Add HalfPrecision class for float representation

Week06/halfprecision_recepkadir_altıntaş.py

@@ -0,0 +1,46 @@
+import math
+
+class HalfPrecision:
+    def __init__(self, number):
+        if not isinstance(number, float):
+            raise TypeError("Input must be a float")
+        self.number = number
+
+    def __str__(self):
+        if math.copysign(1, self.number) == -1:
+            sign_bit = "1"
+        else:
+            sign_bit = "0"
+
+        value = abs(self.number)
+
+        if value == 0.0:
+            return sign_bit + "00000" + "0000000000"
+
+        if math.isinf(value):
+            return sign_bit + "11111" + "0000000000"
+
+        m, e = math.frexp(value)
+        normalized_mantissa = m * 2
+        exponent = e - 1
+        biased_exponent = exponent + 15
+
+        if biased_exponent >= 31:
+            return sign_bit + "11111" + "0000000000"
+
+        if biased_exponent <= 0:
+            return sign_bit + "00000" + "0000000000"
+
+        fractional_part = normalized_mantissa - 1
+        mantissa_int = int(round(fractional_part * 1024))
+
+        if mantissa_int == 1024:
+            mantissa_int = 0
+            biased_exponent += 1
+            if biased_exponent >= 31:
+                return sign_bit + "11111" + "0000000000"
+
+        exponent_str = f"{biased_exponent:05b}"
+        mantissa_str = f"{mantissa_int:010b}"
+
+        return sign_bit + exponent_str + mantissa_str