Merge pull request #124 from snipsco/fix/parser-group-issues

Release 0.61.1 - Fix bug with regex patterns containing duplicated slot names
snipsco · Mar 1, 2019 · c8025bc · c8025bc
2 parents 3f5a57a + 1a12a6e
commit c8025bc
Show file tree

Hide file tree

Showing 10 changed files with 110 additions and 25 deletions.
diff --git a/CHANGELOG.md b/CHANGELOG.md
@@ -1,6 +1,10 @@
 # Changelog
 All notable changes to this project will be documented in this file.
 
+## [0.64.1] - 2019-03-01
+### Fixed
+- Fix bug with regex patterns containing duplicated slot names
+
 ## [0.64.0] - 2019-02-28
 ### Changed
 - Bumped `snips-nlu-ontology` to 0.64.4
@@ -174,6 +178,7 @@ being statically hardcoded, reducing the binary size by 31Mb.
 - Improve support for japanese
 - Rename python package to `snips_nlu_rust`
 
+[0.64.1]: https://github.com/snipsco/snips-nlu-rs/compare/0.64.0...0.64.1
 [0.64.0]: https://github.com/snipsco/snips-nlu-rs/compare/0.63.1...0.64.0
 [0.63.1]: https://github.com/snipsco/snips-nlu-rs/compare/0.63.0...0.63.1
 [0.63.0]: https://github.com/snipsco/snips-nlu-rs/compare/0.62.0...0.63.0

diff --git a/Cargo.toml b/Cargo.toml
@@ -1,6 +1,6 @@
 [package]
 name = "snips-nlu-lib"
-version = "0.64.0"
+version = "0.64.1"
 authors = [
     "Adrien Ball <[email protected]>",
     "Clement Doumouro <[email protected]>",

diff --git a/ffi/Cargo.toml b/ffi/Cargo.toml
@@ -1,6 +1,6 @@
 [package]
 name = "snips-nlu-ffi"
-version = "0.64.0"
+version = "0.64.1"
 edition = "2018"
 authors = [
     "Adrien Ball <[email protected]>",

diff --git a/platforms/c/libsnips_nlu.h b/platforms/c/libsnips_nlu.h
@@ -5,7 +5,7 @@
 #include <stdlib.h>
 #include <stdbool.h>
 
-#define SNIPS_NLU_VERSION "0.64.0"
+#define SNIPS_NLU_VERSION "0.64.1"
 
 /*
  * Enum representing the grain of a resolved date related value

diff --git a/platforms/kotlin/build.gradle b/platforms/kotlin/build.gradle
@@ -11,7 +11,7 @@ buildscript {
 
 apply plugin: 'kotlin'
 
-version = "0.64.0"
+version = "0.64.1"
 group = "ai.snips"
 
 repositories {

diff --git a/platforms/python/ffi/Cargo.toml b/platforms/python/ffi/Cargo.toml
@@ -1,6 +1,6 @@
 [package]
 name = "snips-nlu-python-ffi"
-version = "0.64.0"
+version = "0.64.1"
 authors = ["Adrien Ball <[email protected]>"]
 edition = "2018"
 
@@ -11,4 +11,4 @@ crate-type = ["cdylib"]
 [dependencies]
 libc = "0.2"
 ffi-utils = { git = "https://github.com/snipsco/snips-utils-rs", rev = "4292ad9" }
-snips-nlu-ffi = { git = "https://github.com/snipsco/snips-nlu-rs", tag = "0.64.0" }
+snips-nlu-ffi = { git = "https://github.com/snipsco/snips-nlu-rs", tag = "0.64.1" }
diff --git a/platforms/python/snips_nlu_rust/__version__ b/platforms/python/snips_nlu_rust/__version__
@@ -1 +1 @@
-0.64.0
+0.64.1
diff --git a/platforms/swift/SnipsNlu/Dependencies/build.sh b/platforms/swift/SnipsNlu/Dependencies/build.sh
@@ -4,7 +4,7 @@
 
 set -e
 
-VERSION="0.64.0"
+VERSION="0.64.1"
 SYSTEM=$(echo $1 | tr '[:upper:]' '[:lower:]')
 LIBRARY_NAME=libsnips_nlu_ffi
 LIBRARY_NAME_A=${LIBRARY_NAME}.a

diff --git a/post_release.sh b/post_release.sh
@@ -0,0 +1,15 @@
+#!/bin/sh
+
+VERSION=$1
+
+if [[ -z "$VERSION" ]]
+then
+    echo "Usage: $0 <version>"
+    exit 1
+fi
+
+set -ex
+
+./update_version.sh ${VERSION}
+
+git commit . -m "Set post-release version to $VERSION"
diff --git a/src/intent_parser/deterministic_intent_parser.rs b/src/intent_parser/deterministic_intent_parser.rs
@@ -255,6 +255,7 @@ impl DeterministicIntentParser {
                     }
                 })
                 .map(|(a_match, group_name)| {
+                    let group_name = group_name.split('_').collect::<Vec<&str>>()[0];
                     let slot_name = self.group_names_to_slot_names[group_name].to_string();
                     let entity = self.slot_names_to_entities[intent][&slot_name].to_string();
                     let byte_range = a_match.start()..a_match.end();
@@ -368,29 +369,33 @@ mod tests {
             language_code: "en".to_string(),
             patterns: hashmap![
                 "dummy_intent_1".to_string() => vec![
-                    r"^\s*This\s*is\s*a\s*(?P<group_1>%DUMMY_ENTITY_1%)\s*query\s*with\s*another\s*(?P<group_2>%DUMMY_ENTITY_2%)\s*$".to_string(),
-                    r"^\s*(?P<group_5>%DUMMY_ENTITY_1%)\s*$".to_string(),
-                    r"^\s*This\s*is\s*another\s*(?P<group_3>%DUMMY_ENTITY_2%)\s*query\s*$".to_string(),
-                    r"^\s*This\s*is\s*another\s*über\s*(?P<group_3>%DUMMY_ENTITY_2%)\s*query.\s*$".to_string(),
-                    r"^\s*This\s*is\s*another\s*(?P<group_4>%DUMMY_ENTITY_2%)?\s*$*".to_string(),
+                    r"^\s*This\s*is\s*a\s*(?P<group1>%DUMMY_ENTITY_1%)\s*query\s*with\s*another\s*(?P<group2>%DUMMY_ENTITY_2%)\s*$".to_string(),
+                    r"^\s*(?P<group5>%DUMMY_ENTITY_1%)\s*$".to_string(),
+                    r"^\s*This\s*is\s*another\s*(?P<group3>%DUMMY_ENTITY_2%)\s*query\s*$".to_string(),
+                    r"^\s*This\s*is\s*another\s*über\s*(?P<group3>%DUMMY_ENTITY_2%)\s*query.\s*$".to_string(),
+                    r"^\s*This\s*is\s*another\s*(?P<group4>%DUMMY_ENTITY_2%)?\s*$*".to_string(),
                 ],
                 "dummy_intent_2".to_string() => vec![
-                    r"^\s*This\s*is\s*a\s*(?P<group_0>%DUMMY_ENTITY_1%)\s*query\s*from\s*another\s*intent\s*$".to_string()
+                    r"^\s*This\s*is\s*a\s*(?P<group0>%DUMMY_ENTITY_1%)\s*query\s*from\s*another\s*intent\s*$".to_string()
                 ],
                 "dummy_intent_3".to_string() => vec![
-                    r"^\s*Send\s*(?P<group_6>%SNIPSAMOUNTOFMONEY%)\s*to\s*john\s*$".to_string(),
-                    r"^\s*Send\s*(?P<group_6>%SNIPSAMOUNTOFMONEY%)\s*to\s*john\s*at\s*(?P<group_7>%DUMMY_ENTITY_2%)\s*$".to_string()
+                    r"^\s*Send\s*(?P<group6>%SNIPSAMOUNTOFMONEY%)\s*to\s*john\s*$".to_string(),
+                    r"^\s*Send\s*(?P<group6>%SNIPSAMOUNTOFMONEY%)\s*to\s*john\s*at\s*(?P<group7>%DUMMY_ENTITY_2%)\s*$".to_string()
+                ],
+                "dummy_intent_4".to_string() => vec![
+                    r"^\s*what\s*is\s*(?P<group8>%SNIPSNUMBER%)\s*plus\s*(?P<group8_2>%SNIPSNUMBER%)\s*$".to_string()
                 ],
             ],
             group_names_to_slot_names: hashmap![
-                "group_0".to_string() => "dummy_slot_name".to_string(),
-                "group_1".to_string() => "dummy_slot_name".to_string(),
-                "group_2".to_string() => "dummy_slot_name2".to_string(),
-                "group_3".to_string() => "dummy_slot_name2".to_string(),
-                "group_4".to_string() => "dummy_slot_name3".to_string(),
-                "group_5".to_string() => "dummy_slot_name".to_string(),
-                "group_6".to_string() => "dummy_slot_name4".to_string(),
-                "group_7".to_string() => "dummy_slot_name2".to_string(),
+                "group0".to_string() => "dummy_slot_name".to_string(),
+                "group1".to_string() => "dummy_slot_name".to_string(),
+                "group2".to_string() => "dummy_slot_name2".to_string(),
+                "group3".to_string() => "dummy_slot_name2".to_string(),
+                "group4".to_string() => "dummy_slot_name3".to_string(),
+                "group5".to_string() => "dummy_slot_name".to_string(),
+                "group6".to_string() => "dummy_slot_name4".to_string(),
+                "group7".to_string() => "dummy_slot_name2".to_string(),
+                "group8".to_string() => "dummy_slot_name5".to_string(),
             ],
             slot_names_to_entities: hashmap![
                 "dummy_intent_1".to_string() => hashmap![
@@ -405,6 +410,9 @@ mod tests {
                     "dummy_slot_name2".to_string() => "dummy_entity_2".to_string(),
                     "dummy_slot_name4".to_string() => "snips/amountOfMoney".to_string(),
                 ],
+                "dummy_intent_4".to_string() => hashmap![
+                    "dummy_slot_name5".to_string() => "snips/number".to_string(),
+                ],
             ],
             config: DeterministicParserConfig {
                 ignore_stop_words: true,
@@ -560,7 +568,7 @@ mod tests {
             intent_name: Some("dummy_intent_1".to_string()),
             confidence_score: 1.0,
         };
-        assert_eq!(4, intents.len());
+        assert_eq!(5, intents.len());
         assert_eq!(&first_intent, &intents[0]);
         assert_eq!(0.0, intents[1].confidence_score);
         assert_eq!(0.0, intents[2].confidence_score);
@@ -603,6 +611,63 @@ mod tests {
         assert_eq!(intent, expected_intent);
     }
 
+    #[test]
+    fn test_parse_utterance_with_duplicated_slot_name() {
+        // Given
+        let text = "what is one plus one";
+        let mocked_builtin_entity_parser = MockedBuiltinEntityParser::from_iter(vec![(
+            text.to_string(),
+            vec![
+                BuiltinEntity {
+                    value: "one".to_string(),
+                    range: 8..11,
+                    entity: SlotValue::Number(NumberValue { value: 1. }),
+                    entity_kind: BuiltinEntityKind::Number,
+                },
+                BuiltinEntity {
+                    value: "one".to_string(),
+                    range: 17..20,
+                    entity: SlotValue::Number(NumberValue { value: 1. }),
+                    entity_kind: BuiltinEntityKind::Number,
+                },
+            ],
+        )]);
+        let shared_resources = SharedResourcesBuilder::default()
+            .builtin_entity_parser(mocked_builtin_entity_parser)
+            .build();
+        let parser =
+            DeterministicIntentParser::new(test_configuration(), Arc::new(shared_resources))
+                .unwrap();
+
+        // When
+        let parsing_result = parser.parse(text, None).unwrap();
+        let intent = parsing_result.intent;
+        let slots = parsing_result.slots;
+
+        // Then
+        let expected_intent = IntentClassifierResult {
+            intent_name: Some("dummy_intent_4".to_string()),
+            confidence_score: 1.0,
+        };
+        let expected_slots = vec![
+            InternalSlot {
+                value: "one".to_string(),
+                char_range: 8..11,
+                entity: "snips/number".to_string(),
+                slot_name: "dummy_slot_name5".to_string(),
+            },
+            InternalSlot {
+                value: "one".to_string(),
+                char_range: 17..20,
+                entity: "snips/number".to_string(),
+                slot_name: "dummy_slot_name5".to_string(),
+            },
+        ];
+
+        assert_eq!(expected_intent, intent);
+        assert_eq!(expected_slots, slots);
+    }
+
     #[test]
     fn test_parse_intent_by_ignoring_stop_words() {
         // Given