* Make TopCollector generic
Make TopCollector take a generic type instead of only being tied to
score. This will allow for sharing code between a TopCollector that
sorts results by Score and a TopCollector that sorts documents by a fast
field. This commit makes no functional changes to TopCollector.
* Add TopFieldCollector and TopScoreCollector
Create two new collectors that use the refactored TopCollector.
TopFieldCollector has the same functionality that TopCollector
originally had. TopFieldCollector allows for sorting results by a given
fast field. Closestantivy-search/tantivy#388
* Make TopCollector private
Make TopCollector package private and export TopFieldCollector as
TopCollector to maintain backwards compatibility. Mark TopCollector
as deprecated to encourage use of the non-aliased TopFieldCollector.
Remove Collector implementation for TopCollector since it is not longer
used.
`try_add_token` will now update the stop_offset as well.
`FragmentCandidate::new` now just takes `start_offset`,
it expects `try_add_token` to be called to add a token.
* Moving Range and All to Leaves
* Parsing OR/AND
* Simplify user input ast
* AND and OR supported. Returning an error when mixing syntax
Closes#246
* Added support for NOT
* Updated changelog
* add position_length to Token
refer #291
* Add term offset to `PhraseQuery`
ref #291
* Add new constructor for `PhraseQuery` that allows custom offset
* fix the method name as per pr comment
* Closes#291
Added unit test.
Using offsets from the analyzer in QueryParser.
* Moved NUM_SEARCHERS into a local variable
dynamically determined as the number of available cpus.
var name in lowercase (not a constant anymore).
updated it in docstring
* lowercased the varnames
* User can set number of logical cores in create_from_metas
* cargo fmt
* Num_searchers as Arc<AtomicUsize>
Retrieving the value with Relaxed ordering
Reverted create_from_metas signature. However, it calls num_cpus and
sets the Arc val
