Spaces:
Runtime error
Runtime error
Upload online demo code
Browse files- .gitignore +50 -0
- LICENSE +21 -0
- assets/assistant.pic.jpg +0 -0
- assets/user.pic.jpg +0 -0
- chatbot.py +376 -0
- code_interpreter/BaseCodeInterpreter.py +29 -0
- code_interpreter/JupyterClient.py +85 -0
- code_interpreter/OpenCodeInterpreter.py +80 -0
- requirements.txt +32 -0
- utils/cleaner.py +31 -0
- utils/const.py +120 -0
.gitignore
ADDED
|
@@ -0,0 +1,50 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
# Ignore .ckpt files
|
| 2 |
+
ckpt
|
| 3 |
+
|
| 4 |
+
# Ignore Python compiled files
|
| 5 |
+
__pycache__/
|
| 6 |
+
*.py[cod]
|
| 7 |
+
|
| 8 |
+
# Ignore Python virtual environment
|
| 9 |
+
venv/
|
| 10 |
+
|
| 11 |
+
# Ignore Jupyter notebook checkpoints
|
| 12 |
+
.ipynb_checkpoints/
|
| 13 |
+
.git/
|
| 14 |
+
.vscode/
|
| 15 |
+
|
| 16 |
+
# Ignore .DS_Store on MacOS
|
| 17 |
+
.DS_Store
|
| 18 |
+
|
| 19 |
+
rilab_key.txt
|
| 20 |
+
gpt4_custom_code_interpreter/rilab_key.txt
|
| 21 |
+
openai_api_key.txt
|
| 22 |
+
|
| 23 |
+
gpt4_custom_code_interpreter/
|
| 24 |
+
tmp/
|
| 25 |
+
output/
|
| 26 |
+
wandb/
|
| 27 |
+
|
| 28 |
+
utils/const.py
|
| 29 |
+
utils/hf_model_upload.py
|
| 30 |
+
gpt_data_gen/
|
| 31 |
+
*.json
|
| 32 |
+
*.txt
|
| 33 |
+
*.sh
|
| 34 |
+
*.pt
|
| 35 |
+
*.pth
|
| 36 |
+
*.ckpt
|
| 37 |
+
*.tokenizer
|
| 38 |
+
|
| 39 |
+
# eval data
|
| 40 |
+
eval/ds1000_data
|
| 41 |
+
eval/grade-school-math
|
| 42 |
+
|
| 43 |
+
# gradio features
|
| 44 |
+
chatbot_feat.py
|
| 45 |
+
chatbot_feat2.py
|
| 46 |
+
gradio_test.py
|
| 47 |
+
|
| 48 |
+
cache/
|
| 49 |
+
env/
|
| 50 |
+
json_dataset/
|
LICENSE
ADDED
|
@@ -0,0 +1,21 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
MIT License
|
| 2 |
+
|
| 3 |
+
Copyright (c) 2023 Magnetic2014
|
| 4 |
+
|
| 5 |
+
Permission is hereby granted, free of charge, to any person obtaining a copy
|
| 6 |
+
of this software and associated documentation files (the "Software"), to deal
|
| 7 |
+
in the Software without restriction, including without limitation the rights
|
| 8 |
+
to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
|
| 9 |
+
copies of the Software, and to permit persons to whom the Software is
|
| 10 |
+
furnished to do so, subject to the following conditions:
|
| 11 |
+
|
| 12 |
+
The above copyright notice and this permission notice shall be included in all
|
| 13 |
+
copies or substantial portions of the Software.
|
| 14 |
+
|
| 15 |
+
THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
|
| 16 |
+
IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
|
| 17 |
+
FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
|
| 18 |
+
AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
|
| 19 |
+
LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
|
| 20 |
+
OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
|
| 21 |
+
SOFTWARE.
|
assets/assistant.pic.jpg
ADDED
|
assets/user.pic.jpg
ADDED
|
chatbot.py
ADDED
|
@@ -0,0 +1,376 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
import ast
|
| 2 |
+
import gradio as gr
|
| 3 |
+
import os
|
| 4 |
+
import re
|
| 5 |
+
import json
|
| 6 |
+
import logging
|
| 7 |
+
|
| 8 |
+
import torch
|
| 9 |
+
from datetime import datetime
|
| 10 |
+
|
| 11 |
+
from threading import Thread
|
| 12 |
+
from typing import Optional
|
| 13 |
+
from transformers import TextIteratorStreamer
|
| 14 |
+
from functools import partial
|
| 15 |
+
from huggingface_hub import CommitScheduler
|
| 16 |
+
from uuid import uuid4
|
| 17 |
+
from pathlib import Path
|
| 18 |
+
|
| 19 |
+
from code_interpreter.JupyterClient import JupyterNotebook
|
| 20 |
+
|
| 21 |
+
MAX_INPUT_TOKEN_LENGTH = int(os.getenv("MAX_INPUT_TOKEN_LENGTH", "4096"))
|
| 22 |
+
|
| 23 |
+
import warnings
|
| 24 |
+
|
| 25 |
+
warnings.filterwarnings("ignore", category=UserWarning, module="transformers")
|
| 26 |
+
os.environ["TF_CPP_MIN_LOG_LEVEL"] = "2"
|
| 27 |
+
|
| 28 |
+
|
| 29 |
+
from code_interpreter.OpenCodeInterpreter import OpenCodeInterpreter
|
| 30 |
+
|
| 31 |
+
JSON_DATASET_DIR = Path("json_dataset")
|
| 32 |
+
JSON_DATASET_DIR.mkdir(parents=True, exist_ok=True)
|
| 33 |
+
|
| 34 |
+
upvote_button_value = "👍 Upvote Conversation"
|
| 35 |
+
downvote_button_value = "👎 Downvote Conversation"
|
| 36 |
+
|
| 37 |
+
scheduler = CommitScheduler(
|
| 38 |
+
repo_id="opencodeinterpreter_user_data",
|
| 39 |
+
repo_type="dataset",
|
| 40 |
+
folder_path=JSON_DATASET_DIR,
|
| 41 |
+
path_in_repo="data",
|
| 42 |
+
private=True
|
| 43 |
+
)
|
| 44 |
+
|
| 45 |
+
logging.basicConfig(level=logging.INFO)
|
| 46 |
+
|
| 47 |
+
class StreamingOpenCodeInterpreter(OpenCodeInterpreter):
|
| 48 |
+
streamer: Optional[TextIteratorStreamer] = None
|
| 49 |
+
|
| 50 |
+
# overwirte generate function
|
| 51 |
+
@torch.inference_mode()
|
| 52 |
+
def generate(
|
| 53 |
+
self,
|
| 54 |
+
prompt: str = "",
|
| 55 |
+
max_new_tokens = 1024,
|
| 56 |
+
do_sample: bool = False,
|
| 57 |
+
top_p: float = 0.95,
|
| 58 |
+
top_k: int = 50,
|
| 59 |
+
) -> str:
|
| 60 |
+
# Get the model and tokenizer, and tokenize the user text.
|
| 61 |
+
|
| 62 |
+
self.streamer = TextIteratorStreamer(
|
| 63 |
+
self.tokenizer, skip_prompt=True, Timeout=5
|
| 64 |
+
)
|
| 65 |
+
|
| 66 |
+
inputs = self.tokenizer([prompt], return_tensors="pt", truncation=True, max_length=MAX_INPUT_TOKEN_LENGTH)
|
| 67 |
+
inputs = inputs.to(self.model.device)
|
| 68 |
+
|
| 69 |
+
kwargs = dict(
|
| 70 |
+
**inputs,
|
| 71 |
+
streamer=self.streamer,
|
| 72 |
+
max_new_tokens=max_new_tokens,
|
| 73 |
+
do_sample=do_sample,
|
| 74 |
+
top_p=top_p,
|
| 75 |
+
top_k=top_k,
|
| 76 |
+
eos_token_id=self.tokenizer.eos_token_id
|
| 77 |
+
)
|
| 78 |
+
|
| 79 |
+
thread = Thread(target=self.model.generate, kwargs=kwargs)
|
| 80 |
+
thread.start()
|
| 81 |
+
|
| 82 |
+
return ""
|
| 83 |
+
|
| 84 |
+
def save_json(dialog, mode, json_file_path, flag, dialog_id) -> None:
|
| 85 |
+
with scheduler.lock:
|
| 86 |
+
with json_file_path.open("a") as f:
|
| 87 |
+
json.dump({"id": dialog_id, "dialog": dialog, "mode": mode, "flag": flag, "datetime": datetime.now().isoformat()}, f, ensure_ascii=False)
|
| 88 |
+
f.write("\n")
|
| 89 |
+
|
| 90 |
+
def convert_history(gradio_history: list[list], interpreter_history: list[dict]):
|
| 91 |
+
interpreter_history = [interpreter_history[0]] if interpreter_history and interpreter_history[0]["role"] == "system" else []
|
| 92 |
+
if not gradio_history:
|
| 93 |
+
return interpreter_history
|
| 94 |
+
for item in gradio_history:
|
| 95 |
+
if item[0] is not None:
|
| 96 |
+
interpreter_history.append({"role": "user", "content": item[0]})
|
| 97 |
+
if item[1] is not None:
|
| 98 |
+
interpreter_history.append({"role": "assistant", "content": item[1]})
|
| 99 |
+
return interpreter_history
|
| 100 |
+
|
| 101 |
+
def reset_dialog_info(dialog_info):
|
| 102 |
+
new_uuid = str(uuid4())
|
| 103 |
+
logging.info(f"allocating new uuid {new_uuid} for conversation...")
|
| 104 |
+
return [new_uuid, None]
|
| 105 |
+
|
| 106 |
+
def is_valid_python_code(code):
|
| 107 |
+
try:
|
| 108 |
+
ast.parse(code)
|
| 109 |
+
return True
|
| 110 |
+
except SyntaxError:
|
| 111 |
+
return False
|
| 112 |
+
|
| 113 |
+
|
| 114 |
+
class InputFunctionVisitor(ast.NodeVisitor):
|
| 115 |
+
def __init__(self):
|
| 116 |
+
self.found_input = False
|
| 117 |
+
|
| 118 |
+
def visit_Call(self, node):
|
| 119 |
+
if isinstance(node.func, ast.Name) and node.func.id == 'input':
|
| 120 |
+
self.found_input = True
|
| 121 |
+
self.generic_visit(node)
|
| 122 |
+
|
| 123 |
+
def has_input_function_calls(code):
|
| 124 |
+
try:
|
| 125 |
+
tree = ast.parse(code)
|
| 126 |
+
except SyntaxError:
|
| 127 |
+
return False
|
| 128 |
+
visitor = InputFunctionVisitor()
|
| 129 |
+
visitor.visit(tree)
|
| 130 |
+
return visitor.found_input
|
| 131 |
+
|
| 132 |
+
def gradio_launch(model_path: str, MAX_TRY: int = 3):
|
| 133 |
+
with gr.Blocks() as demo:
|
| 134 |
+
gr.Markdown("# Online Demo of OpenCodeInterpreter Models")
|
| 135 |
+
gr.Markdown("**NOTE: Please read the disclaimer section in [README.md](https://huggingface.co/spaces/m-a-p/OpenCodeInterpreter_demo/blob/main/README.md) before using this demo!**")
|
| 136 |
+
gr.Markdown("**By using this demo, you acknowledge that you have read this disclaimer, understand its terms, and agree to be bound by them.**")
|
| 137 |
+
chatbot = gr.Chatbot(height=600, label="OpenCodeInterpreter", avatar_images=["assets/user.pic.jpg", "assets/assistant.pic.jpg"], show_copy_button=True)
|
| 138 |
+
with gr.Group():
|
| 139 |
+
with gr.Row():
|
| 140 |
+
msg = gr.Textbox(
|
| 141 |
+
container=False,
|
| 142 |
+
show_label=False,
|
| 143 |
+
label="Message",
|
| 144 |
+
placeholder="Type a message...",
|
| 145 |
+
scale=7,
|
| 146 |
+
autofocus=True
|
| 147 |
+
)
|
| 148 |
+
sub = gr.Button(
|
| 149 |
+
"Submit",
|
| 150 |
+
variant="primary",
|
| 151 |
+
scale=1,
|
| 152 |
+
min_width=150
|
| 153 |
+
)
|
| 154 |
+
# stop = gr.Button(
|
| 155 |
+
# "Stop",
|
| 156 |
+
# variant="stop",
|
| 157 |
+
# visible=False,
|
| 158 |
+
# scale=1,
|
| 159 |
+
# min_width=150
|
| 160 |
+
# )
|
| 161 |
+
|
| 162 |
+
with gr.Row():
|
| 163 |
+
# retry = gr.Button("🔄 Retry", variant="secondary")
|
| 164 |
+
# undo = gr.Button("↩️ Undo", variant="secondary")
|
| 165 |
+
upvote = gr.Button(upvote_button_value, variant="secondary")
|
| 166 |
+
downvote = gr.Button(downvote_button_value, variant="secondary")
|
| 167 |
+
clear = gr.Button("🗑️ Clear", variant="secondary")
|
| 168 |
+
|
| 169 |
+
session_state = gr.State([])
|
| 170 |
+
jupyter_state = gr.State(JupyterNotebook())
|
| 171 |
+
dialog_info = gr.State(["", None])
|
| 172 |
+
demo.load(reset_dialog_info, dialog_info, dialog_info)
|
| 173 |
+
|
| 174 |
+
def bot(user_message, history, jupyter_state, dialog_info, interpreter):
|
| 175 |
+
logging.info(f"user message: {user_message}")
|
| 176 |
+
interpreter.dialog = convert_history(gradio_history=history, interpreter_history=interpreter.dialog)
|
| 177 |
+
history.append([user_message, None])
|
| 178 |
+
|
| 179 |
+
interpreter.dialog.append({"role": "user", "content": user_message})
|
| 180 |
+
|
| 181 |
+
# setup
|
| 182 |
+
HAS_CODE = False # For now
|
| 183 |
+
prompt = interpreter.dialog_to_prompt(dialog=interpreter.dialog)
|
| 184 |
+
|
| 185 |
+
_ = interpreter.generate(prompt)
|
| 186 |
+
history[-1][1] = ""
|
| 187 |
+
generated_text = ""
|
| 188 |
+
for character in interpreter.streamer:
|
| 189 |
+
history[-1][1] += character
|
| 190 |
+
history[-1][1] = history[-1][1].replace("<|EOT|>","")
|
| 191 |
+
generated_text += character
|
| 192 |
+
yield history, history, jupyter_state, dialog_info
|
| 193 |
+
|
| 194 |
+
if is_valid_python_code(history[-1][1].strip()):
|
| 195 |
+
history[-1][1] = f"```python\n{history[-1][1].strip()}\n```"
|
| 196 |
+
generated_text = history[-1][1]
|
| 197 |
+
|
| 198 |
+
HAS_CODE, generated_code_block = interpreter.extract_code_blocks(
|
| 199 |
+
generated_text
|
| 200 |
+
)
|
| 201 |
+
|
| 202 |
+
interpreter.dialog.append(
|
| 203 |
+
{
|
| 204 |
+
"role": "assistant",
|
| 205 |
+
"content": generated_text.replace("<unk>_", "")
|
| 206 |
+
.replace("<unk>", "")
|
| 207 |
+
.replace("<|EOT|>", ""),
|
| 208 |
+
}
|
| 209 |
+
)
|
| 210 |
+
|
| 211 |
+
logging.info(f"saving current dialog to file {dialog_info[0]}.json...")
|
| 212 |
+
logging.info(f"current dialog: {interpreter.dialog}")
|
| 213 |
+
save_json(interpreter.dialog, mode="openci_only", flag=dialog_info[1], json_file_path=JSON_DATASET_DIR/f"{dialog_info[0]}.json", dialog_id=dialog_info[0])
|
| 214 |
+
|
| 215 |
+
attempt = 1
|
| 216 |
+
while HAS_CODE:
|
| 217 |
+
if attempt > MAX_TRY:
|
| 218 |
+
break
|
| 219 |
+
# if no code then doesn't have to execute it
|
| 220 |
+
generated_text = "" # clear generated text
|
| 221 |
+
|
| 222 |
+
yield history, history, jupyter_state, dialog_info
|
| 223 |
+
|
| 224 |
+
# replace unknown thing to none ''
|
| 225 |
+
generated_code_block = generated_code_block.replace(
|
| 226 |
+
"<unk>_", ""
|
| 227 |
+
).replace("<unk>", "")
|
| 228 |
+
|
| 229 |
+
if has_input_function_calls(generated_code_block):
|
| 230 |
+
code_block_output = "Please directly assign the value of inputs instead of using input() function in your code."
|
| 231 |
+
else:
|
| 232 |
+
(
|
| 233 |
+
code_block_output,
|
| 234 |
+
error_flag,
|
| 235 |
+
) = interpreter.execute_code_and_return_output(
|
| 236 |
+
f"{generated_code_block}",
|
| 237 |
+
jupyter_state
|
| 238 |
+
)
|
| 239 |
+
if error_flag == "Timeout":
|
| 240 |
+
logging.info(f"{dialog_info[0]}: Restart jupyter kernel due to timeout")
|
| 241 |
+
jupyter_state = JupyterNotebook()
|
| 242 |
+
code_block_output = interpreter.clean_code_output(code_block_output)
|
| 243 |
+
|
| 244 |
+
if code_block_output.strip():
|
| 245 |
+
code_block_output = "Execution result: \n" + code_block_output
|
| 246 |
+
else:
|
| 247 |
+
code_block_output = "Code is executed, but result is empty. Please make sure that you include test case in your code."
|
| 248 |
+
|
| 249 |
+
history.append([code_block_output, ""])
|
| 250 |
+
|
| 251 |
+
interpreter.dialog.append({"role": "user", "content": code_block_output})
|
| 252 |
+
|
| 253 |
+
yield history, history, jupyter_state, dialog_info
|
| 254 |
+
|
| 255 |
+
prompt = interpreter.dialog_to_prompt(dialog=interpreter.dialog)
|
| 256 |
+
|
| 257 |
+
logging.info(f"generating answer for dialog {dialog_info[0]}")
|
| 258 |
+
_ = interpreter.generate(prompt)
|
| 259 |
+
for character in interpreter.streamer:
|
| 260 |
+
history[-1][1] += character
|
| 261 |
+
history[-1][1] = history[-1][1].replace("<|EOT|>","")
|
| 262 |
+
generated_text += character
|
| 263 |
+
yield history, history, jupyter_state, dialog_info
|
| 264 |
+
logging.info(f"finish generating answer for dialog {dialog_info[0]}")
|
| 265 |
+
|
| 266 |
+
HAS_CODE, generated_code_block = interpreter.extract_code_blocks(
|
| 267 |
+
history[-1][1]
|
| 268 |
+
)
|
| 269 |
+
|
| 270 |
+
interpreter.dialog.append(
|
| 271 |
+
{
|
| 272 |
+
"role": "assistant",
|
| 273 |
+
"content": generated_text.replace("<unk>_", "")
|
| 274 |
+
.replace("<unk>", "")
|
| 275 |
+
.replace("<|EOT|>", ""),
|
| 276 |
+
}
|
| 277 |
+
)
|
| 278 |
+
|
| 279 |
+
attempt += 1
|
| 280 |
+
|
| 281 |
+
logging.info(f"saving current dialog to file {dialog_info[0]}.json...")
|
| 282 |
+
logging.info(f"current dialog: {interpreter.dialog}")
|
| 283 |
+
save_json(interpreter.dialog, mode="openci_only", flag=dialog_info[1], json_file_path=JSON_DATASET_DIR/f"{dialog_info[0]}.json", dialog_id=dialog_info[0])
|
| 284 |
+
|
| 285 |
+
if generated_text.endswith("<|EOT|>"):
|
| 286 |
+
continue
|
| 287 |
+
|
| 288 |
+
return history, history, jupyter_state, dialog_info
|
| 289 |
+
|
| 290 |
+
|
| 291 |
+
def reset_textbox():
|
| 292 |
+
return gr.update(value="")
|
| 293 |
+
|
| 294 |
+
def set_button_variant(upvote_button_variant, downvote_button_variant):
|
| 295 |
+
return gr.Button(upvote_button_value, variant=upvote_button_variant), gr.Button(downvote_button_value, variant=downvote_button_variant)
|
| 296 |
+
|
| 297 |
+
def reset_button_and_flag(dialog_info):
|
| 298 |
+
return (*set_button_variant("secondary", "secondary"), [dialog_info[0], None])
|
| 299 |
+
|
| 300 |
+
def clear_history(history, jupyter_state, dialog_info, interpreter):
|
| 301 |
+
interpreter.dialog = []
|
| 302 |
+
jupyter_state.close()
|
| 303 |
+
return ([], [], JupyterNotebook(), reset_dialog_info(dialog_info), *set_button_variant("secondary", "secondary"))
|
| 304 |
+
|
| 305 |
+
def toggle_preference(button, dialog_info):
|
| 306 |
+
if button == upvote_button_value:
|
| 307 |
+
dialog_info[1] = True
|
| 308 |
+
elif button == downvote_button_value:
|
| 309 |
+
dialog_info[1] = False
|
| 310 |
+
else:
|
| 311 |
+
raise ValueError(button)
|
| 312 |
+
logging.info(f"{button} is clicked by {dialog_info[0]}, current flag: {dialog_info[1]}")
|
| 313 |
+
|
| 314 |
+
if dialog_info[1] is None:
|
| 315 |
+
return (*set_button_variant("secondary", "secondary"), dialog_info)
|
| 316 |
+
elif dialog_info[1]:
|
| 317 |
+
return (*set_button_variant("primary", "secondary"), dialog_info)
|
| 318 |
+
else:
|
| 319 |
+
return (*set_button_variant("secondary", "primary"), dialog_info)
|
| 320 |
+
|
| 321 |
+
def save_preference(dialog_info, interpreter):
|
| 322 |
+
if interpreter.dialog:
|
| 323 |
+
save_json(interpreter.dialog, mode="openci_only", flag=dialog_info[1], json_file_path=JSON_DATASET_DIR/f"{dialog_info[0]}.json", dialog_id=dialog_info[0])
|
| 324 |
+
return dialog_info
|
| 325 |
+
|
| 326 |
+
interpreter = StreamingOpenCodeInterpreter(model_path=model_path)
|
| 327 |
+
|
| 328 |
+
sub.click(reset_button_and_flag, dialog_info, [upvote, downvote, dialog_info])
|
| 329 |
+
sub.click(partial(bot, interpreter=interpreter), [msg, session_state, jupyter_state, dialog_info], [chatbot, session_state, jupyter_state, dialog_info])
|
| 330 |
+
sub.click(reset_textbox, [], [msg])
|
| 331 |
+
|
| 332 |
+
clear.click(
|
| 333 |
+
partial(clear_history, interpreter=interpreter),
|
| 334 |
+
[session_state, jupyter_state, dialog_info],
|
| 335 |
+
[chatbot, session_state, jupyter_state, dialog_info, upvote, downvote],
|
| 336 |
+
queue=False
|
| 337 |
+
)
|
| 338 |
+
|
| 339 |
+
upvote.click(
|
| 340 |
+
toggle_preference,
|
| 341 |
+
[upvote, dialog_info],
|
| 342 |
+
[upvote, downvote, dialog_info]
|
| 343 |
+
).then(
|
| 344 |
+
partial(save_preference, interpreter=interpreter),
|
| 345 |
+
dialog_info,
|
| 346 |
+
dialog_info
|
| 347 |
+
)
|
| 348 |
+
|
| 349 |
+
downvote.click(
|
| 350 |
+
toggle_preference,
|
| 351 |
+
[downvote, dialog_info],
|
| 352 |
+
[upvote, downvote, dialog_info]
|
| 353 |
+
).then(
|
| 354 |
+
partial(save_preference, interpreter=interpreter),
|
| 355 |
+
dialog_info,
|
| 356 |
+
dialog_info
|
| 357 |
+
)
|
| 358 |
+
|
| 359 |
+
demo.queue(max_size=20)
|
| 360 |
+
demo.launch(share=True)
|
| 361 |
+
|
| 362 |
+
|
| 363 |
+
if __name__ == "__main__":
|
| 364 |
+
import argparse
|
| 365 |
+
|
| 366 |
+
parser = argparse.ArgumentParser()
|
| 367 |
+
parser.add_argument(
|
| 368 |
+
"--path",
|
| 369 |
+
type=str,
|
| 370 |
+
required=False,
|
| 371 |
+
help="Path to the OpenCodeInterpreter Model.",
|
| 372 |
+
default="m-a-p/OpenCodeInterpreter-DS-6.7B",
|
| 373 |
+
)
|
| 374 |
+
args = parser.parse_args()
|
| 375 |
+
|
| 376 |
+
gradio_launch(model_path=args.path)
|
code_interpreter/BaseCodeInterpreter.py
ADDED
|
@@ -0,0 +1,29 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
import os
|
| 2 |
+
import sys
|
| 3 |
+
import re
|
| 4 |
+
|
| 5 |
+
prj_root_path = os.path.dirname(os.path.dirname(os.path.abspath(__file__)))
|
| 6 |
+
sys.path.append(prj_root_path)
|
| 7 |
+
|
| 8 |
+
|
| 9 |
+
from utils.const import *
|
| 10 |
+
|
| 11 |
+
class BaseCodeInterpreter:
|
| 12 |
+
def __init__(self):
|
| 13 |
+
self.dialog = [
|
| 14 |
+
{
|
| 15 |
+
"role": "system",
|
| 16 |
+
"content": CODE_INTERPRETER_SYSTEM_PROMPT,
|
| 17 |
+
},
|
| 18 |
+
]
|
| 19 |
+
|
| 20 |
+
@staticmethod
|
| 21 |
+
def extract_code_blocks(text: str):
|
| 22 |
+
pattern = r"```(?:python\n)?(.*?)```" # Match optional 'python\n' but don't capture it
|
| 23 |
+
code_blocks = re.findall(pattern, text, re.DOTALL)
|
| 24 |
+
return [block.strip() for block in code_blocks]
|
| 25 |
+
|
| 26 |
+
def execute_code_and_return_output(self, code_str: str, nb):
|
| 27 |
+
_, _ = nb.add_and_run(GUARD_CODE)
|
| 28 |
+
outputs, error_flag = nb.add_and_run(code_str)
|
| 29 |
+
return outputs, error_flag
|
code_interpreter/JupyterClient.py
ADDED
|
@@ -0,0 +1,85 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
from jupyter_client import KernelManager
|
| 2 |
+
import threading
|
| 3 |
+
import re
|
| 4 |
+
from utils.const import *
|
| 5 |
+
|
| 6 |
+
|
| 7 |
+
class JupyterNotebook:
|
| 8 |
+
def __init__(self):
|
| 9 |
+
self.km = KernelManager()
|
| 10 |
+
self.km.start_kernel()
|
| 11 |
+
self.kc = self.km.client()
|
| 12 |
+
_ = self.add_and_run(TOOLS_CODE)
|
| 13 |
+
|
| 14 |
+
def clean_output(self, outputs):
|
| 15 |
+
outputs_only_str = list()
|
| 16 |
+
for i in outputs:
|
| 17 |
+
if type(i) == dict:
|
| 18 |
+
if "text/plain" in list(i.keys()):
|
| 19 |
+
outputs_only_str.append(i["text/plain"])
|
| 20 |
+
elif type(i) == str:
|
| 21 |
+
outputs_only_str.append(i)
|
| 22 |
+
elif type(i) == list:
|
| 23 |
+
error_msg = "\n".join(i)
|
| 24 |
+
error_msg = re.sub(r"\x1b\[.*?m", "", error_msg)
|
| 25 |
+
outputs_only_str.append(error_msg)
|
| 26 |
+
|
| 27 |
+
return "\n".join(outputs_only_str).strip()
|
| 28 |
+
|
| 29 |
+
def add_and_run(self, code_string):
|
| 30 |
+
# This inner function will be executed in a separate thread
|
| 31 |
+
def run_code_in_thread():
|
| 32 |
+
nonlocal outputs, error_flag
|
| 33 |
+
|
| 34 |
+
# Execute the code and get the execution count
|
| 35 |
+
msg_id = self.kc.execute(code_string)
|
| 36 |
+
|
| 37 |
+
while True:
|
| 38 |
+
try:
|
| 39 |
+
msg = self.kc.get_iopub_msg(timeout=20)
|
| 40 |
+
|
| 41 |
+
msg_type = msg["header"]["msg_type"]
|
| 42 |
+
content = msg["content"]
|
| 43 |
+
|
| 44 |
+
if msg_type == "execute_result":
|
| 45 |
+
outputs.append(content["data"])
|
| 46 |
+
elif msg_type == "stream":
|
| 47 |
+
outputs.append(content["text"])
|
| 48 |
+
elif msg_type == "error":
|
| 49 |
+
error_flag = True
|
| 50 |
+
outputs.append(content["traceback"])
|
| 51 |
+
|
| 52 |
+
# If the execution state of the kernel is idle, it means the cell finished executing
|
| 53 |
+
if msg_type == "status" and content["execution_state"] == "idle":
|
| 54 |
+
break
|
| 55 |
+
except:
|
| 56 |
+
break
|
| 57 |
+
|
| 58 |
+
outputs = []
|
| 59 |
+
error_flag = False
|
| 60 |
+
|
| 61 |
+
# Start the thread to run the code
|
| 62 |
+
thread = threading.Thread(target=run_code_in_thread)
|
| 63 |
+
thread.start()
|
| 64 |
+
|
| 65 |
+
# Wait for 20 seconds for the thread to finish
|
| 66 |
+
thread.join(timeout=20)
|
| 67 |
+
|
| 68 |
+
# If the thread is still alive after 20 seconds, it's a timeout
|
| 69 |
+
if thread.is_alive():
|
| 70 |
+
outputs = ["Execution timed out."]
|
| 71 |
+
# outputs = ["Error"]
|
| 72 |
+
error_flag = "Timeout"
|
| 73 |
+
|
| 74 |
+
return self.clean_output(outputs), error_flag
|
| 75 |
+
|
| 76 |
+
def close(self):
|
| 77 |
+
"""Shutdown the kernel."""
|
| 78 |
+
self.km.shutdown_kernel()
|
| 79 |
+
|
| 80 |
+
def __deepcopy__(self, memo):
|
| 81 |
+
if id(self) in memo:
|
| 82 |
+
return memo[id(self)]
|
| 83 |
+
new_copy = type(self)()
|
| 84 |
+
memo[id(self)] = new_copy
|
| 85 |
+
return new_copy
|
code_interpreter/OpenCodeInterpreter.py
ADDED
|
@@ -0,0 +1,80 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
import sys
|
| 2 |
+
import os
|
| 3 |
+
|
| 4 |
+
prj_root_path = os.path.dirname(os.path.dirname(os.path.abspath(__file__)))
|
| 5 |
+
sys.path.append(prj_root_path)
|
| 6 |
+
|
| 7 |
+
from code_interpreter.BaseCodeInterpreter import BaseCodeInterpreter
|
| 8 |
+
from utils.const import *
|
| 9 |
+
|
| 10 |
+
from typing import List, Tuple, Dict
|
| 11 |
+
import re
|
| 12 |
+
|
| 13 |
+
import torch
|
| 14 |
+
from transformers import AutoModelForCausalLM, AutoTokenizer
|
| 15 |
+
|
| 16 |
+
|
| 17 |
+
sys.path.append(os.path.dirname(__file__))
|
| 18 |
+
sys.path.append(os.path.dirname(os.path.abspath(__file__)))
|
| 19 |
+
|
| 20 |
+
import warnings
|
| 21 |
+
|
| 22 |
+
warnings.filterwarnings("ignore", category=UserWarning, module="transformers")
|
| 23 |
+
os.environ["TF_CPP_MIN_LOG_LEVEL"] = "2"
|
| 24 |
+
|
| 25 |
+
|
| 26 |
+
class OpenCodeInterpreter(BaseCodeInterpreter):
|
| 27 |
+
def __init__(
|
| 28 |
+
self,
|
| 29 |
+
model_path: str,
|
| 30 |
+
load_in_8bit: bool = False,
|
| 31 |
+
load_in_4bit: bool = False,
|
| 32 |
+
):
|
| 33 |
+
# build tokenizer
|
| 34 |
+
self.tokenizer = AutoTokenizer.from_pretrained(
|
| 35 |
+
model_path,
|
| 36 |
+
padding_side="right",
|
| 37 |
+
trust_remote_code=True
|
| 38 |
+
)
|
| 39 |
+
|
| 40 |
+
self.model = AutoModelForCausalLM.from_pretrained(
|
| 41 |
+
model_path,
|
| 42 |
+
device_map="auto",
|
| 43 |
+
load_in_4bit=load_in_4bit,
|
| 44 |
+
load_in_8bit=load_in_8bit,
|
| 45 |
+
torch_dtype=torch.float16,
|
| 46 |
+
trust_remote_code=True
|
| 47 |
+
)
|
| 48 |
+
|
| 49 |
+
self.model.resize_token_embeddings(len(self.tokenizer))
|
| 50 |
+
|
| 51 |
+
self.model = self.model.eval()
|
| 52 |
+
|
| 53 |
+
self.dialog = []
|
| 54 |
+
self.MAX_CODE_OUTPUT_LENGTH = 1000
|
| 55 |
+
|
| 56 |
+
|
| 57 |
+
def dialog_to_prompt(self, dialog: List[Dict]) -> str:
|
| 58 |
+
full_str = self.tokenizer.apply_chat_template(dialog, tokenize=False)
|
| 59 |
+
|
| 60 |
+
return full_str
|
| 61 |
+
|
| 62 |
+
def extract_code_blocks(self, prompt: str) -> Tuple[bool, str]:
|
| 63 |
+
pattern = re.escape("```python") + r"(.*?)" + re.escape("```")
|
| 64 |
+
matches = re.findall(pattern, prompt, re.DOTALL)
|
| 65 |
+
|
| 66 |
+
if matches:
|
| 67 |
+
# Return the last matched code block
|
| 68 |
+
return True, matches[-1].strip()
|
| 69 |
+
else:
|
| 70 |
+
return False, ""
|
| 71 |
+
|
| 72 |
+
def clean_code_output(self, output: str) -> str:
|
| 73 |
+
if self.MAX_CODE_OUTPUT_LENGTH < len(output):
|
| 74 |
+
return (
|
| 75 |
+
output[: self.MAX_CODE_OUTPUT_LENGTH // 5]
|
| 76 |
+
+ "\n...(truncated due to length)...\n"
|
| 77 |
+
+ output[-self.MAX_CODE_OUTPUT_LENGTH // 5 :]
|
| 78 |
+
)
|
| 79 |
+
|
| 80 |
+
return output
|
requirements.txt
ADDED
|
@@ -0,0 +1,32 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
accelerate==0.21.0
|
| 2 |
+
bitsandbytes==0.41.1
|
| 3 |
+
colorama==0.4.6
|
| 4 |
+
coloredlogs==15.0.1
|
| 5 |
+
colorlog==6.7.0
|
| 6 |
+
datasets==2.12.0
|
| 7 |
+
deepspeed==0.10.1
|
| 8 |
+
diffusers==0.20.0
|
| 9 |
+
einops==0.6.1
|
| 10 |
+
gradio==3.48.0
|
| 11 |
+
ipykernel==6.25.1
|
| 12 |
+
ipython==8.12.2
|
| 13 |
+
jupyter_client==8.3.0
|
| 14 |
+
jupyter_core==5.3.0
|
| 15 |
+
Markdown==3.4.3
|
| 16 |
+
nbclient==0.8.0
|
| 17 |
+
nbconvert==7.7.1
|
| 18 |
+
nbformat==5.8.0
|
| 19 |
+
omegaconf==2.3.0
|
| 20 |
+
openai==0.27.7
|
| 21 |
+
rich==13.7.0
|
| 22 |
+
scikit-learn==1.4.0
|
| 23 |
+
scipy==1.12.0
|
| 24 |
+
seaborn==0.13.2
|
| 25 |
+
sentencepiece==0.1.99
|
| 26 |
+
termcolor==2.3.0
|
| 27 |
+
tqdm==4.66.1
|
| 28 |
+
transformers==4.37.1
|
| 29 |
+
triton==2.0.0
|
| 30 |
+
yfinance==0.2.28
|
| 31 |
+
retrying==1.3.4
|
| 32 |
+
pydantic<2.0.0
|
utils/cleaner.py
ADDED
|
@@ -0,0 +1,31 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
import re
|
| 2 |
+
import os
|
| 3 |
+
|
| 4 |
+
PYTHON_PREFIX = os.environ.get("CONDA_PREFIX", "/usr/local")
|
| 5 |
+
|
| 6 |
+
SITE_PKG_ERROR_PREFIX = f'File {PYTHON_PREFIX}/lib/python3.10/'
|
| 7 |
+
|
| 8 |
+
def get_error_header(traceback_str):
|
| 9 |
+
lines = traceback_str.split('\n')
|
| 10 |
+
for line in lines:
|
| 11 |
+
if 'Error:' in line:
|
| 12 |
+
return line
|
| 13 |
+
return '' # Return None if no error message is found
|
| 14 |
+
|
| 15 |
+
def clean_error_msg(error_str:str =''):
|
| 16 |
+
filtered_error_msg = error_str.__str__().split('An error occurred while executing the following cell')[-1].split("\n------------------\n")[-1]
|
| 17 |
+
raw_error_msg = "".join(filtered_error_msg)
|
| 18 |
+
|
| 19 |
+
# Remove escape sequences for colored text
|
| 20 |
+
ansi_escape = re.compile(r'\x1b\[[0-?]*[ -/]*[@-~]')
|
| 21 |
+
error_msg = ansi_escape.sub('', raw_error_msg)
|
| 22 |
+
|
| 23 |
+
error_str_out = ''
|
| 24 |
+
error_msg_only_cell = error_msg.split(SITE_PKG_ERROR_PREFIX)
|
| 25 |
+
|
| 26 |
+
error_str_out += f'{error_msg_only_cell[0]}\n'
|
| 27 |
+
error_header = get_error_header(error_msg_only_cell[-1])
|
| 28 |
+
if error_header not in error_str_out:
|
| 29 |
+
error_str_out += get_error_header(error_msg_only_cell[-1])
|
| 30 |
+
|
| 31 |
+
return error_str_out
|
utils/const.py
ADDED
|
@@ -0,0 +1,120 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
TOOLS_CODE = """
|
| 2 |
+
import numpy as np
|
| 3 |
+
import pandas as pd
|
| 4 |
+
import matplotlib.pyplot as plt
|
| 5 |
+
import seaborn as sns
|
| 6 |
+
from scipy import stats
|
| 7 |
+
import os,sys
|
| 8 |
+
import re
|
| 9 |
+
from datetime import datetime
|
| 10 |
+
from sympy import symbols, Eq, solve
|
| 11 |
+
import torch
|
| 12 |
+
import requests
|
| 13 |
+
from bs4 import BeautifulSoup
|
| 14 |
+
import json
|
| 15 |
+
import math
|
| 16 |
+
import yfinance
|
| 17 |
+
import time
|
| 18 |
+
"""
|
| 19 |
+
|
| 20 |
+
write_denial_function = 'lambda *args, **kwargs: (_ for _ in ()).throw(PermissionError("Writing to disk operation is not permitted due to safety reasons. Please do not try again!"))'
|
| 21 |
+
read_denial_function = 'lambda *args, **kwargs: (_ for _ in ()).throw(PermissionError("Reading from disk operation is not permitted due to safety reasons. Please do not try again!"))'
|
| 22 |
+
class_denial = """Class Denial:
|
| 23 |
+
def __getattr__(self, name):
|
| 24 |
+
def method(*args, **kwargs):
|
| 25 |
+
return "Using this class is not permitted due to safety reasons. Please do not try again!"
|
| 26 |
+
return method
|
| 27 |
+
"""
|
| 28 |
+
|
| 29 |
+
GUARD_CODE = f"""
|
| 30 |
+
import builtins
|
| 31 |
+
|
| 32 |
+
_original_open = open
|
| 33 |
+
|
| 34 |
+
def custom_open(file, mode='r', buffering=-1, encoding=None, errors=None, newline=None, closefd=True, opener=None):
|
| 35 |
+
if 'w' in mode or 'a' in mode or 'x' in mode or '+' in mode:
|
| 36 |
+
raise PermissionError("Writing operation is not permitted due to safety reasons. Please do not try again!")
|
| 37 |
+
return _original_open(file, mode, buffering, encoding, errors, newline, closefd, opener)
|
| 38 |
+
|
| 39 |
+
builtins.open = custom_open
|
| 40 |
+
|
| 41 |
+
builtins.exit = {write_denial_function}
|
| 42 |
+
builtins.quit = {write_denial_function}
|
| 43 |
+
|
| 44 |
+
import sys
|
| 45 |
+
|
| 46 |
+
blocked_modules = ['pathlib', 'glob', 'ctypes']
|
| 47 |
+
|
| 48 |
+
for module in blocked_modules:
|
| 49 |
+
sys.modules[module] = PermissionError
|
| 50 |
+
|
| 51 |
+
import os
|
| 52 |
+
|
| 53 |
+
os.listdir = {read_denial_function}
|
| 54 |
+
os.scandir = {read_denial_function}
|
| 55 |
+
os.walk = {read_denial_function}
|
| 56 |
+
os.stat = {read_denial_function}
|
| 57 |
+
os.kill = {write_denial_function}
|
| 58 |
+
os.system = {write_denial_function}
|
| 59 |
+
os.putenv = {write_denial_function}
|
| 60 |
+
os.remove = {write_denial_function}
|
| 61 |
+
os.removedirs = {write_denial_function}
|
| 62 |
+
os.rmdir = {write_denial_function}
|
| 63 |
+
os.fchdir = {write_denial_function}
|
| 64 |
+
os.setuid = {write_denial_function}
|
| 65 |
+
os.fork = {write_denial_function}
|
| 66 |
+
os.forkpty = {write_denial_function}
|
| 67 |
+
os.killpg = {write_denial_function}
|
| 68 |
+
os.rename = {write_denial_function}
|
| 69 |
+
os.renames = {write_denial_function}
|
| 70 |
+
os.truncate = {write_denial_function}
|
| 71 |
+
os.replace = {write_denial_function}
|
| 72 |
+
os.unlink = {write_denial_function}
|
| 73 |
+
os.fchmod = {write_denial_function}
|
| 74 |
+
os.fchown = {write_denial_function}
|
| 75 |
+
os.chmod = {write_denial_function}
|
| 76 |
+
os.chown = {write_denial_function}
|
| 77 |
+
os.chroot = {write_denial_function}
|
| 78 |
+
os.fchdir = {write_denial_function}
|
| 79 |
+
os.lchflags = {write_denial_function}
|
| 80 |
+
os.lchmod = {write_denial_function}
|
| 81 |
+
os.lchown = {write_denial_function}
|
| 82 |
+
os.getcwd = {write_denial_function}
|
| 83 |
+
os.chdir = {write_denial_function}
|
| 84 |
+
os.popen = {write_denial_function}
|
| 85 |
+
os.environ = {{}}
|
| 86 |
+
os.getenv = {write_denial_function}
|
| 87 |
+
builtins.open = {write_denial_function}
|
| 88 |
+
|
| 89 |
+
import shutil
|
| 90 |
+
|
| 91 |
+
shutil.rmtree = {write_denial_function}
|
| 92 |
+
shutil.move = {write_denial_function}
|
| 93 |
+
shutil.chown = {write_denial_function}
|
| 94 |
+
|
| 95 |
+
import subprocess
|
| 96 |
+
|
| 97 |
+
subprocess.Popen = {write_denial_function} # type: ignore
|
| 98 |
+
|
| 99 |
+
__builtins__["help"] = {write_denial_function}
|
| 100 |
+
|
| 101 |
+
import sys
|
| 102 |
+
|
| 103 |
+
sys.modules["ipdb"] = {write_denial_function}
|
| 104 |
+
sys.modules["joblib"] = {write_denial_function}
|
| 105 |
+
sys.modules["resource"] = {write_denial_function}
|
| 106 |
+
sys.modules["psutil"] = {write_denial_function}
|
| 107 |
+
sys.modules["tkinter"] = {write_denial_function}
|
| 108 |
+
|
| 109 |
+
get_ipython().system = lambda *args, **kwargs: (_ for _ in ()).throw(PermissionError("Sorry, magic command is disabled due to safety reasons. Please do not try again!"))
|
| 110 |
+
"""
|
| 111 |
+
|
| 112 |
+
CODE_INTERPRETER_SYSTEM_PROMPT = """You are an AI code interpreter.
|
| 113 |
+
Your goal is to help users do a variety of jobs by executing Python code.
|
| 114 |
+
|
| 115 |
+
You should:
|
| 116 |
+
1. Comprehend the user's requirements carefully & to the letter.
|
| 117 |
+
2. Give a brief description for what you plan to do & call the provided function to run code.
|
| 118 |
+
3. Provide results analysis based on the execution output.
|
| 119 |
+
4. If error occurred, try to fix it.
|
| 120 |
+
5. Response in the same language as the user."""
|