Dynamic Browsing: Support dynamic download of site entirely through browser

[davidfstr]

Currently Crystal is designed for the user to use the GUI to both (1) define URLs and URL patterns to download and (2) request that these URLs actually be downloaded. Then with the URLs already downloaded the user can start Crystal's server and browse them.

It would be nice to support a workflow where the user immediately starts Crystal's server and interactively browses to webpages that are desired to be downloaded. Whenever trying to visit an HTML page that isn't downloaded, Crystal should offer an in-browser UI to interactively create a new Root Resource or Resource Group to download the new page.

Such an interactive browser-based means of downloading a site is likely to be more intuitive that the current GUI-based approach, especially for a browser-based login workflow.

Outcomes:

• [ ] When Crystal serves a HTTP 404 Not in Archive page, it should present a browser-based UI to interactively create a Root Resource or Resource Group.
• [ ] An automated UI test does exist that validates a full browser-based workflow for downloading a site, without using the current wxPython GUI.
  • Such a test is similar in flavor to the current test_can_download_and_serve_a_site_requiring_dynamic_url_discovery stub but differs in that it expects the workflow to be 100% browser-based

[davidfstr]

Prior thinking on such a workflow: