123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288289290291292293294295296297298299300301302303304305306307308309310311312313314315316317318319320321322323324325326327328329330 |
- using HTEXScreen.Service;
- using Microsoft.AspNetCore.Mvc;
-
- using PuppeteerSharp;
- using System.Runtime.InteropServices;
- using System.Text;
- using System.Text.Json;
- using System.Text.RegularExpressions;
- using System.Web;
- namespace HTEXScreen.Controllers
- {
- [ApiController]
- [Route("screen")]
- public class ScreenController : ControllerBase
- {
- private readonly AzureStorageFactory _azureStorage;
- private readonly HttpClient _httpClient;
- public ScreenController(HttpClient httpClient, AzureStorageFactory azureStorage) {
- _httpClient = httpClient;
- _azureStorage = azureStorage;
- }
- [HttpPost("testbase64")]
- public async Task<IActionResult> TestBase64(JsonElement json ) {
- JsonSerializerOptions option = new System.Text.Json.JsonSerializerOptions
- {
- Encoder = System.Text.Encodings.Web.JavaScriptEncoder.UnsafeRelaxedJsonEscaping
- };
- string name = "dataid";
- var reg = $"(?<=\\b{name}=)[^&]*";
- Regex regex = new Regex(reg);
- string decode= HttpUtility.UrlDecode("https://localhost:7298/swagger/index.html?dataid=%E5%BC%A0%E5%90%8C%E5%AD%A6&bs64=1234567&dataid=%E6%9D%8E%E5%90%8C%E5%AD%A6");
- Match match= Regex.Match(decode, reg);
- string a = "";
- while (match.Success)
- {
- a =a+$"{match.Value}";
- match = match.NextMatch();
- }
- string s = JsonSerializer.Serialize(json, option);
- string das= HttpUtility.UrlEncode(s,Encoding.UTF8);
- byte[] byts = Encoding.UTF8.GetBytes(s);
- var rdata = Convert.ToBase64String(byts);
- byte[] bs = Convert.FromBase64String(rdata);
- string data = Encoding.UTF8.GetString(bs);
- return Ok(new { a, rdata,das });
- }
- [HttpGet("download")]
- public async Task<IActionResult> Download([FromQuery] ScreenshotDto screenshot) {
- try {
- HttpResponseMessage response = await _httpClient.GetAsync(screenshot.url);
- if (!string.IsNullOrWhiteSpace(screenshot?.url))
- {
- string? url = screenshot?.url;
- string[] path = url.Split("/");
- string fileName = path[path.Length - 1];
- fileName=HttpUtility.UrlDecode(fileName);
- Stream stream = response.Content.ReadAsStream();
- if (!Directory.Exists("Download")){
- Directory.CreateDirectory("Download");
- }
- FileStream fs = new FileStream($"Download/{fileName}", FileMode.Create);
- byte[] bytes = new byte[stream.Length];
- stream.Read(bytes, 0, bytes.Length);
- stream.Seek(0, SeekOrigin.Begin);
- BinaryWriter bw = new BinaryWriter(fs);
- bw.Write(bytes);
- bw.Close();
- fs.Close();
- return Ok(screenshot);
- }
- else {
- return Ok("链接为空");
- }
- } catch (Exception ex) {
- return Ok($"{ex.Message}{ex.StackTrace}");
- }
- }
- /// <summary>
- /// C#使用Puppeteer http://t.zoukankan.com/zhaotianff-p-13528507.html
- /// 文档https://learnku.com/docs/puppeteer/3.1.0/class-request/8559
- /// https://www.w3cschool.cn/puppeteer/puppeteer-gp1737se.html
- /// </summary>
- /// <param name="screenshot"></param>
- /// <returns></returns>
- [HttpGet("screenshot-png")]
- public async Task<IActionResult> ScreenshotPng([FromQuery] ScreenshotDto screenshot)
- {
- // 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";
- 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;
- var browser = await Puppeteer.LaunchAsync(new LaunchOptions
- {
- ExecutablePath = BrowserExecutablePath,
- Headless = true,
- Args = new string[] { "--no-sandbox", "--disable-setuid-sandbox" }
- });
- var page = await browser.NewPageAsync();
- bool fullPage = true;
- await page.SetViewportAsync(new ViewPortOptions
- {
- Width = screenshot.width,
- Height = screenshot.height
- });;
- fullPage = false;
- await page.GoToAsync(System.Web.HttpUtility.UrlDecode(screenshot.url));
- Thread.Sleep(screenshot.delay);
- string base64 = await page.ScreenshotBase64Async(new ScreenshotOptions { FullPage = fullPage, BurstMode = true });
- //关闭浏览器
- await browser.CloseAsync();
- await browser.DisposeAsync();
- return Ok(new { url = base64 ,type= "data:image/png;base64," });
- }
- catch (Exception ex)
- {
- return BadRequest($"{ex.Message}\n{ex.StackTrace}");
- }
- }
- [HttpPost("screenshot-pdf")]
- public async Task<IActionResult> ScreenshotPngs(ScreenshotDto screenshot)
- {
- //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";
- 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;
- var browser = await Puppeteer.LaunchAsync(new LaunchOptions
- {
- ExecutablePath = BrowserExecutablePath,
- Headless = true,
- Args = new string[] { "--no-sandbox", "--disable-setuid-sandbox" }
- });
-
- ViewPortOptions viewPortOptions = new ViewPortOptions
- {
- Width = screenshot.width,
- Height = screenshot.height
- };
- PdfOptions pdfOptions = new PdfOptions { DisplayHeaderFooter = true, PreferCSSPageSize =true};
- //ScreenshotOptions screenshotOptions= new ScreenshotOptions { FullPage = fullPage, BurstMode = true };
-
- int pagesize = 20;
- if (screenshot.urls.Count <= pagesize)
- {
- await PageToPdfStream(screenshot.urls,screenshot.fileNameKey,screenshot.cnt,browser,viewPortOptions,pdfOptions);
- }
- else
- {
- List<Task> tasks = new List<Task>();
- int pages = (screenshot.urls.Count + pagesize) / pagesize;
- for (int i = 0; i < pages; i++)
- {
- var lists = screenshot.urls.Skip((i) * pagesize).Take(pagesize).ToList();
- tasks.Add(PageToPdfStream(lists, screenshot.fileNameKey, screenshot.cnt, browser, viewPortOptions, pdfOptions));
- }
- await Task.WhenAll(tasks);
- }
-
- //browser.NewPageAsync();
- //foreach (var url in screenshot.urls) {
- // var page = await browser.NewPageAsync();
- // await page.SetViewportAsync(viewPortOptions);
- // string file = $"E://pdfs//{Guid.NewGuid().ToString()}.pdf";
- // var respons = await page.GoToAsync(System.Web.HttpUtility.UrlDecode(url), WaitUntilNavigation.Networkidle2);
- // if (respons.Ok)
- // {
- // await page.PdfAsync(file, pdfOptions);
- // // string base64 = await page.ScreenshotBase64Async(screenshotOptions);
- // }
- //}
- //关闭浏览器
- await browser.CloseAsync();
- await browser.DisposeAsync();
- return Ok();
- }
- catch (Exception ex)
- {
- return BadRequest($"{ex.Message}\n{ex.StackTrace}");
- }
- }
- private async Task PageToPdfStream(List<string> urls,string fileNameKey,string cnt , Browser browser, ViewPortOptions viewPortOptions, PdfOptions pdfOptions)
- {
- 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(System.Web.HttpUtility.UrlDecode(urls[i]), 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>> uploads = new List<Task<string>>();
- 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").UploadFileByContainer(stream, "artreport", $"{id}.pdf", true));
- }
- else {
- uploads.Add(_azureStorage.GetBlobContainerClient(cnt).UploadFileByContainer(stream, "artreport", $"{id}.pdf", true));
- }
- }
- string[] uploadUrls= await Task.WhenAll(uploads);
- page_tasks.ToList().ForEach(x => {
- tasks.Add(x.DisposeAsync().AsTask());
- });
- await Task.WhenAll(tasks);
- }
- private async Task PageToPdf(List<string> urls, Browser browser, ViewPortOptions viewPortOptions, PdfOptions pdfOptions)
- {
- 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(System.Web.HttpUtility.UrlDecode(urls[i]), WaitUntilNavigation.Networkidle2));
- }
- var responses_tasks = await Task.WhenAll(responses);
- List<Task> tasks = new List<Task>();
- page_tasks.ToList().ForEach(x => {
- string file = $"{Guid.NewGuid()}.pdf";
- tasks.Add(x.PdfAsync(file, pdfOptions));
- });
- await Task.WhenAll(tasks);
- tasks.Clear();
- page_tasks.ToList().ForEach(x => {
- tasks.Add(x.DisposeAsync().AsTask());
- });
- await Task.WhenAll(tasks);
- }
- }
- }
- /// <summary>
- /// 网页截图参数
- /// </summary>
- 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; }
- }
|