新文件解析服务

# 安装最新版本
pip install zai-sdk

# 或指定版本
pip install zai-sdk==0.2.0

from zai import ZhipuAiClient

client = ZhipuAiClient(api_key="your api key")
# 用于上传发起文件解析任务
# 返回task_id
response = client.file_parser.create(file=open('example.pdf', 'rb'), file_type='pdf', tool_type='lite')
task_id = getattr(response, "task_id", None)

# 获取文件内容抽取: format_type = text / download_link
# text模式最长返回1m以内的文本内容，download_link响应更快
res_response = client.file_parser.content(task_id=task_id, format_type="download_link")

print(response.json())  # 新版推荐用法
print(response.content.decode('utf-8')) # 旧版解码字节流用法依然支持

# 安装最新版本
pip install zhipuai

# 或指定版本
pip install zhipuai==2.1.5.20250825

from pathlib import Path
from zhipuai import ZhipuAI

client = ZhipuAI(api_key="your api key")
# 用于上传发起文件解析任务
# 返回task_id
response = client.file_parser.create(file=open('example.pdf', 'rb'), file_type='pdf', tool_type='lite')
print(response)

# 获取文件内容抽取
response = client.file_parser.content(task_id="your task_id", format_type="text")
print(response.content.decode('utf-8'))

<dependency>
    <groupId>ai.z.openapi</groupId>
    <artifactId>zai-sdk</artifactId>
    <version>0.3.0</version>
</dependency>

import ai.z.openapi.ZhipuAiClient;
import ai.z.openapi.service.fileparsing.FileParsingDownloadReq;
import ai.z.openapi.service.fileparsing.FileParsingDownloadResponse;
import ai.z.openapi.service.fileparsing.FileParsingResponse;
import ai.z.openapi.service.fileparsing.FileParsingUploadReq;
import ai.z.openapi.utils.StringUtils;

public class FileParsingExample {

    public static void main(String[] args) {
        // 初始化客户端
        ZhipuAiClient client = ZhipuAiClient.builder().ofZHIPU()
             .apiKey("your api key")
             .build();

        try {
            // 示例1: 创建解析任务
            System.out.println("=== 文件解析任务创建示例 ===");
            String filePath = "your file path";
            String taskId = createFileParsingTaskExample(client, filePath, "pdf", "lite");

            // 示例2: 获取解析结果
            System.out.println("\n=== 获取解析结果示例 ===");
            getFileParsingResultExample(client, taskId);

        } catch (Exception e) {
            System.err.println("发生异常: " + e.getMessage());
            e.printStackTrace();
        }
    }

    /**
    * 示例：创建解析任务（上传文件并解析）
    *
    * @param client ZhipuAiClient 实例
    * @return 解析任务的 taskId
    */
    private static String createFileParsingTaskExample(ZhipuAiClient client, String filePath, String fileType, String toolType) {
        if (StringUtils.isEmpty(filePath)) {
            System.err.println("无效的文件路径。");
            return null;
        }
        try {
            FileParsingUploadReq uploadReq = FileParsingUploadReq.builder()
                    .filePath(filePath)
                    .fileType(fileType)  // 支持: pdf, docx 等
                    .toolType(toolType) // 解析工具类型: lite, prime, expert
                    .build();

            System.out.println("正在上传并创建解析任务...");
            FileParsingResponse response = client.fileParsing().createParseTask(uploadReq);
            if (response.isSuccess()) {
                if (null != response.getData().getTaskId()) {
                    String taskId = response.getData().getTaskId();
                    System.out.println("解析任务创建成功，TaskId: " + taskId);
                    return taskId;
                } else {
                    System.err.println("解析任务创建失败: " + response.getData().getMessage());
                }
            } else {
                System.err.println("解析任务创建失败: " + response.getMsg());
            }
        } catch (Exception e) {
            System.err.println("文件解析任务错误: " + e.getMessage());
        }
        // 返回 null 表示创建失败
        return null;
    }

    /**
    * 示例：获取解析结果
    *
    * @param client ZhipuAiClient 实例
    * @param taskId 解析任务ID
    */
    private static void getFileParsingResultExample(ZhipuAiClient client, String taskId) {
        if (taskId == null || taskId.isEmpty()) {
            System.err.println("无效的任务ID，无法获取解析结果。");
            return;
        }

        try {
            int maxRetry = 100;      // 最多轮询100次
            int intervalMs = 3000;  // 每次间隔3秒
            for (int i = 0; i < maxRetry; i++) {
                FileParsingDownloadReq downloadReq = FileParsingDownloadReq.builder()
                        .taskId(taskId)
                        .formatType("text")
                        .build();

                FileParsingDownloadResponse response = client.fileParsing().getParseResult(downloadReq);

                if (response.isSuccess()) {
                    String status = response.getData().getStatus();
                    System.out.println("当前任务状态: " + status);

                    if ("succeeded".equalsIgnoreCase(status)) {
                        System.out.println("解析结果获取成功！");
                        System.out.println("解析内容: " + response.getData().getContent());
                        System.out.println("内容下载链接: " + response.getData().getParsingResultUrl());
                        return;
                    } else if ("processing".equalsIgnoreCase(status)) {
                        System.out.println("解析进行中，请稍候...");
                        Thread.sleep(intervalMs);
                    } else {
                        System.out.println("解析任务异常，状态: " + status + "，消息: " + response.getData().getMessage());
                        return;
                    }
                } else {
                    System.err.println("解析结果获取失败: " + response.getMsg());
                    return;
                }
            }
            System.out.println("等待超时，请稍后自行查询解析结果。");
        } catch (Exception e) {
            System.err.println("获取解析结果时异常: " + e.getMessage());
        }
    }
}

{
    "message": "任务创建成功",
    "success": true,
    "task_id": "task_id"
}

{
    "status": "succeeded",
    "message": "结果获取成功",
    "content": "parsed result text",
    "task_id": "your task_id",
    "parsing_result_url": "download url"
}