Files
FileFlowsPlugins/ImageNodes/Tools/WebsiteImageScraper.cs
2024-01-11 15:00:16 +13:00

502 lines
18 KiB
C#

using System.Text.RegularExpressions;
using HtmlAgilityPack;
using ImageMagick;
namespace FileFlows.ImageNodes.Tools;
/// <summary>
/// Scrapes a website, follows links within a specified depth, and saves images over a certain resolution.
/// </summary>
public class WebImageScraper : Node
{
private HashSet<string> savedImages = new ();
private static readonly HttpClient httpClient = new HttpClient();
private static readonly SemaphoreSlim semaphore = new SemaphoreSlim(5); // Limit to 5 concurrent image downloads
private int PageCount = 0;
/// <summary>
/// Gets or sets the maximum depth of links to follow.
/// </summary>
[NumberInt(1)]
public int MaxDepth { get; set; }
/// <summary>
/// Gets or sets the minimum width of images to save.
/// </summary>
[NumberInt(2)]
public int MinWidth { get; set; }
/// <summary>
/// Gets or sets the minimum height of images to save.
/// </summary>
[NumberInt(3)]
public int MinHeight { get; set; }
/// <summary>
/// Gets or sets the maximum number of pages to process
/// </summary>
[NumberInt(4)]
public int MaxPages { get; set; } = 20;
/// <summary>
/// Starts the web scraping process.
/// </summary>
/// <param name="args">the node parameters used during execution</param>
public override int Execute(NodeParameters args)
{
string textFile = args.WorkingFile;
if (File.Exists(textFile) == false)
{
args.Logger?.ELog($"File '{textFile}' does not exist.");
return -1; // Indicate failure
}
string[] lines = File.ReadAllLines(textFile);
if (lines.Length > 1000)
{
args.Logger?.WLog("The file contains more than 1000 lines of URLs.");
}
lines = lines.Take(1000).Where(x => string.IsNullOrWhiteSpace(x) == false).Select(x => x.Trim()).ToArray();
var allImageUrls = new List<(string, List<string>)>();
foreach (string line in lines) // Limiting to process only 1000 lines
{
if (string.IsNullOrWhiteSpace(line))
continue;
string path = string.Empty;
if (lines.Length > 1)
{
Uri.TryCreate(line, UriKind.Absolute, out Uri uri);
if (uri != null)
{
string uriPath = uri.AbsolutePath.Trim('/');
if (!string.IsNullOrEmpty(uriPath))
{
string[] segments = uriPath.Split('/');
path = segments.Last();
}
}
}
path = string.IsNullOrWhiteSpace(path) ? args.TempPath : Path.Combine(args.TempPath, path);
PageCount = 0;
var imageUrls = ProcessUrlAndGetImageUrls(args.Logger!, line.Trim(), 0, null).GetAwaiter().GetResult();
allImageUrls.Add((path, imageUrls));
}
SaveUrlsToFile(allImageUrls.SelectMany(x => x.Item2).ToList(), Path.Combine(args.TempPath, "image_urls.txt")).GetAwaiter().GetResult();
int totalDownloaded = DownloadImages(args.Logger!, allImageUrls, args.TempPath).GetAwaiter().GetResult();
args.Logger?.ILog($"Total images downloaded: {totalDownloaded}");
return totalDownloaded > 0 ? 1 : 2;
}
/// <summary>
/// Saves a list of URLs to a text file asynchronously.
/// </summary>
/// <param name="urls">The list of URLs to be saved.</param>
/// <param name="fileName">The name of the file to save the URLs.</param>
/// <returns>A task representing the asynchronous saving operation.</returns>
private async Task SaveUrlsToFile(List<string> urls, string fileName)
{
try
{
await File.WriteAllLinesAsync(fileName, urls);
}
catch (Exception ex)
{
// Handle file writing error
}
}
/// <summary>
/// Downloads images asynchronously from a list of image URLs.
/// </summary>
/// <param name="logger">The logger implementation for logging.</param>
/// <param name="imageUrls">The list of image URLs to download.</param>
/// <param name="outputPath">the base path to save the images to</param>
/// <returns>The total count of downloaded images.</returns>
private async Task<int> DownloadImages(ILogger logger, List<(string path, List<string> urls)> imageUrls, string outputPath)
{
int totalDownloaded = 0;
var downloadTasks = new List<Task>();
foreach (var imageData in imageUrls)
{
foreach (var imageUrl in imageData.urls)
{
downloadTasks.Add(ProcessImageAsync(logger, imageUrl, Path.Combine(outputPath, imageData.path)));
}
}
await Task.WhenAll(downloadTasks);
totalDownloaded = downloadTasks.Count(task => task.IsCompletedSuccessfully);
if (totalDownloaded > 1)
{
var paths = imageUrls.Select(x => x.path).Distinct().ToList();
if (paths.Count > 1)
{
foreach(var path in paths)
DeleteDuplicates(logger, path);
}
else
{
DeleteDuplicates(logger, outputPath);
}
}
return totalDownloaded;
}
/// <summary>
/// Processes a URL to extract image URLs and follows links within a specified depth.
/// </summary>
/// <param name="logger">The logger implementation for logging.</param>
/// <param name="url">The URL to be processed.</param>
/// <param name="depth">The current depth of link following.</param>
/// <param name="baseUrl">The base URL for domain comparison.</param>
/// <returns>A list of image URLs extracted from the URL and its sub-links.</returns>
private async Task<List<string>> ProcessUrlAndGetImageUrls(ILogger logger, string url, int depth, string baseUrl)
{
var imageUrls = new List<string>();
if (++PageCount > MaxPages)
return imageUrls;
if (string.IsNullOrEmpty(baseUrl))
{
Uri.TryCreate(url, UriKind.Absolute, out Uri uri);
baseUrl = uri?.GetLeftPart(UriPartial.Path); // Retrieve URL up to path level
}
// Ensure the base URL doesn't contain query parameters
if (!string.IsNullOrEmpty(baseUrl) && baseUrl.Contains("?"))
{
baseUrl = baseUrl.Substring(0, baseUrl.IndexOf('?'));
}
if (baseUrl == null)
return imageUrls;
baseUrl = baseUrl.TrimEnd('/');
try
{
HttpResponseMessage response = await httpClient.GetAsync(url);
if (response.IsSuccessStatusCode)
{
string htmlContent = await response.Content.ReadAsStringAsync();
var doc = new HtmlDocument();
doc.LoadHtml(htmlContent);
var currentImageUrls = doc.DocumentNode.SelectNodes("//img[@src]")?
.Select(node => node.Attributes["src"]?.Value)
.Where(i => string.IsNullOrWhiteSpace(i) == false && IsImageUrl(i))
.ToList() ?? new ();
var dataSrc = doc.DocumentNode.SelectNodes("//*[@data-src]")?
.Select(node =>node.Attributes["data-src"]?.Value)
.Where(i => string.IsNullOrWhiteSpace(i) == false && IsImageUrl(i))
.ToList() ?? new ();
currentImageUrls.AddRange(dataSrc);
var hrefSrc = doc.DocumentNode.SelectNodes("//a[@href]")?
.Select(node =>node.Attributes["href"]?.Value)
.Where(i => string.IsNullOrWhiteSpace(i) == false && IsImageUrl(i))
.ToList() ?? new ();
currentImageUrls.AddRange(hrefSrc);
if (currentImageUrls?.Any() == true)
{
foreach (var imageUrl in currentImageUrls)
{
if (!savedImages.Contains(imageUrl))
{
imageUrls.Add(imageUrl);
savedImages.Add(imageUrl);
}
}
}
var links = doc.DocumentNode.SelectNodes("//a[@href]")
?.Select(a => a.Attributes["href"].Value)
.Where(a => string.IsNullOrWhiteSpace(a) == false && IsSameDomain(a, baseUrl))
.ToList();
if (links != null && depth < MaxDepth)
{
foreach (var link in links)
{
if (string.IsNullOrWhiteSpace(link) || link.StartsWith("#"))
continue;
// Convert relative URLs to absolute URLs
var absoluteLink = ConvertToAbsoluteUrl(baseUrl, link);
if (baseUrl == absoluteLink)
continue;
if (absoluteLink.StartsWith(baseUrl) == false)
continue;
try
{
var urlsFromLink = await ProcessUrlAndGetImageUrls(logger, absoluteLink, depth + 1, baseUrl);
imageUrls.AddRange(urlsFromLink);
}
catch (Exception)
{
// ignored
}
}
}
}
else
{
logger.WLog($"Failed to download images from URL: {url}. Status code: {response.StatusCode}");
}
}
catch (Exception ex)
{
logger.ELog($"Error processing URL: {url}. {ex.Message}");
}
return imageUrls;
}
/// <summary>
/// Converts a relative URL to an absolute URL based on the base URL.
/// </summary>
/// <param name="baseUrl">The base URL used for domain comparison.</param>
/// <param name="link">The URL to be converted to an absolute URL.</param>
/// <returns>The absolute URL.</returns>
private string ConvertToAbsoluteUrl(string baseUrl, string link)
{
if (Uri.IsWellFormedUriString(link, UriKind.Absolute))
{
// Absolute URL
return link;
}
else if (link.StartsWith("/"))
{
// Root-relative URL
if (!baseUrl.EndsWith("/"))
{
// Ensure the base URL ends with a "/"
baseUrl += "/";
}
// Combine the base URL and root-relative link manually
return $"{baseUrl.TrimEnd('/')}/{link.TrimStart('/')}";
}
else
{
// Path-relative URL
var baseUri = new Uri(baseUrl);
var combinedUri = new Uri(baseUri, link);
return combinedUri.ToString();
}
}
/// <summary>
/// Checks if the provided URL is a valid image URL based on common image extensions.
/// </summary>
/// <param name="imageUrl">The URL to be checked.</param>
/// <returns>True if the URL ends with common image extensions; otherwise, false.</returns>
private bool IsImageUrl(string imageUrl)
{
if (imageUrl == null || imageUrl.EndsWith("1px.png") || imageUrl.EndsWith("svg"))
return false;
// Check if the URL ends with common image extensions (e.g., jpg, png, gif, etc.)
var imageExtensionPattern = @"\.(jpg|jpeg|jpe|png|gif|bmp|webp)$";
var regex = new Regex(imageExtensionPattern, RegexOptions.IgnoreCase);
return regex.IsMatch(imageUrl);
}
/// <summary>
/// Checks if a URL is from the same domain as the specified base URL.
/// </summary>
/// <param name="url">The URL to be checked.</param>
/// <param name="baseUrl">The base URL used for comparison.</param>
/// <returns>True if the URL is from the same domain; otherwise, false.</returns>
private bool IsSameDomain(string url, string baseUrl)
{
if (url?.StartsWith("#") == true)
return false;
if (string.IsNullOrEmpty(url) || url.ToLowerInvariant().StartsWith("http") == false)
return true;
if (Uri.TryCreate(url, UriKind.Absolute, out Uri? uri) &&
Uri.TryCreate(baseUrl, UriKind.Absolute, out Uri? baseUri))
{
return uri.Host == baseUri.Host;
}
else
{
// Handling case when URLs are not in a valid format
return false;
}
}
/// <summary>
/// Saves an image from the specified URL if it meets the minimum width and height criteria.
/// </summary>
/// <param name="logger">The logger implementation for logging.</param>
/// <param name="imageUrl">The URL of the image to be saved.</param>
/// <param name="outputPath">the base path to save the images to</param>
/// <returns>true if the image downloaded and the resolution was acceptable</returns>
private async Task<bool> ProcessImageAsync(ILogger logger, string imageUrl, string outputPath)
{
try
{
await semaphore.WaitAsync(); // Wait for a semaphore slot
HttpResponseMessage response = await httpClient.GetAsync(imageUrl);
if (response.IsSuccessStatusCode)
{
byte[] imageData = await response.Content.ReadAsByteArrayAsync();
using (MagickImage image = new MagickImage(imageData))
{
if (image.Width >= MinWidth || image.Height >= MinHeight)
{
var fileName = Path.GetFileName(imageUrl);
var savePath = Path.Combine(outputPath, fileName);
try
{
if (Directory.Exists(outputPath) == false)
Directory.CreateDirectory(outputPath);
}
catch (Exception)
{
}
await image.WriteAsync(savePath); // Save the image directly
savedImages.Add(imageUrl);
logger.ILog($"Saved image: {fileName}");
return true; // Signal successful download
}
}
}
else
{
logger.ELog($"Failed to download image from {imageUrl}. Status code: {response.StatusCode}");
}
}
catch (Exception ex)
{
logger.ELog($"Error saving image from {imageUrl}: {ex.Message}");
}
finally
{
semaphore.Release(); // Release semaphore slot
}
return false; // Signal unsuccessful download
}
/// <summary>
/// Deletes duplicate images within a directory.
/// </summary>
/// <param name="logger">the logger to write logging to</param>
/// <param name="directoryPath">The path to the directory containing images.</param>
private void DeleteDuplicates(ILogger logger, string directoryPath)
{
Dictionary<string, List<string>> hashes = new Dictionary<string, List<string>>();
// Load all image files in the directory
string[] imageFiles = Directory.GetFiles(directoryPath, "*.*");
foreach (string filePath in imageFiles)
{
using (var image = new MagickImage(filePath))
{
image.Resize(300, 300);
// Calculate histogram
int[] histogram = CalculateHistogram(image);
// Convert histogram to a string for hash storage
string hash = string.Join(",", histogram);
if (!hashes.ContainsKey(hash))
{
hashes[hash] = new List<string>();
}
hashes[hash].Add(filePath);
}
}
// Delete duplicates, keeping the largest file
foreach (var kvp in hashes)
{
List<string> duplicateFiles = kvp.Value;
if (duplicateFiles.Count > 1)
{
string largestFile = GetLargestFile(duplicateFiles);
foreach (string file in duplicateFiles)
{
if (file != largestFile)
{
File.Delete(file);
logger?.ILog($"Deleted duplicate: {file}");
}
}
}
}
}
// Calculate the histogram of an image
static int[] CalculateHistogram(MagickImage image)
{
int[] histogram = new int[256]; // Assuming grayscale images
using (var pixels = image.GetPixels())
{
foreach (var pixel in pixels)
{
var grayscale = (int)(pixel.GetChannel((int)Channels.Red) * 0.3 +
pixel.GetChannel((int)Channels.Green) * 0.59 +
pixel.GetChannel((int)Channels.Blue) * 0.11);
histogram[grayscale]++;
}
}
return histogram;
}
private string GetLargestFile(List<string> files)
{
long maxSize = 0;
string largestFile = "";
foreach (string file in files)
{
FileInfo fileInfo = new FileInfo(file);
if (fileInfo.Length > maxSize)
{
maxSize = fileInfo.Length;
largestFile = file;
}
}
return largestFile;
}
}