"Hello World" encoded as Base64

>>> helloB64
MkEnc Proxy () "SGVsbG8gV29ybGQ="

>>> displ helloB64
"MkEnc '[enc-B64] () (ByteString SGVsbG8gV29ybGQ=)"

>>> encodeAll . toEncoding () $ "Hello World" :: Enc '["enc-B64"] () B.ByteString
MkEnc Proxy () "SGVsbG8gV29ybGQ="

Previous text decoded from Base64

>>> fromEncoding . decodeAll $ helloB64
"Hello World"

recreateFAll allows for recovering data at program boundaries (for example, when parsing JSON input). It makes sure that the content satisfies specified encodings.

>>> recreateFAll . toEncoding () $ "SGVsbG8gV29ybGQ=" :: Either RecreateEx (Enc '["enc-B64"] () B.ByteString)
Right (MkEnc Proxy () "SGVsbG8gV29ybGQ=")

>>> recreateFAll . toEncoding () $ "SGVsbG8gV29ybGQ" :: Either RecreateEx (Enc '["enc-B64"] () B.ByteString)
Left (RecreateEx "enc-B64" ("invalid padding"))

The above example start by placing payload in zero-encoded Enc '[] () type and then apply recreateFAll this is a good way to recreate encoded type if encoding is known.

If is it not, UncheckedEnc type can be used.

(See ToEncString for better example).

This module is concerned only with the first approach.

>>> let unchecked = toUncheckedEnc ["enc-B64"] () ("SGVsbG8gV29ybGQ=" :: T.Text)
>>> verifyUncheckedEnc' @'["enc-B64"] unchecked
Just (Right (MkEnc Proxy () "SGVsbG8gV29ybGQ="))

"Hello World" double-Base64 encoded. Notice the same code used as in single encoding, the game is played at type level.

>>> encodeAll . toEncoding () $ "Hello World" :: Enc '["enc-B64","enc-B64"] () B.ByteString
MkEnc Proxy () "U0dWc2JHOGdWMjl5YkdRPQ=="

>>> displ helloB64B64
"MkEnc '[enc-B64,enc-B64] () (ByteString U0dWc2JHOGdWMjl5YkdRPQ==)"

Double Base64 encoded "Hello World" with one layer of encoding removed

>>> decodePart @'["enc-B64"] $ helloB64B64 :: Enc '["enc-B64"] () B.ByteString
MkEnc Proxy () "SGVsbG8gV29ybGQ="

>>> helloB64B64PartDecode == helloB64
True

helloB64B64 all the way to ByteString

Notice a similar polymorphism in decoding.

>>> fromEncoding . decodeAll $ helloB64B64 :: B.ByteString
"Hello World"

We can also decode all the parts:

>>> fromEncoding . decodePart @'["enc-B64","enc-B64"] $ helloB64B64
"Hello World"

what happens when we try to recover encoded once text to Enc '["enc-B64", "enc-B64"].

Again, notice the same expression is used as in previous recovery.

>>> recreateFAll . toEncoding () $ "SGVsbG8gV29ybGQ=" :: Either RecreateEx (Enc '["enc-B64", "enc-B64"] () B.ByteString)
Left (RecreateEx "enc-B64" ("invalid padding"))

"do-UPPER" (from Sample module) encoding applied to "Hello World"

Notice a namespace thing going on, "enc-" is encoding, "do-" is some transformation. These are typically not reversible, some could be recoverable.

The same code is used as in "enc-" examples to encode (now transform).

>>> encodeAll . toEncoding () $ "Hello World" :: Enc '["do-UPPER"] () T.Text
MkEnc Proxy () "HELLO WORLD"

Sample compound transformation

>>> encodeAll . toEncoding () $ "HeLLo world" :: Enc '["do-reverse", "do-Title"] () T.Text
MkEnc Proxy () "dlroW olleH" 

Example configuration

helloTitle is needed in following examples

Configuration can be used to impact the encoding process.

So far we had used () as configuration of all encodings. But since both "do-reverse", "do-Title" are polymorphic in configuration we can also do this:

>>> encodeAll . toEncoding exampleConf $ "HeLLo world" :: Enc '["do-reverse", "do-Title"] Config T.Text
MkEnc Proxy (Config {sizeLimit = SizeLimit {unSizeLimit = 8}}) "dlroW olleH"

>>> encodeAll . toEncoding exampleConf $ "HeLlo world" :: Enc '["do-size-limit", "do-reverse", "do-Title"] Config T.Text
MkEnc Proxy (Config {sizeLimit = SizeLimit {unSizeLimit = 8}}) "dlroW ol"

Instead, encode previously defined helloTitle by reversing it and adding size limit

>>> encodePart @'["do-size-limit", "do-reverse"] helloTitle :: Enc '["do-size-limit", "do-reverse", "do-Title"] Config T.Text
MkEnc Proxy (Config {sizeLimit = SizeLimit {unSizeLimit = 8}}) "dlroW ol"

... and we unwrap the B64 part only

>>> decodePart @'["enc-B64"] $ helloLimitB64
MkEnc Proxy (Config {sizeLimit = SizeLimit {unSizeLimit = 8}}) "HeLlo wo"

ASCII char set ByteStrings are sequences of Bytes (Word8). The type is very permissive, it may contain binary data such as jpeg picture.

"r-ASCII" encoding acts as partial identity function it does not change any bytes in bytestring but it fails if a byte is outside of ASCII range (in Either monad).

Note naming thing: "r-" is partial identity ("r-" is from restriction).

>>> encodeFAll . toEncoding () $ "HeLlo world" :: Either EncodeEx (Enc '["r-ASCII"] () B.ByteString)
Right (MkEnc Proxy () "HeLlo world")

Arguably the type we used for helloB64 was too permissive. a better version is here:

>>> encodeFAll . toEncoding () $ "Hello World" :: Either EncodeEx (Enc '["enc-B64", "r-ASCII"] () B.ByteString)
Right (MkEnc Proxy () "SGVsbG8gV29ybGQ=") 

>>> decodePart @'["enc-B64"] <$> helloAsciiB64
Right (MkEnc Proxy () "Hello World")