +++ b/website/docs/scripting.md @@ -1,4 +1,5 @@ --- +parent: More info nav_order: 900 description: You can script aider via the command line or python. --- diff --git a/website/docs/troubleshooting/edit-errors.md b/website/docs/troubleshooting/edit-errors.md index 7c7a4a359..dbbd527dd 100644 --- a/website/docs/troubleshooting/edit-errors.md +++ b/website/docs/troubleshooting/edit-errors.md @@ -33,8 +33,8 @@ so editing errors are probably unavoidable. ## Reduce distractions Many LLM now have very large context windows, -but filling them with irrelevant code often -cofuses the model. +but filling them with irrelevant code or conversation +can cofuse the model. - Don't add too many files to the chat, *just* add the files you think need to be edited. Aider also sends the LLM a [map of your entire git repo](https://aider.chat/docs/repomap.html), so other relevant code will be included automatically. diff --git a/website/docs/troubleshooting/token-limits.md b/website/docs/troubleshooting/token-limits.md new file mode 100644 index 000000000..dfd624549 --- /dev/null +++ b/website/docs/troubleshooting/token-limits.md @@ -0,0 +1,86 @@ +--- +parent: Troubleshooting +nav_order: 25 +--- + +# Token limits + +Every LLM has limits on how many tokens it can process for each request: + +- The model's **context window** limits how many total tokens of +*input and output* it can process. +- Each model has limit on how many **output tokens** it can +produce. + +Aider will report an error if a model responds indicating that +it has exceeded a token limit. +The error will include suggested actions to try and +avoid hitting token limits. +Here's an example error: + +``` +Model gpt-3.5-turbo has hit a token limit! + +Input tokens: 768 of 16385 +Output tokens: 4096 of 4096 -- exceeded output limit! +Total tokens: 4864 of 16385 + +To reduce output tokens: +- Ask for smaller changes in each request. +- Break your code into smaller source files. +- Try using a stronger model like gpt-4o or opus that can return diffs. + +For more info: https://aider.chat/docs/token-limits.html +``` + +## Input tokens & context window size + +The most common problem is trying to send too much data to a +model, +overflowing its context window. +Technically you can exhaust the context window if the input is +too large or if the input plus output are too large. + +Strong models like GPT-4o and Opus have quite +large context windows, so this sort of error is +typically only an issue when working with weaker models. + +The easiest solution is to try and reduce the input tokens +by removing files from the chat. +It's best to only add the files that aider will need to *edit* +to complete your request. + +- Use `/tokens` to see token usage. +- Use `/drop` to remove unneeded files from the chat session. +- Use `/clear` to clear the chat history. +- Break your code into smaller source files. + +## Output token limits + +Most models have quite small output limits, often as low +as 4k tokens. +If you ask aider to make a large change that affects a lot +of code, the LLM may hit output token limits +as it tries to send back all the changes. + +To avoid hitting output token limits: + +- Ask for smaller changes in each request. +- Break your code into smaller source files. +- Try using a stronger model like gpt-4o or opus that can return diffs. + + +## Other causes + +Sometimes token limit errors are caused by +non-compliant API proxy servers +or bugs in the API server you are using to host a local model. +Aider has been well tested when directly connecting to +major +[LLM provider cloud APIs](https://aider.chat/docs/llms.html). +For serving local models, +[Ollama](https://aider.chat/docs/llms/ollama.html) is known to work well with aider. + +Try using aider without an API proxy server +or directly with one of the recommended cloud APIs +and see if your token limit problems resolve. diff --git a/website/docs/usage.md b/website/docs/usage.md index e0406fe2d..7451f9b2a 100644 --- a/website/docs/usage.md +++ b/website/docs/usage.md @@ -10,11 +10,31 @@ Run `aider` with the source code files you want to edit. These files will be "added to the chat session", so that aider can see their contents and edit them for you. +They can be existing files or the name of files you want +aider to create for you. ``` aider ... ``` +At the aider `>` prompt, ask for code changes and aider +will edit those files to accomplish your request. + + +``` +$ aider factorial.py + +Aider v0.37.1-dev +Models: gpt-4o with diff edit format, weak model gpt-3.5-turbo +Git repo: .git with 258 files +Repo-map: using 1024 tokens +Use /help to see in-chat commands, run with --help to see cmd line args +─────────────────────────────────────────────────────────────────────── +>`Make a program that asks for a number and prints its factorial + +... +``` + ## Adding files Just add the files that the aider will need to *edit*. diff --git a/website/index.md b/website/index.md index 0f0865f68..cbf8498df 100644 --- a/website/index.md +++ b/website/index.md @@ -4,16 +4,24 @@ nav_order: 1 --- + + # Aider is AI pair programming in your terminal Aider lets you pair program with LLMs, to edit code in your local git repository. Start a new project or work with an existing git repo. -Aider works best with GPT-4o and Claude 3 Opus -and can [connect to almost any LLM](https://aider.chat/docs/llms.html). +Aider can [connect to almost any LLM](https://aider.chat/docs/llms.html). +and works best with GPT-4o, Claude 3.5 Sonnet, Claude 3 Opus and DeepSeek Coder V2.

## Getting started + +You can get started quickly like this: -{% include get-started.md %} +``` +$ pip install aider-chat -**See the +# Change directory into a git repo +$ cd /to/your/git/repo + +# Work with GPT-4o on your repo +$ export OPENAI_API_KEY=your-key-goes-here +$ aider + +# Or, work with Anthropic's models +$ export ANTHROPIC_API_KEY=your-key-goes-here + +# Claude 3 Opus +$ aider --opus + +# Claude 3.5 Sonnet +$ aider --sonnet +``` + + +See the [installation instructions](https://aider.chat/docs/install.html) and other [documentation](https://aider.chat/docs/usage.html) -for more details.** +for more details. ## Features -- Chat with aider about your code: `aider ...` +- Run aider with the files you want to edit: `aider ...` - Ask for changes: - - New features, test cases, improvements. - - Bug fixes, updated docs or code refactors. - - Paste in a GitHub issue that needs to be solved. + - Add new features or test cases. + - Describe a bug. + - Paste in an error message or or GitHub issue URL. + - Refactor code. + - Update docs. - Aider will edit your files to complete your request. - Aider [automatically git commits](https://aider.chat/docs/git.html) changes with a sensible commit message. - Aider works with [most popular languages](https://aider.chat/docs/languages.html): python, javascript, typescript, php, html, css, and more... - Aider works best with GPT-4o and Claude 3 Opus and can [connect to almost any LLM](https://aider.chat/docs/llms.html). -- Aider can make coordinated changes across multiple files at once. +- Aider can edit multiple files at once for complex requests. - Aider uses a [map of your entire git repo](https://aider.chat/docs/repomap.html), which helps it work well in larger codebases. -- You can also edit files in your editor while chatting with aider. -Aider will notice and always use the latest version. -So you can bounce back and forth between aider and your editor, to collaboratively code with AI. -- Images can be added to the chat (GPT-4o, GPT-4 Turbo, etc). -- URLs can be added to the chat and aider will read their content. -- [Code with your voice](https://aider.chat/docs/voice.html) using speech recognition. +- Edit files in your editor while chatting with aider, +and it will always use the latest version. +Pair program with AI. +- Add images to the chat (GPT-4o, GPT-4 Turbo, etc). +- Add URLs to the chat and aider will read their content. +- [Code with your voice](https://aider.chat/docs/voice.html). -## State of the art +## Top tier performance -Aider has the -[top score on SWE Bench](https://aider.chat/2024/06/02/main-swe-bench.html). +[Aider has the one of the top scores on SWE Bench](https://aider.chat/2024/06/02/main-swe-bench.html). SWE Bench is a challenging software engineering benchmark where aider solved *real* GitHub issues from popular open source projects like django, scikitlearn, matplotlib, etc. -

- - aider swe bench - -

- -## Documentation +## More info +- [Documentation](https://aider.chat/) - [Installation](https://aider.chat/docs/install.html) - [Usage](https://aider.chat/docs/usage.html) - [Tutorial videos](https://aider.chat/docs/tutorials.html) - [Connecting to LLMs](https://aider.chat/docs/llms.html) +- [Configuration](https://aider.chat/docs/config.html) +- [Troubleshooting](https://aider.chat/docs/troubleshooting.html) - [LLM Leaderboards](https://aider.chat/docs/leaderboards/) -- [FAQ](https://aider.chat/docs/faq.html) +- [GitHub](https://github.com/paul-gauthier/aider) - [Discord](https://discord.gg/Tv2uQnR88V) - [Blog](https://aider.chat/blog/)