commit d2c7c4139a8078aa08127b5dd513924061d66158 Author: bendtherules Date: Sun May 11 10:59:51 2014 +0530 Lynda single page scraping done diff --git a/output.json b/output.json new file mode 100644 index 0000000..887c7ef --- /dev/null +++ b/output.json @@ -0,0 +1,197 @@ +[{"course_link": "http://www.lynda.com/ASPNET-tutorials/Up-Running-ASPNET/158377-2.html", "course_title": "Up and Running with ASP.NET"}, +{"course_link": "http://www.lynda.com/PHP-tutorials/Exporting-Data-Files-PHP/158375-2.html", "course_title": "Exporting Data to Files with PHP"}, +{"course_link": "http://www.lynda.com/Developer-Web-Design-tutorials/Up-Running-CoffeeScript/154415-2.html", "course_title": "Up and Running with CoffeeScript"}, +{"course_link": "http://www.lynda.com/Visual-Studio-tutorials/Visual-Studio-2013-Web-Developers/157134-2.html", "course_title": "Visual Studio 2013 for Web Developers"}, +{"course_link": "http://www.lynda.com/Developer-Programming-Languages-tutorials/Building-Web-Services-Java-EE/149837-2.html", "course_title": "Building Web Services with Java EE"}, +{"course_link": "http://www.lynda.com/PHP-tutorials/Uploading-Files-Securely-PHP/158374-2.html", "course_title": "Uploading Files Securely with PHP"}, +{"course_link": "http://www.lynda.com/Developer-Web-Development-tutorials/Foundations-Programming-Web-Security/133330-2.html", "course_title": "Foundations of Programming: Web Security"}, +{"course_link": "http://www.lynda.com/Brackets-tutorials/Adobe-Edge-Code-Brackets-First-Look/149704-2.html", "course_title": "Adobe Edge Code and Brackets: First Look"}, +{"course_link": "http://www.lynda.com/Ruby-Rails-tutorials/Ruby-Rails-4-Essential-Training/139989-2.html", "course_title": "Ruby on Rails 4 Essential Training"}, +{"course_link": "http://www.lynda.com/PHP-tutorials/PHP-Web-Designers/146234-2.html", "course_title": "PHP for Web Designers"}, +{"course_link": "http://www.lynda.com/CakePHP-tutorials/Up-Running-CakePHP/126123-2.html", "course_title": "Up and Running with CakePHP"}, +{"course_link": "http://www.lynda.com/PHP-tutorials/Choosing-PHP-Editor/139995-2.html", "course_title": "Choosing a PHP Editor"}, +{"course_link": "http://www.lynda.com/JavaScript-tutorials/JavaScript-Events/140780-2.html", "course_title": "JavaScript: Events"}, +{"course_link": "http://www.lynda.com/Joomla-tutorials/Joomla-3-Essential-Training/110275-2.html", "course_title": "Joomla! 3 Essential Training"}, +{"course_link": "http://www.lynda.com/Shopify-tutorials/Up-Running-Shopify/133352-2.html", "course_title": "Up and Running with Shopify"}, +{"course_link": "http://www.lynda.com/OData-tutorials/Foundations-Programming-Web-Services/126131-2.html", "course_title": "Foundations of Programming: Web Services"}, +{"course_link": "http://www.lynda.com/CSS-tutorials/Working-Data-Web/133326-2.html", "course_title": "Working with Data on the Web"}, +{"course_link": "http://www.lynda.com/Azure-tutorials/Up-Running-Azure/128285-2.html", "course_title": "Up and Running with Azure"}, +{"course_link": "http://www.lynda.com/Azure-tutorials/Up-Running-Azure-Mobile-Services/135357-2.html", "course_title": "Up and Running with Azure Mobile Services"}, +{"course_link": "http://www.lynda.com/Bootstrap-tutorials/Bootstrap-3-Advanced-Web-Development/124079-2.html", "course_title": "Bootstrap 3: Advanced Web Development"}, +{"course_link": "http://www.lynda.com/Ruby-Rails-tutorials/Ruby-Rails-3-Essential-Training/55960-2.html", "course_title": "Ruby on Rails 3 Essential Training"}, +{"course_link": "http://www.lynda.com/Emmet-tutorials/Coding-Faster-Emmet/133353-2.html", "course_title": "Coding Faster with Emmet"}, +{"course_link": "http://www.lynda.com/jQuery-tutorials/Create-Interactive-Animated-Timeline-jQuery/124092-2.html", "course_title": "Create an Interactive Animated Timeline with jQuery"}, +{"course_link": "http://www.lynda.com/WordPress-tutorials/Start-Theme-Food-Blogs-WordPress/133348-2.html", "course_title": "Start with a Theme: Food Blogs in WordPress"}, +{"course_link": "http://www.lynda.com/CodeIgniter-tutorials/Up-Running-PHP-CodeIgniter/126122-2.html", "course_title": "Up and Running with PHP CodeIgniter"}, +{"course_link": "http://www.lynda.com/CSS-tutorials/HTML5-Projects-Creating-Responsive-Presentation/129007-2.html", "course_title": "HTML5 Projects: Creating a Responsive Presentation"}, +{"course_link": "http://www.lynda.com/Dreamweaver-tutorials/Dreamweaver-CC-Essential-Training/124084-2.html", "course_title": "Dreamweaver CC Essential Training"}, +{"course_link": "http://www.lynda.com/HTML-tutorials/JavaScript-Enhancing-DOM/122462-2.html", "course_title": "JavaScript: Enhancing the DOM"}, +{"course_link": "http://www.lynda.com/MySQL-tutorials/PHP-MySQL-Essential-Training/119003-2.html", "course_title": "PHP with MySQL Essential Training"}, +{"course_link": "http://www.lynda.com/Composer-tutorials/Up-Running-PHP-PEAR-PECL-Composer/119001-2.html", "course_title": "Up and Running with PHP: PEAR, PECL, and Composer"}, +{"course_link": "http://www.lynda.com/jQuery-tutorials/jQuery-Essential-Training/48370-2.html", "course_title": "jQuery Essential Training"}, +{"course_link": "http://www.lynda.com/Foundation-tutorials/Up-Running-Foundation/122442-2.html", "course_title": "Up and Running with Foundation"}, +{"course_link": "http://www.lynda.com/Bootstrap-tutorials/Bootstrap-Adding-Interactivity-Your-Site/122456-2.html", "course_title": "Bootstrap: Adding Interactivity to Your Site"}, +{"course_link": "http://www.lynda.com/WordPress-tutorials/WordPress-Ecommerce-Easy-Digital-Downloads/122465-2.html", "course_title": "WordPress Ecommerce: Easy Digital Downloads"}, +{"course_link": "http://www.lynda.com/Edge-Animate-tutorials/Creating-Animated-Infographic-Edge-Animate/124091-2.html", "course_title": "Creating an Animated Infographic with Edge Animate"}, +{"course_link": "http://www.lynda.com/Developer-Databases-tutorials/Validating-Processing-Forms-JavaScript-PHP/120466-2.html", "course_title": "Validating and Processing Forms with JavaScript and PHP"}, +{"course_link": "http://www.lynda.com/WordPress-tutorials/WordPress-Ecommerce-Core-Concepts/122463-2.html", "course_title": "WordPress Ecommerce: Core Concepts"}, +{"course_link": "http://www.lynda.com/Backbonejs-tutorials/Up-Running-Backbonejs/120900-2.html", "course_title": "Up and Running with Backbone.js"}, +{"course_link": "http://www.lynda.com/Edge-Animate-tutorials/Creating-Interactive-Projects-Edge-Animate/124307-2.html", "course_title": "Creating Interactive Projects with Edge Animate"}, +{"course_link": "http://www.lynda.com/jQuery-tutorials/Create-Sortable-Photo-Gallery-jQuery/114897-2.html", "course_title": "Create a Sortable Photo Gallery with jQuery"}, +{"course_link": "http://www.lynda.com/JavaScript-tutorials/JavaScript-JSON/114901-2.html", "course_title": "JavaScript and JSON"}, +{"course_link": "http://www.lynda.com/Edge-Animate-tutorials/Create-Homepage-Marquee-Edge-Animate/114898-2.html", "course_title": "Create a Homepage Marquee with Edge Animate"}, +{"course_link": "http://www.lynda.com/Visual-Studio-Express-tutorials/Building-Windows-Store-Game-Using-HTML-JavaScript/114892-2.html", "course_title": "Building a Windows Store Game Using HTML and JavaScript"}, +{"course_link": "http://www.lynda.com/CSS-tutorials/HTML5-Projects-Geolocation/118953-2.html", "course_title": "HTML5 Projects: Integrating Google Store Locator"}, +{"course_link": "http://www.lynda.com/ASP-NET-3-5-tutorials/asp-net-essential-training/784-2.html", "course_title": "ASP.NET Essential Training"}, +{"course_link": "http://www.lynda.com/Sublime-Text-tutorials/Up-Running-Sublime-Text-2/114325-2.html", "course_title": "Up and Running with Sublime Text 2"}, +{"course_link": "http://www.lynda.com/CSS-tutorials/CSS-Frameworks-Grids/112675-2.html", "course_title": "CSS: Frameworks and Grids"}, +{"course_link": "http://www.lynda.com/Developer-tutorials/JavaScript-and-AJAX/114900-2.html", "course_title": "JavaScript and AJAX"}, +{"course_link": "http://www.lynda.com/Facebook-tutorials/Building-Facebook-Applications-PHP-MySQL/107060-2.html", "course_title": "Building Facebook Applications with PHP and MySQL"}, +{"course_link": "http://www.lynda.com/Drupal-tutorials/Responsive-Design-Drupal/107419-2.html", "course_title": "Responsive Design with Drupal"}][{"course_link": "http://www.lynda.com/ASPNET-tutorials/Up-Running-ASPNET/158377-2.html", "course_title": "Up and Running with ASP.NET"}, +{"course_link": "http://www.lynda.com/PHP-tutorials/Exporting-Data-Files-PHP/158375-2.html", "course_title": "Exporting Data to Files with PHP"}, +{"course_link": "http://www.lynda.com/Developer-Web-Design-tutorials/Up-Running-CoffeeScript/154415-2.html", "course_title": "Up and Running with CoffeeScript"}, +{"course_link": "http://www.lynda.com/Visual-Studio-tutorials/Visual-Studio-2013-Web-Developers/157134-2.html", "course_title": "Visual Studio 2013 for Web Developers"}, +{"course_link": "http://www.lynda.com/Developer-Programming-Languages-tutorials/Building-Web-Services-Java-EE/149837-2.html", "course_title": "Building Web Services with Java EE"}, +{"course_link": "http://www.lynda.com/PHP-tutorials/Uploading-Files-Securely-PHP/158374-2.html", "course_title": "Uploading Files Securely with PHP"}, +{"course_link": "http://www.lynda.com/Developer-Web-Development-tutorials/Foundations-Programming-Web-Security/133330-2.html", "course_title": "Foundations of Programming: Web Security"}, +{"course_link": "http://www.lynda.com/Brackets-tutorials/Adobe-Edge-Code-Brackets-First-Look/149704-2.html", "course_title": "Adobe Edge Code and Brackets: First Look"}, +{"course_link": "http://www.lynda.com/Ruby-Rails-tutorials/Ruby-Rails-4-Essential-Training/139989-2.html", "course_title": "Ruby on Rails 4 Essential Training"}, +{"course_link": "http://www.lynda.com/PHP-tutorials/PHP-Web-Designers/146234-2.html", "course_title": "PHP for Web Designers"}, +{"course_link": "http://www.lynda.com/CakePHP-tutorials/Up-Running-CakePHP/126123-2.html", "course_title": "Up and Running with CakePHP"}, +{"course_link": "http://www.lynda.com/PHP-tutorials/Choosing-PHP-Editor/139995-2.html", "course_title": "Choosing a PHP Editor"}, +{"course_link": "http://www.lynda.com/JavaScript-tutorials/JavaScript-Events/140780-2.html", "course_title": "JavaScript: Events"}, +{"course_link": "http://www.lynda.com/Joomla-tutorials/Joomla-3-Essential-Training/110275-2.html", "course_title": "Joomla! 3 Essential Training"}, +{"course_link": "http://www.lynda.com/Shopify-tutorials/Up-Running-Shopify/133352-2.html", "course_title": "Up and Running with Shopify"}, +{"course_link": "http://www.lynda.com/OData-tutorials/Foundations-Programming-Web-Services/126131-2.html", "course_title": "Foundations of Programming: Web Services"}, +{"course_link": "http://www.lynda.com/CSS-tutorials/Working-Data-Web/133326-2.html", "course_title": "Working with Data on the Web"}, +{"course_link": "http://www.lynda.com/Azure-tutorials/Up-Running-Azure/128285-2.html", "course_title": "Up and Running with Azure"}, +{"course_link": "http://www.lynda.com/Azure-tutorials/Up-Running-Azure-Mobile-Services/135357-2.html", "course_title": "Up and Running with Azure Mobile Services"}, +{"course_link": "http://www.lynda.com/Bootstrap-tutorials/Bootstrap-3-Advanced-Web-Development/124079-2.html", "course_title": "Bootstrap 3: Advanced Web Development"}, +{"course_link": "http://www.lynda.com/Ruby-Rails-tutorials/Ruby-Rails-3-Essential-Training/55960-2.html", "course_title": "Ruby on Rails 3 Essential Training"}, +{"course_link": "http://www.lynda.com/Emmet-tutorials/Coding-Faster-Emmet/133353-2.html", "course_title": "Coding Faster with Emmet"}, +{"course_link": "http://www.lynda.com/jQuery-tutorials/Create-Interactive-Animated-Timeline-jQuery/124092-2.html", "course_title": "Create an Interactive Animated Timeline with jQuery"}, +{"course_link": "http://www.lynda.com/WordPress-tutorials/Start-Theme-Food-Blogs-WordPress/133348-2.html", "course_title": "Start with a Theme: Food Blogs in WordPress"}, +{"course_link": "http://www.lynda.com/CodeIgniter-tutorials/Up-Running-PHP-CodeIgniter/126122-2.html", "course_title": "Up and Running with PHP CodeIgniter"}, +{"course_link": "http://www.lynda.com/CSS-tutorials/HTML5-Projects-Creating-Responsive-Presentation/129007-2.html", "course_title": "HTML5 Projects: Creating a Responsive Presentation"}, +{"course_link": "http://www.lynda.com/Dreamweaver-tutorials/Dreamweaver-CC-Essential-Training/124084-2.html", "course_title": "Dreamweaver CC Essential Training"}, +{"course_link": "http://www.lynda.com/HTML-tutorials/JavaScript-Enhancing-DOM/122462-2.html", "course_title": "JavaScript: Enhancing the DOM"}, +{"course_link": "http://www.lynda.com/MySQL-tutorials/PHP-MySQL-Essential-Training/119003-2.html", "course_title": "PHP with MySQL Essential Training"}, +{"course_link": "http://www.lynda.com/Composer-tutorials/Up-Running-PHP-PEAR-PECL-Composer/119001-2.html", "course_title": "Up and Running with PHP: PEAR, PECL, and Composer"}, +{"course_link": "http://www.lynda.com/jQuery-tutorials/jQuery-Essential-Training/48370-2.html", "course_title": "jQuery Essential Training"}, +{"course_link": "http://www.lynda.com/Foundation-tutorials/Up-Running-Foundation/122442-2.html", "course_title": "Up and Running with Foundation"}, +{"course_link": "http://www.lynda.com/Bootstrap-tutorials/Bootstrap-Adding-Interactivity-Your-Site/122456-2.html", "course_title": "Bootstrap: Adding Interactivity to Your Site"}, +{"course_link": "http://www.lynda.com/WordPress-tutorials/WordPress-Ecommerce-Easy-Digital-Downloads/122465-2.html", "course_title": "WordPress Ecommerce: Easy Digital Downloads"}, +{"course_link": "http://www.lynda.com/Edge-Animate-tutorials/Creating-Animated-Infographic-Edge-Animate/124091-2.html", "course_title": "Creating an Animated Infographic with Edge Animate"}, +{"course_link": "http://www.lynda.com/Developer-Databases-tutorials/Validating-Processing-Forms-JavaScript-PHP/120466-2.html", "course_title": "Validating and Processing Forms with JavaScript and PHP"}, +{"course_link": "http://www.lynda.com/WordPress-tutorials/WordPress-Ecommerce-Core-Concepts/122463-2.html", "course_title": "WordPress Ecommerce: Core Concepts"}, +{"course_link": "http://www.lynda.com/Backbonejs-tutorials/Up-Running-Backbonejs/120900-2.html", "course_title": "Up and Running with Backbone.js"}, +{"course_link": "http://www.lynda.com/Edge-Animate-tutorials/Creating-Interactive-Projects-Edge-Animate/124307-2.html", "course_title": "Creating Interactive Projects with Edge Animate"}, +{"course_link": "http://www.lynda.com/jQuery-tutorials/Create-Sortable-Photo-Gallery-jQuery/114897-2.html", "course_title": "Create a Sortable Photo Gallery with jQuery"}, +{"course_link": "http://www.lynda.com/JavaScript-tutorials/JavaScript-JSON/114901-2.html", "course_title": "JavaScript and JSON"}, +{"course_link": "http://www.lynda.com/Edge-Animate-tutorials/Create-Homepage-Marquee-Edge-Animate/114898-2.html", "course_title": "Create a Homepage Marquee with Edge Animate"}, +{"course_link": "http://www.lynda.com/Visual-Studio-Express-tutorials/Building-Windows-Store-Game-Using-HTML-JavaScript/114892-2.html", "course_title": "Building a Windows Store Game Using HTML and JavaScript"}, +{"course_link": "http://www.lynda.com/CSS-tutorials/HTML5-Projects-Geolocation/118953-2.html", "course_title": "HTML5 Projects: Integrating Google Store Locator"}, +{"course_link": "http://www.lynda.com/ASP-NET-3-5-tutorials/asp-net-essential-training/784-2.html", "course_title": "ASP.NET Essential Training"}, +{"course_link": "http://www.lynda.com/Sublime-Text-tutorials/Up-Running-Sublime-Text-2/114325-2.html", "course_title": "Up and Running with Sublime Text 2"}, +{"course_link": "http://www.lynda.com/CSS-tutorials/CSS-Frameworks-Grids/112675-2.html", "course_title": "CSS: Frameworks and Grids"}, +{"course_link": "http://www.lynda.com/Developer-tutorials/JavaScript-and-AJAX/114900-2.html", "course_title": "JavaScript and AJAX"}, +{"course_link": "http://www.lynda.com/Facebook-tutorials/Building-Facebook-Applications-PHP-MySQL/107060-2.html", "course_title": "Building Facebook Applications with PHP and MySQL"}, +{"course_link": "http://www.lynda.com/Drupal-tutorials/Responsive-Design-Drupal/107419-2.html", "course_title": "Responsive Design with Drupal"}][{"course_link": "http://www.lynda.com/ASPNET-tutorials/Up-Running-ASPNET/158377-2.html", "course_title": "Up and Running with ASP.NET"}, +{"course_link": "http://www.lynda.com/PHP-tutorials/Exporting-Data-Files-PHP/158375-2.html", "course_title": "Exporting Data to Files with PHP"}, +{"course_link": "http://www.lynda.com/Developer-Web-Design-tutorials/Up-Running-CoffeeScript/154415-2.html", "course_title": "Up and Running with CoffeeScript"}, +{"course_link": "http://www.lynda.com/Visual-Studio-tutorials/Visual-Studio-2013-Web-Developers/157134-2.html", "course_title": "Visual Studio 2013 for Web Developers"}, +{"course_link": "http://www.lynda.com/Developer-Programming-Languages-tutorials/Building-Web-Services-Java-EE/149837-2.html", "course_title": "Building Web Services with Java EE"}, +{"course_link": "http://www.lynda.com/PHP-tutorials/Uploading-Files-Securely-PHP/158374-2.html", "course_title": "Uploading Files Securely with PHP"}, +{"course_link": "http://www.lynda.com/Developer-Web-Development-tutorials/Foundations-Programming-Web-Security/133330-2.html", "course_title": "Foundations of Programming: Web Security"}, +{"course_link": "http://www.lynda.com/Brackets-tutorials/Adobe-Edge-Code-Brackets-First-Look/149704-2.html", "course_title": "Adobe Edge Code and Brackets: First Look"}, +{"course_link": "http://www.lynda.com/Ruby-Rails-tutorials/Ruby-Rails-4-Essential-Training/139989-2.html", "course_title": "Ruby on Rails 4 Essential Training"}, +{"course_link": "http://www.lynda.com/PHP-tutorials/PHP-Web-Designers/146234-2.html", "course_title": "PHP for Web Designers"}, +{"course_link": "http://www.lynda.com/CakePHP-tutorials/Up-Running-CakePHP/126123-2.html", "course_title": "Up and Running with CakePHP"}, +{"course_link": "http://www.lynda.com/PHP-tutorials/Choosing-PHP-Editor/139995-2.html", "course_title": "Choosing a PHP Editor"}, +{"course_link": "http://www.lynda.com/JavaScript-tutorials/JavaScript-Events/140780-2.html", "course_title": "JavaScript: Events"}, +{"course_link": "http://www.lynda.com/Joomla-tutorials/Joomla-3-Essential-Training/110275-2.html", "course_title": "Joomla! 3 Essential Training"}, +{"course_link": "http://www.lynda.com/Shopify-tutorials/Up-Running-Shopify/133352-2.html", "course_title": "Up and Running with Shopify"}, +{"course_link": "http://www.lynda.com/OData-tutorials/Foundations-Programming-Web-Services/126131-2.html", "course_title": "Foundations of Programming: Web Services"}, +{"course_link": "http://www.lynda.com/CSS-tutorials/Working-Data-Web/133326-2.html", "course_title": "Working with Data on the Web"}, +{"course_link": "http://www.lynda.com/Azure-tutorials/Up-Running-Azure/128285-2.html", "course_title": "Up and Running with Azure"}, +{"course_link": "http://www.lynda.com/Azure-tutorials/Up-Running-Azure-Mobile-Services/135357-2.html", "course_title": "Up and Running with Azure Mobile Services"}, +{"course_link": "http://www.lynda.com/Bootstrap-tutorials/Bootstrap-3-Advanced-Web-Development/124079-2.html", "course_title": "Bootstrap 3: Advanced Web Development"}, +{"course_link": "http://www.lynda.com/Ruby-Rails-tutorials/Ruby-Rails-3-Essential-Training/55960-2.html", "course_title": "Ruby on Rails 3 Essential Training"}, +{"course_link": "http://www.lynda.com/Emmet-tutorials/Coding-Faster-Emmet/133353-2.html", "course_title": "Coding Faster with Emmet"}, +{"course_link": "http://www.lynda.com/jQuery-tutorials/Create-Interactive-Animated-Timeline-jQuery/124092-2.html", "course_title": "Create an Interactive Animated Timeline with jQuery"}, +{"course_link": "http://www.lynda.com/WordPress-tutorials/Start-Theme-Food-Blogs-WordPress/133348-2.html", "course_title": "Start with a Theme: Food Blogs in WordPress"}, +{"course_link": "http://www.lynda.com/CodeIgniter-tutorials/Up-Running-PHP-CodeIgniter/126122-2.html", "course_title": "Up and Running with PHP CodeIgniter"}, +{"course_link": "http://www.lynda.com/CSS-tutorials/HTML5-Projects-Creating-Responsive-Presentation/129007-2.html", "course_title": "HTML5 Projects: Creating a Responsive Presentation"}, +{"course_link": "http://www.lynda.com/Dreamweaver-tutorials/Dreamweaver-CC-Essential-Training/124084-2.html", "course_title": "Dreamweaver CC Essential Training"}, +{"course_link": "http://www.lynda.com/HTML-tutorials/JavaScript-Enhancing-DOM/122462-2.html", "course_title": "JavaScript: Enhancing the DOM"}, +{"course_link": "http://www.lynda.com/MySQL-tutorials/PHP-MySQL-Essential-Training/119003-2.html", "course_title": "PHP with MySQL Essential Training"}, +{"course_link": "http://www.lynda.com/Composer-tutorials/Up-Running-PHP-PEAR-PECL-Composer/119001-2.html", "course_title": "Up and Running with PHP: PEAR, PECL, and Composer"}, +{"course_link": "http://www.lynda.com/jQuery-tutorials/jQuery-Essential-Training/48370-2.html", "course_title": "jQuery Essential Training"}, +{"course_link": "http://www.lynda.com/Foundation-tutorials/Up-Running-Foundation/122442-2.html", "course_title": "Up and Running with Foundation"}, +{"course_link": "http://www.lynda.com/Bootstrap-tutorials/Bootstrap-Adding-Interactivity-Your-Site/122456-2.html", "course_title": "Bootstrap: Adding Interactivity to Your Site"}, +{"course_link": "http://www.lynda.com/WordPress-tutorials/WordPress-Ecommerce-Easy-Digital-Downloads/122465-2.html", "course_title": "WordPress Ecommerce: Easy Digital Downloads"}, +{"course_link": "http://www.lynda.com/Edge-Animate-tutorials/Creating-Animated-Infographic-Edge-Animate/124091-2.html", "course_title": "Creating an Animated Infographic with Edge Animate"}, +{"course_link": "http://www.lynda.com/Developer-Databases-tutorials/Validating-Processing-Forms-JavaScript-PHP/120466-2.html", "course_title": "Validating and Processing Forms with JavaScript and PHP"}, +{"course_link": "http://www.lynda.com/WordPress-tutorials/WordPress-Ecommerce-Core-Concepts/122463-2.html", "course_title": "WordPress Ecommerce: Core Concepts"}, +{"course_link": "http://www.lynda.com/Backbonejs-tutorials/Up-Running-Backbonejs/120900-2.html", "course_title": "Up and Running with Backbone.js"}, +{"course_link": "http://www.lynda.com/Edge-Animate-tutorials/Creating-Interactive-Projects-Edge-Animate/124307-2.html", "course_title": "Creating Interactive Projects with Edge Animate"}, +{"course_link": "http://www.lynda.com/jQuery-tutorials/Create-Sortable-Photo-Gallery-jQuery/114897-2.html", "course_title": "Create a Sortable Photo Gallery with jQuery"}, +{"course_link": "http://www.lynda.com/JavaScript-tutorials/JavaScript-JSON/114901-2.html", "course_title": "JavaScript and JSON"}, +{"course_link": "http://www.lynda.com/Edge-Animate-tutorials/Create-Homepage-Marquee-Edge-Animate/114898-2.html", "course_title": "Create a Homepage Marquee with Edge Animate"}, +{"course_link": "http://www.lynda.com/Visual-Studio-Express-tutorials/Building-Windows-Store-Game-Using-HTML-JavaScript/114892-2.html", "course_title": "Building a Windows Store Game Using HTML and JavaScript"}, +{"course_link": "http://www.lynda.com/CSS-tutorials/HTML5-Projects-Geolocation/118953-2.html", "course_title": "HTML5 Projects: Integrating Google Store Locator"}, +{"course_link": "http://www.lynda.com/ASP-NET-3-5-tutorials/asp-net-essential-training/784-2.html", "course_title": "ASP.NET Essential Training"}, +{"course_link": "http://www.lynda.com/Sublime-Text-tutorials/Up-Running-Sublime-Text-2/114325-2.html", "course_title": "Up and Running with Sublime Text 2"}, +{"course_link": "http://www.lynda.com/CSS-tutorials/CSS-Frameworks-Grids/112675-2.html", "course_title": "CSS: Frameworks and Grids"}, +{"course_link": "http://www.lynda.com/Developer-tutorials/JavaScript-and-AJAX/114900-2.html", "course_title": "JavaScript and AJAX"}, +{"course_link": "http://www.lynda.com/Facebook-tutorials/Building-Facebook-Applications-PHP-MySQL/107060-2.html", "course_title": "Building Facebook Applications with PHP and MySQL"}, +{"course_link": "http://www.lynda.com/Drupal-tutorials/Responsive-Design-Drupal/107419-2.html", "course_title": "Responsive Design with Drupal"}][{"course_link": "http://www.lynda.com/ASPNET-tutorials/Up-Running-ASPNET/158377-2.html", "course_title": "Up and Running with ASP.NET"}, +{"course_link": "http://www.lynda.com/PHP-tutorials/Exporting-Data-Files-PHP/158375-2.html", "course_title": "Exporting Data to Files with PHP"}, +{"course_link": "http://www.lynda.com/Developer-Web-Design-tutorials/Up-Running-CoffeeScript/154415-2.html", "course_title": "Up and Running with CoffeeScript"}, +{"course_link": "http://www.lynda.com/Visual-Studio-tutorials/Visual-Studio-2013-Web-Developers/157134-2.html", "course_title": "Visual Studio 2013 for Web Developers"}, +{"course_link": "http://www.lynda.com/Developer-Programming-Languages-tutorials/Building-Web-Services-Java-EE/149837-2.html", "course_title": "Building Web Services with Java EE"}, +{"course_link": "http://www.lynda.com/PHP-tutorials/Uploading-Files-Securely-PHP/158374-2.html", "course_title": "Uploading Files Securely with PHP"}, +{"course_link": "http://www.lynda.com/Developer-Web-Development-tutorials/Foundations-Programming-Web-Security/133330-2.html", "course_title": "Foundations of Programming: Web Security"}, +{"course_link": "http://www.lynda.com/Brackets-tutorials/Adobe-Edge-Code-Brackets-First-Look/149704-2.html", "course_title": "Adobe Edge Code and Brackets: First Look"}, +{"course_link": "http://www.lynda.com/Ruby-Rails-tutorials/Ruby-Rails-4-Essential-Training/139989-2.html", "course_title": "Ruby on Rails 4 Essential Training"}, +{"course_link": "http://www.lynda.com/PHP-tutorials/PHP-Web-Designers/146234-2.html", "course_title": "PHP for Web Designers"}, +{"course_link": "http://www.lynda.com/CakePHP-tutorials/Up-Running-CakePHP/126123-2.html", "course_title": "Up and Running with CakePHP"}, +{"course_link": "http://www.lynda.com/PHP-tutorials/Choosing-PHP-Editor/139995-2.html", "course_title": "Choosing a PHP Editor"}, +{"course_link": "http://www.lynda.com/JavaScript-tutorials/JavaScript-Events/140780-2.html", "course_title": "JavaScript: Events"}, +{"course_link": "http://www.lynda.com/Joomla-tutorials/Joomla-3-Essential-Training/110275-2.html", "course_title": "Joomla! 3 Essential Training"}, +{"course_link": "http://www.lynda.com/Shopify-tutorials/Up-Running-Shopify/133352-2.html", "course_title": "Up and Running with Shopify"}, +{"course_link": "http://www.lynda.com/OData-tutorials/Foundations-Programming-Web-Services/126131-2.html", "course_title": "Foundations of Programming: Web Services"}, +{"course_link": "http://www.lynda.com/CSS-tutorials/Working-Data-Web/133326-2.html", "course_title": "Working with Data on the Web"}, +{"course_link": "http://www.lynda.com/Azure-tutorials/Up-Running-Azure/128285-2.html", "course_title": "Up and Running with Azure"}, +{"course_link": "http://www.lynda.com/Azure-tutorials/Up-Running-Azure-Mobile-Services/135357-2.html", "course_title": "Up and Running with Azure Mobile Services"}, +{"course_link": "http://www.lynda.com/Bootstrap-tutorials/Bootstrap-3-Advanced-Web-Development/124079-2.html", "course_title": "Bootstrap 3: Advanced Web Development"}, +{"course_link": "http://www.lynda.com/Ruby-Rails-tutorials/Ruby-Rails-3-Essential-Training/55960-2.html", "course_title": "Ruby on Rails 3 Essential Training"}, +{"course_link": "http://www.lynda.com/Emmet-tutorials/Coding-Faster-Emmet/133353-2.html", "course_title": "Coding Faster with Emmet"}, +{"course_link": "http://www.lynda.com/jQuery-tutorials/Create-Interactive-Animated-Timeline-jQuery/124092-2.html", "course_title": "Create an Interactive Animated Timeline with jQuery"}, +{"course_link": "http://www.lynda.com/WordPress-tutorials/Start-Theme-Food-Blogs-WordPress/133348-2.html", "course_title": "Start with a Theme: Food Blogs in WordPress"}, +{"course_link": "http://www.lynda.com/CodeIgniter-tutorials/Up-Running-PHP-CodeIgniter/126122-2.html", "course_title": "Up and Running with PHP CodeIgniter"}, +{"course_link": "http://www.lynda.com/CSS-tutorials/HTML5-Projects-Creating-Responsive-Presentation/129007-2.html", "course_title": "HTML5 Projects: Creating a Responsive Presentation"}, +{"course_link": "http://www.lynda.com/Dreamweaver-tutorials/Dreamweaver-CC-Essential-Training/124084-2.html", "course_title": "Dreamweaver CC Essential Training"}, +{"course_link": "http://www.lynda.com/HTML-tutorials/JavaScript-Enhancing-DOM/122462-2.html", "course_title": "JavaScript: Enhancing the DOM"}, +{"course_link": "http://www.lynda.com/MySQL-tutorials/PHP-MySQL-Essential-Training/119003-2.html", "course_title": "PHP with MySQL Essential Training"}, +{"course_link": "http://www.lynda.com/Composer-tutorials/Up-Running-PHP-PEAR-PECL-Composer/119001-2.html", "course_title": "Up and Running with PHP: PEAR, PECL, and Composer"}, +{"course_link": "http://www.lynda.com/jQuery-tutorials/jQuery-Essential-Training/48370-2.html", "course_title": "jQuery Essential Training"}, +{"course_link": "http://www.lynda.com/Foundation-tutorials/Up-Running-Foundation/122442-2.html", "course_title": "Up and Running with Foundation"}, +{"course_link": "http://www.lynda.com/Bootstrap-tutorials/Bootstrap-Adding-Interactivity-Your-Site/122456-2.html", "course_title": "Bootstrap: Adding Interactivity to Your Site"}, +{"course_link": "http://www.lynda.com/WordPress-tutorials/WordPress-Ecommerce-Easy-Digital-Downloads/122465-2.html", "course_title": "WordPress Ecommerce: Easy Digital Downloads"}, +{"course_link": "http://www.lynda.com/Edge-Animate-tutorials/Creating-Animated-Infographic-Edge-Animate/124091-2.html", "course_title": "Creating an Animated Infographic with Edge Animate"}, +{"course_link": "http://www.lynda.com/Developer-Databases-tutorials/Validating-Processing-Forms-JavaScript-PHP/120466-2.html", "course_title": "Validating and Processing Forms with JavaScript and PHP"}, +{"course_link": "http://www.lynda.com/WordPress-tutorials/WordPress-Ecommerce-Core-Concepts/122463-2.html", "course_title": "WordPress Ecommerce: Core Concepts"}, +{"course_link": "http://www.lynda.com/Backbonejs-tutorials/Up-Running-Backbonejs/120900-2.html", "course_title": "Up and Running with Backbone.js"}, +{"course_link": "http://www.lynda.com/Edge-Animate-tutorials/Creating-Interactive-Projects-Edge-Animate/124307-2.html", "course_title": "Creating Interactive Projects with Edge Animate"}, +{"course_link": "http://www.lynda.com/jQuery-tutorials/Create-Sortable-Photo-Gallery-jQuery/114897-2.html", "course_title": "Create a Sortable Photo Gallery with jQuery"}, +{"course_link": "http://www.lynda.com/JavaScript-tutorials/JavaScript-JSON/114901-2.html", "course_title": "JavaScript and JSON"}, +{"course_link": "http://www.lynda.com/Edge-Animate-tutorials/Create-Homepage-Marquee-Edge-Animate/114898-2.html", "course_title": "Create a Homepage Marquee with Edge Animate"}, +{"course_link": "http://www.lynda.com/Visual-Studio-Express-tutorials/Building-Windows-Store-Game-Using-HTML-JavaScript/114892-2.html", "course_title": "Building a Windows Store Game Using HTML and JavaScript"}, +{"course_link": "http://www.lynda.com/CSS-tutorials/HTML5-Projects-Geolocation/118953-2.html", "course_title": "HTML5 Projects: Integrating Google Store Locator"}, +{"course_link": "http://www.lynda.com/ASP-NET-3-5-tutorials/asp-net-essential-training/784-2.html", "course_title": "ASP.NET Essential Training"}, +{"course_link": "http://www.lynda.com/Sublime-Text-tutorials/Up-Running-Sublime-Text-2/114325-2.html", "course_title": "Up and Running with Sublime Text 2"}, +{"course_link": "http://www.lynda.com/CSS-tutorials/CSS-Frameworks-Grids/112675-2.html", "course_title": "CSS: Frameworks and Grids"}, +{"course_link": "http://www.lynda.com/Developer-tutorials/JavaScript-and-AJAX/114900-2.html", "course_title": "JavaScript and AJAX"}, +{"course_link": "http://www.lynda.com/Facebook-tutorials/Building-Facebook-Applications-PHP-MySQL/107060-2.html", "course_title": "Building Facebook Applications with PHP and MySQL"}, +{"course_link": "http://www.lynda.com/Drupal-tutorials/Responsive-Design-Drupal/107419-2.html", "course_title": "Responsive Design with Drupal"}] \ No newline at end of file diff --git a/scrapy.cfg b/scrapy.cfg new file mode 100644 index 0000000..b2970f1 --- /dev/null +++ b/scrapy.cfg @@ -0,0 +1,11 @@ +# Automatically created by: scrapy startproject +# +# For more information about the [deploy] section see: +# http://doc.scrapy.org/en/latest/topics/scrapyd.html + +[settings] +default = try_scrapy.settings + +[deploy] +#url = http://localhost:6800/ +project = try_scrapy diff --git a/try_scrapy/__init__.py b/try_scrapy/__init__.py new file mode 100644 index 0000000..e69de29 diff --git a/try_scrapy/__init__.pyc b/try_scrapy/__init__.pyc new file mode 100644 index 0000000..5ba95c0 Binary files /dev/null and b/try_scrapy/__init__.pyc differ diff --git a/try_scrapy/items.py b/try_scrapy/items.py new file mode 100644 index 0000000..2bc89b8 --- /dev/null +++ b/try_scrapy/items.py @@ -0,0 +1,13 @@ +# Define here the models for your scraped items +# +# See documentation in: +# http://doc.scrapy.org/en/latest/topics/items.html + +from scrapy.item import Item, Field + + +class LyndaItem(Item): + # define the fields for your item here like: + # name = Field() + course_title = Field() + course_link = Field() diff --git a/try_scrapy/items.pyc b/try_scrapy/items.pyc new file mode 100644 index 0000000..6d439d1 Binary files /dev/null and b/try_scrapy/items.pyc differ diff --git a/try_scrapy/pipelines.py b/try_scrapy/pipelines.py new file mode 100644 index 0000000..977ffbe --- /dev/null +++ b/try_scrapy/pipelines.py @@ -0,0 +1,8 @@ +# Define your item pipelines here +# +# Don't forget to add your pipeline to the ITEM_PIPELINES setting +# See: http://doc.scrapy.org/en/latest/topics/item-pipeline.html + +class TryScrapyPipeline(object): + def process_item(self, item, spider): + return item diff --git a/try_scrapy/settings.py b/try_scrapy/settings.py new file mode 100644 index 0000000..775eec5 --- /dev/null +++ b/try_scrapy/settings.py @@ -0,0 +1,15 @@ +# Scrapy settings for try_scrapy project +# +# For simplicity, this file contains only the most important settings by +# default. All the other settings are documented here: +# +# http://doc.scrapy.org/en/latest/topics/settings.html +# + +BOT_NAME = 'try_scrapy' + +SPIDER_MODULES = ['try_scrapy.spiders'] +NEWSPIDER_MODULE = 'try_scrapy.spiders' + +# Crawl responsibly by identifying yourself (and your website) on the user-agent +#USER_AGENT = 'try_scrapy (+http://www.yourdomain.com)' diff --git a/try_scrapy/settings.pyc b/try_scrapy/settings.pyc new file mode 100644 index 0000000..bdd84fe Binary files /dev/null and b/try_scrapy/settings.pyc differ diff --git a/try_scrapy/spiders/__init__.py b/try_scrapy/spiders/__init__.py new file mode 100644 index 0000000..ebd689a --- /dev/null +++ b/try_scrapy/spiders/__init__.py @@ -0,0 +1,4 @@ +# This package will contain the spiders of your Scrapy project +# +# Please refer to the documentation for information on how to create and manage +# your spiders. diff --git a/try_scrapy/spiders/__init__.pyc b/try_scrapy/spiders/__init__.pyc new file mode 100644 index 0000000..7787c1e Binary files /dev/null and b/try_scrapy/spiders/__init__.pyc differ diff --git a/try_scrapy/spiders/test_spider.py b/try_scrapy/spiders/test_spider.py new file mode 100644 index 0000000..b7e9231 --- /dev/null +++ b/try_scrapy/spiders/test_spider.py @@ -0,0 +1,34 @@ +from scrapy.spider import Spider +from scrapy.selector import Selector +from try_scrapy.items import LyndaItem +from scrapy.contrib.spiders import CrawlSpider, Rule +from scrapy.contrib.linkextractors.sgml import SgmlLinkExtractor + +class lynda_spider(CrawlSpider): + name="get_lynda" + # allowed_domains = ["dmoz.org"] + start_urls=[r"file:///C:/Users/b/Downloads/Learn%20Web%20Development%20with%20Video%20Courses%20and%20Tutorials%20from%20lynda.com.htm"] + rules = ( + Rule(SgmlLinkExtractor(attrs=["href","data-url"],allow=[r".*?page=\d+.*"])), + ) + + def parse(self,response): + sel = Selector(response) + course_a=sel.xpath(r'//*[@class="course-list"]//li//*[@class="details-row"]/h3/a') + # print course_a.extract() + titles=course_a.xpath("text()").extract() + links=course_a.xpath("@href").extract() + items=[] + for i in range(len(course_a)): + tmp_item=LyndaItem() + tmp_item["course_title"]=titles[i].strip() + tmp_item["course_link"]=links[i] + items.append(tmp_item) + # print titles[i].strip() + # print links[i] + return items + +# class NextPageLinkExtractor(SgmlLinkExtractor): +# # allow_domains=[r"http://www.lynda.com/"] +# attrs=["href","data-url"] +# allow=[r".*?page=\d+.*"] diff --git a/try_scrapy/spiders/test_spider.pyc b/try_scrapy/spiders/test_spider.pyc new file mode 100644 index 0000000..e9db007 Binary files /dev/null and b/try_scrapy/spiders/test_spider.pyc differ