rustfmt groups imports (#660)

This enables the `group_imports` option. With this rustfmt automatically groups the imports in `std`, `external`, and `crate.
This is almost what we already wanted but we lose control of some things:
* it is not able to differentiate between workspace and external so all our crates will go in that section,
* it does not understand local import: 
```rust
mod foo {
    struct Foo;
}

use foo::Foo;
use rand::random;
```
instead of having a separate group for `foo`. 
* It does not respect groups that we create manually
```rust
use tokio;

use itertools;
```
the imports will be merged.

Looking at the changeset we are not too bad at maintaining this style. Most of the changes are the position of `crate` wr.t. to `self` or `super`, fusing groups, and moving our workspace crates to the external one.
But this will be our only option to enforce this style with a tool.

https://rust-lang.github.io/rustfmt/?version=v1.5.1&search=#group_imports
rust-lang/rustfmt#4709
rust-lang/rustfmt#4693

discovery_engine_core/ai/ai/src/coi/context.rs

@@ -211,13 +211,12 @@ mod tests {
     use ndarray::arr1;
     use xayn_discovery_engine_test_utils::assert_approx_eq;
 
+    use super::*;
     use crate::{
         coi::point::tests::{create_neg_cois, create_pos_cois},
         utils::SECONDS_PER_DAY_F32,
     };
 
-    use super::*;
-
     #[test]
     fn test_has_enough_cois() {
         let user_interests = UserInterests::default();

discovery_engine_core/ai/ai/src/coi/point.rs

@@ -164,7 +164,6 @@ where
 #[cfg(test)]
 pub(crate) mod tests {
     use ndarray::{arr1, FixedInitializer};
-
     use xayn_discovery_engine_test_utils::assert_approx_eq;
 
     use super::*;

discovery_engine_core/ai/ai/src/coi/stats.rs

@@ -128,10 +128,10 @@ pub(super) fn compute_coi_decay_factor(
 
 #[cfg(test)]
 mod tests {
-    use crate::coi::{config::Config, point::tests::create_pos_cois};
     use xayn_discovery_engine_test_utils::assert_approx_eq;
 
     use super::*;
+    use crate::coi::{config::Config, point::tests::create_pos_cois};
 
     #[test]
     fn test_compute_relevances_empty_cois() {

discovery_engine_core/ai/ai/src/embedding.rs

@@ -17,7 +17,6 @@ use std::ops::RangeInclusive;
 use itertools::Itertools;
 use ndarray::{Array2, ArrayBase, ArrayView1, Data, Ix1};
 use xayn_discovery_engine_bert::Embedding1;
-
 // Re-export `Embedding` specific errors.
 pub use xayn_discovery_engine_bert::MalformedBytesEmbedding;
 
@@ -109,9 +108,9 @@ pub fn cosine_similarity(a: ArrayView1<'_, f32>, b: ArrayView1<'_, f32>) -> f32
 #[cfg(test)]
 mod tests {
     use ndarray::{arr1, arr2};
+    use xayn_discovery_engine_test_utils::assert_approx_eq;
 
     use super::*;
-    use xayn_discovery_engine_test_utils::assert_approx_eq;
 
     #[test]
     fn test_l2_norm() {

discovery_engine_core/ai/ai/src/kps/key_phrase.rs

@@ -467,13 +467,12 @@ mod tests {
     use ndarray::arr2;
     use xayn_discovery_engine_test_utils::assert_approx_eq;
 
+    use super::*;
     use crate::{
         coi::{config::Config as CoiConfig, point::tests::create_pos_cois},
         kps::config::Config as KpsConfig,
     };
 
-    use super::*;
-
     impl KeyPhrases {
         pub(crate) fn new<'a, const N: usize>(
             iter: impl IntoIterator<Item = (CoiId, (&'a str, &'a str), &'a str, [f32; N])>,

discovery_engine_core/ai/ai/src/lib.rs

@@ -38,6 +38,8 @@ mod error;
 mod kps;
 pub mod utils;
 
+#[cfg(doc)]
+pub use crate::embedding::COSINE_SIMILARITY_RANGE;
 pub use crate::{
     coi::{
         config::{Config as CoiConfig, Error as CoiConfigError},
@@ -62,6 +64,3 @@ pub use crate::{
     },
     utils::{nan_safe_f32_cmp, nan_safe_f32_cmp_desc, system_time_now},
 };
-
-#[cfg(doc)]
-pub use crate::embedding::COSINE_SIMILARITY_RANGE;

discovery_engine_core/ai/ai/src/utils.rs

@@ -117,9 +117,9 @@ mod tests {
 
     use serde::{Deserialize, Serialize};
     use serde_json::{from_str, to_string};
+    use xayn_discovery_engine_test_utils::assert_approx_eq;
 
     use super::*;
-    use xayn_discovery_engine_test_utils::assert_approx_eq;
 
     #[test]
     fn test_nan_safe_f32_cmp_sorts_in_the_right_order() {

discovery_engine_core/ai/bert/benches/bert.rs

@@ -19,7 +19,6 @@ use std::path::Path;
 use criterion::{black_box, criterion_group, criterion_main, Criterion};
 use ndarray::s;
 use onnxruntime::{environment::Environment, GraphOptimizationLevel};
-
 use xayn_discovery_engine_bert::{tokenizer::Tokenizer, Config, Embedding2, NonePooler};
 use xayn_discovery_engine_test_utils::asset::smbert;
 

discovery_engine_core/ai/bert/examples/validate.rs

@@ -27,7 +27,6 @@ use csv::Reader;
 use indicatif::ProgressBar;
 use ndarray::{s, Array1, Array2, ArrayView1, Axis};
 use onnxruntime::{environment::Environment, session::Session, GraphOptimizationLevel};
-
 use xayn_discovery_engine_bert::{
     tokenizer::Tokenizer,
     Config,

discovery_engine_core/ai/test-utils/src/lib.rs

@@ -34,6 +34,7 @@ mod approx_eq;
 pub mod asset;
 pub mod uuid;
 
-pub use crate::approx_eq::ApproxEqIter;
 #[doc(hidden)] // required for standalone export of assert_approx_eq!
 pub use float_cmp::approx_eq;
+
+pub use crate::approx_eq::ApproxEqIter;

discovery_engine_core/core/src/config.rs

@@ -20,7 +20,6 @@ use figment::{
 };
 use serde::{Deserialize, Serialize};
 use tokio::sync::RwLock;
-
 use xayn_discovery_engine_ai::{CoiConfig, KpsConfig};
 use xayn_discovery_engine_providers::Market;
 

discovery_engine_core/core/src/document.rs

@@ -25,14 +25,12 @@ use serde_repr::{Deserialize_repr, Serialize_repr};
 use thiserror::Error;
 use url::Url;
 use uuid::Uuid;
-
+pub use xayn_discovery_engine_ai::Embedding;
 use xayn_discovery_engine_ai::{Document as AiDocument, DocumentId};
 use xayn_discovery_engine_providers::GenericArticle;
 
 use crate::stack::Id as StackId;
 
-pub use xayn_discovery_engine_ai::Embedding;
-
 /// Errors that could happen when constructing a [`Document`].
 #[derive(Error, Debug, DisplayDoc)]
 pub enum Error {
@@ -307,7 +305,6 @@ pub struct WeightedSource {
 #[cfg(test)]
 pub(crate) mod tests {
     use chrono::TimeZone;
-
     use xayn_discovery_engine_providers::{Rank, UrlWithDomain};
 
     use super::*;

discovery_engine_core/core/src/engine.rs

@@ -29,7 +29,6 @@ use rayon::iter::{Either, IntoParallelIterator, ParallelIterator};
 use thiserror::Error;
 use tokio::sync::RwLock;
 use tracing::{debug, error, info, instrument, Level};
-
 use xayn_discovery_engine_ai::{
     self,
     cosine_similarity,
@@ -1408,12 +1407,11 @@ pub(crate) mod tests {
         MockServer,
         ResponseTemplate,
     };
-
-    use crate::{document::tests::mock_generic_article, stack::ops::MockOps};
     use xayn_discovery_engine_providers::{Rank, UrlWithDomain};
     use xayn_discovery_engine_test_utils::asset::{smbert_mocked, smbert_quantized};
 
     use super::*;
+    use crate::{document::tests::mock_generic_article, stack::ops::MockOps};
 
     #[test]
     fn test_usize_not_to_small() {

discovery_engine_core/core/src/stack.rs

@@ -299,9 +299,8 @@ mod tests {
     use uuid::Uuid;
     use xayn_discovery_engine_test_utils::assert_approx_eq;
 
-    use crate::stack::ops::MockOps;
-
     use super::*;
+    use crate::stack::ops::MockOps;
 
     // assert that `f` returns ok if the argument contains only documents valid `stack_id`
     fn assert_valid_document<T, F>(f: F, stack_id: Id)

discovery_engine_core/core/src/stack/exploration.rs

@@ -15,6 +15,7 @@
 //! Exploration stack.
 
 use std::collections::HashSet;
+
 use uuid::uuid;
 use xayn_discovery_engine_ai::UserInterests;
 

discovery_engine_core/core/src/stack/exploration/selection.rs

@@ -184,9 +184,8 @@ mod tests {
     use xayn_discovery_engine_ai::CoiId;
     use xayn_discovery_engine_test_utils::{assert_approx_eq, uuid::mock_uuid};
 
-    use crate::document::Id;
-
     use super::*;
+    use crate::document::Id;
 
     fn new_doc() -> Document {
         Document {

discovery_engine_core/core/src/stack/filters/article.rs

@@ -12,12 +12,13 @@
 // You should have received a copy of the GNU Affero General Public License
 // along with this program.  If not, see <https://www.gnu.org/licenses/>.
 
+use xayn_discovery_engine_ai::GenericError;
+use xayn_discovery_engine_providers::GenericArticle;
+
 use crate::{
     document::{Document, HistoricDocument},
     stack::filters::DuplicateFilter,
 };
-use xayn_discovery_engine_ai::GenericError;
-use xayn_discovery_engine_providers::GenericArticle;
 
 pub(crate) trait ArticleFilter {
     fn apply(
@@ -76,12 +77,11 @@ impl SourcesFilter {
 mod tests {
     use std::{collections::HashMap, iter::FromIterator};
 
-    use crate::document::Document;
     use itertools::Itertools;
-
     use xayn_discovery_engine_providers::{NewscatcherArticle, Rank};
 
     use super::*;
+    use crate::document::Document;
 
     pub(crate) fn to_generic_article(articles: Vec<NewscatcherArticle>) -> Vec<GenericArticle> {
         articles

discovery_engine_core/core/src/stack/filters/deduplication.rs

@@ -13,6 +13,7 @@
 // along with this program.  If not, see <https://www.gnu.org/licenses/>.
 
 use std::collections::{HashMap, HashSet};
+
 use xayn_discovery_engine_providers::GenericArticle;
 
 use crate::document::{Document, HistoricDocument, NewsResource};

discovery_engine_core/core/src/stack/filters/semantic.rs

@@ -20,9 +20,8 @@ use kodama::{linkage, Dendrogram, Method};
 use ndarray::ArrayView1;
 use xayn_discovery_engine_ai::{cosine_similarity, nan_safe_f32_cmp, pairwise_cosine_similarity};
 
-use crate::document::{Document, WeightedSource};
-
 use super::source_weight;
+use crate::document::{Document, WeightedSource};
 
 /// Computes the condensed cosine similarity matrix of the documents' embeddings.
 fn condensed_cosine_similarity<'a, I>(embeddings: I) -> Vec<f32>
@@ -304,9 +303,8 @@ mod tests {
     use xayn_discovery_engine_bert::{AveragePooler, AvgBert, Config as BertConfig};
     use xayn_discovery_engine_test_utils::{assert_approx_eq, asset::smbert_quantized};
 
-    use crate::document::NewsResource;
-
     use super::*;
+    use crate::document::NewsResource;
 
     #[test]
     fn test_condensed_cosine_similarity() {

discovery_engine_core/core/src/stack/filters/source.rs

@@ -44,9 +44,10 @@ pub(crate) fn source_weight(document: &Document, sources: &[WeightedSource]) ->
 
 #[cfg(test)]
 mod tests {
+    use xayn_discovery_engine_providers::{GenericArticle, NewscatcherArticle};
+
     use super::*;
     use crate::document::{Document, WeightedSource};
-    use xayn_discovery_engine_providers::{GenericArticle, NewscatcherArticle};
 
     impl WeightedSource {
         fn new(source_domain: &str, weight: i32) -> Self {

discovery_engine_core/core/src/stack/ops.rs

@@ -17,7 +17,6 @@ use displaydoc::Display;
 #[cfg(test)]
 use mockall::automock;
 use thiserror::Error;
-
 use xayn_discovery_engine_ai::{GenericError, KeyPhrase};
 use xayn_discovery_engine_providers::{GenericArticle, Market};
 

discovery_engine_core/core/src/stack/ops/breaking.rs

@@ -27,6 +27,7 @@ use xayn_discovery_engine_providers::{
     RankLimit,
 };
 
+use super::{common::request_min_new_items, NewItemsError, Ops};
 use crate::{
     config::EndpointConfig,
     document::{Document, HistoricDocument},
@@ -36,8 +37,6 @@ use crate::{
     },
 };
 
-use super::{common::request_min_new_items, NewItemsError, Ops};
-
 /// Stack operations customized for breaking news items.
 pub struct BreakingNews {
     client: Arc<dyn HeadlinesProvider>,

discovery_engine_core/core/src/stack/ops/personalized.rs

@@ -27,6 +27,7 @@ use xayn_discovery_engine_providers::{
     SimilarSearchQuery,
 };
 
+use super::{common::request_min_new_items, NewItemsError, Ops};
 use crate::{
     config::EndpointConfig,
     document::{Document, HistoricDocument},
@@ -36,8 +37,6 @@ use crate::{
     },
 };
 
-use super::{common::request_min_new_items, NewItemsError, Ops};
-
 /// Stack operations customized for personalized news items.
 pub struct PersonalizedNews {
     client: Arc<dyn SimilarSearchProvider>,

discovery_engine_core/core/src/stack/ops/trusted.rs

@@ -27,6 +27,7 @@ use xayn_discovery_engine_providers::{
     TrustedHeadlinesQuery,
 };
 
+use super::{common::request_min_new_items, NewItemsError, Ops};
 use crate::{
     config::EndpointConfig,
     document::{Document, HistoricDocument},
@@ -36,8 +37,6 @@ use crate::{
     },
 };
 
-use super::{common::request_min_new_items, NewItemsError, Ops};
-
 /// Stack operations customized for trusted news.
 pub struct TrustedNews {
     client: Arc<dyn TrustedHeadlinesProvider>,

discovery_engine_core/core/src/state.rs

@@ -105,16 +105,18 @@ impl Engine {
 }
 
 mod naive_date_time_migration {
-    use crate::{
-        document::{Document, Id, NewsResource, UserReaction},
-        stack::{Data, Id as StackId},
-    };
+    use std::collections::HashMap;
+
     use chrono::{DateTime, NaiveDateTime, Utc};
     use serde::{Deserialize, Serialize};
-    use std::collections::HashMap;
     use url::Url;
     use xayn_discovery_engine_ai::Embedding;
 
+    use crate::{
+        document::{Document, Id, NewsResource, UserReaction},
+        stack::{Data, Id as StackId},
+    };
+
     #[derive(Serialize, Deserialize)]
     pub(super) struct DataWithNaiveDateTime {
         pub(super) alpha: f32,

discovery_engine_core/core/src/storage.rs

@@ -25,14 +25,13 @@ use async_trait::async_trait;
 use chrono::{DateTime, Utc};
 use displaydoc::Display;
 use thiserror::Error;
-use xayn_discovery_engine_ai::{GenericError, MalformedBytesEmbedding};
+use xayn_discovery_engine_ai::{Embedding, GenericError, MalformedBytesEmbedding};
 
 use crate::{
     document::{self, HistoricDocument, NewsResource, UserReaction, ViewMode, WeightedSource},
     stack,
     storage::models::{ApiDocumentView, NewDocument, Search, TimeSpentDocumentView},
 };
-use xayn_discovery_engine_ai::Embedding;
 
 pub(crate) type BoxedStorage = Box<dyn Storage + Send + Sync>;