Xingyao Wang 31b244f95e
[Refactor, Evaluation] Refactor and clean up evaluation harness to remove global config and use EventStreamRuntime (#3230)
* move multi-line bash tests to test_runtime;
support multi-line bash for esruntime;

* add testcase to handle PS2 prompt

* use bashlex for bash parsing to handle multi-line commands;
add testcases for multi-line commands

* revert ghcr runtime change

* Apply stash

* fix run as other user;
make test async;

* fix test runtime for run as od

* add run-as-devin to all the runtime tests

* handle the case when username is root

* move all run-as-devin tests from sandbox;
only tests a few cases on different user to save time;

* move over multi-line echo related tests to test_runtime

* fix user-specific jupyter by fixing the pypoetry virtualenv folder

* make plugin's init async;
chdir at initialization of jupyter plugin;
move ipy simple testcase to test runtime;

* support agentskills import in
move tests for jupyter pwd tests;
overload `add_env_vars` for EventStreamRuntime to update env var also in Jupyter;
make agentskills read env var lazily, in case env var is updated;

* fix ServerRuntime agentskills issue

* move agnostic image test to test_runtime

* merge runtime tests in CI

* fix enable auto lint as env var

* update warning message

* update warning message

* test for different container images

* change parsing output as debug

* add exception handling for update_pwd_decorator

* fix unit test indentation

* add plugins as default input to Runtime class;
remove init_sandbox_plugins;
implement add_env_var (include jupyter) in the base class;

* fix server runtime auto lint

* Revert "add exception handling for update_pwd_decorator"

This reverts commit 2b668b1506e02145cb8f87e321aad62febca3d50.

* tries to print debugging info for agentskills

* explictly setting uid (try fix permission issue)

* Revert "tries to print debugging info for agentskills"

This reverts commit 8be4c86756f0e3fc62957b327ba2ac4999c419de.

* set sandbox user id during testing to hopefully fix the permission issue

* add browser tools for server runtime

* try to debug for old pwd

* update debug cmd

* only test agnostic runtime when TEST_RUNTIME is Server

* fix temp dir mkdir

* load TEST_RUNTIME at the beginning

* remove ipython tests

* only log to file when DEBUG

* default logging to project root

* temporarily remove log to file

* fix LLM logger dir

* fix logger

* make set pwd an optional aux action

* fix prev pwd

* fix infinity recursion

* simplify

* do not import the whole od library to avoid logger folder by jupyter

* fix browsing

* increase timeout

* attempt to fix agentskills yet again

* clean up in testcases, since CI maybe run as non-root

* add _cause attribute for event.id

* remove parent

* add a bunch of debugging statement again for CI :(

* fix temp_dir fixture

* change all temp dir to follow pytest's tmp_path_factory

* remove extra bracket

* clean up error printing a bit

* jupyter chdir to self.config.workspace_mount_path_in_sandbox on initialization

* jupyter chdir to self.config.workspace_mount_path_in_sandbox on initialization

* add typing for tmp dir fixture

* clear the directory before running the test to avoid weird CI temp dir

* remove agnostic test case for server runtime

* Revert "remove agnostic test case for server runtime"

This reverts commit 30e2181c3fc1410e69596c2dcd06be01f1d016b3.

* disable agnostic tests in CI

* fix test

* make sure plugin arg is not passed when no plugin is specified;
remove redundant on_event function;

* move mock prompt

* rename runtime

* remove extra logging

* refactor run_controller's interface;
support multiple runtime for integration test;
filter out hostname for prompt

* uncomment other tests

* pass the right runtime to controller

* log runtime when start

* uncomment tests

* improve symbol filters

* add intergration test prompts that seemd ok

* add integration test workflow

* add python3 to default ubuntu image

* symlink python and fix permission to jupyter pip

* add retry for jupyter execute server

* fix jupyter pip install;
add post-process for jupyter pip install;
simplify init by add agent_skills path to PYTHONPATH;
add testcase to tests jupyter pip install;

* fix bug

* use ubuntu:22.04 for eventstream integration tests

* add todo

* update testcase

* remove redundant code

* fix unit test

* reduce dependency for runtime

* try making llama-index an optional dependency that's not installed by default

* remove pip install since it seemd not needed

* log ipython execution;
await write message since it returns a future

* update ipy testcase

* do not install llama-index in CI

* do not install llama-index in the app docker as well

* set sandbox container image in the integration test script

* log plugins & env var for runtime

* update conftest for sha256

* add git

* remove all non-alphanumeric chalracters

* add working ipy module tests!

* default to use host network

* remove is_async from browser to make thing a little more reliable;
retry loading browser when error;

* add sleep to wait a bit for http server

* kill http server before regenerate browsing tests

* fix browsing

* only set sandbox container image if undefined

* skip empty config value

* update evaluation to use the latest run_controller

* revert logger in execute_server to be compatible with server runtime

* revert logging level to fix jupyter

* set logger level

* revert the logging

* chmod for workspace to fix permission

* support getting timeout from action

* update test for server runtime

* try to fix file permission

* fix test_cmd_run_action_serialization_deserialization test (added timeout)

* poetry: pip 24.2, torch 2.2.2

* revert adding pip to pyproject.toml

* add build to dependencies in pyproject.toml

* forgot poetry lock --no-update

* fix a DelegatorAgent prompt_002.log (timeout)

* fix a DelegatorAgent prompt_003.log (timeout)

* couple more timeout attribs in prompt files

* some more prompt files

* prompts galore

* add clarification comment for timeout

* default timeout to config

* add assert

* update integraton tests for eventstream

* update integration tests

* fix timeout for action<->dict

* remove redundant on_event

* default to use instance image

* update run_controller interface

* add logging for copy

* refactor swe_bench for the new design

* fix action execution timeout

* updatelock

* remove build sandbox locally

* fix runtime

* use plain for-loop for single process

* remove extra print

* get swebench inference working

* print whole `test_result` dict

* got swebench patch post-process working

* update swe-bench evaluation readme

* refactor using shared reset_logger function

* move messy swebench prompt to a different file

* support the ability to specify whether to keep prompt

* support the ability to specify whether to keep prompt

* fix dockerfile

* fix import and remove unnecessary strip logic

* fix action serialization

* get agentbench running

* remove extra ls for agent bench

* fix agentbench metric

* factor out common documentation for eval

* update biocoder doc

* remove swe_env_box since it is no longer needed

* get biocoder working

* add func timeout for bird

* fix jupyter pwd with ~ as user name

* fix jupyter pwd with ~ as user name

* get bird working

* get browsing evaluation working

* make eda runnable

* fix id column

* fix eda run_infer

* unify eval output using a structured format;
make swebench coompatible with that format;
update client source code for every swebench run;
do not inject testcmd for swebench

* standardize existing benchs for the new eval output

* set update source code = true

* get gaia standardized

* fix gaia

* gorilla refactored but stuck at language.so to test

* refactor and make gpqa work

* refactor humanevalfix and get it working

* refactor logic reasoning and get it working

* refactor browser env so it works with eventstream runtime for eval

* add initial version of miniwob refactor

* fix browsergym environment

* get miniwob working!!

* allowing injecting additional dependency to OD runtime docker image

* allowing injecting additional dependency to OD runtime docker image

* support logic reasoning with pre-injected dependency

* get mint working

* update runtime build

* fix mint docker

* add test for keep_prompt;
add missing await close for some tests

* update integration tests for eventstream runtime

* fix integration tests for server runtime

* refactor ml bench and toolqa

* refactor webarena

* fix default factory

* Update run_infer.py

* add APIError to retry

* increase timeout for swebench

* make sure to hide api key when dump eval output

* update the behavior of put source code to put files instead of tarball

* add dishash to dependency

* sendintr when timeout

* fix dockerfile copy

* reduce timeout

* use dirhash to avoid repeat building for update source

* fix runtime_build testcase

* add dir_hash to docker build pipeline

* revert api error

* update poetry lock

* add retries for swebench run infer

* fix git patch

* update poetry lock

* adjust config order

* fix mount volumns

* enforce all eval to use "instance_id"

* remove file store from runtime

* make file_store public inside eventstream

* move the runtime logic inside `main` out

* support using async function for process_instance_fn

* refactor run_infer with the create_time

* fix file store

* Update evaluation/toolqa/utils.py

Co-authored-by: Graham Neubig <neubig@gmail.com>

* fix typo

---------

Co-authored-by: tobitege <tobitege@gmx.de>
Co-authored-by: super-dainiu <78588128+super-dainiu@users.noreply.github.com>
Co-authored-by: Graham Neubig <neubig@gmail.com>
2024-08-06 17:21:45 +00:00

694 lines
26 KiB
Python

import atexit
import os
import re
import sys
import tarfile
import tempfile
import time
import uuid
from glob import glob
import docker
from pexpect import exceptions, pxssh
from tenacity import retry, stop_after_attempt, wait_fixed
from opendevin.core.config import SandboxConfig
from opendevin.core.const.guide_url import TROUBLESHOOTING_URL
from opendevin.core.logger import opendevin_logger as logger
from opendevin.core.schema import CancellableStream
from opendevin.runtime.plugins import AgentSkillsRequirement, JupyterRequirement
from opendevin.runtime.plugins.requirement import PluginRequirement
from opendevin.runtime.sandbox import Sandbox
from opendevin.runtime.utils import find_available_tcp_port, split_bash_commands
from opendevin.runtime.utils.image_agnostic import get_od_sandbox_image
class SSHExecCancellableStream(CancellableStream):
def __init__(self, ssh, cmd, timeout):
super().__init__(self.read_output())
self.ssh = ssh
self.cmd = cmd
self.timeout = timeout
def close(self):
self.closed = True
def exit_code(self):
marker = f'EXIT_CODE_MARKER_{uuid.uuid4().hex}'
self.ssh.sendline(f'echo "{marker}$?{marker}"')
if not self.ssh.prompt(timeout=self.timeout):
return None # Timeout occurred
output = self.ssh.before
match = re.search(f'{marker}(\\d+){marker}', output)
if match:
try:
return int(match.group(1))
except ValueError:
# Log the unexpected format
logger.error(f'Unexpected exit code format: {match.group(1)}')
return None
else:
# If we can't find our marked exit code, log the output and return None
logger.error(f'Could not find exit code in output: {output}')
return None
def read_output(self):
st = time.time()
buf = ''
crlf = '\r\n'
lf = '\n'
prompt_len = len(self.ssh.PROMPT)
while True:
try:
if self.closed:
break
_output = self.ssh.read_nonblocking(timeout=1)
if not _output:
continue
buf += _output
if len(buf) < prompt_len:
continue
match = re.search(self.ssh.PROMPT, buf)
if match:
idx, _ = match.span()
yield buf[:idx].replace(crlf, lf)
buf = ''
break
res = buf[:-prompt_len]
if len(res) == 0 or res.find(crlf) == -1:
continue
buf = buf[-prompt_len:]
yield res.replace(crlf, lf)
except exceptions.TIMEOUT:
if time.time() - st < self.timeout:
match = re.search(self.ssh.PROMPT, buf)
if match:
idx, _ = match.span()
yield buf[:idx].replace(crlf, lf)
break
continue
else:
yield buf.replace(crlf, lf)
break
except exceptions.EOF:
break
class DockerSSHBox(Sandbox):
instance_id: str
container_image: str
container_name_prefix = 'opendevin-sandbox-'
container_name: str
container: docker.models.containers.Container
docker_client: docker.DockerClient
_ssh_password: str
_ssh_port: int
ssh: pxssh.pxssh | None = None
def __init__(
self,
config: SandboxConfig,
persist_sandbox: bool,
workspace_mount_path: str | None,
sandbox_workspace_dir: str,
cache_dir: str,
run_as_devin: bool,
ssh_hostname: str = 'host.docker.internal',
ssh_password: str | None = None,
ssh_port: int = 22,
sid: str | None = None,
):
self.config = config
self.workspace_mount_path = workspace_mount_path
self.sandbox_workspace_dir = sandbox_workspace_dir
self.cache_dir = cache_dir
self.use_host_network = config.use_host_network
self.run_as_devin = run_as_devin
logger.info(
f'SSHBox is running as {"opendevin" if self.run_as_devin else "root"} user with USER_ID={config.user_id} in the sandbox'
)
# Initialize docker client. Throws an exception if Docker is not reachable.
try:
self.docker_client = docker.from_env()
except Exception as ex:
logger.exception(
f'Error creating controller. Please check Docker is running and visit `{TROUBLESHOOTING_URL}` for more debugging information.',
exc_info=False,
)
raise ex
if persist_sandbox:
if not self.run_as_devin:
raise Exception(
'Persistent sandbox is currently designed for opendevin user only. Please set run_as_devin=True in your config.toml'
)
self.instance_id = 'persisted'
else:
self.instance_id = (sid or '') + str(uuid.uuid4())
self.container_image = get_od_sandbox_image(
config.container_image, self.docker_client
)
self.container_name = self.container_name_prefix + self.instance_id
# set up random user password
self.persist_sandbox = persist_sandbox
self.ssh_hostname = ssh_hostname
if persist_sandbox:
if not ssh_password:
raise ValueError('ssh_password is required for persistent sandbox')
self._ssh_password = ssh_password
self._ssh_port = ssh_port
else:
self._ssh_password = str(uuid.uuid4())
self._ssh_port = find_available_tcp_port()
try:
docker.DockerClient().containers.get(self.container_name)
self.is_initial_session = False
except docker.errors.NotFound:
self.is_initial_session = True
logger.info('Detected initial session.')
if not persist_sandbox or self.is_initial_session:
logger.info('Creating new Docker container')
n_tries = 5
while n_tries > 0:
try:
self.restart_docker_container()
break
except Exception as e:
logger.exception(
'Failed to start Docker container, retrying...', exc_info=False
)
n_tries -= 1
if n_tries == 0:
raise e
time.sleep(5)
self.setup_user()
else:
self.container = self.docker_client.containers.get(self.container_name)
logger.info('Using existing Docker container')
self.start_docker_container()
try:
self.start_ssh_session()
except Exception as e:
self.close()
raise e
time.sleep(1)
# make sure /tmp always exists
self.execute('mkdir -p /tmp')
# set git config
self.execute('git config --global user.name "OpenDevin"')
self.execute('git config --global user.email "opendevin@all-hands.dev"')
atexit.register(self.close)
super().__init__(config)
def setup_user(self):
time.sleep(2)
# Make users sudoers passwordless
# TODO(sandbox): add this line in the Dockerfile for next minor version of docker image
exit_code, logs = self.container.exec_run(
['/bin/bash', '-c', r"echo '%sudo ALL=(ALL) NOPASSWD:ALL' >> /etc/sudoers"],
workdir=self.sandbox_workspace_dir,
environment=self._env,
)
if exit_code != 0:
raise Exception(
f'Failed to make all users passwordless sudoers in sandbox: {logs}'
)
# Check if the opendevin user exists
exit_code, logs = self.container.exec_run(
['/bin/bash', '-c', 'id -u opendevin'],
workdir=self.sandbox_workspace_dir,
environment=self._env,
)
if exit_code == 0:
# User exists, delete it
exit_code, logs = self.container.exec_run(
['/bin/bash', '-c', 'userdel -r opendevin'],
workdir=self.sandbox_workspace_dir,
environment=self._env,
)
if exit_code != 0:
raise Exception(f'Failed to remove opendevin user in sandbox: {logs}')
if self.run_as_devin:
# Create the opendevin user
exit_code, logs = self.container.exec_run(
[
'/bin/bash',
'-c',
f'useradd -rm -d /home/opendevin -s /bin/bash -g root -G sudo -u {self.config.user_id} opendevin',
],
workdir=self.sandbox_workspace_dir,
environment=self._env,
)
if exit_code != 0:
raise Exception(f'Failed to create opendevin user in sandbox: {logs}')
exit_code, logs = self.container.exec_run(
[
'/bin/bash',
'-c',
f"echo 'opendevin:{self._ssh_password}' | chpasswd",
],
workdir=self.sandbox_workspace_dir,
environment=self._env,
)
if exit_code != 0:
raise Exception(f'Failed to set password in sandbox: {logs}')
# chown the home directory
exit_code, logs = self.container.exec_run(
['/bin/bash', '-c', 'chown opendevin:root /home/opendevin'],
workdir=self.sandbox_workspace_dir,
environment=self._env,
)
if exit_code != 0:
raise Exception(
f'Failed to chown home directory for opendevin in sandbox: {logs}'
)
# check the miniforge3 directory exist
exit_code, logs = self.container.exec_run(
[
'/bin/bash',
'-c',
'[ -d "/opendevin/miniforge3" ] && exit 0 || exit 1',
],
workdir=self.sandbox_workspace_dir,
environment=self._env,
)
if exit_code != 0:
if exit_code == 1:
raise Exception(
'OPENDEVIN_PYTHON_INTERPRETER is not usable. Please pull the latest Docker image: docker pull ghcr.io/opendevin/sandbox:main'
)
else:
raise Exception(
f'An error occurred while checking if miniforge3 directory exists: {logs}'
)
exit_code, logs = self.container.exec_run(
[
'/bin/bash',
'-c',
f'chown opendevin:root {self.sandbox_workspace_dir}',
],
workdir=self.sandbox_workspace_dir,
environment=self._env,
)
if exit_code != 0:
# This is not a fatal error, just a warning
logger.warning(
f'Failed to chown workspace directory for opendevin in sandbox: {logs}. But this should be fine if the {self.sandbox_workspace_dir=} is mounted by the app docker container.'
)
else:
exit_code, logs = self.container.exec_run(
# change password for root
['/bin/bash', '-c', f"echo 'root:{self._ssh_password}' | chpasswd"],
workdir=self.sandbox_workspace_dir,
environment=self._env,
)
if exit_code != 0:
raise Exception(f'Failed to set password for root in sandbox: {logs}')
exit_code, logs = self.container.exec_run(
['/bin/bash', '-c', "echo 'opendevin-sandbox' > /etc/hostname"],
workdir=self.sandbox_workspace_dir,
environment=self._env,
)
# Use the retry decorator, with a maximum of 5 attempts and a fixed wait time of 5 seconds between attempts
@retry(stop=stop_after_attempt(5), wait=wait_fixed(5))
def __ssh_login(self):
time.sleep(2)
try:
self.ssh = pxssh.pxssh(
echo=False,
timeout=self.config.timeout,
encoding='utf-8',
codec_errors='replace',
)
hostname = self.ssh_hostname
username = 'opendevin' if self.run_as_devin else 'root'
if self.persist_sandbox:
password_msg = 'using your SSH password'
else:
password_msg = f"using the password '{self._ssh_password}'"
logger.info('Connecting to SSH session...')
hostname_to_log = hostname.replace('host.docker.internal', 'localhost')
ssh_cmd = f'`ssh -v -p {self._ssh_port} {username}@{hostname_to_log}`'
logger.info(
f'You can debug the SSH connection by running: {ssh_cmd} {password_msg}'
)
self.ssh.login(hostname, username, self._ssh_password, port=self._ssh_port)
logger.info('Connected to SSH session')
except pxssh.ExceptionPxssh as e:
logger.exception(
'Failed to login to SSH session, retrying...', exc_info=False
)
raise e
def start_ssh_session(self):
time.sleep(3)
self.__ssh_login()
assert self.ssh is not None
# Fix: https://github.com/pexpect/pexpect/issues/669
self.ssh.sendline("bind 'set enable-bracketed-paste off'")
self.ssh.prompt()
time.sleep(1)
# cd to workspace
self.ssh.sendline(f'cd {self.sandbox_workspace_dir}')
self.ssh.prompt()
def get_exec_cmd(self, cmd: str) -> list[str]:
if self.run_as_devin:
return ['su', 'opendevin', '-c', cmd]
else:
return ['/bin/bash', '-c', cmd]
def _send_interrupt(
self,
cmd: str,
prev_output: str = '',
ignore_last_output: bool = False,
) -> tuple[int, str]:
assert self.ssh is not None
logger.exception(
f'Command "{cmd}" timed out, killing process...', exc_info=False
)
# send a SIGINT to the process
self.ssh.sendintr()
self.ssh.prompt()
command_output = prev_output
if not ignore_last_output:
command_output += '\n' + self.ssh.before
return (
-1,
f'Command: "{cmd}" timed out. Sent SIGINT to the process: {command_output}',
)
def execute(
self, cmd: str, stream: bool = False, timeout: int | None = None
) -> tuple[int, str | CancellableStream]:
assert self.ssh is not None
timeout = timeout or self.config.timeout
commands = split_bash_commands(cmd)
if len(commands) > 1:
all_output = ''
for command in commands:
exit_code, output = self.execute(command)
if all_output:
all_output += '\r\n'
all_output += str(output)
if exit_code != 0:
return exit_code, all_output
return 0, all_output
self.ssh.sendline(cmd)
if stream:
return 0, SSHExecCancellableStream(self.ssh, cmd, self.config.timeout)
success = self.ssh.prompt(timeout=timeout)
if not success:
return self._send_interrupt(cmd)
command_output = self.ssh.before
# once out, make sure that we have *every* output, we while loop until we get an empty output
while True:
self.ssh.sendline('\n')
timeout_not_reached = self.ssh.prompt(timeout=1)
if not timeout_not_reached:
logger.debug('TIMEOUT REACHED')
break
output = self.ssh.before
if isinstance(output, str) and output.strip() == '':
break
command_output += output
command_output = command_output.removesuffix('\r\n')
# get the exit code
self.ssh.sendline('echo $?')
self.ssh.prompt()
exit_code_str = self.ssh.before.strip()
_start_time = time.time()
while not exit_code_str:
self.ssh.prompt(timeout=1)
exit_code_str = self.ssh.before.strip()
if time.time() - _start_time > timeout:
return self._send_interrupt(
cmd, command_output, ignore_last_output=True
)
cleaned_exit_code_str = exit_code_str.replace('echo $?', '').strip().split()[0]
try:
exit_code = int(cleaned_exit_code_str)
except ValueError:
logger.error(f'Invalid exit code: {cleaned_exit_code_str}')
# Handle the invalid exit code appropriately (e.g., raise an exception or set a default value)
exit_code = -1 # or some other appropriate default value
return exit_code, command_output
def copy_to(self, host_src: str, sandbox_dest: str, recursive: bool = False):
if not os.path.exists(host_src):
raise FileNotFoundError(f'Source file {host_src} does not exist')
# mkdir -p sandbox_dest if it doesn't exist
exit_code, logs = self.container.exec_run(
['/bin/bash', '-c', f'mkdir -p {sandbox_dest}'],
workdir=self.sandbox_workspace_dir,
environment=self._env,
)
if exit_code != 0:
raise Exception(
f'Failed to create directory {sandbox_dest} in sandbox: {logs}'
)
# use temp directory to store the tar file to avoid
# conflict of filename when running multi-processes
with tempfile.TemporaryDirectory() as tmp_dir:
if recursive:
assert os.path.isdir(
host_src
), 'Source must be a directory when recursive is True'
files = glob(host_src + '/**/*', recursive=True)
srcname = os.path.basename(host_src)
tar_filename = os.path.join(tmp_dir, srcname + '.tar')
with tarfile.open(tar_filename, mode='w') as tar:
for file in files:
tar.add(
file,
arcname=os.path.relpath(file, os.path.dirname(host_src)),
)
else:
assert os.path.isfile(
host_src
), 'Source must be a file when recursive is False'
srcname = os.path.basename(host_src)
tar_filename = os.path.join(tmp_dir, srcname + '.tar')
with tarfile.open(tar_filename, mode='w') as tar:
tar.add(host_src, arcname=srcname)
with open(tar_filename, 'rb') as f:
data = f.read()
self.container.put_archive(sandbox_dest, data)
def start_docker_container(self):
try:
container = self.docker_client.containers.get(self.container_name)
logger.info('Container status: %s', container.status)
if container.status != 'running':
container.start()
logger.info('Container started')
elapsed = 0
while container.status != 'running':
time.sleep(1)
elapsed += 1
if elapsed > self.config.timeout:
break
container = self.docker_client.containers.get(self.container_name)
except Exception:
logger.exception('Failed to start container')
def remove_docker_container(self):
try:
container = self.docker_client.containers.get(self.container_name)
container.stop()
logger.info('Container stopped')
container.remove()
logger.info('Container removed')
elapsed = 0
while container.status != 'exited':
time.sleep(1)
elapsed += 1
if elapsed > self.config.timeout:
break
container = self.docker_client.containers.get(self.container_name)
except docker.errors.NotFound:
pass
def get_working_directory(self):
exit_code, result = self.execute('pwd')
if exit_code != 0:
raise Exception('Failed to get working directory')
return str(result).strip()
def is_container_running(self):
try:
container = self.docker_client.containers.get(self.container_name)
if container.status == 'running':
self.container = container
return True
return False
except docker.errors.NotFound:
return False
@property
def volumes(self):
mount_volumes = {
# mount cache directory to /home/opendevin/.cache for pip cache reuse
self.cache_dir: {
'bind': (
'/home/opendevin/.cache' if self.run_as_devin else '/root/.cache'
),
'mode': 'rw',
},
}
if self.workspace_mount_path is not None:
mount_volumes[self.workspace_mount_path] = {
'bind': self.sandbox_workspace_dir,
'mode': 'rw',
}
return mount_volumes
def restart_docker_container(self):
try:
self.remove_docker_container()
except docker.errors.DockerException as ex:
logger.exception('Failed to remove container', exc_info=False)
raise ex
try:
network_kwargs: dict[str, str | dict[str, int]] = {}
if self.use_host_network:
network_kwargs['network_mode'] = 'host'
else:
# FIXME: This is a temporary workaround for Windows where host network mode has bugs.
# FIXME: Docker Desktop for Mac OS has experimental support for host network mode
network_kwargs['ports'] = {f'{self._ssh_port}/tcp': self._ssh_port}
logger.warning(
(
'Using port forwarding till the enable host network mode of Docker is out of experimental mode.'
'Check the 897th issue on https://github.com/OpenDevin/OpenDevin/issues/ for more information.'
)
)
# start the container
logger.info(f'Mounting volumes: {self.volumes}')
self.container = self.docker_client.containers.run(
self.container_image,
# allow root login
command=f"/usr/sbin/sshd -D -p {self._ssh_port} -o 'PermitRootLogin=yes'",
**network_kwargs,
working_dir=self.sandbox_workspace_dir,
name=self.container_name,
detach=True,
volumes=self.volumes,
)
logger.info('Container started')
except Exception as ex:
logger.exception('Failed to start container: ' + str(ex), exc_info=False)
raise ex
# wait for container to be ready
elapsed = 0
while self.container.status != 'running':
if self.container.status == 'exited':
logger.info('container exited')
logger.info('container logs:')
logger.info(self.container.logs())
break
time.sleep(1)
elapsed += 1
self.container = self.docker_client.containers.get(self.container_name)
logger.info(
f'waiting for container to start: {elapsed}, container status: {self.container.status}'
)
if elapsed > self.config.timeout:
break
if self.container.status != 'running':
raise Exception('Failed to start container')
# clean up the container, cannot do it in __del__ because the python interpreter is already shutting down
def close(self):
containers = self.docker_client.containers.list(all=True)
for container in containers:
try:
if container.name.startswith(self.container_name):
if self.persist_sandbox:
container.stop()
else:
# only remove the container we created
# otherwise all other containers with the same prefix will be removed
# which will mess up with parallel evaluation
container.remove(force=True)
except docker.errors.NotFound:
pass
self.docker_client.close()
if __name__ == '__main__':
try:
ssh_box = DockerSSHBox(
config=SandboxConfig(),
run_as_devin=False,
workspace_mount_path='/path/to/workspace',
cache_dir='/path/to/cache',
sandbox_workspace_dir='/sandbox',
persist_sandbox=False,
)
except Exception as e:
logger.exception('Failed to start Docker container: %s', e)
sys.exit(1)
logger.info(
"Interactive Docker container started. Type 'exit' or use Ctrl+C to exit."
)
# Initialize required plugins
plugins: list[PluginRequirement] = [AgentSkillsRequirement(), JupyterRequirement()]
ssh_box.init_plugins(plugins)
logger.info(
'--- AgentSkills COMMAND DOCUMENTATION ---\n'
f'{AgentSkillsRequirement().documentation}\n'
'---'
)
sys.stdout.flush()
try:
while True:
try:
user_input = input('$ ')
except EOFError:
logger.info('Exiting...')
break
if user_input.lower() == 'exit':
logger.info('Exiting...')
break
exit_code, output = ssh_box.execute(user_input)
logger.info('exit code: %d', exit_code)
logger.info(output)
sys.stdout.flush()
except KeyboardInterrupt:
logger.info('Exiting...')
ssh_box.close()