chore(deps): update cargo dependencies

Updated packages:
 0.10 (major)
 0.10 (major)

Fixed breaking changes:
- rand 0.10 moved random()/random_range() from Rng to RngExt trait
 Rng (in rand_core)
- rand 0.10 removed RngCore re-export from rand crate
 rand::make_rng()

Supersedes: #100, #101

Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>

Author: tclem

crates/bpe/Cargo.toml

@@ -21,7 +21,7 @@ aneubeck-daachorse = "1.1.1"
 base64 = { version = "0.22", optional = true }
 fnv = "1.0"
 itertools = "0.14"
-rand = { version = "0.9", optional = true }
+rand = { version = "0.10", optional = true }
 serde = { version = "1", features = ["derive"] }
 
 [dev-dependencies]

crates/bpe/benchmarks/Cargo.toml

@@ -21,6 +21,6 @@ test = true
 bpe = { path = "../../bpe", features = ["rand", "tiktoken"] }
 bpe-openai = { path = "../../bpe-openai" }
 criterion = "0.8"
-rand = "0.9"
+rand = "0.10"
 tiktoken-rs = "0.9"
 tokenizers = { version = "0.22", features = ["http"] }

crates/bpe/benchmarks/performance.rs

@@ -9,9 +9,7 @@ use bpe_benchmarks::*;
 use criterion::{
     criterion_group, criterion_main, AxisScale, BenchmarkId, Criterion, PlotConfiguration,
 };
-use rand::rngs::StdRng;
-use rand::SeedableRng;
-use rand::{rng, Rng};
+use rand::{rng, RngExt};
 
 fn counting_benchmark(c: &mut Criterion) {
     for (name, bpe, _, _) in TOKENIZERS.iter() {

crates/bpe/src/byte_pair_encoding.rs

@@ -168,7 +168,7 @@ pub fn find_hash_factor_for_tiktoken(data: &str) -> Result<u64, base64::DecodeEr
 pub fn find_hash_factor_for_dictionary(tokens: impl IntoIterator<Item = Vec<u8>>) -> u64 {
     use std::collections::HashSet;
 
-    use rand::Rng;
+    use rand::RngExt;
 
     let all_tokens = tokens.into_iter().collect_vec();
     let mut rnd = rand::rng();
@@ -573,7 +573,7 @@ impl BytePairEncoding {
     //     and hence may include previously discarded token later down the byte stream. At the sentence level though we don't expect it to make much difference.
     //     Also, this implementation of BPE constructs merges on the fly from the set of tokens, hence might come up with a different set of merges with the same dictionary.
     #[cfg(feature = "rand")]
-    pub fn encode_minimal_dropout<R: rand::Rng>(
+    pub fn encode_minimal_dropout<R: rand::RngExt>(
         &self,
         text: &[u8],
         dropout: f32,
@@ -627,7 +627,7 @@ pub fn create_test_string_with_predicate(
     min_bytes: usize,
     predicate: impl Fn(&str) -> bool,
 ) -> String {
-    use rand::{rng, Rng};
+    use rand::{rng, RngExt};
     // the string we accumulated thus far
     let mut result = String::new();
     // the tokens we added so we can backtrack
@@ -662,7 +662,7 @@ pub fn create_test_string_with_predicate(
 
 #[cfg(feature = "rand")]
 pub fn select_test_string(text: &str, min_bytes: usize) -> &str {
-    use rand::{rng, Rng};
+    use rand::{rng, RngExt};
     let mut start = rng().random_range(0..text.len() - min_bytes);
     while !text.is_char_boundary(start) {
         start -= 1;
@@ -677,7 +677,7 @@ pub fn select_test_string(text: &str, min_bytes: usize) -> &str {
 /// Generate test bytes by concatenating random tokens.
 #[cfg(feature = "rand")]
 pub fn create_test_bytes(bpe: &BytePairEncoding, min_bytes: usize) -> Vec<u8> {
-    use rand::{rng, Rng};
+    use rand::{rng, RngExt};
     let mut result = Vec::new();
     while result.len() < min_bytes {
         let i = rng().random_range(0..bpe.num_tokens());

crates/bpe/tests/Cargo.toml

@@ -6,5 +6,5 @@ edition = "2021"
 bpe = { path = "../../bpe", features = ["rand"] }
 bpe-openai = { path = "../../bpe-openai" }
 itertools = "0.14"
-rand = "0.9"
+rand = "0.10"
 tiktoken-rs = "0.9"

crates/bpe/tests/src/lib.rs

@@ -1,7 +1,7 @@
 #[cfg(test)]
 mod tests {
     use itertools::Itertools;
-    use rand::{rng, Rng};
+    use rand::{rng, RngExt};
     use tiktoken_rs::cl100k_base_singleton;
 
     use bpe::appendable_encoder::AppendableEncoder;

crates/geo_filters/Cargo.toml

@@ -30,17 +30,17 @@ fnv = "1.0"
 hyperloglogplus = { version = "0.4", optional = true }
 itertools = "0.14"
 once_cell = "1.18"
-rand = { version = "0.9", optional = true }
+rand = { version = "0.10", optional = true }
 rayon = { version = "1.7", optional = true }
 regex = { version = "1", optional = true }
 serde = { version = "1.0", default-features = false, optional = true }
-rand_chacha = { version = "0.9", optional = true }
+rand_chacha = { version = "0.10", optional = true }
 
 [dev-dependencies]
 criterion = "0.8"
 geo_filters = { path = ".", features = ["evaluation"] }
-rand = "0.9"
-rand_chacha = "0.9"
+rand = "0.10"
+rand_chacha = "0.10"
 rayon = "1.7"
 
 [[bench]]

crates/geo_filters/src/config.rs

@@ -353,7 +353,7 @@ pub(crate) fn take_ref<I: Iterator>(iter: &mut I, n: usize) -> impl Iterator<Ite
 
 #[cfg(test)]
 pub(crate) mod tests {
-    use rand::RngCore;
+    use rand::Rng as RngCore;
     use rand_chacha::ChaCha12Rng;
 
     use crate::{Count, Method};

crates/geo_filters/src/config/lookup.rs

@@ -45,7 +45,7 @@ impl HashToBucketLookup {
 
 #[cfg(test)]
 mod tests {
-    use rand::RngCore;
+    use rand::Rng as RngCore;
     use rand_chacha::ChaCha12Rng;
 
     use crate::{

crates/geo_filters/src/diff_count.rs

@@ -362,8 +362,7 @@ impl<'a, C: GeoConfig<Diff>> GeoDiffCount<'a, C> {
     /// the resulting geofilter should always be the same.
     #[cfg(any(test, feature = "test-support"))]
     pub fn pseudorandom_filter_with_config(config: C, items: usize) -> Self {
-        use rand::RngCore;
-        use rand_chacha::rand_core::SeedableRng;
+        use rand_chacha::rand_core::{Rng, SeedableRng};
 
         let mut rng = rand_chacha::ChaCha12Rng::seed_from_u64(items as u64);
         let mut filter = Self::new(config);
@@ -468,7 +467,7 @@ mod tests {
     use std::io::Write;
 
     use itertools::Itertools;
-    use rand::{seq::IteratorRandom, RngCore};
+    use rand::{seq::IteratorRandom, Rng as RngCore};
     use rand_chacha::ChaCha12Rng;
 
     use crate::{