We compute the string index of each query word which matches the result's name in query/name_cost.ml, but then only keep the total of the costs. If we also preserved the indexes, we could explain the results by highlighting the matching parts of the names with <mark>.
Doing the same for the documentation comments will require do a similar search but on html (which isn't done yet, as we just assume a query word not found in the name will be in the comment!). It'll be even harder to have a similar highlight for types, since their pretty-pretting is handled by odoc so we don't know which substrings to highlight... We can probably do follow-up PRs for those, name matching would already be a great first step.
We compute the string index of each query word which matches the result's name in query/name_cost.ml, but then only keep the total of the costs. If we also preserved the indexes, we could explain the results by highlighting the matching parts of the names with
<mark>.Doing the same for the documentation comments will require do a similar search but on html (which isn't done yet, as we just assume a query word not found in the name will be in the comment!). It'll be even harder to have a similar highlight for types, since their pretty-pretting is handled by odoc so we don't know which substrings to highlight... We can probably do follow-up PRs for those, name matching would already be a great first step.