Add coerce and output format options for numeric fields

Author: guilload

docs/configuration/index-config.md

@@ -164,7 +164,7 @@ Indexing with position is required to run phrase queries.
 
 Quickwit handles three numeric types: `i64`, `u64`, and `f64`.
 
-Numeric values can be stored in a fast field (the equivalent of Lucene's `DocValues`) which is a column-oriented storage.
+Numeric values can be stored in a fast field (the equivalent of Lucene's `DocValues`), which is a column-oriented storage used for range queries and aggregations.
 
 Example of a mapping for an u64 field:
 
@@ -179,12 +179,14 @@ fast: true
 
 **Parameters for i64, u64 and f64 field**
 
-| Variable      | Description   | Default value |
-| ------------- | ------------- | ------------- |
-| `description` | Optional description for the field. | `None` |
-| `stored`    | Whether the field values are stored in the document store | `true` |
-| `indexed`   | Whether the field values are indexed | `true` |
-| `fast`      | Whether the field values are stored in a fast field | `false` |
+| Variable        | Description   | Default value |
+| --------------- | ------------- | ------------- |
+| `description`   | Optional description for the field. | `None` |
+| `stored`        | Whether the field values are stored in the document store. | `true` |
+| `indexed`       | Whether the field values are indexed. | `true` |
+| `fast`          | Whether the field values are stored in a fast field. | `false` |
+| `coerce`        | Whether to convert numbers passed as strings to integers or floats. | `true` |
+| `output_format` | JSON type used to return numbers in search results. Possible values are `number` or `string`. | `number` |
 
 #### `datetime` type
 

quickwit/quickwit-doc-mapper/Cargo.toml

@@ -38,11 +38,12 @@ quickwit-query = { workspace = true }
 criterion = { workspace = true }
 matches = { workspace = true }
 proptest = { workspace = true }
-quickwit-proto = { workspace = true }
-quickwit-query = { workspace = true, features = ["testsuite"] }
 serde_yaml = { workspace = true }
 time = { workspace = true }
 
+quickwit-proto = { workspace = true }
+quickwit-query = { workspace = true, features = ["testsuite"] }
+
 [features]
 multilang = ["quickwit-query/multilang"]
 testsuite = ["multilang"]

quickwit/quickwit-doc-mapper/src/default_doc_mapper/field_mapping_entry.rs

@@ -95,9 +95,40 @@ pub struct QuickwitNumericOptions {
     pub indexed: bool,
     #[serde(default)]
     pub fast: bool,
+    #[serde(default = "default_as_true")]
+    pub coerce: bool,
+    #[serde(default)]
+    pub output_format: NumericOutputFormat,
 }
 
 impl Default for QuickwitNumericOptions {
+    fn default() -> Self {
+        Self {
+            description: None,
+            indexed: true,
+            stored: true,
+            fast: false,
+            coerce: true,
+            output_format: NumericOutputFormat::default(),
+        }
+    }
+}
+
+#[derive(Clone, Serialize, Deserialize, Debug, PartialEq, utoipa::ToSchema)]
+#[serde(deny_unknown_fields)]
+pub struct QuickwitBoolOptions {
+    #[serde(default)]
+    #[serde(skip_serializing_if = "Option::is_none")]
+    pub description: Option<String>,
+    #[serde(default = "default_as_true")]
+    pub stored: bool,
+    #[serde(default = "default_as_true")]
+    pub indexed: bool,
+    #[serde(default)]
+    pub fast: bool,
+}
+
+impl Default for QuickwitBoolOptions {
     fn default() -> Self {
         Self {
             description: None,
@@ -150,15 +181,15 @@ pub enum BinaryFormat {
 impl BinaryFormat {
     pub fn as_str(&self) -> &str {
         match self {
-            BinaryFormat::Base64 => "base64",
-            BinaryFormat::Hex => "hex",
+            Self::Base64 => "base64",
+            Self::Hex => "hex",
         }
     }
 
     pub fn format_to_json(&self, value: &[u8]) -> JsonValue {
         match self {
-            BinaryFormat::Base64 => BASE64_STANDARD.encode(value).into(),
-            BinaryFormat::Hex => hex::encode(value).into(),
+            Self::Base64 => BASE64_STANDARD.encode(value).into(),
+            Self::Hex => hex::encode(value).into(),
         }
     }
 
@@ -172,21 +203,27 @@ impl BinaryFormat {
             ));
         };
         let payload = match self {
-            BinaryFormat::Base64 => {
-                BASE64_STANDARD
-                    .decode(&byte_str)
-                    .map_err(|base64_decode_err| {
-                        format!("Expected base64 string, got `{byte_str}`: {base64_decode_err}")
-                    })?
-            }
-            BinaryFormat::Hex => hex::decode(&byte_str).map_err(|hex_decode_err| {
+            Self::Base64 => BASE64_STANDARD
+                .decode(&byte_str)
+                .map_err(|base64_decode_err| {
+                    format!("Expected base64 string, got `{byte_str}`: {base64_decode_err}")
+                })?,
+            Self::Hex => hex::decode(&byte_str).map_err(|hex_decode_err| {
                 format!("Expected hex string, got `{byte_str}`: {hex_decode_err}")
             })?,
         };
         Ok(TantivyValue::Bytes(payload))
     }
 }
 
+#[derive(Clone, Copy, Debug, Eq, PartialEq, Hash, Default, Serialize, Deserialize)]
+#[serde(rename_all = "snake_case")]
+pub enum NumericOutputFormat {
+    #[default]
+    Number,
+    String,
+}
+
 #[derive(Clone, Debug, Serialize, Deserialize, PartialEq, utoipa::ToSchema)]
 #[serde(deny_unknown_fields)]
 pub struct QuickwitIpAddrOptions {
@@ -618,8 +655,8 @@ fn deserialize_mapping_type(
             Ok(FieldMappingType::F64(numeric_options, cardinality))
         }
         Type::Bool => {
-            let numeric_options: QuickwitNumericOptions = serde_json::from_value(json)?;
-            Ok(FieldMappingType::Bool(numeric_options, cardinality))
+            let bool_options: QuickwitBoolOptions = serde_json::from_value(json)?;
+            Ok(FieldMappingType::Bool(bool_options, cardinality))
         }
         Type::IpAddr => {
             let ip_addr_options: QuickwitIpAddrOptions = serde_json::from_value(json)?;
@@ -685,8 +722,8 @@ fn typed_mapping_to_json_params(
         FieldMappingType::Text(text_options, _) => serialize_to_map(&text_options),
         FieldMappingType::U64(options, _)
         | FieldMappingType::I64(options, _)
-        | FieldMappingType::F64(options, _)
-        | FieldMappingType::Bool(options, _) => serialize_to_map(&options),
+        | FieldMappingType::F64(options, _) => serialize_to_map(&options),
+        FieldMappingType::Bool(options, _) => serialize_to_map(&options),
         FieldMappingType::Bytes(options, _) => serialize_to_map(&options),
         FieldMappingType::IpAddr(options, _) => serialize_to_map(&options),
         FieldMappingType::DateTime(date_time_options, _) => serialize_to_map(&date_time_options),
@@ -1062,20 +1099,21 @@ mod tests {
 
     #[test]
     fn test_deserialize_i64_parsing_error_with_text_options() {
-        let result = serde_json::from_str::<FieldMappingEntry>(
+        let error = serde_json::from_str::<FieldMappingEntry>(
             r#"
             {
                 "name": "my_field_name",
                 "type": "i64",
                 "tokenizer": "basic"
             }
             "#,
-        );
-        let error = result.unwrap_err();
+        )
+        .unwrap_err();
+
         assert_eq!(
             error.to_string(),
             "Error while parsing field `my_field_name`: unknown field `tokenizer`, expected one \
-             of `description`, `stored`, `indexed`, `fast`"
+             of `description`, `stored`, `indexed`, `fast`, `coerce`, `output_format`"
         );
     }
 
@@ -1146,6 +1184,8 @@ mod tests {
                 "stored": true,
                 "fast": false,
                 "indexed": true,
+                "coerce": true,
+                "output_format": "number"
             })
         );
         Ok(())
@@ -1165,7 +1205,7 @@ mod tests {
             .unwrap_err()
             .to_string(),
             "Error while parsing field `my_field_name`: unknown field `tokenizer`, expected one \
-             of `description`, `stored`, `indexed`, `fast`"
+             of `description`, `stored`, `indexed`, `fast`, `coerce`, `output_format`"
         );
     }
 
@@ -1232,6 +1272,8 @@ mod tests {
                 "stored": true,
                 "fast": false,
                 "indexed": true,
+                "coerce": true,
+                "output_format": "number"
             })
         );
     }
@@ -1256,6 +1298,8 @@ mod tests {
                 "stored": true,
                 "fast": false,
                 "indexed": true,
+                "coerce": true,
+                "output_format": "number"
             })
         );
     }
@@ -1616,6 +1660,8 @@ mod tests {
                 "stored": true,
                 "fast": false,
                 "indexed": true,
+                "coerce": true,
+                "output_format": "number"
             })
         );
     }

quickwit/quickwit-doc-mapper/src/default_doc_mapper/field_mapping_type.rs

@@ -20,6 +20,7 @@
 use tantivy::schema::Type;
 
 use super::date_time_type::QuickwitDateTimeOptions;
+use super::field_mapping_entry::QuickwitBoolOptions;
 use crate::default_doc_mapper::field_mapping_entry::{
     QuickwitBytesOptions, QuickwitIpAddrOptions, QuickwitJsonOptions, QuickwitNumericOptions,
     QuickwitObjectOptions, QuickwitTextOptions,
@@ -41,7 +42,7 @@ pub(crate) enum FieldMappingType {
     /// 64-bit float mapping type configuration.
     F64(QuickwitNumericOptions, Cardinality),
     /// Bool mapping type configuration.
-    Bool(QuickwitNumericOptions, Cardinality),
+    Bool(QuickwitBoolOptions, Cardinality),
     /// IP Address mapping type configuration.
     IpAddr(QuickwitIpAddrOptions, Cardinality),
     /// Bytes mapping type configuration.