Skip to content
Open
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
361 changes: 361 additions & 0 deletions rust/types/src/collection_configuration.rs
Original file line number Diff line number Diff line change
Expand Up @@ -1079,4 +1079,365 @@ mod tests {
))
);
}

#[cfg(feature = "testing")]
mod proptests {
use super::*;
use crate::hnsw_configuration::Space;
use crate::proptest_utils::strategies::{
embedding_function_strategy, internal_collection_configuration_strategy,
internal_hnsw_configuration_strategy, internal_spann_configuration_strategy,
knn_index_strategy,
};
use crate::{HnswConfiguration, MetadataValue, SpannConfiguration};
use proptest::prelude::*;
use proptest::test_runner::TestCaseError;

fn space_to_metadata_str(space: &Space) -> &'static str {
match space {
Space::L2 => "l2",
Space::Cosine => "cosine",
Space::Ip => "ip",
}
}

fn metadata_from_hnsw_config(config: &InternalHnswConfiguration) -> Metadata {
let mut metadata = Metadata::new();
metadata.insert(
"hnsw:space".to_string(),
MetadataValue::Str(space_to_metadata_str(&config.space).to_string()),
);
metadata.insert(
"hnsw:construction_ef".to_string(),
MetadataValue::Int(config.ef_construction as i64),
);
metadata.insert(
"hnsw:search_ef".to_string(),
MetadataValue::Int(config.ef_search as i64),
);
metadata.insert(
"hnsw:M".to_string(),
MetadataValue::Int(config.max_neighbors as i64),
);
metadata.insert(
"hnsw:num_threads".to_string(),
MetadataValue::Int(config.num_threads as i64),
);
metadata.insert(
"hnsw:resize_factor".to_string(),
MetadataValue::Float(config.resize_factor),
);
metadata.insert(
"hnsw:sync_threshold".to_string(),
MetadataValue::Int(config.sync_threshold as i64),
);
metadata.insert(
"hnsw:batch_size".to_string(),
MetadataValue::Int(config.batch_size as i64),
);
metadata
}

fn metadata_hnsw_strategy() -> impl Strategy<Value = (InternalHnswConfiguration, Metadata)>
{
internal_hnsw_configuration_strategy().prop_map(|config| {
let metadata = metadata_from_hnsw_config(&config);
(config, metadata)
})
}

fn assert_spann_is_default_with_space(
config: &InternalSpannConfiguration,
expected_space: Space,
) -> Result<(), TestCaseError> {
let default_config = InternalSpannConfiguration {
space: expected_space,
..InternalSpannConfiguration::default()
};
prop_assert_eq!(config, &default_config);
Ok(())
}

fn assert_hnsw_is_default_with_space(
config: &InternalHnswConfiguration,
expected_space: Space,
) -> Result<(), TestCaseError> {
let default_config = InternalHnswConfiguration {
space: expected_space,
..InternalHnswConfiguration::default()
};
prop_assert_eq!(config, &default_config);
Ok(())
}

proptest! {
#[test]
fn try_from_config_roundtrip_internal(
internal_config in internal_collection_configuration_strategy()
) {
let collection_config: CollectionConfiguration = internal_config.clone().into();
let default_knn = match &internal_config.vector_index {
VectorIndexConfiguration::Hnsw(_) => KnnIndex::Hnsw,
VectorIndexConfiguration::Spann(_) => KnnIndex::Spann,
};

let result = InternalCollectionConfiguration::try_from_config(
collection_config.clone(),
default_knn,
None,
)
.expect("conversion should succeed");

let embedding_function = internal_config.embedding_function.clone();
let expected_vector_index = match &internal_config.vector_index {
VectorIndexConfiguration::Hnsw(original) => {
let external: HnswConfiguration = original.clone().into();
let expected_internal: InternalHnswConfiguration = external.clone().into();
match &result.vector_index {
VectorIndexConfiguration::Hnsw(converted) => {
prop_assert_eq!(converted, &expected_internal);
}
_ => prop_assert!(false, "expected HNSW configuration"),
}
VectorIndexConfiguration::Hnsw(expected_internal)
}
VectorIndexConfiguration::Spann(original) => {
let external: SpannConfiguration = original.clone().into();
let expected_internal: InternalSpannConfiguration = external.clone().into();
match &result.vector_index {
VectorIndexConfiguration::Spann(converted) => {
prop_assert_eq!(converted, &expected_internal);
}
_ => prop_assert!(false, "expected SPANN configuration"),
}
VectorIndexConfiguration::Spann(expected_internal)
}
};

prop_assert_eq!(
result.embedding_function.clone(),
embedding_function.clone()
);
let expected = InternalCollectionConfiguration {
vector_index: expected_vector_index,
embedding_function: embedding_function.clone(),
};
prop_assert_eq!(result, expected);

let opposite_knn = match &internal_config.vector_index {
VectorIndexConfiguration::Hnsw(_) => KnnIndex::Spann,
VectorIndexConfiguration::Spann(_) => KnnIndex::Hnsw,
};
let opposite_result = InternalCollectionConfiguration::try_from_config(
collection_config,
opposite_knn,
None,
)
.expect("conversion for opposite default should succeed");

prop_assert_eq!(
opposite_result.embedding_function.clone(),
internal_config.embedding_function.clone()
);

match (&internal_config.vector_index, &opposite_result.vector_index) {
(VectorIndexConfiguration::Hnsw(original), VectorIndexConfiguration::Spann(spann)) => {
let expected_space = original.space.clone();
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

[BestPractice]

The Space enum derives Copy, so calling .clone() on it is redundant. You can remove the .clone() calls for cleaner, more idiomatic code. This applies to several places in this file where .clone() is called on a value of type Space.

For example, this line:

let expected_space = original.space.clone();

can be simplified to:

let expected_space = original.space;

Other instances are on lines 1249, 1282, 1283, 1340, 1343, 1344, 1372, 1375, and 1376.

Context for Agents
[**BestPractice**]

The `Space` enum derives `Copy`, so calling `.clone()` on it is redundant. You can remove the `.clone()` calls for cleaner, more idiomatic code. This applies to several places in this file where `.clone()` is called on a value of type `Space`.

For example, this line:
```rust
let expected_space = original.space.clone();
```
can be simplified to:
```rust
let expected_space = original.space;
```

Other instances are on lines 1249, 1282, 1283, 1340, 1343, 1344, 1372, 1375, and 1376.

File: rust/types/src/collection_configuration.rs
Line: 1245

assert_spann_is_default_with_space(spann, expected_space)?;
}
(VectorIndexConfiguration::Spann(original), VectorIndexConfiguration::Hnsw(hnsw)) => {
let expected_space = original.space.clone();
assert_hnsw_is_default_with_space(hnsw, expected_space)?;
}
_ => prop_assert!(false, "unexpected opposite conversion result"),
}
}
}

proptest! {
#[test]
fn try_from_config_uses_metadata_when_configs_absent(
(expected_hnsw, metadata) in metadata_hnsw_strategy(),
embedding in embedding_function_strategy(),
knn in knn_index_strategy(),
) {
let collection_config = CollectionConfiguration {
hnsw: None,
spann: None,
embedding_function: embedding.clone(),
};

let result = InternalCollectionConfiguration::try_from_config(
collection_config,
knn,
Some(metadata.clone()),
)
.expect("conversion should succeed");

match (knn, &result.vector_index) {
(KnnIndex::Hnsw, VectorIndexConfiguration::Hnsw(hnsw)) => {
prop_assert_eq!(hnsw, &expected_hnsw);
}
(KnnIndex::Spann, VectorIndexConfiguration::Spann(spann)) => {
prop_assert_eq!(spann.space.clone(), expected_hnsw.space.clone());
assert_spann_is_default_with_space(spann, expected_hnsw.space.clone())?;
}
_ => prop_assert!(false, "unexpected vector index variant"),
}
prop_assert_eq!(result.embedding_function.clone(), embedding);
}
}

proptest! {
#[test]
fn try_from_config_uses_metadata_when_hnsw_config_is_default_values(
(expected_hnsw, metadata) in metadata_hnsw_strategy(),
embedding in embedding_function_strategy(),
) {
let collection_config = CollectionConfiguration {
hnsw: Some(HnswConfiguration::default()),
spann: None,
embedding_function: embedding.clone(),
};

let result = InternalCollectionConfiguration::try_from_config(
collection_config,
KnnIndex::Hnsw,
Some(metadata.clone()),
)
.expect("conversion should succeed");

match &result.vector_index {
VectorIndexConfiguration::Hnsw(hnsw) => {
prop_assert_eq!(hnsw, &expected_hnsw);
}
_ => prop_assert!(false, "expected hnsw configuration"),
}
prop_assert_eq!(result.embedding_function.clone(), embedding);
}
}

proptest! {
#[test]
fn try_from_config_prefers_spann_when_default_is_spann(
hnsw_config in internal_hnsw_configuration_strategy(),
embedding in embedding_function_strategy(),
) {
let embedding_clone = embedding.clone();
let collection_config = CollectionConfiguration {
hnsw: Some(hnsw_config.clone().into()),
spann: None,
embedding_function: embedding,
};

let result = InternalCollectionConfiguration::try_from_config(
collection_config,
KnnIndex::Spann,
None,
)
.expect("conversion should succeed");

let expected_space = hnsw_config.space.clone();
match &result.vector_index {
VectorIndexConfiguration::Spann(spann) => {
prop_assert_eq!(spann.space.clone(), expected_space.clone());
assert_spann_is_default_with_space(spann, expected_space)?;
}
_ => prop_assert!(false, "expected spann configuration"),
}
prop_assert_eq!(result.embedding_function.clone(), embedding_clone);
}
}

proptest! {
#[test]
fn try_from_config_prefers_hnsw_when_default_is_hnsw(
spann_config in internal_spann_configuration_strategy(),
embedding in embedding_function_strategy(),
) {
let embedding_clone = embedding.clone();
let collection_config = CollectionConfiguration {
hnsw: None,
spann: Some(spann_config.clone().into()),
embedding_function: embedding,
};

let result = InternalCollectionConfiguration::try_from_config(
collection_config,
KnnIndex::Hnsw,
None,
)
.expect("conversion should succeed");

let expected_space = spann_config.space.clone();
match &result.vector_index {
VectorIndexConfiguration::Hnsw(hnsw) => {
prop_assert_eq!(hnsw.space.clone(), expected_space.clone());
assert_hnsw_is_default_with_space(hnsw, expected_space)?;
}
_ => prop_assert!(false, "expected hnsw configuration"),
}
prop_assert_eq!(result.embedding_function.clone(), embedding_clone);
}
}

proptest! {
#[test]
fn try_from_config_defaults_when_configs_absent(
embedding in embedding_function_strategy(),
knn in knn_index_strategy(),
) {
let collection_config = CollectionConfiguration {
hnsw: None,
spann: None,
embedding_function: embedding.clone(),
};

let result = InternalCollectionConfiguration::try_from_config(
collection_config,
knn,
None,
)
.expect("conversion should succeed");

match (knn, &result.vector_index) {
(KnnIndex::Hnsw, VectorIndexConfiguration::Hnsw(hnsw)) => {
prop_assert_eq!(hnsw, &InternalHnswConfiguration::default());
}
(KnnIndex::Spann, VectorIndexConfiguration::Spann(spann)) => {
prop_assert_eq!(spann, &InternalSpannConfiguration::default());
}
_ => prop_assert!(false, "unexpected vector index variant"),
}
prop_assert_eq!(result.embedding_function.clone(), embedding);
}
}

proptest! {
#[test]
fn try_from_config_errors_on_multiple_configs(
hnsw_config in internal_hnsw_configuration_strategy(),
spann_config in internal_spann_configuration_strategy(),
embedding in embedding_function_strategy(),
knn in knn_index_strategy(),
) {
let collection_config = CollectionConfiguration {
hnsw: Some(hnsw_config.into()),
spann: Some(spann_config.into()),
embedding_function: embedding,
};

let result = InternalCollectionConfiguration::try_from_config(
collection_config,
knn,
None,
);

prop_assert!(matches!(
result,
Err(CollectionConfigurationToInternalConfigurationError::MultipleVectorIndexConfigurations)
));
}
}
}
}
Loading