PDF转HTML工具——用springboot包装pdf2htmlEX命令行工具

Convert PDF to HTML without losing text or format.

springbootpdf2htmlEX命令行工具包装为web服务, 使得PDFHTML更方便。

pdf2htmlEX命令行工具详情见:
//github.com/pdf2htmlEX/pdf2htmlEX

pdf2html-service源码见:
//github.com/iflyendless/pdf2html-service

快速开始

# 拉取镜像
docker pull iflyendless/pdf2html-service:1.0.1

# 启动
docker run --name pdf2html -p 8686:8686 -d --rm iflyendless/pdf2html-service:1.0.1

使用:

curl -o html.zip --request POST 'localhost:8686/api/pdf2html' --form '[email protected]/pdfs/example.pdf'

提醒一下: /pdfs/example.pdf指的是pdf文件所在的绝对路径

在当前目录解压html.zip, 即可看到转换后的html文件以及000-task.txt

构建镜像

# 下载代码
git clone //github.com/iflyendless/pdf2html-service.git

# 进入项目
cd pdf2html-service

# 跳过单元测试打包
mvn clean package -DskipTests

# build docker image
docker build -t pdf2html-service:1.0.1 .

如果构建镜像失败,请检查 //enos.itcollege.ee/~jpoial/allalaadimised/jdk8/ 该站点下jdk版本是否与Dockerfile中的下载版本一致。

启动

docker run --name pdf2html -p 8686:8686 -d --rm pdf2html-service:1.0.1

如果需要格外设置一些参数的话, 可以启动docker的时候通过-e传进去:

# 同时启动的最大子进程数, 需要根据系统的资源合理设置(默认15)
-e PDF2HTML_MAX_PROCESS=15

# 执行/usr/local/bin/pdf2htmlEX命令时最大超时时间,单位s表示秒(默认600s)
-e PDF2HTML_COMMAND_TIMEOUT=600s

即:

docker run --name pdf2html -p 8686:8686 -e PDF2HTML_MAX_PROCESS=10 -e PDF2HTML_COMMAND_TIMEOUT=60s -d --rm pdf2html-service:1.0.1

更多配置见: resources目录下的application.yml文件。

Http接口

(1) 查看版本

curl //localhost:8686/api/version

(2) 查看配置

curl //localhost:8686/api/config

(3) 上传多个pdf, 并下载html压缩包

curl -o html.zip --request POST 'localhost:8686/api/pdf2html' --form '[email protected]/pdfs/001.pdf' --form '[email protected]/pdfs/002.pdf' --form '[email protected]/pdfs/003.pdf'

提醒一下: /pdfs/001.pdf指的是pdf文件所在的绝对路径

(4) 查询程序暴露出来的metric

curl //localhost:8686/api/metric

问题排查

# 进入容器
docker exec -it pdf2html bash

# 查看日志目录
cd /opt/pdf2html-service/logs

# 查看转换失败的pdf
cd /tmp/pdf2html-service/failed-pdfs

# 手动调用pdf2htmlEX命令转换pdf
pdf2htmlEX --help

实现

每次手动调用pdf2htmlEX命令行工具不太方便,在此基础上包装成一个web服务更加方便使用。完整源码见:
//github.com/iflyendless/pdf2html-service

思路

由于pdf2htmlEX命令行工具的依赖较为复杂,编译也比较麻烦,所以可直接在官方提供的Docker Image中安装JDK,然后用springboot快速编写一个web应用,接收用户http请求,后台调用pdf2htmlEX命令行工具将多个PDF都转为HTML,然后压缩生成的HTMLzip包,让用户下载。

Dockerfile如下:

# pdf2htmlex image
FROM pdf2htmlex/pdf2htmlex:0.18.8.rc1-master-20200630-Ubuntu-bionic-x86_64

ENV TZ='CST-8'
ENV LANG C.UTF-8

# apt
RUN sed -i [email protected]/archive.ubuntu.com/@/mirrors.aliyun.com/@g /etc/apt/sources.list
RUN apt-get clean && apt-get update
RUN apt-get install -y vim curl htop net-tools

# vim
RUN echo "set fileencodings=utf-8,ucs-bom,gb18030,gbk,gb2312,cp936" >> /etc/vim/vimrc
RUN echo "set termencoding=utf-8" >> /etc/vim/vimrc
RUN echo "set encoding=utf-8" >> /etc/vim/vimrc

# jdk
ADD //enos.itcollege.ee/~jpoial/allalaadimised/jdk8/jdk-8u291-linux-x64.tar.gz /tmp/
RUN tar -zxf /tmp/jdk-*.tar.gz -C /opt/ && rm -f /tmp/jdk-*.tar.gz && mv /opt/jdk* /opt/jdk

ENV JAVA_HOME /opt/jdk
ENV PATH ${JAVA_HOME}/bin:$PATH

# pdf2html-service
COPY target/pdf2html-service-*.tar.gz /tmp/
RUN tar -zxf /tmp/pdf2html-service-*.tar.gz -C /opt/ && rm -f /tmp/pdf2html-service-*.tar.gz

ENTRYPOINT [""]
WORKDIR /opt/pdf2html-service
CMD ["bash","-c","./start.sh && tail -f /dev/null"]

引入依赖

<properties>
    <project.build.sourceEncoding>UTF-8</project.build.sourceEncoding>
    <maven.compiler.source>1.8</maven.compiler.source>
    <maven.compiler.target>1.8</maven.compiler.target>
    <java.version>1.8</java.version>
    <maven.build.timestamp.format>yyyyMMdd</maven.build.timestamp.format>
    <hutool.version>5.6.3</hutool.version>
</properties>

<dependencies>

    <dependency>
        <groupId>org.springframework.boot</groupId>
        <artifactId>spring-boot-starter-web</artifactId>
    </dependency>

    <dependency>
        <groupId>org.springframework.boot</groupId>
        <artifactId>spring-boot-configuration-processor</artifactId>
    </dependency>

    <dependency>
        <groupId>org.apache.commons</groupId>
        <artifactId>commons-exec</artifactId>
        <version>1.3</version>
    </dependency>

    <dependency>
        <groupId>cn.hutool</groupId>
        <artifactId>hutool-all</artifactId>
        <version>${hutool.version}</version>
    </dependency>

    <dependency>
        <groupId>org.projectlombok</groupId>
        <artifactId>lombok</artifactId>
    </dependency>

</dependencies>

这是一个springboot应用:

import org.springframework.boot.SpringApplication;
import org.springframework.boot.autoconfigure.SpringBootApplication;
import org.springframework.boot.context.properties.ConfigurationPropertiesScan;

@SpringBootApplication
@ConfigurationPropertiesScan
public class Pdf2HtmlService {

    public static void main(String[] args) {
        SpringApplication.run(Pdf2HtmlService.class);
    }
}

程序配置

application.yml大致如下:

server:
  port: ${APP_PORT:8686}
  servlet.context-path: /

pdf2html:
  # /usr/local/bin/pdf2htmlEX --zoom 1.3
  command: ${PDF2HTML_COMMAND:/usr/local/bin/pdf2htmlEX --zoom 1.3 --quiet 1}
  command-timeout: ${PDF2HTML_COMMAND_TIMEOUT:600s}
  work-dir: ${PDF2HTML_WORK_DIR:/tmp/pdf2html-service}
  max-process: ${PDF2HTML_MAX_PROCESS:15}

spring:
  application:
    name: pdf2html-service

对应的Pdf2HtmlProperties如下:

import lombok.Data;
import org.springframework.boot.context.properties.ConfigurationProperties;

import java.time.Duration;

@Data
@ConfigurationProperties(prefix = "pdf2html")
public class Pdf2HtmlProperties {

    private String command;

    private String workDir;

    private Duration commandTimeout;

    // 同时启动的最大子进程数, 需要根据系统的性能合理设置
    private int maxProcess;
}

下面简单解释一下这几个配置的含义:

  • command:指的是调用pdf2htmlEX命令行工具的具体command,详细参数见pdf2htmlEX --help
  • command-timeout:使用的apachecommons-exec工具包,异步调用命令行,可设置最大超时时间。commons-exec的使用详情见://commons.apache.org/proper/commons-exec/tutorial.html
  • work-dir:该web应用程序的工作目录,也就是接收到用户的request时,先将pdf文件写入该目录的一个子目录下,调用pdf2htmlEX生成的html默认也是在该目录下,然后压缩该目录下生成的html文件,写入response。另外注意的是:转换失败的pdf会写入到该work-dir下的failed-pdfs下。方便复现、排查问题。
  • max-process:由于我的实现中调用命令行工具是全异步操作,必须对同时启动的命令行个数加以限制,避免短时间内产生大量子进程,不仅严重影响程序性能,而且可能导致系统瞬间卡死。所以该配置限制了同时启动的最大子进程数, 需要根据系统的性能合理设置。这里是用JDK自带的java.util.concurrent.Semaphore来限制子进程数量。

接口实现

接口实现并不复杂,关键地方也加了一些注释。如下:

import cn.hutool.core.collection.CollUtil;
import cn.hutool.core.io.FileTypeUtil;
import cn.hutool.core.io.FileUtil;
import cn.hutool.core.util.ArrayUtil;
import cn.hutool.core.util.CharsetUtil;
import cn.hutool.core.util.IdUtil;
import cn.hutool.core.util.ZipUtil;
import com.github.iflyendless.config.Pdf2HtmlProperties;
import lombok.Getter;
import lombok.extern.slf4j.Slf4j;
import org.apache.commons.exec.*;
import org.springframework.web.bind.annotation.*;
import org.springframework.web.multipart.MultipartFile;

import javax.annotation.PostConstruct;
import javax.annotation.Resource;
import javax.servlet.ServletOutputStream;
import javax.servlet.http.HttpServletResponse;
import java.io.File;
import java.io.FileFilter;
import java.net.URLEncoder;
import java.util.*;
import java.util.concurrent.ConcurrentHashMap;
import java.util.concurrent.CountDownLatch;
import java.util.concurrent.Semaphore;

@Slf4j
@RestController
@RequestMapping("/api")
public class Pdf2HtmlController {

    private static final String PDF = "pdf";
    private static final String FAILED_PDF_DIR = "failed-pdfs";
    private static final String TASK_FILE = "000-task.txt";

    @Resource
    private Pdf2HtmlProperties pdf2HtmlProperties;

    // 为了限制同时启动pdf2htmlEX命令行工具的子进程数
    private static Semaphore semaphore;

    // 转换html失败的pdf写到这个目录, 方便后面手动转换排查原因
    private static File failedPdfDir;

    @PostConstruct
    public void init() {
        semaphore = new Semaphore(pdf2HtmlProperties.getMaxProcess());
        failedPdfDir = FileUtil.mkdir(FileUtil.file(pdf2HtmlProperties.getWorkDir(), FAILED_PDF_DIR));
    }

    @GetMapping("/version")
    public Object version() {
        return "1.0.1";
    }

    @GetMapping("/config")
    public Object config() {
        return pdf2HtmlProperties;
    }

    @GetMapping("/metric")
    public Object metric() {
        Map<String, Object> semaphoreMap = new LinkedHashMap<>();
        semaphoreMap.put("availablePermits", semaphore.availablePermits());
        semaphoreMap.put("queueLength", semaphore.getQueueLength());

        Map<String, Object> metricMap = new LinkedHashMap<>();
        metricMap.put("semaphore", semaphoreMap);

        return metricMap;
    }

    @PostMapping("/pdf2html")
    public void pdf2html(@RequestParam("files") MultipartFile[] files,
                         HttpServletResponse response) {
        if (ArrayUtil.isEmpty(files)) {
            log.warn("文件数为0");
            return;
        }

        File dir = FileUtil.mkdir(FileUtil.file(pdf2HtmlProperties.getWorkDir(), IdUtil.simpleUUID()));

        try (ServletOutputStream outputStream = response.getOutputStream()) {
            List<File> fileList = new ArrayList<>(files.length);
            for (MultipartFile f : files) {
                if (f == null || f.isEmpty()) {
                    continue;
                }
                // 写入本地工作目录
                File localFile = FileUtil.writeFromStream(f.getInputStream(), FileUtil.file(dir, f.getOriginalFilename()));
                // 只处理pdf文件
                if (isPdf(localFile)) {
                    fileList.add(localFile);
                }
            }

            if (CollUtil.isEmpty(fileList)) {
                return;
            }

            long start = System.currentTimeMillis();

            int size = fileList.size();
            CountDownLatch latch = new CountDownLatch(size);
            // 处理失败的pdf统计
            Map<String, Throwable> failedMap = new ConcurrentHashMap<>();

            for (File file : fileList) {
                // 这里限制启动子进程的数量
                // 因为后面的调用是异步的, 防止瞬间产生大量子进程
                semaphore.acquire();
                // 异步调用pdf2htmlEX命令行工具
                invokeCommand(dir, file, latch, failedMap);
            }

            // 等待所有子进程结束
            latch.await();

            log.info("pdf2html一共耗时{}ms, pdf数量为{}", System.currentTimeMillis() - start, size);

            // 记录 统计数据写入文件000-task.txt, 转换html失败的pdf写入固定目录
            recordTaskResult(size, failedMap, dir, fileList);

            // 将生成的html文件以及task.txt压缩, 并写入response
            ZipUtil.zip(outputStream, CharsetUtil.CHARSET_UTF_8, true, new FileFilter() {
                @Override
                public boolean accept(File pathname) {
                    if (pathname.isDirectory()) {
                        return true;
                    }
                    String name = pathname.getName().toLowerCase();
                    return name.endsWith(".html") || name.endsWith(".txt");
                }
            }, dir);

            response.addHeader("Content-Disposition",
                    "attachment;fileName=" + URLEncoder.encode(dir.getName() + ".zip", "UTF-8"));
            response.addHeader("Content-type", "application/zip");
        } catch (Throwable e) {
            log.error("pdf2html error", e);
        } finally {
            FileUtil.del(dir);
        }
    }

    /**
     * 这里使用apache的commons-exec执行pdf2htmlEX命令行工具
     * 详情见: //commons.apache.org/proper/commons-exec/tutorial.html
     */
    public void invokeCommand(File workDir, File file, CountDownLatch latch, Map<String, Throwable> failedMap) {
        String filePath = file.getAbsolutePath();

        String line = String.format("%s --dest-dir %s %s", pdf2HtmlProperties.getCommand(), workDir.getAbsolutePath(), filePath);
        CommandLine commandLine = CommandLine.parse(line);

        // 命令行的超时处理
        ExecuteWatchdog watchdog = new ExecuteWatchdog(1000 * pdf2HtmlProperties.getCommandTimeout().getSeconds());
        // 命令行 执行完成的回调
        ResultHandler resultHandler = new ResultHandler(file, latch, failedMap);

        Executor executor = new DefaultExecutor();
        executor.setExitValue(0);
        executor.setWatchdog(watchdog);

        try {
            executor.execute(commandLine, resultHandler);
        } catch (Throwable e) {
            semaphore.release();
            String fileName = file.getName();
            if (!failedMap.containsKey(fileName)) {
                failedMap.put(fileName, e);
            }
            latch.countDown();

            log.error("invokeCommand failed, command: {}, error:{}", line, e);
        }
    }

    public static boolean isPdf(File file) {
        try {
            return PDF.equalsIgnoreCase(FileTypeUtil.getType(file));
        } catch (Exception e) {
            log.error("识别pdf类型失败, 文件名:{}, error: {}", file.getAbsolutePath(), e);
            return false;
        }
    }

    public static void recordTaskResult(int total, Map<String, Throwable> failedMap, File workDir, List<File> pdfs) {
        List<String> list = new ArrayList<>();
        list.add("total:" + total);
        list.add("success:" + (total - failedMap.size()));
        list.add("failed:" + failedMap.size());

        list.add("");
        list.add("failed-pdfs:");
        list.add("");

        Set<String> failedNames = failedMap.keySet();
        list.addAll(failedNames);

        // 记录任务完成大致情况
        FileUtil.writeLines(list, FileUtil.file(workDir, TASK_FILE), CharsetUtil.CHARSET_UTF_8);

        // 转换失败的pdf写入其他目录,后续可能需要进一步处理
        if (CollUtil.isNotEmpty(failedNames)) {
            for (File pdf : pdfs) {
                String name = pdf.getName();
                if (failedNames.contains(name)) {
                    File dest = FileUtil.file(failedPdfDir, name);
                    if (dest.exists()) {
                        dest = FileUtil.file(failedPdfDir, IdUtil.simpleUUID() + "-" + name);
                    }
                    FileUtil.copyFile(pdf, dest);
                }
            }
        }
    }

    /**
     * 根据具体的业务逻辑做相应的实现, 这里会打印一下错误日志
     */
    public static class ResultHandler implements ExecuteResultHandler {

        private final File file;
        private final CountDownLatch latch;
        private final Map<String, Throwable> failedMap;

        @Getter
        private int exitValue = -8686;

        public ResultHandler(File file, CountDownLatch latch, Map<String, Throwable> failedMap) {
            this.file = file;
            this.latch = latch;
            this.failedMap = failedMap;
        }

        @Override
        public void onProcessComplete(int exitValue) {
            semaphore.release();
            this.latch.countDown();

            this.exitValue = exitValue;
        }

        @Override
        public void onProcessFailed(ExecuteException e) {
            semaphore.release();
            this.failedMap.put(this.file.getName(), e);
            this.latch.countDown();

            log.error("pdf2html failed, file: {}, error:{}", this.file.getAbsolutePath(), e);
        }
    }
}

写在后面

由于本人对前端开发不太熟悉,就没有花时间做个简单的页面了。如果你了解前端开发而且对此工具有点兴趣,可以顺手写个页面出来,那就更好了!!!另外,如果你知道PDFHTML有更好的工具或实现,欢迎评论区留言!!!

随手记录,方便你我他。

Tags: