The new fast field code, based on columnar, had a larger minimum memory
footprint, causing the first docuemnt to trigger a flush of the asegment
in this unit test.
This PR prevents the allocation of a large capacity for the different hashmap tables
using in the columnar writer.
Closes#1859
Fixes an issue in the skip list deserialization, which deserialized the byte start offset incorrectly as u32.
`get_doc` will fail for any docs that live in a block with start offset larger than u32::MAX (~4GB).
Causes index corruption, if a segment with a doc store larger 4GB is merged.
tantivy version 0.19 is affected
Prev: thread 'main' panicked at 'called `Result::unwrap()` on an `Err` value: IoError(Custom { kind: InvalidData, error: "Reach end of buffer while reading VInt" })', src/main.rs:46:14
Now: Automatic downgrade to next available level
* Improvement on the scalar / random bitpacker code.
Added proptesting
Added simple benchmark
Added assert and comments on the very non trivial hidden contract
Remove the need for an extra padding.
The last point introduces a small performance regression (~10%).
* Fixing unit tests
closes#1766
Finding tantivy tokenizers is a frustrating experience currently, since
they need be updated for each tantivy version. That's unnecessary since
the API is rather stable anyway.
Introduce MakeZero trait, remove make_zero from FastValue
Merge two multivalue fastfield implementations into one
prepare range query on fastfield for different types
* Make nightly Clippy mostly happy.
* Document how to produce TermSetQuery queries using QueryParser.
* Enable construction of queries using FuzzyTermQuery via the QueryParser
* Use FxHashMap instead of HashMap in the QueryParser as these hash tables are not exposed to DoS attacks.
* Use a struct instead of a tuple to improve readability.
* Refactoring to prepare for the addition of dynamic fast field
- Exposing insert_key / insert_value
- Renamed SSTable::{Reader/Writer}-> SSTable::{ValueReader/ValueWriter}
- Added a generic Dictionary object in the sstable crate
- Removing the TermDictionary wrapper from tantivy, relying directly on
an alias of the generic Dictionary object.
- dropped the use of byteorder in sstable.
- Stopped scanning / reading the entire dictionary when streaming a range.
* Added a benchmark for streaming sstable ranges.
* CR comments.
Rename deserialize_u64 -> deserialize_vint_u64
* Removed needless allocation, split serialize into serialize and clear.
* Fix max bucket limit in histogram
The max bucket limit in histogram was broken, since some code introduced temporary filtering of buckets, which then resulted into an incorrect increment on the bucket count.
The provided solution covers more scenarios, but there are still some scenarios unhandled (See #1702).
* Apply suggestions from code review
Co-authored-by: Paul Masurel <paul@quickwit.io>
Co-authored-by: Paul Masurel <paul@quickwit.io>
* prepare fastfield format for null index
* add format version for fastfield
* Update fastfield_codecs/src/compact_space/mod.rs
* switch to variable size footer
* serialize delta of end
