Normalize for the search the facets values

milli/src/heed_codec/beu16_str_codec.rs

@@ -0,0 +1,27 @@
+use std::borrow::Cow;
+use std::convert::TryInto;
+use std::str;
+
+pub struct BEU16StrCodec;
+
+impl<'a> heed::BytesDecode<'a> for BEU16StrCodec {
+    type DItem = (u16, &'a str);
+
+    fn bytes_decode(bytes: &'a [u8]) -> Option<Self::DItem> {
+        let (n_bytes, str_bytes) = bytes.split_at(2);
+        let n = n_bytes.try_into().map(u16::from_be_bytes).ok()?;
+        let s = str::from_utf8(str_bytes).ok()?;
+        Some((n, s))
+    }
+}
+
+impl<'a> heed::BytesEncode<'a> for BEU16StrCodec {
+    type EItem = (u16, &'a str);
+
+    fn bytes_encode((n, s): &Self::EItem) -> Option<Cow<[u8]>> {
+        let mut bytes = Vec::with_capacity(s.len() + 2);
+        bytes.extend_from_slice(&n.to_be_bytes());
+        bytes.extend_from_slice(s.as_bytes());
+        Some(Cow::Owned(bytes))
+    }
+}

milli/src/heed_codec/mod.rs

@@ -1,3 +1,4 @@
+mod beu16_str_codec;
 mod beu32_str_codec;
 mod byte_slice_ref;
 pub mod facet;
@@ -14,6 +15,7 @@ mod str_str_u8_codec;
 pub use byte_slice_ref::ByteSliceRefCodec;
 pub use str_ref::StrRefCodec;
 
+pub use self::beu16_str_codec::BEU16StrCodec;
 pub use self::beu32_str_codec::BEU32StrCodec;
 pub use self::field_id_word_count_codec::FieldIdWordCountCodec;
 pub use self::fst_set_codec::FstSetCodec;