frontend/dataset_generate_page.py

@@ -1,41 +1,9 @@
 import gradio as gr
-from global_var import docs, scan_docs_directory, prompt_store
 
 def dataset_generate_page():
     with gr.Blocks() as demo:
         gr.Markdown("## 数据集生成")
         with gr.Row():
             with gr.Column():
-                # 获取文档列表并设置初始值
+                pass
-                docs_list = [str(doc.name) for doc in scan_docs_directory("workdir")]
-                initial_doc = docs_list[0] if docs_list else None
-                
-                doc_dropdown = gr.Dropdown(
-                    choices=docs_list,
-                    value=initial_doc,  # 设置初始选中项
-                    label="选择文档",
-                    allow_custom_value=True,
-                    interactive=True
-                )
-                doc_state = gr.State(value=initial_doc)  # 用文档初始值初始化状态
-
-            with gr.Column():
-                # 获取模板列表并设置初始值
-                prompts = prompt_store.all()
-                prompt_choices = [f"{p['id']} {p['name']}" for p in prompts]
-                initial_prompt = prompt_choices[0] if prompt_choices else None
-                
-                prompt_dropdown = gr.Dropdown(
-                    choices=prompt_choices,
-                    value=initial_prompt,  # 设置初始选中项
-                    label="选择模板",
-                    allow_custom_value=True,
-                    interactive=True
-                )
-                prompt_state = gr.State(value=initial_prompt)  # 用模板初始值初始化状态
-
-        # 绑定事件（保留原有逻辑，确保交互时更新）
-        doc_dropdown.change(lambda x: x, inputs=doc_dropdown, outputs=doc_state)
-        prompt_dropdown.change(lambda x: x, inputs=prompt_dropdown, outputs=prompt_state)
-
     return demo

global_var.py

@@ -1,6 +1,4 @@
 from db import get_sqlite_engine,get_prompt_tinydb
-from tools import scan_docs_directory
 
 prompt_store = get_prompt_tinydb("workdir")
-sql_engine = get_sqlite_engine("workdir")
+sql_engine = get_sqlite_engine("workdir")
-docs = scan_docs_directory("workdir")

tools/convert_json_to_dataset.py

@@ -1,35 +0,0 @@
-from typing import List
-from schema.dataset import dataset, dataset_item, Q_A
-import json
-
-def convert_json_to_dataset(json_data: List[dict]) -> dataset:
-    # 将JSON数据转换为dataset格式
-    dataset_items = []
-    item_id = 1  # 自增ID计数器
-    for item in json_data:
-        qa = Q_A(question=item["question"], answer=item["answer"])
-        dataset_item_obj = dataset_item(id=item_id, message=[qa])
-        dataset_items.append(dataset_item_obj)
-        item_id += 1  # ID自增
-    
-    # 创建dataset对象
-    result_dataset = dataset(
-        name="Converted Dataset",
-        model_id=None,
-        description="Dataset converted from JSON",
-        dataset_items=dataset_items
-    )
-    return result_dataset
-
-# 示例：从文件读取JSON并转换
-if __name__ == "__main__":
-    # 假设JSON数据存储在文件中
-    with open(r"workdir\dataset_old\llamafactory.json", "r", encoding="utf-8") as file:
-        json_data = json.load(file)
-    
-    # 转换为dataset格式
-    converted_dataset = convert_json_to_dataset(json_data)
-    
-    # 输出结果到文件
-    with open("output.json", "w", encoding="utf-8") as file:
-        file.write(converted_dataset.model_dump_json(indent=4))