Merge pull request #38 from allo-media/master

Dutch support.

Cargo.toml

@@ -1,10 +1,10 @@
 [package]
 name = "text2num"
-version = "2.3.0"
+version = "2.4.0"
 authors = ["Allo-Media <contact@allo-media.fr>"]
 edition = "2021"
 license = "MIT"
-description = "Parse and convert numbers written in English, Spanish, German, Italian or French into their digit representation."
+description = "Parse and convert numbers written in English, Dutch, Spanish, German, Italian or French into their digit representation."
 keywords = ["NLP", "words-to-numbers"]
 categories = ["text-processing"]
 repository = "https://github.com/allo-media/text2num-rs"

README.md

@@ -1,4 +1,4 @@
-# Parse and convert numbers written in English, Spanish or French into their digit representation.
+# Parse and convert numbers written in English, Dutch, Spanish, German, Italian or French into their digit representation.
 
 This crate provides a library for recognizing, parsing and transcribing into digits (base 10) numbers expressed in natural language.
 

src/digit_string.rs

@@ -41,6 +41,7 @@ impl DigitString {
         self.frozen = false;
         self.marker = MorphologicalMarker::None;
         self.buffer.clear();
+        self.flags = 0;
     }
 
     /// Freeze the DigitSring to signal the number is complete.

src/lang/fr/mod.rs

@@ -165,7 +165,7 @@ impl LangInterpretor for French {
             }
             "mille" | "mil" | "millième" if b.is_range_free(3, 5) => {
                 let peek = b.peek(2);
-                if peek == b"1" || peek == b"01" {
+                if peek == b"1" {
                     Err(Error::Overlap)
                 } else {
                     b.shift(3)
@@ -317,6 +317,7 @@ mod tests {
         assert_text2digits!("quinze", "15");
 
         assert_text2digits!("soixante quinze mille", "75000");
+        assert_text2digits!("cent un mille", "101000");
         assert_text2digits!("un milliard vingt-cinq millions", "1025000000");
     }
 

src/lang/it/mod.rs

@@ -335,7 +335,6 @@ mod tests {
     fn test_basic() {
         assert_text2digits!("due", "2");
         assert_text2digits!("dieci", "10");
-        assert_text2digits!("dieci", "10");
         assert_text2digits!("tredici", "13");
         assert_text2digits!("diciassette", "17");
         assert_text2digits!("venti", "20");

src/lang/mod.rs

@@ -27,6 +27,7 @@ mod en;
 mod es;
 mod fr;
 mod it;
+mod nl;
 
 use crate::digit_string::DigitString;
 
@@ -37,6 +38,7 @@ pub use en::English;
 pub use es::Spanish;
 pub use fr::French;
 pub use it::Italian;
+pub use nl::Dutch;
 
 /// Model the Morphological markers that differenciate ordinals or fractions from cardinals,
 /// and that must be retained on the digit form.
@@ -128,6 +130,7 @@ pub enum Language {
     German(German),
     Italian(Italian),
     Spanish(Spanish),
+    Dutch(Dutch),
 }
 
 impl Language {
@@ -150,6 +153,10 @@ impl Language {
     pub fn spanish() -> Self {
         Language::Spanish(Spanish::default())
     }
+
+    pub fn dutch() -> Self {
+        Language::Dutch(Dutch::default())
+    }
 }
 
 macro_rules! delegate {
@@ -219,5 +226,5 @@ macro_rules! delegate {
 }
 
 impl LangInterpretor for Language {
-    delegate!(French, English, German, Italian, Spanish);
+    delegate!(Dutch, French, English, German, Italian, Spanish);
 }