-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathsession.py
292 lines (224 loc) · 8.93 KB
/
session.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
from roles import Analyst, Coder, Tester
from utils import find_method_name
import time
from utils import code_truncate
class Session(object):
def __init__(self, TEAM, ANALYST, PYTHON_DEVELOPER, TESTER, requirement, model='gpt-3.5-turbo-0301', majority=1, max_tokens=512,
temperature=0.0, top_p=1.0, max_round=4, before_func=''):
self.session_history = {}
self.max_round = max_round
self.before_func = before_func
self.requirement = requirement
self.analyst = Analyst(TEAM, ANALYST, requirement, model, majority, max_tokens, temperature, top_p)
self.coder = Coder(TEAM, PYTHON_DEVELOPER, requirement, model, majority, max_tokens, temperature, top_p)
self.tester = Tester(TEAM, TESTER, requirement, model, majority, max_tokens, temperature, top_p)
def run_session(self):
plan = self.analyst.analyze()
report = plan
is_init=True
self.session_history["plan"] = plan
code = ""
for i in range(self.max_round):
naivecode = self.coder.implement(report, is_init)
method_name = find_method_name(naivecode)
if method_name:
code = naivecode
if code.strip() == "":
if i == 0:
code = "error"
else:
code = self.session_history['Round_{}'.format(i-1)]["code"]
break
if i == self.max_round-1:
self.session_history['Round_{}'.format(i)] = {"code": code}
break
tests = self.tester.test(code)
test_report = code_truncate(tests)
answer_report = unsafe_execute(self.before_func+code+'\n'+test_report+'\n'+f'check({method_name})', '')
report = f'The compilation output of the preceding code is: {answer_report}'
is_init = False
self.session_history['Round_{}'.format(i)] = {"code": code, "report": report}
if (plan == "error") or (code == "error") or (report == "error"):
code = "error"
break
if answer_report == "Code Test Passed.":
break
self.analyst.itf.clear_history()
self.coder.itf.clear_history()
self.tester.itf.clear_history()
return code, self.session_history
def run_analyst_coder(self):
plan = self.analyst.analyze()
is_init=True
self.session_history["plan"] = plan
code = self.coder.implement(plan, is_init)
if (plan == "error") or (code == "error"):
code = "error"
self.analyst.itf.clear_history()
self.coder.itf.clear_history()
self.tester.itf.clear_history()
return code, self.session_history
def run_coder_tester(self):
report = ""
is_init=True
code = ""
for i in range(self.max_round):
naivecode = self.coder.implement(report, is_init)
if find_method_name(naivecode):
code = naivecode
if code.strip() == "":
if i == 0:
code = self.coder.implement(report, is_init=True)
else:
code = self.session_history['Round_{}'.format(i-1)]["code"]
break
if i == self.max_round-1:
self.session_history['Round_{}'.format(i)] = {"code": code}
break
tests = self.tester.test(code)
test_report = code_truncate(tests)
answer_report = unsafe_execute(self.before_func+code+'\n'+test_report+'\n'+f'check({method_name})', '')
report = f'The compilation output of the preceding code is: {answer_report}'
is_init = False
self.session_history['Round_{}'.format(i)] = {"code": code, "report": report}
if (code == "error") or (report == "error"):
code = "error"
break
if report == "Code Test Passed.":
break
self.analyst.itf.clear_history()
self.coder.itf.clear_history()
self.tester.itf.clear_history()
return code, self.session_history
def run_coder_only(self):
plan = ""
code = self.coder.implement(plan, is_init=True)
self.coder.itf.clear_history()
return code, self.session_history
import contextlib
import faulthandler
import io
import os
import platform
import signal
import tempfile
def unsafe_execute(code, report):
with create_tempdir():
# These system calls are needed when cleaning up tempdir.
import os
import shutil
rmtree = shutil.rmtree
rmdir = os.rmdir
chdir = os.chdir
# Disable functionalities that can make destructive changes to the test.
reliability_guard()
# Construct the check program and run it.
check_program = (
code + report
)
try:
exec_globals = {}
with swallow_io():
timeout = 10
with time_limit(timeout):
exec(check_program, exec_globals)
result = "Code Test Passed."
except AssertionError as e:
result = f"failed with AssertionError. {e}"
except TimeoutException:
result = "timed out"
except BaseException as e:
result = f"{e}"
# Needed for cleaning up.
shutil.rmtree = rmtree
os.rmdir = rmdir
os.chdir = chdir
return result
def reliability_guard(maximum_memory_bytes = None):
"""
This disables various destructive functions and prevents the generated code
from interfering with the test (e.g. fork bomb, killing other processes,
removing filesystem files, etc.)
WARNING
This function is NOT a security sandbox. Untrusted code, including, model-
generated code, should not be blindly executed outside of one. See the
Codex paper for more information about OpenAI's code sandbox, and proceed
with caution.
"""
if maximum_memory_bytes is not None:
import resource
resource.setrlimit(resource.RLIMIT_AS, (maximum_memory_bytes, maximum_memory_bytes))
resource.setrlimit(resource.RLIMIT_DATA, (maximum_memory_bytes, maximum_memory_bytes))
if not platform.uname().system == 'Darwin':
resource.setrlimit(resource.RLIMIT_STACK, (maximum_memory_bytes, maximum_memory_bytes))
faulthandler.disable()
import builtins
builtins.exit = None
builtins.quit = None
import os
os.environ['OMP_NUM_THREADS'] = '1'
os.rmdir = None
os.chdir = None
import shutil
shutil.rmtree = None
shutil.move = None
shutil.chown = None
import subprocess
subprocess.Popen = None # type: ignore
__builtins__['help'] = None
import sys
sys.modules['ipdb'] = None
sys.modules['joblib'] = None
sys.modules['resource'] = None
sys.modules['psutil'] = None
sys.modules['tkinter'] = None
@contextlib.contextmanager
def time_limit(seconds: float):
def signal_handler(signum, frame):
raise TimeoutException("Timed out!")
signal.setitimer(signal.ITIMER_REAL, seconds)
signal.signal(signal.SIGALRM, signal_handler)
try:
yield
finally:
signal.setitimer(signal.ITIMER_REAL, 0)
@contextlib.contextmanager
def swallow_io():
stream = WriteOnlyStringIO()
with contextlib.redirect_stdout(stream):
with contextlib.redirect_stderr(stream):
with redirect_stdin(stream):
yield
@contextlib.contextmanager
def create_tempdir():
with tempfile.TemporaryDirectory() as dirname:
with chdir(dirname):
yield dirname
class TimeoutException(Exception):
pass
class WriteOnlyStringIO(io.StringIO):
""" StringIO that throws an exception when it's read from """
def read(self, *args, **kwargs):
raise IOError
def readline(self, *args, **kwargs):
raise IOError
def readlines(self, *args, **kwargs):
raise IOError
def readable(self, *args, **kwargs):
""" Returns True if the IO object can be read. """
return False
class redirect_stdin(contextlib._RedirectStream): # type: ignore
_stream = 'stdin'
@contextlib.contextmanager
def chdir(root):
if root == ".":
yield
return
cwd = os.getcwd()
os.chdir(root)
try:
yield
except BaseException as exc:
raise exc
finally:
os.chdir(cwd)