Expose incremental UTF8 decoding APIs

Author: adithyaov

core/src/Streamly/Internal/Unicode/Stream.hs

@@ -190,8 +190,25 @@ encodeLatin1Lax = encodeLatin1
 -- UTF-8 decoding
 -------------------------------------------------------------------------------
 
--- Int helps in cheaper conversion from Int to Char
+-- | CodePoint represents a specific character in the Unicode standard.  The
+-- code point is a numerical value assigned to each character, and UTF-8
+-- encoding uses a variable number of bytes to represent different code points.
+--
+-- Calculate the code point value: Depending on the type of the leading byte,
+-- extract the significant bits from each byte of the sequence and combine them
+-- to form the complete code point value. The specific bit manipulations will
+-- differ based on the number of bytes used.
 type CodePoint = Int
+
+-- | DecodeState refers to the number of bytes remaining to complete the current
+-- UTF-8 character decoding. For ASCII characters (code points 0 to 127), no
+-- decoding state is necessary because they are represented by a single byte.
+-- Therefore, the decoding state for ASCII characters can be considered as 0.
+--
+-- For multi-byte characters, the decoding state indicates the number of bytes
+-- remaining to complete the character. It is usually initialized to a non-zero
+-- value corresponding to the number of bytes in the multi-byte character, e.g
+-- DecodeState will be 1 for 2-bytes char.
 type DecodeState = Word8
 
 -- We can divide the errors in three general categories:
@@ -410,17 +427,24 @@ decodeUtf8EitherD :: Monad m
     => D.Stream m Word8 -> D.Stream m (Either DecodeError Char)
 decodeUtf8EitherD = resumeDecodeUtf8EitherD 0 0
 
--- |
+-- | Decode a bytestream as UTF-8 encoded characters, returning an 'Either'
+-- stream.
+--
+-- This function is similar to 'decodeUtf8', but instead of replacing the
+-- invalid codepoint encountered, it returns a 'Left' 'DecodeError'.
+--
+-- When decoding is successful and a valid character is encountered, the
+-- function returns 'Right Char'.
 --
--- /Pre-release/
 {-# INLINE decodeUtf8Either #-}
 decodeUtf8Either :: Monad m
     => Stream m Word8 -> Stream m (Either DecodeError Char)
 decodeUtf8Either = decodeUtf8EitherD
 
--- |
+-- | Resuming the decoding of a bytestream given a 'DecodeState' and a
+-- 'CodePoint'.
 --
--- /Pre-release/
+-- >>> decodeUtf8Either = resumeDecodeUtf8Either 0 0
 {-# INLINE resumeDecodeUtf8Either #-}
 resumeDecodeUtf8Either
     :: Monad m

core/src/Streamly/Unicode/Stream.hs

@@ -81,6 +81,13 @@ module Streamly.Unicode.Stream
     , decodeUtf8'
     , decodeUtf8Chunks
 
+    -- ** Resumable UTF-8 Decoding
+    , DecodeError(..)
+    , DecodeState
+    , CodePoint
+    , decodeUtf8Either
+    , resumeDecodeUtf8Either
+
     -- * Elimination (Encoding)
     , encodeLatin1
     , encodeLatin1'