MDN JavaScript: make it a Longtail

duckduckgo / zeroclickinfo-fathead

DuckDuckGo Instant Answers based on keyword data files

https://duckduckhack.com/

Other

318 stars 365 forks source link

MDN JavaScript: make it a Longtail #239

Closed MariagraziaAlastra closed 7 years ago

MariagraziaAlastra commented 8 years ago

Currently, MDN JavaScript doesn't trigger nearly enough; it could cover a lot more JS reference queries, but the strict query - article title matching prevents it from doing that.

To Do

Converting the output file either to JSON or XML.

From there we'll be able to test and see if anything else needs to be done.

Read the docs to learn more about Fatheads and Longtails.

Please see the forum topic Improve MDN JavaScript Fathead for more details, and leave a reply before starting to work on this. Thanks!

IA Page: https://duck.co/ia/view/mdnjs Maintainer: @ericedens

Forum topic: Improve MDN JavaScript Fathead

This issue is part of the Programming Mission: help us improve the results for JavaScript related searches!

bfmags commented 8 years ago

Started work around converting the output file to XML (JSON will be the same). Creating a bash script to parse the file (WIP https://github.com/bfmags/output2xml), considering switching to python since there's already some parser code

bfmags commented 8 years ago

got first xml output from output.txt https://github.com/bfmags/zeroclickinfo-longtail/blob/master/lib/longtail/mdnjs/output.xml
based on the existing code to generate the xml (some cleanup/rewrite still to be done) https://github.com/bfmags/zeroclickinfo-longtail/blob/master/lib/longtail/mdnjs/toxml.py
the xml in the docs page http://docs.duckduckhack.com/resources/longtail-overview.html doesn’t seem to be valid?
missing instant_answer_id
~~already have the json writer in place, missing some tweaks, will commit soon~~
update commited: https://github.com/bfmags/zeroclickinfo-longtail/commit/84be7bb0068f979bf8869a0d7c51815abd2cd74c

MariagraziaAlastra commented 8 years ago

the xml in the docs page http://docs.duckduckhack.com/resources/longtail-overview.html doesn’t seem to be valid?

Yeah, missing an </add> at the end...

Thanks @bfmags, you're doing great!

moollaza commented 7 years ago

Closing as this isn't currently an option for us. We're going to continue with the Fathead approach for now. If this comes up again we can reopen.