3 ms·

Have any useful Python 3 functions? Contribute them to Quilt

2 points by Richienb 8y ago

eesmith 8y ago
My Mom does quilting, and one of her specialties is selvage quilts, so I have some idea of what good quilts look like. This collection has some very odd functions. Take "colourcode". Why is there a 'longhex' at all? Why not have a 'destinationtype' which is "longhex" in addition to the existing "hex"? That makes it much easier to implement the code as a dispatch table, rather than a lengthy if/else chain (with multiple lower() calls!). For example: _colourcode_ops = { "hex": lambda c: c.hex, "longhex": lambda c: c.hex_l, "hsl": lambda c: c.hsl, ... } def colourcode(startcolourcode, destinationtype): ... c = colour.Color(str(startcolourcode)) op = _colourcode_ops.get(destinationtype.lower()) if op is not None: return op(c) raise ValueError(f"Unsupported destinationtype: {destinationtype}") The 'destinationtype' docstring is confusing because the code normalizes to lowercase, so why do the examples show "HEX, HSL, RGB, red, blue, green, hue" instead of a all upper or all lower? Oh, and I believe raising a ValueError ("Raised when an operation or function receives an argument that has the right type but an inappropriate value, and the situation is not described by a more precise exception") is better than a RuntimeWarning ("Base class for warnings about dubious runtime behavior") since 1) this isn't a warning, and 2) it doesn't deal with runtime behavior in the way Python intends. As another example, consider: def posnegtoggle(number): if bool(number > 0): return number - number * 2 elif bool(number < 0): return number + abs(number) * 2 elif bool(number == 0): return number There's no need for the bool() of a comparison because the reference manual says "Comparisons yield boolean values". And, what's the point? Why would someone use this instead of unary negation? Note that: >>> math.inf inf >>> -math.inf -inf >>> posnegtoggle(math.inf) nan Furthermore, math.nan is a number, but the above returns None for posnegtoggle(math.nan) when it should return the original nan. The isfib() also doesn't expect math.inf as an input value. I think the following function: def isinfinite(variable): return bool(math.isfinite(variable)) should be written: def isinfinite(variable): return not math.isfinite(variable) but again, I don't see the point of calling that function instead of just using "not math.isfinite()" directly. As far as I can tell, "hcf()" always returns 1. Is it supposed to be the same as math.gcd()? I think the loop is supposed to be range(smaller, 0, -1). fracsimplify() should use the fractions built-in module. Then it wouldn't have the bug that fracsimplify(6, 4) reduces to "6/4" instead of "3/2". This is Python 3 so use numerator//2 instead of int(numerator / 2).

2 points by Richienb 8y ago

My Mom does quilting, and one of her specialties is selvage quilts, so I have some idea of what good quilts look like. This collection has some very odd functions. Take "colourcode". Why is there a 'longhex' at all? Why not have a 'destinationtype' which is "longhex" in addition to the existing "hex"? That makes it much easier to implement the code as a dispatch table, rather than a lengthy if/else chain (with multiple lower() calls!). For example: _colourcode_ops = { "hex": lambda c: c.hex, "longhex": lambda c: c.hex_l, "hsl": lambda c: c.hsl, ... } def colourcode(startcolourcode, destinationtype): ... c = colour.Color(str(startcolourcode)) op = _colourcode_ops.get(destinationtype.lower()) if op is not None: return op(c) raise ValueError(f"Unsupported destinationtype: {destinationtype}") The 'destinationtype' docstring is confusing because the code normalizes to lowercase, so why do the examples show "HEX, HSL, RGB, red, blue, green, hue" instead of a all upper or all lower? Oh, and I believe raising a ValueError ("Raised when an operation or function receives an argument that has the right type but an inappropriate value, and the situation is not described by a more precise exception") is better than a RuntimeWarning ("Base class for warnings about dubious runtime behavior") since 1) this isn't a warning, and 2) it doesn't deal with runtime behavior in the way Python intends. As another example, consider: def posnegtoggle(number): if bool(number > 0): return number - number * 2 elif bool(number < 0): return number + abs(number) * 2 elif bool(number == 0): return number There's no need for the bool() of a comparison because the reference manual says "Comparisons yield boolean values". And, what's the point? Why would someone use this instead of unary negation? Note that: >>> math.inf inf >>> -math.inf -inf >>> posnegtoggle(math.inf) nan Furthermore, math.nan is a number, but the above returns None for posnegtoggle(math.nan) when it should return the original nan. The isfib() also doesn't expect math.inf as an input value. I think the following function: def isinfinite(variable): return bool(math.isfinite(variable)) should be written: def isinfinite(variable): return not math.isfinite(variable) but again, I don't see the point of calling that function instead of just using "not math.isfinite()" directly. As far as I can tell, "hcf()" always returns 1. Is it supposed to be the same as math.gcd()? I think the loop is supposed to be range(smaller, 0, -1). fracsimplify() should use the fractions built-in module. Then it wouldn't have the bug that fracsimplify(6, 4) reduces to "6/4" instead of "3/2". This is Python 3 so use numerator//2 instead of int(numerator / 2).