archivesunleashed / warclight

A Rails engine supporting the discovery of web archives.
https://archivesunleashed.org/warclight/
Other
48 stars 10 forks source link

Advanced Search #22

Closed ianmilligan1 closed 6 years ago

ianmilligan1 commented 7 years ago
Title (Goal) Advanced Search
Primary Actor user
Scope access
Level High
Story This again borrows from UKWA Shine as we seek to bring functionality of Blacklight up to their platform! Researchers may want functionality beyond the simple faceted search provided on the main page of Warclight. For example, a user may want to find mentions of "Twitter" in sites excluding "Twitter.com," or may want to provide more granular date ranges ("public transportation" in July and August 2007, rather than just the year itself), or they may want to find pages that contain multiple words within x words of each other. The user would be overwhelmed if all the facets were provided in the main search, but they may want more granularity in an advanced search.

Word Proximity: One common example we used on WebArchives.ca was the word proximity. We searched for the word "Harper" and the word "fascist" within 25 words of each other, as a way to begin to find sites that may have used extreme rhetoric with the Conservative government. This allowed us to find Green Party blogs that characterized the Conservative government this way.

Here is the example from SHINE:

screen shot 2017-08-23 at 11 28 45 am

Granularity: Right now in the main page, here are the list of facets for date for example:

screen shot 2017-08-23 at 11 29 24 am

For many users, this is enough! But people may want to use only a few months, or exclude certain date ranges (say if a crawl was corrupted, or if they were overwhelmed by a particular event and wanted to ignore it). It would be good to have date range search like so (again borrowing from Shine):

screen shot 2017-08-23 at 11 30 08 am

ianmilligan1 commented 7 years ago

This will likely build off the blacklight_advanced_search plugin. We are waiting on the plugin to update to be compatible with blacklight (7.0.0.alpha).

ruebot commented 7 years ago

Talked to @cbeer in #blacklight code4lib slack, and his advice is to not make the advanced search plugin a direct dependency, but rather generate it into the host application as part of our install generator.

Examples to get us started:

w/r/t advanced search plugin version, we might want to put some cycles into helping update that to 7.0.0

it’s currently lagging, yes. usually, the first person who wants to use it with an incompatible version of blacklight goes and updates it. this one shouldn’t be too bad, so i’d encourage you to maybe take a look at doing that.

...and

it looks like the build was working recently, so if you just run it against the latest 6.x release, fix any deprecation warnings (which ought to be informative, and if they’re not, we should fix that), update it to use 7.x, and bang your head against the wall doing the bootstrap migration.

ianmilligan1 commented 6 years ago

We decided not to implement advanced search: faceting should be enough, and users don't necessarily use it. Let's make our basic search as "advanced" as users need. The advanced search plugin also hasn't seen much development lately, so we're worried about baking it in.