From 2e555b5b79df100a3731eaa14b8f4cadc7b81c19 Mon Sep 17 00:00:00 2001
From: "iliya.saroukha" <iliya.saroukhanian@etu.hesge.ch>
Date: Fri, 18 Apr 2025 17:27:24 +0200
Subject: [PATCH] fix: cleaned up impl
---
dyn_huffman.py | 194 +++++++++++++++++++++++++++++--------------------
1 file changed, 117 insertions(+), 77 deletions(-)
diff --git a/dyn_huffman.py b/dyn_huffman.py
index e7e737b..84e383a 100644
--- a/dyn_huffman.py
+++ b/dyn_huffman.py
@@ -7,16 +7,43 @@ import argparse
import string
LIST_ALPHABET = list(string.ascii_lowercase) + list(string.digits)
-# LIST_ALPHABET = list(string.ascii_lowercase)
DICT_ALPHABET = {item: index + 1 for index, item in enumerate(LIST_ALPHABET)}
+
+def compute_vitter_params(m: int) -> tuple[int, int]:
+ e = math.floor(np.log2(m))
+ r = m - (2 ** e)
+ return (e, r)
+
+
+def compute_k(char: str) -> int:
+ return DICT_ALPHABET[char]
+
+
+def compute_code(char: str):
+ e, r = compute_vitter_params(len(LIST_ALPHABET))
+ k_char = compute_k(char)
+
+ if k_char <= 2*r:
+ bin_val = k_char - 1
+ n_bits = e + 1
+ return bin(bin_val)[2:].zfill(n_bits)
+ else:
+ bin_val = k_char - r - 1
+ return bin(bin_val)[2:].zfill(e)
+
+
+FIXED_CODES = {compute_code(item): item for index,
+ item in enumerate(LIST_ALPHABET)}
+
swap_count = 0
+
@dataclass
class Node:
weight: int = 0
nyt_code: str | None = None
- value: str | None = "NYT"
+ value: str = "NYT"
left: Node | None = None
right: Node | None = None
@@ -25,45 +52,27 @@ def swap(tree: Node):
global swap_count
if tree.left and tree.right:
if tree.left.weight > tree.right.weight:
- # print(f'({tree.left.weight}, {tree.right.weight})')
swap_count += 1
tmp = tree.right
tree.right = tree.left
tree.left = tmp
-# def compute_prefix(tree: Node, char: str) -> str:
-# prefix = ''
-#
-# if tree:
-# if tree.value == char:
-# return ''
-#
-# if tree.left:
-# prefix = '0' + compute_prefix(tree.left, char)
-#
-# if tree.right:
-# prefix = '1' + compute_prefix(tree.right, char)
-#
-# return prefix
-
def compute_prefix(tree: Node, char: str) -> str:
if tree.value == char:
return ''
-
+
if tree.left:
left_result = compute_prefix(tree.left, char)
if left_result is not None:
return '0' + left_result
-
+
if tree.right:
right_result = compute_prefix(tree.right, char)
if right_result is not None:
return '1' + right_result
- return None # char not found in this subtree
-
-
+ return None
def find_char(tree: Node, char: str) -> Node | None:
@@ -82,21 +91,25 @@ def find_char(tree: Node, char: str) -> Node | None:
return n
-
-def insert_char(tree: Node, char: str, nyt_code = ''):
+def insert_char(tree: Node, char: str, nyt_code=''):
if tree.left:
if tree.left.value == char:
tree.left.weight += 1
tree.weight += 1
return
- # Leaf case
- if tree.left == None and tree.right == None:
+ if tree.right:
+ if tree.right.value == char:
+ tree.right.weight += 1
+ tree.weight += 1
+ return
+
+ if tree.left is None and tree.right is None:
tree.left = Node()
tree.right = Node(value=char, weight=1, nyt_code=nyt_code)
tree.weight += 1
return
-
+
if tree.left:
if tree.left.value == 'NYT':
insert_char(tree.left, char, nyt_code + '0')
@@ -108,35 +121,18 @@ def insert_char(tree: Node, char: str, nyt_code = ''):
swap(tree)
-def compute_vitter_params(m: int) -> tuple[int, int]:
- e = math.floor(np.log2(m))
- r = m - (2 ** e)
- return (e, r)
-
-
-def compute_k(char: str) -> int:
- return DICT_ALPHABET[char]
-
-
-def compute_code(char: str):
+def compute_char(fixed: str) -> str:
e, r = compute_vitter_params(len(LIST_ALPHABET))
- k_char = compute_k(char)
- if k_char <= 2*r:
- bin_val = k_char - 1
- n_bits = e + 1
- return bin(bin_val)[2:].zfill(n_bits)
- else:
- bin_val = k_char - r - 1
- return bin(bin_val)[2:].zfill(e)
+ return int(fixed, 2)
-def encode(string: str) -> str:
+def encode(input: str) -> str:
tree = Node()
encoded: list[str] = []
seen: set[str] = set()
- for c in string:
+ for c in input:
insert_char(tree, c)
prefix = find_char(tree, c)
@@ -145,45 +141,89 @@ def encode(string: str) -> str:
encoded.append(prefix.nyt_code)
encoded.append(compute_code(c))
else:
- print(f'{c} already seen')
encoded.append(compute_prefix(tree, c))
-
+ pprint.pp(tree)
pprint.pp(tree)
+
print(f'Nombre de swaps: {swap_count}')
return ' '.join(encoded)[1:]
-def main(string: str) -> None:
- msg = encode(string)
- print(f'Code final: {msg}')
-
-# def main(string: str) -> None:
-# tree = Node()
-# fixed_code = []
-#
-# for c in string:
-# fixed_code.append(compute_code(c))
-#
-# pprint.pp(fixed_code)
-#
-# for c in string:
-# pprint.pp(tree)
-# print(f'\n====== Inserting \'{c}\' ======\n')
-# insert_char(tree, c)
-#
-# pprint.pp(tree)
-#
-# n = find_char(tree, 'o')
-#
-# pprint.pp(n)
+def decode(encoded: str) -> (str, str):
+ bits = encoded.replace(' ', '')
+ tree = Node()
+ result = ''
+ seen = set()
+ i = 0
+
+ message_with_nyt: list[str] = []
+ message_without_nyt: list[str] = []
+
+ while i < len(bits):
+ node = tree
+
+ while node.left or node.right:
+ if bits[i] == '0':
+ node = node.left
+ else:
+ node = node.right
+ i += 1
+
+ if node.value == 'NYT':
+ e, r = compute_vitter_params(len(LIST_ALPHABET))
+ prefix_bits = bits[i:i+e]
+ val = int(prefix_bits, 2)
+
+ if val < r:
+ code_bits = bits[i:i+e+1]
+ k = int(code_bits, 2) + 1
+ i += e + 1
+ else:
+ k = val + r + 1
+ i += e
+
+ char = LIST_ALPHABET[k - 1]
+ result += char
+ seen.add(char)
+ insert_char(tree, char)
+
+ message_with_nyt.append('NYT')
+ message_with_nyt.append(char)
+ message_without_nyt.append(char)
+ else:
+ char = node.value
+ result += char
+ insert_char(tree, char)
-if __name__ == "__main__":
+ message_with_nyt.append(char)
+ message_without_nyt.append(char)
+ pprint.pp(tree)
+
+ pprint.pp(tree)
+
+ return ' '.join(message_with_nyt[1:]), ''.join(message_without_nyt)
+
+
+def main() -> None:
parser = argparse.ArgumentParser(prog='dyn_huffman.py')
- parser.add_argument('string')
+ parser.add_argument('input')
args = parser.parse_args()
- main(args.string)
+ print('=============== Encoding ===============')
+ encoded_input = encode(args.input)
+ print(f'Code final: {encoded_input}')
+
+ print()
+
+ print('=============== Decoding ===============')
+ with_nyt, without_nyt = decode(encoded_input)
+ print(f'Message initial avec NYT : {with_nyt}')
+ print(f'Message initial sans NYT : {without_nyt}')
+
+
+if __name__ == "__main__":
+ main()
--
GitLab