This is an automated email from the ASF dual-hosted git repository.
tisonkun pushed a commit to branch main
in repository https://gitbox.apache.org/repos/asf/datasketches-rust.git
The following commit(s) were added to refs/heads/main by this push:
new 175e20e chore: prepare 0.7.0 release (#379)
175e20e is described below
commit 175e20ea2fe215a4b7a45f2a73d767934d1cda34
Author: tison <[email protected]>
AuthorDate: Mon Nov 16 16:37:52 2026 +0800
chore: prepare 0.7.0 release (#379)
---
CHANGELOG.md | 22 ++++++++++++----------
Cargo.lock | 2 +-
datasketches/Cargo.toml | 2 +-
3 files changed, 14 insertions(+), 12 deletions(-)
diff --git a/CHANGELOG.md b/CHANGELOG.md
index 254a3d4..9845f8c 100644
--- a/CHANGELOG.md
+++ b/CHANGELOG.md
@@ -4,6 +4,8 @@ All significant changes to this project will be documented in
this file.
## Unreleased
+## v0.7.0
+
### Breaking changes
* `BloomFilterBuilder::with_accuracy` now sizes the filter for its integer
hash count, correcting false-positive targets that previously produced
undersized filters. Capacity and hash count can differ from earlier versions;
to combine with an existing filter, use `with_size(existing.capacity(),
existing.num_hashes())` and the same seed.
@@ -12,29 +14,29 @@ All significant changes to this project will be documented
in this file.
* Fix 32-bit (`wasm32`) double hashing in `BloomFilter` to use the full 64-bit
hash consistently with 64-bit targets, Java, C++, and Go. 64-bit targets and
serialized formats are unchanged; rebuild or re-ingest filters created on
32-bit targets instead of mixing them across versions.
* Remove `BloomFilter::is_compatible`; handle `InvalidArgument` from `union`,
`intersect`, or `difference`, or compare `capacity()`, `num_hashes()`, and
`seed()` directly.
* `CountMinSketch::new` and `CountMinSketch::with_seed` now create serial
version 2 sketches, require `num_hashes` in `1..=127`, and emit images that
require a version 2 reader. Version 2 enables cross-language interchange when
using `i64` counters and the documented input encodings, while existing version
1 images remain readable, updatable, and serializable as version 1.
`CountMinSketch::merge` rejects mixing version 1 and version 2 sketches;
upgrade readers before writers and rebuild [...]
+* `CountMinSketch::decay` now returns `Result<(), Error>`, returning
`InvalidArgument` without modifying the sketch when `decay` is NaN or outside
`(0.0, 1.0]`.
* `CpcSketch::max_serialized_bytes` and `ReqSketch::serialized_size_bytes` are
removed; serialize the sketch and inspect `len()` on the returned byte buffer
when the encoded size is needed.
-* `ThetaJaccardSimilarity::compute` and `TupleJaccardSimilarity::compute` now
return `datasketches::common::JaccardSimilarity` instead of
`datasketches::thetacommon::JaccardSimilarity`, and `datasketches::thetacommon`
is no longer part of the public API.
* `FrequentItemsSketch::update_with_count` and
`FrequentItemsSketch::update_with_count_ref` are renamed to
`update_with_weight` and `update_with_weight_ref` to match weighted update
naming across the crate.
-* `CountMinSketch::decay` now returns `Result<(), Error>`, returning
`InvalidArgument` without modifying the sketch when `decay` is NaN or outside
`(0.0, 1.0]`.
+* `FrequentItemsSketch::num_active_items` is renamed to `num_retained` to
match retained-entry naming across the crate; update callers of
`num_active_items()` to call `num_retained()`.
+* `KllSketch::n` and `ReqSketch::n` are renamed to `total_weight` to match
stream-weight naming across the crate; update callers of `n()` to call
`total_weight()`.
* `KllSketch::merge` now returns `()` instead of `Result<(), Error>` and
panics without modifying the sketch if the combined stream weight exceeds
`u64::MAX`.
* `ReqSketch::num_retained` now returns `usize` instead of `u32`, matching
`KllSketch`, `ThetaSketch`, and `TupleSketch`.
* `ReqSketch::rank_lower_bound` and `ReqSketch::rank_upper_bound` now return
`Result<f64, Error>`, returning `InvalidArgument` when `rank` is NaN or outside
`[0.0, 1.0]` even on empty or exact-mode sketches.
-* `SketchSlice::advance` now takes `n: usize` instead of `u64` and panics if
`n > self.remaining().len()`.
+* `ThetaJaccardSimilarity::compute` and `TupleJaccardSimilarity::compute` now
return `datasketches::common::JaccardSimilarity` instead of
`datasketches::thetacommon::JaccardSimilarity`, and `datasketches::thetacommon`
is no longer part of the public API.
* Remove `PartialEq` from `BloomFilter`, `CountMinSketch`, `HllSketch`,
`JaccardSimilarity`, and `KllSketch`.
-* Rename `KllSketch::n` and `ReqSketch::n` to `total_weight` to match
stream-weight naming across the crate; update callers of `n()` to call
`total_weight()`.
-* Rename `FrequentItemsSketch::num_active_items` to `num_retained` to match
retained-entry naming across the crate; update callers of `num_active_items()`
to call `num_retained()`.
+* `SketchSlice::advance` now takes `n: usize` instead of `u64` and panics if
`n > self.remaining().len()`.
### New features
-* Add `KllSketch::estimated_size` and `ReqSketch::estimated_size` to estimate
sketch memory usage in bytes.
-* Add `CpcUnion::serialize`, `HllUnion::serialize`, `ThetaUnion::serialize`,
`ThetaUnion::serialize_compressed`, and `TupleUnion::serialize` to encode union
results without calling `to_sketch`. Direct serialization avoids intermediate
sketch allocations for `HllUnion` and `CpcUnion`, lets `HllUnion::serialize`
target any `HllType` (`Hll4`, `Hll6`, or `Hll8`) at the union's effective
`lg_k`, supports v4 compressed output via `ThetaUnion::serialize_compressed`,
and encodes `TupleUnion` sum [...]
* Add `XorFilter`, `XorFilterBuilder`, and `XorFilterType` behind the `xor`
feature for immutable membership queries on values or precomputed hashes.
Filters support 8-, 16-, and 32-bit fingerprints and can be serialized for
later queries.
+* Add `CpcUnion::serialize`, `HllUnion::serialize`, `ThetaUnion::serialize`,
`ThetaUnion::serialize_compressed`, and `TupleUnion::serialize` to encode union
results without calling `to_sketch`. Direct serialization avoids intermediate
sketch allocations for `HllUnion` and `CpcUnion`, lets `HllUnion::serialize`
target any `HllType` (`Hll4`, `Hll6`, or `Hll8`) at the union's effective
`lg_k`, supports v4 compressed output via `ThetaUnion::serialize_compressed`,
and encodes `TupleUnion` sum [...]
+* Add `KllSketch::estimated_size` and `ReqSketch::estimated_size` to estimate
sketch memory usage in bytes.
* Add `TDigest::num_retained` and `TDigestMut::num_retained` to report the
number of retained entries or centroids in a t-digest.
### Improvements
-* Implement `Clone` for `TDigest`, `ThetaSketchBuilder`, `TupleSketchBuilder`,
and `TupleUnionBuilder`; derive `Eq` and `Hash` across `NumStdDev`,
`ResizeFactor`, and `SearchCriteria`, and `Hash` across `ErrorType`, `HllType`,
and `RankAccuracy`; and implement `Debug` for `ReqSketchIterator`,
`SketchSlice`, and `SketchBytes` (plus `Clone` and `Default` on codec types).
-* Improve the readability of `Debug` output across sketches, filters, views,
and stateful set operators, and implement `Debug` for `TDigest`.
+* Implement `Clone` for `SketchBytes`, `SketchSlice`, `TDigest`,
`ThetaSketchBuilder`, `TupleSketchBuilder`, and `TupleUnionBuilder`; `Default`
for `SketchBytes`; `Eq` and `Hash` for `NumStdDev`, `ResizeFactor`, and
`SearchCriteria`; and `Hash` for `ErrorType`, `HllType`, and `RankAccuracy`.
+* Improve the readability of `Debug` output across sketches, filters, views,
and stateful set operators, and implement `Debug` for `ReqSketchIterator`,
`SketchBytes`, `SketchSlice`, and `TDigest`.
### Performance improvements
@@ -50,9 +52,9 @@ All significant changes to this project will be documented in
this file.
### Bug fixes
+* `FrequentItemsSketch::deserialize` now returns `InvalidData` when an active
item weight is zero or when the sum of active item weights and maximum error
exceeds the declared stream weight.
* `ReqSketch::merge` now cascades compaction across newly grown levels when a
merged sketch overflows the top compactor's nominal capacity.
* `ReqSketch::deserialize` now returns `InvalidData` for out-of-range or
non-finite compactor `section_size_raw` values instead of overflowing
section-size rounding.
-* `FrequentItemsSketch::deserialize` now returns `InvalidData` when an active
item weight is zero or when the sum of active item weights and maximum error
exceeds the declared stream weight.
* `ThetaJaccardSimilarity` and `TupleJaccardSimilarity` now return
`InvalidArgument` instead of `InvalidData` when a non-empty input sketch's seed
hash does not match the configured seed.
* `ThetaANotB::compute` and `TupleANotB::compute` now stamp empty results with
the operator's seed hash instead of copying an unchecked seed hash from an
empty left input.
diff --git a/Cargo.lock b/Cargo.lock
index 6e5efe2..2aee820 100644
--- a/Cargo.lock
+++ b/Cargo.lock
@@ -234,7 +234,7 @@ dependencies = [
[[package]]
name = "datasketches"
-version = "0.6.0"
+version = "0.7.0"
dependencies = [
"googletest",
"insta",
diff --git a/datasketches/Cargo.toml b/datasketches/Cargo.toml
index 621421f..267ffca 100644
--- a/datasketches/Cargo.toml
+++ b/datasketches/Cargo.toml
@@ -17,7 +17,7 @@
[package]
name = "datasketches"
-version = "0.6.0"
+version = "0.7.0"
edition.workspace = true
homepage.workspace = true
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]