adapt meilisearch-http to the new schemaless option

meilisearch-http/src/data.rs

@@ -84,13 +84,13 @@ impl DataInner {
         let mut fields_frequency = HashMap::<_, usize>::new();
         for result in all_documents_fields {
             let (_, attr, _) = result?;
-            *fields_frequency.entry(attr).or_default() += 1;
+            *fields_frequency.entry(schema.indexed_pos_to_field_id(attr).unwrap()).or_default() += 1;
         }
 
         // convert attributes to their names
         let frequency: HashMap<_, _> = fields_frequency
             .into_iter()
-            .map(|(a, c)| (schema.attribute_name(a).to_owned(), c))
+            .map(|(a, c)| (schema.get_name(a).unwrap(), c))
             .collect();
 
         index

meilisearch-http/src/helpers/meilisearch.rs

@@ -1,11 +1,11 @@
-use crate::routes::setting::{RankingOrdering, Setting};
 use indexmap::IndexMap;
-use log::{error, warn};
+use log::error;
 use meilisearch_core::criterion::*;
 use meilisearch_core::Highlight;
 use meilisearch_core::{Index, RankedMap};
 use meilisearch_core::MainT;
-use meilisearch_schema::{Schema, SchemaAttr};
+use meilisearch_core::settings::RankingRule;
+use meilisearch_schema::{Schema, FieldId};
 use serde::{Deserialize, Serialize};
 use serde_json::Value;
 use std::cmp::Ordering;
@@ -172,7 +172,7 @@ impl<'a> SearchBuilder<'a> {
                     let ref_index = &self.index;
                     let value = value.trim().to_lowercase();
 
-                    let attr = match schema.attribute(attr) {
+                    let attr = match schema.get_id(attr) {
                         Some(attr) => attr,
                         None => return Err(Error::UnknownFilteredAttribute),
                     };
@@ -274,75 +274,24 @@ impl<'a> SearchBuilder<'a> {
         ranked_map: &'a RankedMap,
         schema: &Schema,
     ) -> Result<Option<Criteria<'a>>, Error> {
-        let current_settings = match self.index.main.customs(reader).unwrap() {
-            Some(bytes) => bincode::deserialize(bytes).unwrap(),
-            None => Setting::default(),
-        };
-
-        let ranking_rules = &current_settings.ranking_rules;
-        let ranking_order = &current_settings.ranking_order;
+        let ranking_rules = self.index.main.ranking_rules(reader).unwrap();
 
         if let Some(ranking_rules) = ranking_rules {
             let mut builder = CriteriaBuilder::with_capacity(7 + ranking_rules.len());
-            if let Some(ranking_rules_order) = ranking_order {
-                for rule in ranking_rules_order {
-                    match rule.as_str() {
-                        "_typo" => builder.push(Typo),
-                        "_words" => builder.push(Words),
-                        "_proximity" => builder.push(Proximity),
-                        "_attribute" => builder.push(Attribute),
-                        "_words_position" => builder.push(WordsPosition),
-                        "_exact" => builder.push(Exact),
-                        _ => {
-                            let order = match ranking_rules.get(rule.as_str()) {
-                                Some(o) => o,
-                                None => continue,
-                            };
-
-                            let custom_ranking = match order {
-                                RankingOrdering::Asc => {
-                                    SortByAttr::lower_is_better(&ranked_map, &schema, &rule)
-                                        .unwrap()
-                                }
-                                RankingOrdering::Dsc => {
-                                    SortByAttr::higher_is_better(&ranked_map, &schema, &rule)
-                                        .unwrap()
-                                }
-                            };
-
-                            builder.push(custom_ranking);
-                        }
-                    }
-                }
-                builder.push(DocumentId);
-                return Ok(Some(builder.build()));
-            } else {
-                builder.push(Typo);
-                builder.push(Words);
-                builder.push(Proximity);
-                builder.push(Attribute);
-                builder.push(WordsPosition);
-                builder.push(Exact);
-                for (rule, order) in ranking_rules.iter() {
-                    let custom_ranking = match order {
-                        RankingOrdering::Asc => {
-                            SortByAttr::lower_is_better(&ranked_map, &schema, &rule)
-                        }
-                        RankingOrdering::Dsc => {
-                            SortByAttr::higher_is_better(&ranked_map, &schema, &rule)
-                        }
-                    };
-                    if let Ok(custom_ranking) = custom_ranking {
-                        builder.push(custom_ranking);
-                    } else {
-                        // TODO push this warning to a log tree
-                        warn!("Custom ranking cannot be added; Attribute {} not registered for ranking", rule)
-                    }
-
-                }
-                builder.push(DocumentId);
-                return Ok(Some(builder.build()));
+            for rule in ranking_rules {
+                match rule {
+                    RankingRule::Typo => builder.push(Typo),
+                    RankingRule::Words => builder.push(Words),
+                    RankingRule::Proximity => builder.push(Proximity),
+                    RankingRule::Attribute => builder.push(Attribute),
+                    RankingRule::WordsPosition => builder.push(WordsPosition),
+                    RankingRule::Exact => builder.push(Exact),
+                    RankingRule::Asc(field) => builder.push(SortByAttr::lower_is_better(&ranked_map, &schema, &field).unwrap()),
+                    RankingRule::Dsc(field) => builder.push(SortByAttr::higher_is_better(&ranked_map, &schema, &field).unwrap()),
+                };
             }
+            builder.push(DocumentId);
+            return Ok(Some(builder.build()));
         }
 
         Ok(None)
@@ -421,14 +370,14 @@ fn crop_document(
     matches.sort_unstable_by_key(|m| (m.char_index, m.char_length));
 
     for (field, length) in fields {
-        let attribute = match schema.attribute(field) {
+        let attribute = match schema.get_id(field) {
             Some(attribute) => attribute,
             None => continue,
         };
 
         let selected_matches = matches
             .iter()
-            .filter(|m| SchemaAttr::new(m.attribute) == attribute)
+            .filter(|m| FieldId::new(m.attribute) == attribute)
             .cloned();
 
         if let Some(Value::String(ref mut original_text)) = document.get_mut(field) {
@@ -437,7 +386,7 @@ fn crop_document(
 
             *original_text = cropped_text;
 
-            matches.retain(|m| SchemaAttr::new(m.attribute) != attribute);
+            matches.retain(|m| FieldId::new(m.attribute) != attribute);
             matches.extend_from_slice(&cropped_matches);
         }
     }
@@ -450,26 +399,25 @@ fn calculate_matches(
 ) -> MatchesInfos {
     let mut matches_result: HashMap<String, Vec<MatchPosition>> = HashMap::new();
     for m in matches.iter() {
-        let attribute = schema
-            .attribute_name(SchemaAttr::new(m.attribute))
-            .to_string();
-        if let Some(attributes_to_retrieve) = attributes_to_retrieve.clone() {
-            if !attributes_to_retrieve.contains(attribute.as_str()) {
-                continue;
+        if let Some(attribute) = schema.get_name(FieldId::new(m.attribute)) {
+            if let Some(attributes_to_retrieve) = attributes_to_retrieve.clone() {
+                if !attributes_to_retrieve.contains(attribute.as_str()) {
+                    continue;
+                }
+            };
+            if let Some(pos) = matches_result.get_mut(&attribute) {
+                pos.push(MatchPosition {
+                    start: m.char_index as usize,
+                    length: m.char_length as usize,
+                });
+            } else {
+                let mut positions = Vec::new();
+                positions.push(MatchPosition {
+                    start: m.char_index as usize,
+                    length: m.char_length as usize,
+                });
+                matches_result.insert(attribute, positions);
             }
-        };
-        if let Some(pos) = matches_result.get_mut(&attribute) {
-            pos.push(MatchPosition {
-                start: m.char_index as usize,
-                length: m.char_length as usize,
-            });
-        } else {
-            let mut positions = Vec::new();
-            positions.push(MatchPosition {
-                start: m.char_index as usize,
-                length: m.char_length as usize,
-            });
-            matches_result.insert(attribute, positions);
         }
     }
     for (_, val) in matches_result.iter_mut() {

meilisearch-http/src/models/mod.rs

@@ -1,3 +1,2 @@
-pub mod schema;
 pub mod token;
 pub mod update_operation;

meilisearch-http/src/models/schema.rs

@@ -1,118 +0,0 @@
-use std::collections::HashSet;
-
-use indexmap::IndexMap;
-use meilisearch_schema::{Schema, SchemaBuilder, SchemaProps};
-use serde::{Deserialize, Serialize};
-
-#[derive(Debug, Clone, Eq, PartialEq, Hash, Serialize, Deserialize)]
-#[serde(rename_all = "camelCase")]
-pub enum FieldProperties {
-    Identifier,
-    Indexed,
-    Displayed,
-    Ranked,
-}
-
-#[derive(Debug, Clone, Eq, PartialEq, Serialize, Deserialize)]
-pub struct SchemaBody(IndexMap<String, HashSet<FieldProperties>>);
-
-impl From<Schema> for SchemaBody {
-    fn from(value: Schema) -> SchemaBody {
-        let mut map = IndexMap::new();
-        for (name, _attr, props) in value.iter() {
-            let old_properties = map.entry(name.to_owned()).or_insert(HashSet::new());
-            if props.is_indexed() {
-                old_properties.insert(FieldProperties::Indexed);
-            }
-            if props.is_displayed() {
-                old_properties.insert(FieldProperties::Displayed);
-            }
-            if props.is_ranked() {
-                old_properties.insert(FieldProperties::Ranked);
-            }
-        }
-        let old_properties = map
-            .entry(value.identifier_name().to_string())
-            .or_insert(HashSet::new());
-        old_properties.insert(FieldProperties::Identifier);
-        old_properties.insert(FieldProperties::Displayed);
-        SchemaBody(map)
-    }
-}
-
-impl Into<Schema> for SchemaBody {
-    fn into(self) -> Schema {
-        let mut identifier = "documentId".to_string();
-        let mut attributes = IndexMap::new();
-        for (field, properties) in self.0 {
-            let mut indexed = false;
-            let mut displayed = false;
-            let mut ranked = false;
-            for property in properties {
-                match property {
-                    FieldProperties::Indexed => indexed = true,
-                    FieldProperties::Displayed => displayed = true,
-                    FieldProperties::Ranked => ranked = true,
-                    FieldProperties::Identifier => identifier = field.clone(),
-                }
-            }
-            attributes.insert(
-                field,
-                SchemaProps {
-                    indexed,
-                    displayed,
-                    ranked,
-                },
-            );
-        }
-
-        let mut builder = SchemaBuilder::with_identifier(identifier);
-        for (field, props) in attributes {
-            builder.new_attribute(field, props);
-        }
-        builder.build()
-    }
-}
-
-#[cfg(test)]
-mod tests {
-    use super::*;
-
-    #[test]
-    fn test_schema_body_conversion() {
-        let schema_body = r#"
-        {
-            "id": ["identifier", "indexed", "displayed"],
-            "title": ["indexed", "displayed"],
-            "date": ["displayed"]
-        }
-        "#;
-
-        let schema_builder = r#"
-        {
-            "identifier": "id",
-            "attributes": {
-                "id": {
-                    "indexed": true,
-                    "displayed": true
-                },
-                "title": {
-                    "indexed": true,
-                    "displayed": true
-                },
-                "date": {
-                    "displayed": true
-                }
-            }
-        }
-        "#;
-
-        let schema_body: SchemaBody = serde_json::from_str(schema_body).unwrap();
-        let schema_builder: SchemaBuilder = serde_json::from_str(schema_builder).unwrap();
-
-        let schema_from_body: Schema = schema_body.into();
-        let schema_from_builder: Schema = schema_builder.build();
-
-        assert_eq!(schema_from_body, schema_from_builder);
-    }
-}

meilisearch-http/src/routes/document.rs

@@ -1,3 +1,4 @@
+
 use std::collections::{BTreeSet, HashSet};
 
 use http::StatusCode;
@@ -7,6 +8,7 @@ use serde_json::Value;
 use tide::querystring::ContextExt as QSContextExt;
 use tide::response::IntoResponse;
 use tide::{Context, Response};
+use meilisearch_core::settings::Settings;
 
 use crate::error::{ResponseError, SResult};
 use crate::helpers::tide::ContextExt;
@@ -117,27 +119,13 @@ pub async fn get_all_documents(ctx: Context<Data>) -> SResult<Response> {
     Ok(tide::response::json(response_body))
 }
 
-fn infered_schema(document: &IndexMap<String, Value>, identifier: Option<String>) -> Option<meilisearch_schema::Schema> {
-    use meilisearch_schema::{SchemaBuilder, DISPLAYED, INDEXED};
-
-    let mut identifier = identifier;
+fn find_identifier(document: &IndexMap<String, Value>) -> Option<String> {
     for key in document.keys() {
-        if identifier.is_none() && key.to_lowercase().contains("id") {
-            identifier = Some(key.to_string());
-            break;
+        if key.to_lowercase().contains("id") {
+            return Some(key.to_string())
         }
     }
-
-    match identifier {
-        Some(identifier) => {
-            let mut builder = SchemaBuilder::with_identifier(identifier);
-            for key in document.keys() {
-                builder.new_attribute(key, DISPLAYED | INDEXED);
-            }
-            Some(builder.build())
-        }
-        None => None,
-    }
+    return None
 }
 
 #[derive(Default, Deserialize)]
@@ -165,14 +153,22 @@ async fn update_multiple_documents(mut ctx: Context<Data>, is_partial: bool) ->
         .schema(&reader)
         .map_err(ResponseError::internal)?;
     if current_schema.is_none() {
-        match data.first().and_then(|docs| infered_schema(docs, query.identifier)) {
-            Some(schema) => {
-                index
-                    .schema_update(&mut update_writer, schema)
-                    .map_err(ResponseError::internal)?;
+        let id = match query.identifier {
+            Some(id) => id,
+            None => {
+                match data.first().and_then(|docs| find_identifier(docs)) {
+                    Some(id) => id,
+                    None => return Err(ResponseError::bad_request("Could not infer a schema")),
+                }
             }
-            None => return Err(ResponseError::bad_request("Could not infer a schema")),
-        }
+        };
+        let settings = Settings {
+            attribute_identifier: Some(id),
+            ..Settings::default()
+        };
+        index
+            .settings_update(&mut update_writer, settings.into())
+            .map_err(ResponseError::internal)?;
     }
 
     let mut document_addition = if is_partial {

meilisearch-http/src/routes/index.rs

@@ -2,19 +2,16 @@ use chrono::{DateTime, Utc};
 use http::StatusCode;
 use log::error;
 use meilisearch_core::ProcessedUpdateResult;
-use meilisearch_schema::{Schema, SchemaBuilder};
+// use meilisearch_schema::Schema;
 use rand::seq::SliceRandom;
 use serde::{Deserialize, Serialize};
 use serde_json::json;
-use tide::querystring::ContextExt as QSContextExt;
 use tide::response::IntoResponse;
 use tide::{Context, Response};
 
 use crate::error::{ResponseError, SResult};
 use crate::helpers::tide::ContextExt;
-use crate::models::schema::SchemaBody;
 use crate::models::token::ACL::*;
-use crate::routes::document::IndexUpdateResponse;
 use crate::Data;
 
 fn generate_uid() -> String {
@@ -124,7 +121,7 @@ pub async fn get_index(ctx: Context<Data>) -> SResult<Response> {
 struct IndexCreateRequest {
     name: String,
     uid: Option<String>,
-    schema: Option<SchemaBody>,
+    // schema: Option<SchemaBody>,
 }
 
 #[derive(Debug, Serialize)]
@@ -132,9 +129,9 @@ struct IndexCreateRequest {
 struct IndexCreateResponse {
     name: String,
     uid: String,
-    schema: Option<SchemaBody>,
-    #[serde(skip_serializing_if = "Option::is_none")]
-    update_id: Option<u64>,
+    // schema: Option<SchemaBody>,
+    // #[serde(skip_serializing_if = "Option::is_none")]
+    // update_id: Option<u64>,
     created_at: DateTime<Utc>,
     updated_at: DateTime<Utc>,
 }
@@ -165,30 +162,29 @@ pub async fn create_index(mut ctx: Context<Data>) -> SResult<Response> {
     };
 
     let mut writer = db.main_write_txn().map_err(ResponseError::internal)?;
-    let mut update_writer = db.update_write_txn().map_err(ResponseError::internal)?;
 
     created_index
         .main
         .put_name(&mut writer, &body.name)
         .map_err(ResponseError::internal)?;
 
-    let schema: Option<Schema> = body.schema.clone().map(Into::into);
-    let mut response_update_id = None;
-    if let Some(schema) = schema {
-        let update_id = created_index
-            .schema_update(&mut update_writer, schema)
-            .map_err(ResponseError::internal)?;
-        response_update_id = Some(update_id)
-    }
+    // let schema: Option<Schema> = body.schema.clone().map(Into::into);
+    // let mut response_update_id = None;
+    // if let Some(schema) = schema {
+    //     let update_id = created_index
+    //         .schema_update(&mut update_writer, schema)
+    //         .map_err(ResponseError::internal)?;
+    //     response_update_id = Some(update_id)
+    // }
 
-    writer.commit().map_err(ResponseError::internal)?;
-    update_writer.commit().map_err(ResponseError::internal)?;
+    // writer.commit().map_err(ResponseError::internal)?;
+    // update_writer.commit().map_err(ResponseError::internal)?;
 
     let response_body = IndexCreateResponse {
         name: body.name,
         uid,
-        schema: body.schema,
-        update_id: response_update_id,
+        // schema: body.schema,
+        // update_id: update_id,
         created_at: Utc::now(),
         updated_at: Utc::now(),
     };
@@ -263,78 +259,6 @@ pub async fn update_index(mut ctx: Context<Data>) -> SResult<Response> {
         .into_response())
 }
 
-#[derive(Default, Deserialize)]
-#[serde(rename_all = "camelCase", deny_unknown_fields)]
-struct SchemaParams {
-    raw: bool,
-}
-
-pub async fn get_index_schema(ctx: Context<Data>) -> SResult<Response> {
-    ctx.is_allowed(IndexesRead)?;
-
-    let index = ctx.index()?;
-
-    // Tide doesn't support "no query param"
-    let params: SchemaParams = ctx.url_query().unwrap_or_default();
-
-    let db = &ctx.state().db;
-    let reader = db.main_read_txn().map_err(ResponseError::internal)?;
-
-    let schema = index
-        .main
-        .schema(&reader)
-        .map_err(ResponseError::open_index)?;
-
-    match schema {
-        Some(schema) => {
-            if params.raw {
-                Ok(tide::response::json(schema))
-            } else {
-                Ok(tide::response::json(SchemaBody::from(schema)))
-            }
-        }
-        None => Err(ResponseError::not_found("missing index schema")),
-    }
-}
-
-pub async fn update_schema(mut ctx: Context<Data>) -> SResult<Response> {
-    ctx.is_allowed(IndexesWrite)?;
-
-    let index_uid = ctx.url_param("index")?;
-
-    let params: SchemaParams = ctx.url_query().unwrap_or_default();
-
-    let schema = if params.raw {
-        ctx.body_json::<SchemaBuilder>()
-            .await
-            .map_err(ResponseError::bad_request)?
-            .build()
-    } else {
-        ctx.body_json::<SchemaBody>()
-            .await
-            .map_err(ResponseError::bad_request)?
-            .into()
-    };
-
-    let db = &ctx.state().db;
-    let mut writer = db.update_write_txn().map_err(ResponseError::internal)?;
-
-    let index = db
-        .open_index(&index_uid)
-        .ok_or(ResponseError::index_not_found(index_uid))?;
-
-    let update_id = index
-        .schema_update(&mut writer, schema.clone())
-        .map_err(ResponseError::internal)?;
-
-    writer.commit().map_err(ResponseError::internal)?;
-
-    let response_body = IndexUpdateResponse { update_id };
-    Ok(tide::response::json(response_body)
-        .with_status(StatusCode::ACCEPTED)
-        .into_response())
-}
-
 pub async fn get_update_status(ctx: Context<Data>) -> SResult<Response> {
     ctx.is_allowed(IndexesRead)?;
 

meilisearch-http/src/routes/mod.rs

@@ -51,10 +51,10 @@ pub fn load_routes(app: &mut tide::App<Data>) {
                     .put(index::update_index)
                     .delete(index::delete_index);
 
-                router
-                    .at("/schema")
-                    .get(index::get_index_schema)
-                    .put(index::update_schema);
+                // router
+                //     .at("/schema")
+                //     .get(index::get_index_schema)
+                //     .put(index::update_schema);
 
                 router.at("/documents").nest(|router| {
                     router

meilisearch-http/src/routes/search.rs

@@ -64,8 +64,9 @@ pub async fn search_with_url_query(ctx: Context<Data>) -> SResult<Response> {
         let crop_length = query.crop_length.unwrap_or(200);
         if attributes_to_crop == "*" {
             let attributes_to_crop = schema
+                .get_displayed_name()
                 .iter()
-                .map(|(attr, ..)| (attr.to_string(), crop_length))
+                .map(|attr| (attr.to_string(), crop_length))
                 .collect();
             search_builder.attributes_to_crop(attributes_to_crop);
         } else {
@@ -79,7 +80,7 @@ pub async fn search_with_url_query(ctx: Context<Data>) -> SResult<Response> {
 
     if let Some(attributes_to_highlight) = query.attributes_to_highlight {
         let attributes_to_highlight = if attributes_to_highlight == "*" {
-            schema.iter().map(|(attr, ..)| attr.to_string()).collect()
+            schema.get_displayed_name()
         } else {
             attributes_to_highlight
                 .split(',')