TEAMModelOS/TEAMModelOS.Extension/HTEX.Screen/Service/ScreenService.cs

using Microsoft.Azure.Cosmos;
using PuppeteerSharp;
using PuppeteerSharp.Media;
using System.Runtime.InteropServices;
using System.Text;
using System.Text.Json;
using System.Text.Json.Nodes;
using System.Text.RegularExpressions;
using System.Web;
using TEAMModelOS.SDK.DI;

namespace HTEXScreen.Service
{
    public class ScreenService
    {
        public static async Task  UpdateStuArtPDF(IEnumerable<string> urls, ScreenshotDto screenshot, AzureRedisFactory _azureRedisFactory, AzureCosmosFactory _azureCosmosFactory) {
            var env = screenshot.env.Equals("release") ? "Default" : "Test";
            long now = DateTimeOffset.UtcNow.ToUnixTimeMilliseconds();
            List<Task< ResponseMessage>> responses = new List<Task<ResponseMessage>>();
            List<Task<bool>> redisSaves = new List<Task<bool>>();
            foreach (var url in urls) 
            {
                //https://teammodeltest.blob.core.chinacloudapi.cn/hbcn/art/e9a5ec36-7299-45dc-9517-7457960346c4/report/202106005.pdf
                var  uri = HttpUtility.UrlDecode(url);
                var paths =  uri.Split("/art/");
                if (paths.Length == 2) {
                    var ps= paths[1].Split("/");
                    if (ps.Length == 3) {
                        Uri uris = new Uri(paths[0]);

                        // 获取URL的Segments属性，这是一个String数组，包含URL中的每个部分
                        string[] segments = uris.Segments;
                        string key = $"ArtPDF:{ps[0]}";
                        // 确保segments数组至少有一个元素
                        if (segments.Length > 0)
                        {
                            // 获取数组中的最后一个元素，即最后一个'/'之后的部分
                            string lastSegment = segments[segments.Length - 1];
                            key = $"ArtPDF:{ps[0]}:{lastSegment}";
                        }
                        string field = ps[2].Replace(".pdf", "",StringComparison.OrdinalIgnoreCase);
                        var value = _azureRedisFactory.GetRedisClient(8, env).HashGet(key, field);
                        if (value.HasValue && !value.IsNullOrEmpty)
                        { 
                            JsonNode node = JsonNode.Parse(value.ToString());
                            if (node != null) {
                                var pdfNode= node["pdf"];
                                var id = node["id"];
                                var code = node["code"];
                                if (pdfNode != null)
                                {
                                    pdfNode["url"] = uri;
                                    pdfNode["blob"] = $"/art/{paths[1]}";
                                    pdfNode["createTime"] = now;
                                    pdfNode["prime"] = true;//此处的作用是判断是否已经生成OK.
                                    node["pdf"] = pdfNode;
                                }
                                else {
                                    pdfNode = new JsonObject();
                                    pdfNode["url"] = uri;
                                    pdfNode["blob"] = $"/art/{paths[1]}";
                                    pdfNode["createTime"] = now;
                                    pdfNode["prime"] = true;//此处的作用是判断是否已经生成OK.
                                    node["pdf"] = pdfNode;
                                   
                                }
                                string json = node.ToJsonString();
                                redisSaves.Add(_azureRedisFactory.GetRedisClient(8, env).HashSetAsync(key, field, json));
                                byte[] bytes = Encoding.UTF8.GetBytes(json);
                                var memoryStream = new MemoryStream(bytes);
                                responses.Add(_azureCosmosFactory.GetCosmosClient(null,env).GetContainer("TEAMModelOS", "Student").ReplaceItemStreamAsync(memoryStream, $"{id}", new  PartitionKey($"{code}")));
                            }
                        }
                    }
                }
            }
            await Task.WhenAll(redisSaves);
            await Task.WhenAll(responses);
        }


        public static async Task<List<(string name ,string url )>> ScreenshotPdf(ScreenshotDto screenshot, AzureStorageFactory _azureStorage) {

            //W3C School教程 https://www.w3cschool.cn/puppeteer/puppeteer-rip537tj.html
            //  进入容器的命令 docker exec -it f9e27d175498 /bin/bash

            //依赖包 https://blog.csdn.net/weixin_45447477/article/details/115188938
            //sudo apt-get install  libgdk-pixbuf2.0-0 libgdk-pixbuf-xlib-2.0-0  libdbusmenu-gtk3-4  libdbusmenu-glib4 libindicator3-7  ca-certificates fonts-liberation libappindicator3-1 libasound2 libatk-bridge2.0-0 libatk1.0-0 libc6 libcairo2 libcups2 libdbus-1-3 libexpat1 libfontconfig1 libgbm1 libgcc1 libglib2.0-0 libgtk-3-0 libnspr4 libnss3 libpango-1.0-0 libpangocairo-1.0-0 libstdc++6 libx11-6 libx11-xcb1 libxcb1 libxcomposite1 libxcursor1 libxdamage1 libxext6 libxfixes3 libxi6 libxrandr2 libxrender1 libxss1 libxtst6 lsb-release wget xdg-utils -y
            //解决ubuntu18上使用puppeteer https://blog.csdn.net/qq_42414062/article/details/114539378
            //https://www.hardkoded.com/blog/running-puppeteer-sharp-azure-functions  使用。
            //string url = "https://teammodelos.blob.core.chinacloudapi.cn/0-public/pie-borderRadius.html";
            Browser browser = null;
            try
            {

                var bfOptions = new BrowserFetcherOptions();
                if (RuntimeInformation.IsOSPlatform(OSPlatform.Linux))
                {
                    string dir = "/app";
                    if (!Directory.Exists(dir))
                    {
                        Directory.CreateDirectory(dir);
                    }
                    bfOptions.Path = dir;
                }
                var bf = new BrowserFetcher(bfOptions);
                var revisionInfo = bf.DownloadAsync(BrowserFetcher.DefaultChromiumRevision).Result;
                string BrowserExecutablePath = revisionInfo.ExecutablePath;
                browser = await Puppeteer.LaunchAsync(new LaunchOptions
                {
                    ExecutablePath = BrowserExecutablePath,
                    Headless = true,
                    Args = new string[] { "--no-sandbox", "--disable-setuid-sandbox" }
                });
                double unitPX = 37.7813;
                ViewPortOptions viewPortOptions = new ViewPortOptions
                {
                    // Width = (int)Math.Ceiling(unitPX * 21),
                    //  Height = (int)Math.Ceiling(unitPX * 29.7 * 1)
                };
                //ViewPortOptions viewPortOptions = new ViewPortOptions
                //{
                //    Width = screenshot.width,
                //    Height = screenshot.height
                //};
                PdfOptions pdfOptions = new PdfOptions { DisplayHeaderFooter = true, FooterTemplate = "", PreferCSSPageSize = true, Format = PaperFormat.A4 };
                //ScreenshotOptions screenshotOptions= new ScreenshotOptions { FullPage = fullPage, BurstMode = true };
                List<(string anme ,string url )> urls = new List<(string name, string url)>();

                if (screenshot.urls.Count <= screenshot.pagesize)
                {
                    urls.AddRange(await PageToPdfStream(_azureStorage,screenshot.urls, screenshot.fileNameKey, screenshot.cnt, screenshot.root, screenshot.env, browser, viewPortOptions, pdfOptions));
                }
                else
                {
                    List<Task<List<(string name ,string url )>>> tasks = new List<Task<List<(string name, string url)>>>();
                    int pages = (screenshot.urls.Count + screenshot.pagesize) / screenshot.pagesize;
                    for (int i = 0; i < pages; i++)
                    {
                        var lists = screenshot.urls.Skip((i) * screenshot.pagesize).Take(screenshot.pagesize).ToList();
                        tasks.Add(PageToPdfStream(_azureStorage,lists, screenshot.fileNameKey, screenshot.cnt, screenshot.root, screenshot.env, browser, viewPortOptions, pdfOptions));
                    }
                    var tsk = await Task.WhenAll(tasks);
                    foreach (var ts in tsk)
                    {
                        urls.AddRange(ts);
                    }
                }

                //browser.NewPageAsync();
                //foreach (var url in screenshot.urls) {
                //    var page = await browser.NewPageAsync();
                //    await page.SetViewportAsync(viewPortOptions);

                //    string file = $"E://pdfs//{Guid.NewGuid().ToString()}.pdf";
                //    var respons = await page.GoToAsync(System.Web.HttpUtility.UrlDecode(url), WaitUntilNavigation.Networkidle2);
                //    if (respons.Ok)
                //    {
                //        await page.PdfAsync(file, pdfOptions);
                //        // string base64 = await page.ScreenshotBase64Async(screenshotOptions);
                //    }
                //}

                //关闭浏览器
                await browser.CloseAsync();
                await browser.DisposeAsync();
                return urls;
            }
            catch (Exception ex)
            {
                StreamWriter file = new StreamWriter("erorr_log.txt", append: true);
                await file.WriteLineAsync($"ScreenService {JsonSerializer.Serialize(screenshot)}-----{ex.Message}----{ex.StackTrace}");
                file.Close();
                return new List<(string name, string url)>();
                //return BadRequest($"{ex.Message}\n{ex.StackTrace}");
            }
            finally
            {
                if (browser != null && !browser.IsClosed)
                {
                    await browser.CloseAsync();
                    await browser.DisposeAsync();
                }
            }

        }

        private static async Task<List<(string name ,string url )>> PageToPdfStream(AzureStorageFactory _azureStorage,List<string> urls, string fileNameKey, string cnt, string root, string env, Browser browser, ViewPortOptions viewPortOptions, PdfOptions pdfOptions)
        {
            string name = env.Equals("release") ? "Default" : "Test";
            List<Task<Page>> pages = new List<Task<Page>>();
            urls.ForEach(x => {
                pages.Add(browser.NewPageAsync());
            });
            var page_tasks = await Task.WhenAll(pages);
            List<Task<Response>> responses = new List<Task<Response>>();
            page_tasks.ToList().ForEach(x => {
                x.SetViewportAsync(viewPortOptions);

            });
            for (int i = 0; i < urls.Count; i++)
            {
                responses.Add(page_tasks[i].GoToAsync(urls[i], 30000, new WaitUntilNavigation[] { WaitUntilNavigation.Networkidle2 }));
            }
            var responses_tasks = await Task.WhenAll(responses);
            //List<Task<Stream>> streams = new List<Task<Stream>>();
            List<Task> tasks = new List<Task>();
            List<Task<(string name , string url )>> uploads = new List<Task<(string name, string url)>>();
            foreach (var page_task in page_tasks)
            {
                string url = page_task.Url;
                string[] paths = HttpUtility.UrlDecode(url).Split("/");
                var reg = $"(?<=\\b{fileNameKey}=)[^&]*";
                Regex regex = new Regex(reg);
                string decode = HttpUtility.UrlDecode(url);
                Match match = Regex.Match(decode, reg);
                string id = "";
                while (match.Success)
                {
                    id = id + $"{match.Value}";
                    match = match.NextMatch();
                }
                //需要解析参数。paths[paths.Length-1]
                Stream stream = await page_task.PdfStreamAsync(pdfOptions);
                if (string.IsNullOrWhiteSpace(cnt))
                {
                    uploads.Add(_azureStorage.GetBlobContainerClient("teammodelos", name).UploadFileByContainerBName(stream, root, $"{id}.pdf", true));
                }
                else
                {
                    uploads.Add(_azureStorage.GetBlobContainerClient(cnt, name).UploadFileByContainerBName(stream, root, $"{id}.pdf", true));
                }
            }
            (string name ,string url )[]  uploadUrls = await Task.WhenAll(uploads);
            page_tasks.ToList().ForEach(x => {
                tasks.Add(x.DisposeAsync().AsTask());
            });
            await Task.WhenAll(tasks);
            return uploadUrls.ToList();
        }
    }
    public class ScreenshotDto
    {
        public int width { get; set; } = 1920;
        public int height { get; set; } = 1080;
        public string? url { get; set; }
        /// <summary>
        /// 批量地址
        /// </summary>
        public List<string> urls { get; set; } = new List<string>();
        /// <summary>
        /// 提取参数的唯一id作为文件名
        /// </summary>
        public string? fileNameKey { get; set; }
        /// <summary>
        /// 存在哪个容器里
        /// </summary>
        public string? cnt { get; set; }
        public int delay { get; set; }
        public int pagesize { get; set; } = 5;
        public string? root { get; set; }
        public string? env { get; set; } = "release";
        public string? msgId { get; set; }
    }
}