Compare commits

...

5 Commits

Author SHA1 Message Date
carry
4b465ec917 chore: 更新 .gitignore 文件
- 修改测试代码注释,扩大至参考代码
- 新增 refer/ 目录到忽略列表
2025-04-10 17:38:29 +08:00
carry
e7cc03297b feat(frontend): 添加了简单聊天机器人页面 2025-04-10 17:38:02 +08:00
carry
051d1a7535 feat(frontend): 添加模型管理页面并初始化模型相关全局变量
- 在 frontend/__init__.py 中添加 model_manage_page 模块引用
- 新增 model_manage_page.py 文件,实现模型管理页面的基本框架
- 在 global_var.py 中添加 model 和 tokenizer 全局变量
- 在 main.py 中集成模型管理页面到主应用的 Tabs 组件中
2025-04-10 17:37:45 +08:00
carry
97172f9596 feat(dataset): 设置问答数据集展示页面的每页显示数量
- 在 dataset_manage_page 函数中添加 samples_per_page 参数
- 设置每页显示的样本数量为 20 条
2025-04-10 16:12:59 +08:00
carry
f582820443 feat(tools): 添加生成 Pydantic V2 模型示例 JSON 的工具脚本
- 新增 json_example.py 脚本,用于生成 Pydantic V2 模型的示例 JSON 数据结构
- 支持列表、字典、可选类型以及基本数据类型(字符串、整数、浮点数、布尔值、日期和时间)的示例生成
- 可递归生成嵌套模型的示例 JSON
- 示例使用了项目中的 Q_A 模型,生成了包含多个 Q_A 对象的列表 JSON 结构
2025-04-10 15:38:28 +08:00
8 changed files with 106 additions and 7 deletions

5
.gitignore vendored
View File

@ -30,5 +30,6 @@ workdir/
# cache
unsloth_compiled_cache
# 测试代码
test.ipynb
# 测试和参考代码
test.ipynb
refer/

View File

@ -1,6 +1,7 @@
from .chat_page import *
from .setting_page import *
from .train_page import *
from .model_manage_page import *
from .dataset_manage_page import *
from .dataset_generate_page import *
from .prompt_manage_page import *

View File

@ -1,9 +1,28 @@
import gradio as gr
from global_var import model,tokenizer
def chat_page():
with gr.Blocks() as demo:
import random
import time
gr.Markdown("## 聊天")
with gr.Row():
with gr.Column():
pass
chatbot = gr.Chatbot(type="messages")
msg = gr.Textbox()
clear = gr.Button("Clear")
def user(user_message, history: list):
return "", history + [{"role": "user", "content": user_message}]
def bot(history: list):
bot_message = random.choice(["How are you?", "I love you", "I'm very hungry"])
history.append({"role": "assistant", "content": ""})
for character in bot_message:
history[-1]['content'] += character
time.sleep(0.1)
yield history
msg.submit(user, [msg, chatbot], [msg, chatbot], queue=False).then(
bot, chatbot, chatbot
)
clear.click(lambda: None, None, chatbot, queue=False)
return demo

View File

@ -23,7 +23,8 @@ def dataset_manage_page():
components=["text", "text"],
label="问答数据",
headers=["问题", "答案"],
samples=[["示例问题", "示例答案"]]
samples=[["示例问题", "示例答案"]],
samples_per_page=20,
)
def update_qa_display(dataset_name):

View File

@ -0,0 +1,9 @@
import gradio as gr
def model_manage_page():
with gr.Blocks() as demo:
gr.Markdown("## 模型管理")
with gr.Row():
with gr.Column():
pass
return demo

View File

@ -4,4 +4,7 @@ from tools import scan_docs_directory
prompt_store = get_prompt_tinydb("workdir")
sql_engine = get_sqlite_engine("workdir")
docs = scan_docs_directory("workdir")
datasets = get_all_dataset("workdir")
datasets = get_all_dataset("workdir")
model = None
tokenizer = None

View File

@ -10,6 +10,8 @@ if __name__ == "__main__":
with gr.Blocks() as app:
gr.Markdown("# 基于文档驱动的自适应编码大模型微调框架")
with gr.Tabs():
with gr.TabItem("模型管理"):
model_manage_page()
with gr.TabItem("模型推理"):
chat_page()
with gr.TabItem("模型微调"):

63
tools/json_example.py Normal file
View File

@ -0,0 +1,63 @@
from pydantic import BaseModel, create_model
from typing import Any, Dict, List, Optional, get_args, get_origin
import json
from datetime import datetime, date
def generate_example_json(model: type[BaseModel]) -> str:
"""
根据 Pydantic V2 模型生成示例 JSON 数据结构
"""
def _generate_example(field_type: Any) -> Any:
origin = get_origin(field_type)
args = get_args(field_type)
if origin is list or origin is List:
if args:
return [_generate_example(args[0])]
else:
return []
elif origin is dict or origin is Dict:
if len(args) == 2 and args[0] is str:
return {"key": _generate_example(args[1])}
else:
return {}
elif origin is Optional or origin is type(None):
if args:
return _generate_example(args[0])
else:
return None
elif field_type is str:
return "string"
elif field_type is int:
return 0
elif field_type is float:
return 0.0
elif field_type is bool:
return True
elif field_type is datetime:
return datetime.now().isoformat()
elif field_type is date:
return date.today().isoformat()
elif issubclass(field_type, BaseModel):
return generate_example_json(field_type)
else:
return "unknown" # 对于未知类型返回 "unknown"
example_data = {}
for field_name, field in model.model_fields.items():
example_data[field_name] = _generate_example(field.annotation)
return json.dumps(example_data, indent=2, default=str)
if __name__ == "__main__":
import sys
from pathlib import Path
# 添加项目根目录到sys.path
sys.path.append(str(Path(__file__).resolve().parent.parent))
from schema import Q_A
class Q_A_list(BaseModel):
Q_As: List[Q_A]
print("示例 JSON:")
print(generate_example_json(Q_A_list))