CryShana / CryCrawler

Cross-platform distributed multi-threaded web crawler

CryCrawler

Portable cross-platform web crawler. Used to crawl websites and download files that match specified criteria.

Note! This is a college project. Don't expect this to be actively maintained. (I also don't recommend using it for any complex and large crawling projects)

Features

• Portable - Single executable file you can run anywhere without any extra requirements
• Cross-platform - Works on Windows/Linux/OSX and other platforms supported by NET Core
• Multi-threaded - Single crawler can use as many threads as specified by the user for crawling.
• Distributed - Connect multiple crawlers to a host; or connect multiple hosts to a master host. (using SSL)
• Extensible - Extend CryCrawler's functionality by using plugins/extensions.
• Web GUI - Check status, change configuration, start/stop program, all through your browser - remotely.
• Breadth/Depth-first crawling - Two different modes for crawling websites (FIFO or LIFO for storing URLs)
• Robots Exclusion Standard - Option for crawler to respect or ignore 'robots.txt' provided by websites
• Custom User-Agent - User can provide a custom user-agent for crawler to use when getting websites.
• File Critera Configuration - Decide which files to download based on extension, media type, file size, filename or URL.
• Domain Whitelist/Blacklist - Force crawler to stay only on certain domains or simply blaclist domains you don't need.
• Duplicate file detection - Files with same names are compared using MD5 checksums to ensure no duplicates.
• Persistent - CryCrawler will keep retrying to crawl failed URLs until they are crawled (up to a certain time limit)

Overview

CryCrawler has two working modes:

• Crawling mode (default) - program will attempt to crawl provided URLs (provided either locally or by host)
• Hosting mode - program will manage multiple connected clients and assign them URLs to crawl

Assuming we are on Linux:

• ./CryCrawler - Will run CryCrawler in the default Crawling mode
• ./CryCrawler -h (or ./CryCrawler --host) - Will run CryCrawler in Hosting mode
• ./CryCrawler --help - Will display a list of supported flags (-d - Debug mode, -n - New session)

Running it for the first time will generate the following:

• config.json - main configuration file for CryCrawler
• plugins/ - folder where you can place your plugins to be loaded on startup
• plugins/PluginTemplate.cs-template - plugin template file for creating plugins
• crycrawler_cache - this file stores all crawled URLs and backups of current URLs to be crawled.

Crawling mode

CryCrawler will start crawling available URLs immediately.

By default, if no previous URLs are loaded, it will use seed Urls (defined in config.json or via WebGUI) to start.

However, if you are using Host for URLs - locally defined URLs will be ignored and any existing crawled data will be erased. This is because Host manages all crawled URLs and work.

Hosting mode

CryCrawler starts listening for connections. URLs will start being crawled only once at least one client is connected.

Loading URLs works the same way as it works in Crawling mode.

Quick Start

1. Run CryCrawler normally (by double-clicking it or running ./CryCrawler)

2. Access CryCrawler's WebGUI by opening your browser and navigating to 127.0.0.1:6001

3. Scroll down to Seed Urls and enter the URLs you wish to start with (one per line)

4. You can adjust any other configuration options here (but if you want a more extensive configuration, check config.json)

5. Click Update Configuration once done and CryCrawler will load up the URLs and start crawling

6. Remember to properly shut it down by using Ctrl + C (to ensure all current uncrawled URLs are backed up to the cache)

Configuration

The config.json has 3 main parts (HostConfig, WorkerConfig and WebGUI)

• HostConfig - Configuration used by the Host (CryCrawler in Hosting mode)
  • ListenerConfiguration - Configure the listening endpoint and optionally define password
  • MaxClientAgeMinutes - Max. allowed time for a client to be offline before Host removes it from it's client list
  • ClientWorkLimit - Max. number of found URLs a client can accumulate before it needs to send them to the host
• WorkerConfig
  • HostEndpoint - Host endpoint to connect to. (set UseHost to true to use this endpoint)
  • DownloadsPath - Relative path to the folder where all files will be downloaded to
  • DontCreateSubfolders - If false, subfolders will be created for each domain and path of file
  • LogEveryCrawl - If true, every URL crawl will be written to the console
  • MaxConcurrency - Max. number of threads the crawler can use for crawling
  • CrawlDelaySeconds - Global crawl delay in seconds (decimal number).
  • DepthSearch - If true, it will crawl URLs depth-first - recommended is false
  • MaxLoggedDownloads - Max. number of recently downloaded items to track and display on Web GUI
  • MaxFileChunkSizekB - Max. file chunk size (in kB) when transferring files to Host in chunk
  • MaxCrawledWorksBeforeCleanHost - Max. number of crawled works to store before clearing them
  • AutoSaveIntervalSeconds - Define interval when to backup uncrawled URLs in memory to disk cache.
  • UserAgent - User-Agent the crawler will use when visiting websites (can not be empty)
  • RespectRobotsExclusionStandard - If true, crawler will try to get robots.txt from every website and respect it.
  • AcceptedExtensions - (File Criteria) List of accepted extensions (ex: .jpg)
  • AcceptedMediaTypes - (File Criteria) List of accepted media types (ex: image/jpeg)
  • ScanTargetsMediaTypes - Media types that should be scanned for more URLs (ex: text/html)
  • DomainWhitelist - List of whitelisted domains. If this list is not empty, only these domains will be crawled.
  • DomainBlacklist - List of blacklisted domains. Any domain listed here will be ignored.
  • BlacklistedURLPatterns - List of blacklisted URL patterns. Any URL that matches any pattern in this list will be ignored. Case insensitive.
  • FilenameMustContainEither - (File Criteria) List of words - filename must contain at least one of them. Case insensitive.
  • URLMustMatchPattern - (File Criteria) List of URL patterns - filename URL must match at least one of them. Case insensitive.
  • MaximumAllowedFileSizekB - (File Criteria) Max. allowed file size in kB (-1 for no limit)
  • MinimumAllowedFileSizekB - (File Criteria) Min. allowed file size in kB (-1 for no limit)
  • AcceptAllFiles - If true, ignores all file criteria and downloads every file it finds.
  • Urls - List of seed URLs that will be loaded when cache is empty (no previous URLs can be loaded)
• WebGUI - Configure the listening endpoint for the Web GUI

Build

Go into CryCrawler directory (path should be CryCrawler/CryCrawler) and run the following command:

dotnet publish -c Release -r win-x64 /p:PublishSingleFile=true /p:PublishReadyToRun=true

Replace the win-x64 with the runtime identifier (RID) that you need:

• Windows (64-bit): win-x64
• Windows (32-bit): win-x86
• Linux (64-bit): linux-x64
• Linux (ARM): linux-arm
• Mac OS: osx-x64

The executable should now be located in CryCrawler\bin\Release\netcoreapp3.0\RID\publish

For other platforms please check the runtime identifier catalog here.

Simple configuration example

In this example we are crawling a website called http://examplewebsite.com/.

Open the generated config.json file (if it's missing, run the crawler once and exit it to generate it).

We need to add the starting URL to Urls like so:

"Urls": [
  "http://examplewebsite.com/"
]

We want to crawl through every JPG and PNG image on this website. Define the file criteria like so:

"AcceptedExtensions": [
  ".jpg",
  ".jpeg",
  ".png"
],
"AcceptedMediaTypes": [
  "image/png",
  "image/jpeg"
]

In case we want to download only large images, we can define the min. and max. file size in kilobytes: (-1 means no limit)

"MaximumAllowedFileSizekB": -1.0,
"MinimumAllowedFileSizekB": 50.0,

In case we want to download only certain images with a certain URL like http://examplewebsite.com/galleryA/image.jpg, we can define URL patterns like so:

"URLMustMatchPattern": [
  "/galleryA/*"
]

If we want to target file names that contain either portrait or landscape, we can define these words like this:

"FilenameMustContainEither": [
  "portrait",
  "landscape"
]

In this example we want to stick to this domain only. We can whitelist this domain like this: (When a whitelist is not empty, any non-whitelisted domain is ignored)

(Witelisted domains include all subdomains)

"DomainWhitelist": [
  "examplewebsite.com"
]

There are cases where you would also want to ignore certain subdomains. We can blacklist them like this:

(Blacklisted domains, unlike whitelisted domains, don't include all subdomains)

"DomainBlacklist": [
  "cdn.examplewebsite.com"
]

A website can also have parts that we don't need and will only slow down the crawling process. In our case we are only interested in the gallery, so we don't care about stuff like blogs, news, etc.

We can blacklist certain URL patterns like so:

"BlacklistedURLPatterns": [
  "/wp-content/*"
]

Other relevant options that we might want to change include the following:

"LogEveryCrawl": true,
"MaxConcurrency": 3,
"CrawlDelaySeconds": 0.0,
"DepthSearch": false,

LogEveryCrawl can be set to false to not spam the console window with every crawled page.

CrawlDelaySeconds is the global crawl delay. We can set this in case we want to slow down the crawling process to avoid being blocked by websites.

DepthSearch should be set to false most of the time as it works better in most cases.