Closed pvanheus closed 7 years ago
Thanks for reporting this. As discussed, I think the tokeniser we run before indexing the documents is failing to strip the XML tags. Should be a simple fix (famous last words!).
Fixed in the search engine by adding a pre-processing step to strip XML tags https://github.com/blahah/yunodb/commit/15a706d5b8a833774e050a6fddf6e3a8c8b3cd63
Seems not to fix things:
is there an index that needs to be rebuilt or a cache that needs to be emptied or something?
ah, yes you'll need to delete the search index:
rm -rf ~/.sciencefair
will nuke the whole thing
With the eLife datasource enabled, search for "Homo naledi" fails.
It should return papers such as DOI 10.7554/eLife.10627.
As discussed on IRC, this might be related to the title having
<italics>Homo naledia</italics>
markup.