HW2 added

lesson_2/main1.py

@@ -0,0 +1,61 @@
+import math
+import sys
+
+def index_to_bytes(value, bits):
+    n = bits // 8
+    if bits % 8 != 0:
+        n += 1
+    return value.to_bytes(n, 'little')
+
+
+if not (2 <= len(sys.argv) <= 3):
+    print('Usage: python', sys.argv[0], 'input_file (output_file)')
+    exit(-1)
+
+with open(sys.argv[1], 'rb') as file:
+    data = file.read()
+
+dictionary = {(i,): i for i in range(256)}
+sequence = list()
+
+outpath = sys.argv[2] if len(sys.argv) == 3 else sys.argv[1] + '.lzw'
+# cond ? a : b
+# a if cond else b
+# if len(sys.argv) == 3:
+#     outpath = sys.argv[2]
+# else:
+#     outpath = sys.argv[1] + '.lzw'
+with open(outpath, 'wb') as file:
+    for sym in data:
+        sequence.append(sym)
+        key = tuple(sequence)
+        if key in dictionary:
+            continue
+
+        n = math.ceil(math.log2(len(dictionary)))
+        index_value = dictionary[key[:-1]]
+        enc_value = index_to_bytes(index_value, n)
+        file.write(enc_value)
+
+        value = len(dictionary)
+        dictionary[key] = value
+        # print(value, '->', key)
+
+        sequence = sequence[-1:]
+
+    n = math.ceil(math.log2(len(dictionary)))
+    index_value = dictionary[tuple(sequence)]
+    enc_value = index_to_bytes(index_value, n)
+    file.write(enc_value)
+
+        # [1, 2, 3]
+        # [3]
+
+        # a
+        # a[b] -> [a] (ab) -> dict
+        # b[e] -> [b] (be) -> dict
+        # e
+
+        # a
+        # ab
+        # ab[c] -> [ab] (abc) -> dict

lesson_2/main1_dec.py

@@ -0,0 +1,45 @@
+import sys
+import math
+
+
+if not (2 <= len(sys.argv) <= 3):
+    print('Usage: python', sys.argv[0], 'input_file (output_file)')
+    exit(-1)
+
+dictionary = {i: (i,) for i in range(256)}
+
+inpath = sys.argv[1]
+outpath = sys.argv[2] if len(sys.argv) > 2 else sys.argv[1][:-4][:sys.argv[1].find('.')] + '_dec' + \
+                                                sys.argv[1][:-4][sys.argv[1].find('.'):]
+
+sequence = []
+
+with open(inpath, 'rb') as file_in:
+    with open(outpath, 'wb') as file_out:
+        bites = math.ceil(math.log2(len(dictionary)))
+        bytes = bites // 8
+        if bytes % 8 == 0:
+            bytes += 1
+        data = file_in.read(bytes)
+        sym = int.from_bytes(data, 'little')
+        file_out.write(bytearray(dictionary[sym]))
+        sequence.append(dictionary[sym][0])
+        sequence.append(sym)
+        sequence = list(dictionary[sym])
+        while True:
+            bites = math.ceil(math.log2(len(dictionary)))
+            bytes_in_cycle = bites // 8
+            bytes_in_cycle += 1
+            sym = file_in.read(bytes_in_cycle)
+            if not sym:
+                break
+            sym = int.from_bytes(sym, 'little')
+            if sym in dictionary:
+                file_out.write(bytearray(dictionary[sym]))
+                sequence.append(dictionary[sym][0])
+                dictionary[len(dictionary)] = tuple(sequence)
+                sequence = list(dictionary[sym])
+            else:
+                sequence.append(sequence[0])
+                file_out.write(bytearray(sequence))
+                dictionary[len(dictionary)] = tuple(sequence)