File size: 7,731 Bytes
77320e4 |
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 |
import time
from io import BytesIO
from typing import Any, Dict, List, Union
from fastapi import UploadFile
from starlette.datastructures import UploadFile as StarletteUploadFile
from werkzeug.datastructures import FileStorage
from ..conversation_sessions import CodeInterpreterSession
from ..exceptions.exceptions import (
DependencyException,
InputErrorException,
InternalErrorException,
ModelMaxIterationsException,
)
from ..schemas import Message, RoleType
from ..utils import get_logger
from ..tools import AsyncPythonSandBoxTool
logger = get_logger()
async def predict(
prompt: str,
model_name: str,
config_path: str,
uploaded_files: Any,
**kwargs: Dict[str, Any]):
start_time = time.time()
# create new session
session = await CodeInterpreterSession.create(
model_name=model_name,
config_path=config_path,
**kwargs
)
files = upload_files(uploaded_files, session.session_id)
logger.info(f"Session Creation Latency: {time.time() - start_time}")
# upload file
if isinstance(files, str):
logger.info(f"Upload {files} as file path")
await session.upload_to_sandbox(files)
# upload list of file
elif isinstance(files, list):
for file in files:
if isinstance(file, str):
await session.upload_to_sandbox(file)
elif isinstance(file, UploadFile) or isinstance(file, StarletteUploadFile):
file_content = file.file.read() # get file content
file_like_object = BytesIO(file_content)
file_storage = FileStorage(
stream=file_like_object,
filename=file.filename,
content_type=file.content_type
)
await session.upload_to_sandbox(file_storage)
else:
raise InputErrorException("The file type {} not supported, can't be uploaded".format(type(file)))
# chat
try:
logger.info(f"Instruction message: {prompt}")
content = None
output_files = []
user_messages = [Message(RoleType.User, prompt)]
async for response in session.chat(user_messages):
logger.info(f'Session Chat Response: {response}')
if content is None:
content = response.output_text
else:
content += response.output_text
output_files.extend([output_file.__dict__() for output_file in response.output_files])
session.messages.append(Message(RoleType.Agent, content))
AsyncPythonSandBoxTool.kill_kernels(session.session_id)
logger.info(f"Release python sandbox {session.session_id}")
logger.info(f"Total Latency: {time.time() - start_time}")
return content
except (ModelMaxIterationsException, DependencyException, InputErrorException, InternalErrorException, Exception) \
as e:
exception_messages = {
ModelMaxIterationsException: "Sorry. The agent didn't find the correct answer after multiple trials, "
"Please try another question.",
DependencyException: "Agent failed to process message due to dependency issue. You can try it later. "
"If it still happens, please contact oncall.",
InputErrorException: "Agent failed to process message due to value issue. If you believe all input are "
"correct, please contact oncall.",
InternalErrorException: "Agent failed to process message due to internal error, please contact oncall.",
Exception: "Agent failed to process message due to unknown error, please contact oncall."
}
err_msg = exception_messages.get(type(e), f"Unknown error occurred: {str(e)}")
logger.error(err_msg, exc_info=True)
raise Exception(err_msg)
import time
from typing import Union, List, Any, Dict
from io import BytesIO
from fastapi import UploadFile
from starlette.datastructures import UploadFile as StarletteUploadFile
from ..conversation_sessions import CodeInterpreterSession
from ..schemas import (
Message,
RoleType
)
from werkzeug.datastructures import FileStorage
from ..exceptions.exceptions import InputErrorException, DependencyException, InternalErrorException, \
ModelMaxIterationsException
from ..utils import get_logger, upload_files
logger = get_logger()
async def predict(
prompt: str,
model_name: str,
uploaded_files: Any,
**kwargs: Dict[str, Any]):
start_time = time.time()
# create new session
session = await CodeInterpreterSession.create(
model_name=model_name,
**kwargs
)
files = upload_files(uploaded_files, session.session_id)
logger.info(f"Session Creation Latency: {time.time() - start_time}")
# upload file
if isinstance(files, str):
logger.info(f"Upload {files} as file path")
await session.upload_to_sandbox(files)
# upload list of file
elif isinstance(files, list):
for file in files:
if isinstance(file, str):
await session.upload_to_sandbox(file)
elif isinstance(file, UploadFile) or isinstance(file, StarletteUploadFile):
file_content = file.file.read() # get file content
file_like_object = BytesIO(file_content)
file_storage = FileStorage(
stream=file_like_object,
filename=file.filename,
content_type=file.content_type
)
await session.upload_to_sandbox(file_storage)
else:
raise InputErrorException("The file type {} not supported, can't be uploaded".format(type(file)))
# chat
try:
logger.info(f"Instruction message: {prompt}")
content = None
output_files = []
user_messages = [Message(RoleType.User, prompt)]
async for response in session.chat(user_messages):
logger.info(f'Session Chat Response: {response}')
if content is None:
content = response.output_text
else:
content += response.output_text
output_files.extend([output_file.__dict__() for output_file in response.output_files])
session.messages.append(Message(RoleType.Agent, content))
logger.info(f"Total Latency: {time.time() - start_time}")
return content
except (ModelMaxIterationsException, DependencyException, InputErrorException, InternalErrorException, Exception) \
as e:
exception_messages = {
ModelMaxIterationsException: "Sorry. The agent didn't find the correct answer after multiple trials, "
"Please try another question.",
DependencyException: "Agent failed to process message due to dependency issue. You can try it later. "
"If it still happens, please contact oncall.",
InputErrorException: "Agent failed to process message due to value issue. If you believe all input are "
"correct, please contact oncall.",
InternalErrorException: "Agent failed to process message due to internal error, please contact oncall.",
Exception: "Agent failed to process message due to unknown error, please contact oncall."
}
err_msg = exception_messages.get(type(e), f"Unknown error occurred: {str(e)}")
logger.error(err_msg, exc_info=True)
raise Exception(err_msg)
|