ql:contains-word now can show the score of the word match in the respective text
#1397
Conversation
|
modified: src/index/IndexImpl.Text.cpp modified: test/QueryPlannerTestHelpers.h modified: test/engine/TextIndexScanForWordTest.cpp Problems: |
|
…adapted unit tests. Missing e2e tests.
Commit doesn't contain all changes necessary for pull request yet.
…x. This is done through passing the words and docsfile as string, and then building the text index as normal. Basic Test is existent (TODO make more edge case tests) and e2e testing is fixed.
…re still unstable because of the way nofContexts are counted. Implemented new more refined tests.
|
Everything should be fixed now. Tests are complete and scores are shown. Also new test methods to build a text index are implemented. |
Codecov ReportAttention: Patch coverage is
Additional details and impacted files@@ Coverage Diff @@
## master #1397 +/- ##
==========================================
+ Coverage 89.67% 89.80% +0.12%
==========================================
Files 383 385 +2
Lines 36942 36995 +53
Branches 4174 4181 +7
==========================================
+ Hits 33129 33223 +94
+ Misses 2512 2477 -35
+ Partials 1301 1295 -6 ☔ View full report in Codecov by Sentry. |
…1665) Each operation now has a `bool` that determines whether the results can be stored in the cache or not (whether it is actually stored depends on other circumstances, like the available cache size). That `bool` does not have to be fixed when the operation is created, but can be changed. For example, this is useful for index scans that only return a subset of their full result (because of another constraining operation, like a join or a filter).
This is a first step towards making QLever compile with C++17. If the compile-time flag `QLEVER_CPP_17` is set, use Eric Niebler's `range-v3` library as a drop-in replacement for `std::ranges`. In the code, we simply write `ql::ranges` instead of `std::ranges` in most places. Some places need special treatment. For example, where `std::ranges` was used as a C++20 concept, we now use the macros `CPP_template` and `CPP_and` (also from the `range-v3` library), which does the right thing for both C++20 and C++17.
… saving nofNonLiterals in the configuration json file.
…g saved in the TextMetaData and instead saving nofNonLiterals in the configuration json file." This reverts commit 1adcecb.
…adapted unit tests. Missing e2e tests.
…x. This is done through passing the words and docsfile as string, and then building the text index as normal. Basic Test is existent (TODO make more edge case tests) and e2e testing is fixed.
…re still unstable because of the way nofContexts are counted. Implemented new more refined tests.
…o the wordsFileContent and docsFileContent strings. Now you can clearly see what lines are added and can writing tests is cleaner
…in the wordsFileContent and docsFileContent as pair contentsOfWordsFileAndDocsFile
Signed-off-by: Johannes Kalmbach <[email protected]>
… saving nofNonLiterals in the configuration json file.
…g saved in the TextMetaData and instead saving nofNonLiterals in the configuration json file." This reverts commit 1adcecb.
… saving nofNonLiterals in the configuration json file.
…tIndex everywhere aswell as num-non-literals to nom-non-literals-text-index
Conformance check passed ✅No test result changes. |
|
joka921
changed the title
ql:contains-word now can show the score of the word match in the respective text
The fulltext index of QLever has forever been able to associate the occurence of a word in a text with a score.
This PR adds the functionality to actually retrieve this score and to use it in the remainder of the query.
Currently the score is bound to a variable the name of which is automatically determined from the involved literals and variables. The easiest way to get the names of these variables is to use
SELECT *or to look at the runtime information tree.