- No changes.
- Add html to response dict when available.
- Upgrade to GitHub Actions
- Upgrade beautifulsoup4 dependency
- Upgrade requests dependency
- Try and return the "best" url. (#75).
- Fix issue where AMP image data was a list of strings. (#75).
- Fix issue where AMP images was a list of dictionaries and being identified as an object.
- Pin requests==2.18.4
- Always get oembed AND html data.
- Fix filters.oembed module once lassie is packaged.
- Add support for OEmbed providers (YouTube)
- Remove owl emoji from README.rst so installs on Windows don't fail.
- Fix issue where a website may have malformed HTML and no <html> tag causing soup.html to be None (#60)
- Updated beautifulsoup4 to 4.5.3
- Update html5lib to 1.0b10
- Added a default fake user agent to use instead of using python-requests/version (some websites will mark certain user agents as bot attempts)
- Updated requests to 2.13.0
- Fix Python 3 support
- Handle empty AMP image lists
- Handle AMP image list of strings vs list of objects
- Handle AMP data that is contained in a list
- Retrieve videos and thumbnails (as images) from AMP VideoObjects
- Fix issue where AMP images could be lists inside an object
- Fix issue where some keys returned (i.e. description) would not be retrieved if the key existed with an empty value already
- Fix issue where AMP images could be images and not objects
- Add support for AMP "description" attribute
- Fix issue where an error would be thrown if width/height of an image weren't strings
- Fix duplicate AMP title request, should have been url
- Add support for links that use AMP
- Add status_code to response dictionary (for "file-like" responses, as well)
- Add support for open graph site_name
- Add status_code to response dictionary
- Pinned requests library to version 2.8.1
- Pinned beautifulsoup4 library to version 4.4.1
- Add Python 3.5 to Travis CI build matrix (officially support 3.5)
- Catch and raise LassieError on HEAD requests when handle_file_content is passed to the Lassie API
- Pinned requests library to version 2.8.0
- Support for secure url image and videos from Open Graph
- Simplified merge_settings and data updating internally
- Handle when a website doesn't set a value on the "keywords" meta tag
- Updated requests and beautifulsoup4 library versions
- Fix issue where headers didn't always have 'Content-Type' key
- Added ability to fetch links that are image files (jpg, gif, png, bmp)
- Renamed _retreive_content to _retrieve_content because I evidently don't know how to spell correctly
- Updated requests and beautifulsoup4 library versions
- Added support for manipulating the request, see Advanced Usage docs
- Fixed issue where lassie.fetch would break if the page had no title
- Lassie is now more lenient when it comes to width and height values of images (now accepts integers (100) or integer with px (100px)
- Image URLs for all images are now absolute
- Added support for locale to be returned. If lang is specified in the html tag and it normalizes to an actual locale, it will be added to the returned data.
- Fixed bug where height was not being returned for body images
- Added test coverage, we're 100% covered! :D
- Remove spaces from the returned keywords list
- Fixed issue where favicon was not being retrieved
- Fixed priority for class level vs method level params
- Fix package error when importing
- Initial Release