init

judgement-text-annotator/__init__.py

@@ -0,0 +1,178 @@
+import json
+from datetime import datetime
+from docx import Document
+import gradio as gr
+import csv
+import os
+import tempfile
+from io import StringIO
+from dotenv import load_dotenv
+
+from openai import OpenAI
+
+load_dotenv()
+
+# Deepseek API
+API_key = os.getenv("DEEPSEEK_API_KEY")
+
+file_path = []
+
+mark_judgment_info = ""
+def extract_info_from_judgment(judgment_text):
+    system_prompt = """
+        角色：
+        你是一位专业且经验丰富的法律文书信息提取助手。
+
+        背景：
+        我们有一系列交通事故相关的法律文书，需要从中提取关键信息。
+
+        知识：
+        法律文书包含了诸如案件类型、文书ID、案件名称、案件编号、裁判日期等多方面的信息。
+
+        任务：
+        仔细阅读输入的法律文书，按照规定的JSON格式准确提取各项信息。若文书中未提及某项信息，则对应字段填写“空”。
+
+        其他：
+        JSON格式不用加上```json```标记，直接填写即可。
+
+        """
+
+    user_prompt = """请从以下法律文书中提取相关信息：
+            "案件类型": "空",
+            "文书ID": "空",
+            "案件名称一": "空",
+            "案件名称二": "空",
+            "案件编号": "空",
+            "裁判日期": "空",
+            "法院名称": "空",
+            "肇事人": "空",
+            "性别": "空",
+            "出生日期": "空",
+            "民族": "空",
+            "文化程度": "空",
+            "户籍所在地": "空",
+            "案发时间": "空",
+            "车辆品牌和车型": "空",
+            "事故发生地": "空",
+            "酒精": "空",
+            "伤亡数量": "空",
+            "驾照实习期开始": "空",
+            "驾驶实习期结束": "空",
+            "驾照类型": "空",
+            "实习期类型": "空",
+            "经济损失": "空",
+            "撤销案件号": "空",
+            "维持案件号": "空"
+        """ + judgment_text
+
+    messages = [
+        {"role": "system", "content": system_prompt},
+        {"role": "user", "content": user_prompt}
+    ]
+
+    client = OpenAI(api_key=API_key, base_url="https://api.deepseek.com")
+
+    try:
+        completion = client.chat.completions.create(
+            model="deepseek-chat",
+            messages=messages
+        )
+        print(completion)
+        content = completion.choices[0].message.content if completion.choices else ""
+        # print(content)
+
+        try:
+            info_dict = json.loads(content)
+            return info_dict
+        except json.JSONDecodeError as e:
+            print(f"解析JSON时出错: {e}")
+            return {}
+
+    except Exception as e:
+        print(f"An error occurred: {e}")
+    return {}
+
+def mark_judgment(judgment_text):
+    global mark_judgment_info
+    if judgment_text == "":
+        return []
+    info_dict = extract_info_from_judgment(judgment_text)
+    mark_judgment_info = [[key, value] for key, value in info_dict.items()]
+    return mark_judgment_info
+
+def download_csv():
+    output = StringIO()
+    writer = csv.writer(output)
+    for key, value in mark_judgment_info:
+        writer.writerow([key, value])
+    output.seek(0)
+
+    # 保存到临时文件
+    timestamp = datetime.now().strftime("%Y%m%d%H%M%S")
+    temp_file_path = os.path.join(tempfile.gettempdir(), f"marked_judgment_{timestamp}.csv")
+
+
+    with open(temp_file_path, 'w', newline='', encoding='utf-8') as f:
+        f.write(output.getvalue())
+
+    file_path.append(temp_file_path)
+
+    return file_path
+
+def clear_input():
+    return "", []
+
+def read_docx(file):
+    doc = Document(file.name)
+    content = []
+    for paragraph in doc.paragraphs:
+        content.append(paragraph.text)
+    return '\n'.join(content)
+
+def read_uploaded_file(file):
+
+    if file is None:
+        return ""
+    print(file.name)
+    is_doc = file.name.endswith(".doc") or file.name.endswith(".docx")
+    if is_doc:
+        connect = read_docx(file)
+        return connect
+    else:
+        try:
+            with open(file.name, 'r', encoding='utf-8') as f:
+                content = f.read()
+                return content
+        except UnicodeDecodeError as e:
+            return "读取文件时出现编码错误，请检查文件编码是否正确"
+
+def main():
+    with gr.Blocks() as demo:
+        gr.Markdown("输入判决书文本内容，系统将输出标记后的信息。")
+
+        with gr.Row():
+            with gr.Column():
+                input_text = gr.Textbox(
+                    lines=10,
+                    placeholder="请输入判决书文本内容...",
+                    label="",
+                    interactive=True
+                )
+                with gr.Row():
+                    submit_btn = gr.Button("提交")
+                    clear_btn = gr.Button("清除")
+            with gr.Column():
+                output_boxes = gr.DataFrame(headers=["Key", "Value"], visible=True, wrap=True, interactive=True)
+                download_btn = gr.Button("转为CSV", elem_id="download-btn")
+        with gr.Row():
+            upload_file = gr.File(label="上传判决书", elem_id="upload-file")
+
+        upload_file.change(fn=read_uploaded_file, inputs=upload_file, outputs=input_text)
+        submit_btn.click(fn=mark_judgment, inputs=input_text, outputs=output_boxes)
+        clear_btn.click(fn=clear_input, outputs=[input_text, output_boxes])
+        download_btn.click(fn=download_csv, outputs=gr.File(label="CSV文件"))
+
+        demo.launch()
+
+if __name__ == "__main__":
+   main()