{-
- BinConv.hs
-
- Paul Sanders, SRD. 1992
-
- This module contains routines for converting numbers to and from a
- number of binary digits.
-
-}
module BinConv (codes_to_ascii, ascii_to_codes, dec_to_binx) where
zeroes = '0' : zeroes
-- dec_to_binx converts a decimal to a fixed number of binary digits
-- dec_to_binx #binary-digits decimal-number = binary-string
dec_to_binx :: Int -> Int -> String
dec_to_binx x y
= take (x - length bin_string) zeroes ++ bin_string
where
bin_string = dec_to_bin y
dec_to_bin = reverse . dec_to_bin'
dec_to_bin' 0 = []
dec_to_bin' x
= (if (x `rem` 2) == 1
then '1'
else '0') : dec_to_bin' (x `div` 2)
codes_to_ascii :: [Int] -> [Int]
codes_to_ascii [] = []
codes_to_ascii (x:y:ns)
= x_div : ((x_rem * 16) + y_div) : y_rem : codes_to_ascii ns
where
(x_div, x_rem) = divRem x 16
(y_div, y_rem) = divRem y 256
codes_to_ascii [n]
= [x_div , x_rem]
where
(x_div, x_rem) = divRem n 16
ascii_to_codes [] = []
ascii_to_codes (x:y:z:ns)
= (x * 16) + y_div : (y_rem * 256) + z : ascii_to_codes ns
where
(y_div, y_rem) = divRem y 16
ascii_to_codes [x,y]
= [(x * 16) + y_rem]
where
(y_div, y_rem) = divRem y 16
divRem x y = (x `div` y, x `rem` y) -- missing from PreludeCore ?
module Main (main) where
main = putStr (show (result 1000))
result 0 = []
result n = codes_to_ascii (3077, 1192) ++ result (n-1)
codes_to_ascii (x,y)
= x_div : ((x_rem * 16) + y_div) : [y_rem]
where
(x_div, x_rem) = divRem x 16
(y_div, y_rem) = divRem y 256
divRem x y = (x `div` y, x `rem` y) -- missing from PreludeCore ?
{-
- Decode.hs
-
- Module containing the code to decode LZW encodings
-
- Paul Sanders, Applications Research Division, BTL 1992
-
- DEC_VERSION 1 uses a list with keys in ascending order as a table, ie.
- entry n is given by table!!n.
-
- DEC_VERSION 2 uses a list with keys in descending order as a table, ie.
- entry n is given by table!!(#table-n). We don't need to calculate the
- length of the table however as this is given by the value of the next
- code to be added.
-
- DEC_VERSION 3 uses a balanced binary tree to store the keys. We can do
- this cheaply by putting the key in the correct place straight away and
- therefore not doing any rebalancing.
-}
module Decode (decode)
where
import Prelude hiding( lookup ) -- lookup defined locally
import Defaults
import BinConv
data Optional a = NONE | SOME a deriving (Eq, Show{-was:Text-})
{- We ideally want to store the table as an array but these are inefficient
- so we use a list instead. We don't use the tree used by encode since we
- can make use of the fact that all our keys (the codes) come in order and
- will be placed at the end of the table, at position 'code'.
-
- An entry of (SOME n, 'c') indicates that this code has prefix code n
- and final character c.
-}
{- Kick off the decoding giving the real function the first code value and
- the initial table.
-}
decode :: [Int] -> String
decode []
= []
decode cs
= decode' cs first_code init_table
{- decode` decodes the first character which is special since no new code
- gets added for it. It is also special in so far as we know that the
- code is a singleton character and thus has prefix NONE. The '@' is a
- dummy character and can be anything.
-}
decode' [] _ _ = []
decode' (c:cs) n t
= ch : do_decode cs n c ch t
where
(NONE, ch) = lookup c t
{- do_decode decodes all the codes bar the first.
-
- If the code is in the table (ie the code is less than the next code to be
- added) then we output the string for that code (using unfold if a prefix
- type) and add a new code to the table with the final character output as
- the extension and the previous code as prefix.
-
- If the code is not one we know about then we give it to decode_special for
- special treatment
-}
do_decode [] _ _ _ _ = []
do_decode (c:cs) n old_n fin_char t
= if c >= n -- we don't have this code in the table yet
then decode_special (c:cs) n old_n fin_char t
else outchs ++ do_decode cs n' c (head outchs) t'
where
outchs = reverse (unfold c (n-1) t)
(n', t') = if n == max_entries
then (n, t)
else (n+1, insert n (SOME old_n, head outchs) t)
{- decode_special decodes a code that isn't in the table.
-
- The algorithm in Welch describes why this works, suffice it to say that
- the output string is given by the last character output and the string
- given by the previous code. An entry is also made in the table for the
- last character output and the old code.
-}
decode_special (c:cs) n old_n fin_char t
= outchs ++ do_decode cs n' c (head outchs) t'
where
outchs = reverse (fin_char : unfold old_n (n-1) t)
(n', t') = if n == max_entries
then (n, t)
else (n+1, insert n (SOME old_n, fin_char) t)
{- unfold a prefix code.
-
- chain back through the prefixes outputting the extension characters as we
- go.
-}
unfold n t_len t
= if prefix == NONE
then [c]
else c : unfold n' t_len t
where
(prefix, c) = lookup n t
SOME n' = prefix
data DecompTable = Branch DecompTable DecompTable | Leaf (Optional Int, Char) deriving (Show{-was:Text-})
{- Insert a code pair into the table. The position of the code is given by
- the breakdown of the key into its binary digits
-}
insert n v t = insert' (dec_to_binx code_bits n) v t
{- We can place a code exactly where it belongs using the following algorithm.
- Take the code's binary rep expanded to the maximum number of bits. Start
- at the first bit, if a 0 then insert the code to the left, if a 1 then
- insert to the right. Carry on with the other bits until we run out and are
- thus at the right place and can construct the node.
-}
insert' [] v (Leaf _)
= Leaf v
insert' ('0' : bs) v (Branch l r)
= Branch (insert' bs v l) r
insert' ('1' : bs) v (Branch l r)
= Branch l (insert' bs v r)
insert' ('0' : bs) v t
= Branch (insert' bs v t) t
insert' ('1' : bs) v t
= Branch t (insert' bs v t)
{- For a lookup we use the same mechanism to locate the position of the item
- in the tree but if we find that the route has not been constructed or the
- node has the dummy value then that code is not yet in the tree. The way
- in which the decode algorithm works this should never happen.
-}
lookup n t = lookup' (dec_to_binx code_bits n) t
lookup' [] (Leaf v)
= v
lookup' ('0' : bs) (Branch l _)
= lookup' bs l
lookup' ('1' : bs) (Branch _ r)
= lookup' bs r
lookup' _ _ = error "tree insert error - seek professional help"
init_table = mk_init_table 0 (Leaf (SOME 99999, '@'))
mk_init_table 256 t = t
mk_init_table n t = mk_init_table (n+1) (insert n (NONE, toEnum n) t)
{-
- Defaults.hs
-
- Contains the tuning values for compress and uncompress
-
-}
module Defaults
where
-- Maximum number of table entries (probably = 2^code_bits)
max_entries :: Int
max_entries = 2 ^ code_bits
-- First code value available
first_code :: Int
first_code = 256
-- Number of bits per output character
ascii_bits :: Int
ascii_bits = 8
-- Number of bits to represent code by
code_bits :: Int
code_bits = 12
{-
- Encode Mk 2, using a prefix table for the codes
-
- Paul Sanders, Systems Research, British Telecom Laboratories 1992
-}
module Encode (encode) where
import Defaults
import PTTrees
-- for convenience we make the code table type explicit
type CodeTable = PrefixTree Char Int
-- encode sets up the arguments for the real function.
encode :: String -> [Int]
encode input = encode' input first_code initial_table
{-
- encode' loops through the input string assembling the codes produced
- by code_string. The first character is treated specially in that it
- is not added to the table; its code is simply its ascii value.
-}
encode' [] _ _
= []
encode' input v t
= case (code_string input 0 v t) of { (input', n, t') ->
n : encode' input' (v + 1) t'
}
{-
- code_string parses enough of the input string to produce one code and
- returns the remaining input, the code and a new code table.
-
- The first character is taken and its place found in the code table. The
- extension code table found for this character is then used as the lookup
- table for the next character.
-
- If a character is not found in the current table then output the code
- of the character associated with the current table and add the current
- character to the current table and assign it the next new code value.
-}
code_string input@(c : input2) old_code next_code (PT p@(PTE k v t) l r)
| c < k = (f1 r1 p r)
| c > k = (f2 r2 p l)
| otherwise = (f3 r3 k v l r)
where
r1 = code_string input old_code next_code l
r2 = code_string input old_code next_code r
r3 = code_string input2 v next_code t
f1 (input_l,nl,l2) p r = (input_l,nl,PT p l2 r)
f2 (input_r,nr,r2) p l = (input_r,nr,PT p l r2)
f3 (input2,n,t2) k v l r = (input2, n, PT (PTE k v t2) l r)
code_string input@(c : input_file2) old_code next_code PTNil
| next_code >= 4096 = (input, old_code, PTNil)
| otherwise = (input, old_code, PT (PTE c next_code PTNil) PTNil PTNil)
code_string [] old_code next_code code_table
= ([], old_code, PTNil)
{-
- We want the inital table to be balanced, but this is expensive to compute
- as a rebalance is needed evert two inserts (yuk!). So we do the ordinary
- infix-order binary tree insert but give the keys in such an order as to
- give a balanced tree.
-
- (I would have defined the tree by hand but the constant was too big
- for hc-0.41)
-}
initial_table :: CodeTable
initial_table = foldr tab_insert PTNil balanced_list
tab_insert n = insert (toEnum n) n
balanced_list
= [128,64,32,16,8,4,2,1,0,3,6,5,7,12,10,9,11,14,13,15,24,20,18,17,19,22,
21,23,28,26,25,27,30,29,31,48,40,36,34,33,35,38,37,39,44,42,41,43,46,
45,47,56,52,50,49,51,54,53,55,60,58,57,59,62,61,63,96,80,72,68,66,65]
++ bal_list2 ++ bal_list3 ++ bal_list4 ++ bal_list5
bal_list2
= [67,70,69,71,76,74,73,75,78,77,79,88,84,82,81,83,86,85,87,92,90,89,91,
94,93,95,112,104,100,98,97,99,102,101,103,108,106,105,107,110,109,111,
120,116,114,113,115,118,117,119,124,122,121,123,126,125,127,192,160]
bal_list3
= [144,136,132,130,129,131,134,133,135,140,138,137,139,142,141,143,152,
148,146,145,147,150,149,151,156,154,153,155,158,157,159,176,168,164,
162,161,163,166,165,167,172,170,169,171,174,173,175,184,180,178,177]
bal_list4
= [179,182,181,183,188,186,185,187,190,189,191,224,208,200,196,194,193,
195,198,197,199,204,202,201,203,206,205,207,216,212,210,209,211,214,
213,215,220,218,217,219,222,221,223,240,232,228,226,225,227,230,229,
231,236,234,233,235,238,237,239,248,244,242,241,243,246,245,247,252]
bal_list5
= [250,249,251,254,253,255]
module Main (main){-export list added by partain-} where {
-- partain: with "ghc -cpp -DSLEAZY_UNBOXING", you get (guess what)?
-- without it, you get the code as originally written.
--
-- Things done here:
-- * The obvious unboxing (e.g., Int ==> Int#).
-- * use quot/rem, not div/mod
-- * inline PrefixElement type into PrefixTree.PT constructor
-- * cvt final clause of 3-way comparison to "otherwise"
-- * use shifts, not quot/rem (not necessary: C compiler converts
-- them just fine)
--
-- Obviously, more egregious hacking could be done:
-- * replace Tuple/List types that mention Ints with specialised
-- variants
#if defined(__GLASGOW_HASKELL__) && defined(SLEAZY_UNBOXING)
#define FAST_INT Int#
#define ILIT(x) (x#)
#define IBOX(x) (I# (x))
#define _ADD_ `plusInt#`
#define _SUB_ `minusInt#`
#define _MUL_ `timesInt#`
#define _DIV_ `divInt#`
#define _QUOT_ `quotInt#`
#define _REM_ `remInt#`
#define _NEG_ negateInt#
#define _EQ_ `eqInt#`
#define _LT_ `ltInt#`
#define _LE_ `leInt#`
#define _GE_ `geInt#`
#define _GT_ `gtInt#`
#define _CHR_ chr#
#define FAST_BOOL Int#
#define _TRUE_ 1#
#define _FALSE_ 0#
#define _IS_TRUE_(x) ((x) `eqInt#` 1#)
#define FAST_CHAR Char#
#define CBOX(x) (C# (x))
data FAST_TRIPLE = TRIP [Char] Int# PrefixTree;
#define _TRIP_(a,b,c) (TRIP (a) (b) (c))
#define PrefixElement FAST_CHAR FAST_INT PrefixTree
#define _PTE_(a,b,c) (a) (b) (c)
#else {- ! __GLASGOW_HASKELL__ -}
#define FAST_INT Int
#define ILIT(x) (x)
#define IBOX(x) (x)
#define _ADD_ +
#define _SUB_ -
#define _MUL_ *
#define _DIV_ `div`
#define _QUOT_ `quot`
#define _REM_ `rem`
#define _NEG_ -
#define _EQ_ ==
#define _LT_ <
#define _LE_ <=
#define _GE_ >=
#define _GT_ >
#define _CHR_ toEnum
#define FAST_BOOL Bool
#define _TRUE_ True
#define _FALSE_ False
#define _IS_TRUE_(x) (x)
#define FAST_CHAR Char
#define CBOX(x) (x)
type FAST_TRIPLE = ([Char], Int, PrefixTree);
#define _TRIP_(a,b,c) ((a), (b), (c))
data PrefixElement = PTE FAST_CHAR FAST_INT PrefixTree;
#define _PTE_(a,b,c) (PTE (a) (b) (c))
#endif {- ! __GLASGOW_HASKELL__ -}
-- end of partain
data PrefixTree = PTNil | PT PrefixElement PrefixTree PrefixTree;
--create_code_table :: PrefixTree; -- partain: sig
create_code_table = create_code_table2 ILIT(0) ILIT(256);
create_code_table2 :: FAST_INT -> FAST_INT -> PrefixTree;
create_code_table2 first_code ILIT(0) = PTNil;
create_code_table2 first_code ILIT(1)
= PT _PTE_((_CHR_ first_code), first_code, PTNil) PTNil PTNil;
create_code_table2 first_code n_codes
= PT _PTE_((_CHR_ m_code), m_code, PTNil) left right
where {
left = create_code_table2 first_code (m_code _SUB_ first_code);
right = create_code_table2 m_code2 ((first_code _ADD_ n_codes) _SUB_ m_code2);
m_code = (first_code _ADD_ (first_code _ADD_ n_codes _SUB_ ILIT(1))) _QUOT_ ILIT(2);
m_code2 = m_code _ADD_ ILIT(1);
};
lzw_code_file :: [Char] -> PrefixTree -> FAST_INT -> [Int];
lzw_code_file [] code_table next_code = [];
lzw_code_file input code_table next_code
= -- partain: case-ified lazy where
case (code_string input ILIT(0) next_code code_table) of {
_TRIP_(input2,n,code_table2) ->
IBOX(n) : lzw_code_file input2 code_table2 (next_code _ADD_ ILIT(1))
};
code_string :: [Char] -> FAST_INT -> FAST_INT -> PrefixTree -> FAST_TRIPLE;
#if defined(__GLASGOW_HASKELL__) && defined(SLEAZY_UNBOXING)
code_string input@(CBOX(c) : input2) old_code next_code (PT k v t {-p@(PTE k v t)-} l r)
| CBOX(c) < CBOX(k) = f1 r1 {-p-} k v t r
| CBOX(c) > CBOX(k) = f2 r2 {-p-} k v t l
| otherwise {- CBOX(c) == CBOX(k) -} = f3 r3 k v l r
#else
code_string input@(CBOX(c) : input2) old_code next_code (PT p@(PTE k v t) l r)
| CBOX(c) < CBOX(k) = f1 r1 p r
| CBOX(c) > CBOX(k) = f2 r2 p l
| otherwise {- CBOX(c) == CBOX(k) -} = f3 r3 k v l r
#endif
where {
r1 = code_string input old_code next_code l;
r2 = code_string input old_code next_code r;
r3 = code_string input2 v next_code t;
#if defined(__GLASGOW_HASKELL__) && defined(SLEAZY_UNBOXING)
f1 _TRIP_(input_l,nl,l2) k v t r = _TRIP_(input_l,nl,PT k v t l2 r);
f2 _TRIP_(input_r,nr,r2) k v t l = _TRIP_(input_r,nr,PT k v t l r2);
#else
f1 _TRIP_(input_l,nl,l2) p r = _TRIP_(input_l,nl,PT p l2 r);
f2 _TRIP_(input_r,nr,r2) p l = _TRIP_(input_r,nr,PT p l r2);
#endif
f3 _TRIP_(input2,n,t2) k v l r = _TRIP_(input2, n, PT _PTE_(k, v, t2) l r);
};
--code_string input@(c : input2) old_code next_code (PT p@(PTE k v t) l r)
-- | c < k = (input_l,nl,PT p l' r)
-- | c > k = (input_r,nr,PT p l r')
-- | c == k = (input',n,PT (PTE k v t') l r)
-- where {
-- (input_l,nl,l') = code_string input old_code next_code l;
-- (input_r,nr,r') = code_string input old_code next_code r;
-- (input',n,t') = code_string input2 v next_code t;
-- };
code_string input@(CBOX(c) : input_file2) old_code next_code PTNil
= if (next_code _GE_ ILIT(4096))
then _TRIP_(input, old_code, PTNil)
else _TRIP_(input, old_code, PT _PTE_(c, next_code, PTNil) PTNil PTNil);
code_string [] old_code next_code code_table = _TRIP_([], old_code, PTNil);
integer_list_to_char_list (IBOX(n) : l)
= CBOX(_CHR_ (n _QUOT_ ILIT(16))) : integer_list_to_char_list2 l n;
integer_list_to_char_list [] = [];
integer_list_to_char_list2 (IBOX(c) : l) n
= CBOX(_CHR_ ((n _MUL_ ILIT(16)) _ADD_ ((c _QUOT_ ILIT(256)) _REM_ ILIT(16))))
: CBOX(_CHR_ c)
: integer_list_to_char_list l;
integer_list_to_char_list2 [] n = CBOX(_CHR_ (n _MUL_ ILIT(16))) : [];
main :: IO ();
main = getContents >>= \ input_string -> main2 input_string;
main2 :: String -> IO ();
main2 input_string
= putStr output_list
where {
output_list = integer_list_to_char_list code_list;
code_list = lzw_code_file input_string create_code_table ILIT(256);
};
}
-- Lzw2.hs looks like an earlier version of Lzw.hs
module Main (main){-export list added by partain-} where {
-- partain: with "ghc -cpp -DSLEAZY_UNBOXING", you get (guess what)?
-- without it, you get the code as originally written.
--
-- Things done here:
-- * The obvious unboxing (e.g., Int ==> Int#).
-- * use quot/rem, not div/mod
-- * inline PrefixElement type into PrefixTree.PT constructor
-- * cvt final clause of 3-way comparison to "otherwise"
-- * use shifts, not quot/rem (not necessary: C compiler converts
-- them just fine)
--
-- Obviously, more egregious hacking could be done:
-- * replace Tuple/List types that mention Ints with specialised
-- variants
#define FAST_INT Int#
#define ILIT(x) (x#)
#define IBOX(x) (I# (x))
#define _ADD_ `plusInt#`
#define _SUB_ `minusInt#`
#define _MUL_ `timesInt#`
#define _DIV_ `divInt#`
#define _QUOT_ `quotInt#`
#define _REM_ `remInt#`
#define _NEG_ negateInt#
#define _EQ_ `eqInt#`
#define _LT_ `ltInt#`
#define _LE_ `leInt#`
#define _GE_ `geInt#`
#define _GT_ `gtInt#`
#define _CHR_ chr#
#define FAST_BOOL Int#
#define _TRUE_ 1#
#define _FALSE_ 0#
#define _IS_TRUE_(x) ((x) `eqInt#` 1#)
#define FAST_CHAR Char#
#define CBOX(x) (C# (x))
data FAST_TRIPLE = TRIP [Char] Int# PrefixTree;
#define _TRIP_(a,b,c) (TRIP (a) (b) (c))
#define PrefixElement FAST_CHAR FAST_INT PrefixTree
#define _PTE_(a,b,c) (a) (b) (c)
-- end of partain
data PrefixTree = PTNil | PT PrefixElement PrefixTree PrefixTree;
create_code_table = create_code_table2 ILIT(0) ILIT(256);
create_code_table2 :: FAST_INT -> FAST_INT -> PrefixTree;
create_code_table2 first_code ILIT(0) = PTNil;
create_code_table2 first_code ILIT(1)
= PT _PTE_((_CHR_ first_code), first_code, PTNil) PTNil PTNil;
create_code_table2 first_code n_codes
= PT _PTE_((_CHR_ m_code), m_code, PTNil) left right
where {
left = create_code_table2 first_code (m_code _SUB_ first_code);
right = create_code_table2 m_code2 ((first_code _ADD_ n_codes) _SUB_ m_code2);
m_code = (first_code _ADD_ (first_code _ADD_ n_codes _SUB_ ILIT(1))) _QUOT_ ILIT(2);
m_code2 = m_code _ADD_ ILIT(1);
};
lzw_code_file :: [Char] -> PrefixTree -> FAST_INT -> [Int];
lzw_code_file [] code_table next_code = [];
lzw_code_file input code_table next_code
= -- partain: case-ified lazy where
case (code_string ILIT(0) next_code input code_table) of {
_TRIP_(input2,n,code_table2) ->
IBOX(n) : lzw_code_file input2 code_table2 (next_code _ADD_ ILIT(1))
};
code_string :: FAST_INT -> FAST_INT -> [Char] -> PrefixTree -> FAST_TRIPLE;
code_string old_code next_code input@(CBOX(c) : input2) (PT k v t {-p@(PTE k v t)-} l r)
| CBOX(c) < CBOX(k) = _scc_ "cs1" (f1 r1 {-p-} k v t r)
| CBOX(c) > CBOX(k) = _scc_ "cs2" (f2 r2 {-p-} k v t l)
| otherwise {- CBOX(c) == CBOX(k) -} = _scc_ "cs3" (f3 r3 k v l r)
where {
r1 = code_string old_code next_code input l;
r2 = code_string old_code next_code input r;
r3 = code_string v next_code input2 t;
f1 _TRIP_(input_l,nl,l2) k v t r = _TRIP_(input_l,nl,PT k v t l2 r);
f2 _TRIP_(input_r,nr,r2) k v t l = _TRIP_(input_r,nr,PT k v t l r2);
f3 _TRIP_(input2,n,t2) k v l r = _TRIP_(input2, n, PT _PTE_(k, v, t2) l r);
};
code_string old_code next_code input@(CBOX(c) : input_file2) PTNil
= if (next_code _GE_ ILIT(4096))
then _scc_ "cs4" _TRIP_(input, old_code, PTNil)
else _scc_ "cs5" _TRIP_(input, old_code, PT _PTE_(c, next_code, PTNil) PTNil PTNil);
code_string old_code next_code [] code_table = _scc_ "cs6" _TRIP_([], old_code, PTNil);
integer_list_to_char_list (IBOX(n) : l)
= CBOX(_CHR_ (n _QUOT_ ILIT(16))) : integer_list_to_char_list2 l n;
integer_list_to_char_list [] = [];
integer_list_to_char_list2 (IBOX(c) : l) n
= CBOX(_CHR_ ((n _MUL_ ILIT(16)) _ADD_ ((c _QUOT_ ILIT(256)) _REM_ ILIT(16))))
: CBOX(_CHR_ c)
: integer_list_to_char_list l;
integer_list_to_char_list2 [] n = CBOX(_CHR_ (n _MUL_ ILIT(16))) : [];
main :: IO ();
main = getContents >>= \input_string -> main2 input_string;
main2 :: String -> IO ();
main2 input_string
= putStr output_list
where {
output_list = integer_list_to_char_list code_list;
code_list = lzw_code_file input_string create_code_table ILIT(256);
};
}
{-
- Compress.hs
-
- This program is a version of the compress utility as defined in
- "A Technique for High Performance Data Compression", Terry A. Welch,
- Computer, vol 17, no 6 1984, pp 8-19
-
- Usage: compress file
-
- Paul Sanders, Systems Research Division, British Telecom Laboratories 1992
-
-}
module Main (main) where
import Defaults
import BinConv -- binary conversion routines
import Encode -- coding routine
main = getContents >>= \ inp ->
putStr (compress inp)
{- To compress a string we first encode it, then convert it to n-bit binaries
- convert back to decimal as ascii-bit values and then to characters
-}
compress = map toEnum . codes_to_ascii . encode
module PTTrees (insert, PrefixTree(..), PrefixElem(..)) where
data PrefixTree a b = PTNil |
PT (PrefixElem a b) (PrefixTree a b) (PrefixTree a b)
data PrefixElem a b = PTE a b (PrefixTree a b)
{-partain-}
--insert :: Char -> Int -> PrefixTree Char Int -> PrefixTree Char Int
{-partain-}
insert k v PTNil =
PT (PTE k v PTNil) PTNil PTNil
insert k v (PT p@(PTE k' v' t) l r)
| k < k' = PT p (insert k v l) r
| k > k' = PT p l (insert k v r)
| otherwise = PT p l r
{-
- Uncompress.hs
-
- This program is a version of the compress utility as defined in
- "A Technique for High Performance Data Compression", Terry A. Welch,
- Computer, vol 17, no 6 1984, pp 8-19
-
-
- Paul Sanders, Systems Research Division, British Telecom Laboratories 1992
-
-}
module Main (main) where
import Defaults
import BinConv -- binary conversion routines
import Decode -- decoding routines
main = getContents >>= \ inp ->
putStr (uncompress inp)
{- To uncompress a string we first convert the characters to n-bit binaries
- and then to decimals which can then be decoded.
-}
uncompress = decode . ascii_to_codes . map fromEnum