server.py

from flask import Flask, request, jsonify
import os
import base64
import subprocess
import copy
from flask import Flask, send_file, abort
from pypdf import PdfWriter, PdfReader
from pypdf.generic import RectangleObject
import sys
import shutil
import string
from flask_cors import CORS


######################################## 默认配置 ########################################
port_num = int(os.environ.get("PORT", 8888))  # 设置端口号: 默认为8888
pdf2zh = "babeldoc"  # 设置pdf2zh指令: 默认为'pdf2zh'

######### 可以在Zotero偏好设置中配置以下参数, Zotero配置会覆盖本文件中的配置参数 #########
thread_num = 4  # 设置线程数: 默认为4
service = "bing"  # 设置翻译服务: 默认为bing
translated_dir = "./translated/"  # 设置翻译文件的输出路径(临时路径, 可以在翻译后删除)
config_path = "./config.toml"  # 设置PDF2zh配置文件路径
source_languages = "en"  # 设置源语言
target_languages = "zh"  # 设置目标语言
global_translated_dir = translated_dir

# 从环境变量读取OpenAI配置
openai_base_url = os.environ.get("OPENAI_BASE_URL", "")
openai_model = os.environ.get("OPENAI_MODEL", "gpt-4o")
openai_api_key = os.environ.get("OPENAI_API_KEY", "")
model_type = openai_model  # 用于判断模型类型
claude_api_key = os.environ.get("CLAUDE_API_KEY", "")
##########################################################################################


class Config:
    def __init__(self, request):
        self.thread_num = request.get_json().get("threadNum")
        if self.thread_num == None or self.thread_num == "":
            self.thread_num = thread_num

        self.service = request.get_json().get("engine")
        if self.service == None or self.service == "":
            self.service = service

        self.source_languages = request.get_json().get("sourceLanguages")
        if self.source_languages == None or self.source_languages == "":
            self.source_languages = source_languages

        self.target_languages = request.get_json().get("targetLanguages")
        if self.target_languages == None or self.target_languages == "":
            self.target_languages = target_languages

        self.translated_dir = request.get_json().get("outputPath")
        if self.translated_dir == None or self.translated_dir == "":
            self.translated_dir = translated_dir
        self.translated_dir = get_absolute_path(self.translated_dir)
        os.makedirs(self.translated_dir, exist_ok=True)

        self.config_path = request.get_json().get("configPath")
        if self.config_path == None or self.config_path == "":
            self.config_path = config_path
        self.config_path = get_absolute_path(self.config_path)

        self.mono_cut = request.get_json().get("mono_cut")
        self.dual_cut = request.get_json().get("dual_cut")
        self.compare = request.get_json().get("compare")

        print("outputPath: ", self.translated_dir)
        print("configPath: ", self.config_path)

        global global_translated_dir
        global_translated_dir = self.translated_dir


def get_absolute_path(path):
    if os.path.isabs(path):
        return path
    else:
        return os.path.abspath(path)


def get_file_from_request(request):
    config = Config(request)
    data = request.get_json()
    path = data.get("filePath")
    print("filePath: ", path)
    path = path.replace('\\', '/') # 把所有反斜杠\替换为正斜杠/ (Windows->Linux/MacOS)
    file_content = data.get("fileContent")
    input_path = os.path.join(config.translated_dir, os.path.basename(path))
    input_path = get_absolute_path(input_path)
    print("input path: ", input_path)
    if file_content:
        if file_content.startswith(
            "data:application/pdf;base64,"
        ):  # 移除 Base64 编码中的前缀(如果有)
            file_content = file_content[len("data:application/pdf;base64,") :]
        file_data = base64.b64decode(file_content)  # 解码 Base64 内容
        with open(input_path, "wb") as f:
            f.write(file_data)
    return input_path, config


def translate_pdf(input_path, config):
    print("\n############# Translating #############")
    print("## translate file path ## : ", input_path)
    print("## translated_dir ## : ", config.translated_dir)
    print("## config_path ## : ", config.config_path)
    
    try:
        # 检查是否存在本地配置文件
        local_config_exists = os.path.exists(config.config_path)
        print("## local_config_exists ## : ", local_config_exists)
        
        # 判断是否在部署环境中运行（通过检查环境变量）
        is_deployed_env = os.environ.get("OPENAI_BASE_URL") or os.environ.get("OPENAI_API_KEY")
        print("## is_deployed_env ## : ", is_deployed_env)
        
        # 如果在部署环境中运行且有环境变量配置，则使用环境变量生成config.toml
        if is_deployed_env and not local_config_exists:
            try:
                # 如果环境变量中有配置，则使用环境变量中的配置生成config.toml
                template_path = os.path.join(os.path.dirname(os.path.abspath(__file__)), "config.toml.template")
                print("## template_path ## : ", template_path)
                
                with open(template_path, "r") as template_file:
                    template_content = template_file.read()
                
                # 替换模板中的环境变量
                template = string.Template(template_content)
                config_content = template.substitute({
                    "OPENAI_BASE_URL": openai_base_url,
                    "OPENAI_MODEL": openai_model,
                    "OPENAI_API_KEY": openai_api_key
                })
                
                # 写入配置文件
                config_file_path = os.path.join(os.path.dirname(os.path.abspath(__file__)), config_path)
                print("## config_file_path ## : ", config_file_path)
                
                with open(config_file_path, "w") as config_file:
                    config_file.write(config_content)
                
                # 确保使用配置文件
                config.config_path = config_file_path
                print("## 配置文件已生成 ## : ", config.config_path)
            except Exception as e:
                print(f"## 生成配置文件时出错 ## : {str(e)}")
                raise

        # 确保翻译目录存在
        os.makedirs(config.translated_dir, exist_ok=True)
        print(f"## 确保翻译目录存在 ## : {config.translated_dir}")
        
        # 执行pdf2zh翻译, 用户可以自定义命令内容:
        if not os.path.exists(config.config_path):
            command = [
                pdf2zh,
                input_path,
                "--t",
                str(config.thread_num),
                "--output",
                config.translated_dir,
                "--service",
                config.service,
                "--lang-in",
                config.source_languages,
                "--lang-out",
                config.target_languages,
            ]
            
            # 如果设置了API密钥，添加到命令中
            if model_type.lower().startswith("gpt") and openai_api_key:
                command.extend(["--openai-api-key", openai_api_key, "--openai-model", model_type])
            elif model_type.lower().startswith("claude") and claude_api_key:
                command.extend(["--claude-api-key", claude_api_key, "--claude-model", model_type])
        else:
            command = [pdf2zh, "-c", config.config_path, "--files", input_path]
        
        print("## 执行命令 ## : ", " ".join(command))
        
        # 使用Popen而不是run，以便实时显示输出（包括进度条）
        process = subprocess.Popen(
            command,
            stdout=subprocess.PIPE,
            stderr=subprocess.PIPE,
            text=True,
            bufsize=1,
            universal_newlines=True
        )
        
        # 实时读取并显示输出
        stdout_lines = []
        stderr_lines = []
        
        while True:
            stdout_line = process.stdout.readline()
            stderr_line = process.stderr.readline()
            
            if stdout_line:
                print(stdout_line.strip())
                stdout_lines.append(stdout_line)
            if stderr_line:
                print(stderr_line.strip())
                stderr_lines.append(stderr_line)
            
            # 检查进程是否结束
            if process.poll() is not None:
                # 读取剩余输出
                for line in process.stdout:
                    print(line.strip())
                    stdout_lines.append(line)
                for line in process.stderr:
                    print(line.strip())
                    stderr_lines.append(line)
                break
        
        # 获取返回码
        returncode = process.returncode
        stdout = ''.join(stdout_lines)
        stderr = ''.join(stderr_lines)
        
        print(f"## 命令执行结果 ## : 返回码={returncode}")
        
        if returncode != 0:
            error_msg = f"命令执行失败，返回码: {returncode}, 错误: {stderr}"
            print(f"## 详细错误信息 ## : {error_msg}")
            raise Exception(error_msg)

        # 检查输出文件
        expected_mono = os.path.join(
            config.translated_dir,
            os.path.basename(input_path).replace(".pdf", ".zh.mono.pdf"),
        )
        expected_dual = os.path.join(
            config.translated_dir,
            os.path.basename(input_path).replace(".pdf", ".zh.dual.pdf"),
        )
        
        print(f"## 检查输出文件 ## : mono={expected_mono}, dual={expected_dual}")
        print(f"## 文件存在检查 ## : mono存在={os.path.exists(expected_mono)}, dual存在={os.path.exists(expected_dual)}")
        
        # 执行 mv 命令
        mono = os.path.join(
            config.translated_dir, os.path.basename(input_path).replace(".pdf", "-mono.pdf")
        )
        dual = os.path.join(
            config.translated_dir, os.path.basename(input_path).replace(".pdf", "-dual.pdf")
        )
        
        try:
            if os.path.exists(expected_mono):
                shutil.move(expected_mono, mono)
                print(f"## 移动文件成功 ## : {expected_mono} -> {mono}")
            else:
                raise Exception(f"源文件不存在: {expected_mono}")
                
            if os.path.exists(expected_dual):
                shutil.move(expected_dual, dual)
                print(f"## 移动文件成功 ## : {expected_dual} -> {dual}")
            else:
                raise Exception(f"源文件不存在: {expected_dual}")
        except Exception as e:
            print(f"## 移动文件时出错 ## : {str(e)}")
            raise
        
        if not os.path.exists(mono) or not os.path.exists(dual):
            raise Exception("[Failed to generate translated files]: " + mono + ", " + dual)
            
        print("[mono file generated]: ", mono)
        print("[dual file generated]: ", dual)
        return mono, dual
    except Exception as e:
        print(f"## translate_pdf函数出错 ## : {str(e)}")
        # 重新抛出异常，以便上层函数可以捕获
        raise


app = Flask(__name__)
CORS(app, resources={r"/*": {"origins": "*"}})  # 允许所有来源的跨域请求


@app.route("/translate", methods=["POST"])
def translate():
    try:
        print("\n############# 开始翻译请求处理 #############")
        input_path, config = get_file_from_request(request)
        print(f"## 获取到输入文件 ## : {input_path}")
        
        mono, dual = translate_pdf(input_path, config)
        print(f"## 翻译完成 ## : mono={mono}, dual={dual}")
        
        if config.mono_cut and config.mono_cut == "true":
            try:
                path = mono.replace("-mono.pdf", "-mono-cut.pdf")
                print(f"## 开始切割mono文件 ## : {mono} -> {path}")
                split_and_merge_pdf(mono, path, compare=False)
                if not os.path.exists(path):
                    raise Exception("[Failed to generate cutted files]: " + path)
                print("[mono-cut file generated]: ", path)
            except Exception as e:
                print(f"## 切割mono文件出错 ## : {str(e)}")
                raise
                
        if config.dual_cut and config.dual_cut == "true":
            try:
                path = dual.replace("-dual.pdf", "-dual-cut.pdf")
                print(f"## 开始切割dual文件 ## : {dual} -> {path}")
                split_and_merge_pdf(dual, path, compare=False)
                if not os.path.exists(path):
                    raise Exception("[Failed to generate cutted files]: " + path)
                print("[dual-cut file generated]: ", path)
            except Exception as e:
                print(f"## 切割dual文件出错 ## : {str(e)}")
                raise
                
        if config.compare and config.compare == "true":
            try:
                path = dual.replace("-dual.pdf", "-compare.pdf")
                print(f"## 开始生成对比文件 ## : {dual} -> {path}")
                split_and_merge_pdf(dual, path, compare=True)
                if not os.path.exists(path):
                    raise Exception("[Failed to generate compare files]: " + path)
                print("[compare file generated]: ", path)
            except Exception as e:
                print(f"## 生成对比文件出错 ## : {str(e)}")
                raise
                
        return jsonify({"status": "success"}), 200
    except Exception as e:
        print(f"[Translate Error]: {e}")
        # 返回更详细的错误信息
        error_message = str(e)
        traceback_info = sys.exc_info()
        if traceback_info[2]:
            import traceback
            traceback_str = "".join(traceback.format_tb(traceback_info[2]))
            print(f"## 错误堆栈 ## : {traceback_str}")
            error_message = f"{error_message}\n{traceback_str}"
        return jsonify({"status": "error", "message": error_message}), 500


@app.route("/translatedFile/<filename>")
def download(filename):
    print("\n############# Downloading #############")
    file_path = os.path.join(get_absolute_path(global_translated_dir), filename)
    if not os.path.isfile(file_path):
        print("[Download File not found]: ", file_path)
        return "[Download File not found]: " + file_path, 404
    print("[Download file]: ", file_path)
    return send_file(file_path, as_attachment=True, download_name=filename)


# 工具函数, 用于切割双栏pdf文件
def split_and_merge_pdf(input_pdf, output_pdf, compare=False):
    writer = PdfWriter()
    if "dual" in input_pdf:
        readers = [PdfReader(input_pdf) for _ in range(4)]
        for i in range(0, len(readers[0].pages), 2):
            original_media_box = readers[0].pages[i].mediabox
            width = original_media_box.width
            height = original_media_box.height

            left_page_1 = readers[0].pages[i]
            for box in ["mediabox", "cropbox", "trimbox", "bleedbox", "artbox"]:
                setattr(left_page_1, box, RectangleObject((0, 0, width / 2, height)))

            left_page_2 = readers[1].pages[i + 1]
            for box in ["mediabox", "cropbox", "trimbox", "bleedbox", "artbox"]:
                setattr(left_page_2, box, RectangleObject((0, 0, width / 2, height)))

            right_page_1 = readers[2].pages[i]
            for box in ["mediabox", "cropbox", "trimbox", "bleedbox", "artbox"]:
                setattr(
                    right_page_1, box, RectangleObject((width / 2, 0, width, height))
                )

            right_page_2 = readers[3].pages[i + 1]
            for box in ["mediabox", "cropbox", "trimbox", "bleedbox", "artbox"]:
                setattr(
                    right_page_2, box, RectangleObject((width / 2, 0, width, height))
                )

            if compare == True:
                blank_page_1 = writer.add_blank_page(width, height)
                blank_page_1.merge_transformed_page(left_page_1, (1, 0, 0, 1, 0, 0))
                blank_page_1.merge_transformed_page(
                    left_page_2, (1, 0, 0, 1, width / 2, 0)
                )
                blank_page_2 = writer.add_blank_page(width, height)
                blank_page_2.merge_transformed_page(
                    right_page_1, (1, 0, 0, 1, -width / 2, 0)
                )
                blank_page_2.merge_transformed_page(right_page_2, (1, 0, 0, 1, 0, 0))
            else:
                writer.add_page(left_page_1)
                writer.add_page(left_page_2)
                writer.add_page(right_page_1)
                writer.add_page(right_page_2)
    else:
        readers = [PdfReader(input_pdf) for _ in range(2)]
        for i in range(len(readers[0].pages)):
            page = readers[0].pages[i]

            original_media_box = page.mediabox
            width = original_media_box.width
            height = original_media_box.height

            left_page = readers[0].pages[i]
            left_page.mediabox = RectangleObject((0, 0, width / 2, height))
            right_page = readers[1].pages[i]
            right_page.mediabox = RectangleObject((width / 2, 0, width, height))

            writer.add_page(left_page)
            writer.add_page(right_page)

    with open(output_pdf, "wb") as output_file:
        writer.write(output_file)


# 用于切割双栏pdf文件
@app.route("/cut", methods=["POST"])
def cut():
    print("\n############# Cutting #############")
    input_path, config = get_file_from_request(request)
    try:
        translated_path = os.path.join(
            config.translated_dir,
            os.path.basename(input_path).replace(".pdf", "-cut.pdf"),
        )
        split_and_merge_pdf(input_path, translated_path)

        if not os.path.exists(translated_path):
            raise Exception("[Failed to generate cut files]: ", translated_path)
        print("[Cut file generated]: ", translated_path)
        return jsonify({"status": "success"}), 200
    except Exception as e:
        print(f"[Cut File Error]: {e}")
        # 返回更详细的错误信息
        error_message = str(e)
        traceback_info = sys.exc_info()
        if traceback_info[2]:
            import traceback
            traceback_str = "".join(traceback.format_tb(traceback_info[2]))
            print(f"## 错误堆栈 ## : {traceback_str}")
            error_message = f"{error_message}\n{traceback_str}"
        return jsonify({"status": "error", "message": error_message}), 500


# 用于生成中英对照文件
@app.route("/cut-compare", methods=["POST"])
def cut_compare():
    print("\n############# Comparing #############")
    try:
        input_path, config = get_file_from_request(request)
        print(f"## 获取到输入文件 ## : {input_path}")
        
        if "dual" in input_path:
            translated_path = os.path.join(
                config.translated_dir,
                os.path.basename(input_path).replace(".pdf", "-compare.pdf"),
            )
            print(f"## 直接生成对比文件 ## : {input_path} -> {translated_path}")
            
            # 确保翻译目录存在
            os.makedirs(os.path.dirname(translated_path), exist_ok=True)
            
            split_and_merge_pdf(input_path, translated_path, compare=True)
        else:
            print(f"## 需要先翻译再生成对比文件 ## : {input_path}")
            _, dual = translate_pdf(input_path, config)
            translated_path = dual.replace("-dual.pdf", "-compare.pdf")
            print(f"## 生成对比文件 ## : {dual} -> {translated_path}")
            split_and_merge_pdf(dual, translated_path, compare=True)

        if not os.path.exists(translated_path):
            raise Exception("[Failed to generate cutted file]: " + translated_path)
        print("[Compare file generated]: ", translated_path)
        return jsonify({"status": "success"}), 200
    except Exception as e:
        print(f"[cut_compare() Error]: {e}")
        # 返回更详细的错误信息
        error_message = str(e)
        traceback_info = sys.exc_info()
        if traceback_info[2]:
            import traceback
            traceback_str = "".join(traceback.format_tb(traceback_info[2]))
            print(f"## 错误堆栈 ## : {traceback_str}")
            error_message = f"{error_message}\n{traceback_str}"
        return jsonify({"status": "error", "message": error_message}), 500


if __name__ == "__main__":
    if len(sys.argv) > 1:
        port_num = int(sys.argv[1])
    app.run(host="0.0.0.0", port=port_num)