I want to create custom error messages for a function.
def tr( launch_speed , launch_angle_deg , num_samples ):
#Error displays
try:
launch_speed>0
except:
raise Exception("Launch speed has to be positive!")
try:
0<launch_angle_deg<90
except:
raise Exception("Launch angle has to be 0 to 90 degrees!")
try:
um_samples = int(input())
except:
raise Exception("Integer amount of samples!")
try:
num_samples >=2
except:
raise Exception("At least 2 samples!")
Essentially, what I want is to get an error message every time a wrong value has been written in the function variables, and I've tried creating these messages based on what I've gathered on the Internet, but it doesn't seem to work.
You can't use try: except: for everything; for example, launch_speed>0 will not raise an error for negative values. Instead, I think you want e.g.
if launch_speed < 0: # note spacing, and if not try
raise ValueError("Launch speed must be positive.") # note specific error
You should also test for and raise more specific errors (see "the evils of except"), e.g.:
try:
num_samples = int(raw_input()) # don't use input in 2.x
except ValueError: # note specific error
raise TypeError("Integer amount of samples!")
You can see the list of built-in errors in the documentation.
Why not go one step further and build your own exception types? There's a quick tutorial in the docs which could be used something like:
class Error(Exception):
"""Base class for exceptions defined in this module"""
pass
class LaunchError(Error):
"""Errors related to the launch"""
pass
class LaunchSpeedError(LaunchError):
"""Launch speed is wrong"""
pass
class LaunchAngleError(LaunchError):
"""Launch angle is wrong"""
pass
class SamplesError(Error):
"""Error relating to samples"""
pass
In this case the default functionality of Exception is fine, but you may be able to get finer granularity in what you catch by defining extra exceptions.
if launch_speed < 0:
raise LaunchSpeedError("Launch speed must be positive")
if 0 <= launch_angle < 90:
raise LaunchAngleError("Launch angle must be between 0 and 90")
um_samples = input()
try:
um_samples = int(um_samples)
except ValueError:
raise SampleError("Samples must be an integer, not {}".format(um_samples))
if um_samples < 2:
raise SampleError("Must include more than one sample, not {}".format(str(um_samples)))
Related
I'm doing a project that is requiring a lot of input validation. I currently have a function defined as follows:
def get_valid_float(inputMessage,errorMessage):
while True:
variableInput = input(inputMessage)
try:
variableInput = float(variableInput)
return variableInput
except ValueError:
print(errorMessage)
This function allows me to choose a custom message to prompt the user. It will then validate that the user input is indeed a float, and will print a custom error message in the event that it is not. It will loop until the user gives a valid input.
However, I would rather not create a function to validate each and every data type. It seems like it would be best to combine these into one get_valid_input() function, and pass a third argument allowing me to choose what data type I am attempting to verify. For example, get_valid_input(complex,inputMessage,errorMessage).
I am obviously unable to pass a keyword as an argument. This makes me think the only way to do this would to be to do something like this:
def get_valid_float(dataType,inputMessage,errorMessage):
if dataType == "float"
while True:
variableInput = input(inputMessage)
try:
variableInput = float(variableInput)
return variableInput
except ValueError:
print(errorMessage)
elif dataType == "integer"
while True:
variableInput = input(inputMessage)
try:
variableInput = int(variableInput)
return variableInput
except ValueError:
print(errorMessage)
And so on, with an elif for every data type. Surely there is an easier way to do this, that somehow allows me to execute the line variableInput = {dataType}(variableInput) to confirm that they input a value of data type "dataType". Any ideas?
Just pass as an argument the actual data type, rather than the name of the data type. E.g:
def get_valid_input(dataType, inputMessage, errorMessage):
while True:
value = input(inputMessage)
try:
value = dataType(value)
break
except ValueError:
print(errorMessage)
You would call it like this:
floatvalue = get_valid_input(float, "enter a float value: ", "that is an invalid float")
intvalue = get_valid_input(int, "enter an integer value: ", "that is an invalid integer")
I am obviously unable to pass a keyword as an argument.
Not sure why you're saying that, but you can! :)
Also no need for error message, just catch all Exceptions (Not recommended but since you are just printing out the error it seems fine here)
The message strings aren't really needed, try using the name of the dataType and the exception's message like this:
def get_valid_data(dataType):
while True:
variableInput = input(f"Put in data of type {dataType.__name__}: ")
try:
variableInput = dataType(variableInput)
return variableInput
except Exception as e:
print(e)
get_valid_data(int)
>>> Put in data of type int: "hi"
>>> invalid literal for int() with base 10: '"hi"'
The following snipped calls parse_args() to identify a parameter --num and that tests that the value is acceptable in separate code after parsing.
import argparse
def cmd_line_opts(cmdline):
parser = argparse.ArgumentParser(description="blah blah",)
parser.add_argument('--num', default=None, type=int,)
return parser.parse_args(cmdline)
p = cmd_line_opts(['--num', '2'])
if p.num < 1:
raise ValueError('--num must be > 0')
p = cmd_line_opts(['--num', '0'])
if p.num < 1:
raise ValueError('--num must be > 0')
I am wondering if it would be possible to include the test and a suitable error message directly in the parsing. I suspect that this involves using the Action API but I can't figure out how to signal a failure when returning from the action.
You just need a custom type for the argument.
def positive_int(s):
try:
s = int(s)
if s <= 0:
raise ValueError()
except ValueError:
raise ArgumentTypeError("Not a valid positive integer: {}".format(s))
# ...
parser.add_argument("--num", type=positive_int, default=0)
The type argument is any callable that accepts a string and returns a value to use for the argument's value. It doesn't have to be an actual type (built-in or otherwise).
Full credit to #chepner for his answer, above. At least on my system you need qualify ArgumentTypeError. I put it in a comment but thought it would be helpful to someone who just wanted to copy and paste the code.
import argparse
def positive_int(s):
try:
s = int(s)
if s <= 0:
raise ValueError()
except ValueError:
raise argparse.ArgumentTypeError("Not a valid positive integer: {}".format(s))
def cmd_line_opts(cmdline):
parser = argparse.ArgumentParser(description="blah blah",)
parser.add_argument("--num", type=positive_int, default=0)
return parser.parse_args(cmdline)
p = cmd_line_opts(['--num', 'x'])
I loop though a list of currencies in order to download price series from an API and it happens that some of them are not supported so that it raises a module defined exception class : ExchangeError: This currency pair is not supported.
When it occurs I would like to continue the loop to the next currency but for some reason I'm unable to handle the module exception.
Here is an example that works fine with a built-in exception :
f = [1,2,3,4,'A',5]
def foo(nb):
return nb /2
for i in f :
try:
print(foo(i))
except TypeError :
continue
As expected it returns :
0.5
1
1.5
2
2.5
But as soon as it is a module (or user defined) exception it throws an error saying the exception is not defined :
#retry(wait_exponential_multiplier=1000, wait_exponential_max=10000)
def apiFetchOHLC(obj, currency, timeframe, option):
ohlcv = obj().fetch_ohlcv(currency, timeframe, since = option)
return ohlc
for c in currencies_list :
...
try :
# Download data
ohlc = apiFetchOHLC(obj, c, tf, maxCandlesLimit)
# except : # works fine
except ExchangeError : # doesn't work
print("Oops! That was no valid currency. Continue...")
continue
This is the error I get when I run the loop :
except ExchangeError:
NameError: name 'ExchangeError' is not defined
To make it works I need to remove the exception type ExchangeError but to me it is not a workaround because it will continue the loop whatever the exception is, and sometimes I need to retry the download.
How can I achieve this with try and except or with the retrying package ? (link)
def foo(count):
try:
while(count < 10):
if(count%2 == 1):
raise Exception()
print(count)
count = count+1
except:
print( str(count) + ' -> Exception')
foo(count+1)
foo(2)
Whenever an exception occurs in a try block, handle it in except block as follows -
To continue the process you are doing in try block, you should reach the try block from an except block - keep the try block in a function, name it foo, so that you could call it from the except block
To continue from the next iteration, you need to know the previous iteration where the exception has been raised - pass an argument to that function
After identifying the problem better I have found that I needed to give the full name space of the exception class I want to catch:
for c in currencies_list :
...
try :
# Download data
ohlc = apiFetchOHLC(obj, c, tf, maxCandlesLimit)
except ccxt.ExchangeError :
print("Oops! That was no valid currency. Continue...")
continue
i have to check if the content of some variable are of some specific type (it is a sanity check i want to implement). If each one of these is not of the specified type then the program has to abort.
I'm not experienced in python, i've a glance to the official guide on the web where it is stated that it is possible to define own exceptions by class implementation. I don't know if that would be hard or not, but since i don't have to do something particularly complicated is it possible to do that in a simple way?
Update
I've implemented the exception int exception in the following way:
class TypeCheckingException(Exception):
def __init__(self,type_in,type_exp):
self.type_in = type_in
self.type_exp = type_exp
The specific line of the implementation its use are:
try:
if isinstance(var1,str) == 0:
raise TypeCheckingException(type(var1),str)
if isinstance(var2,str) == 0:
raise TypeCheckingException(type(var2),str)
if isinstance(var3,int) == 0:
raise TypeCheckingException(type(var3),int)
if isinstance(var4,int) == 0:
raise TypeCheckingException(type(var4),int)
if isinstance(var5,str) == 0:
raise TypeCheckingException(type(var5),str)
if isinstance(var6,float) == 0:
raise TypeCheckingException(type(var6),float)
except TypeCheckingException, tce:
print "Type expected input " + str(tce.type_in) + " while type expected is " + str(tce.type_exp) + "."
In this case works, but the error output is quite verbose... i'd like to make it shorter...
Any clue?
You want to use isinstance(variable, type).
>>> a = 'str'
>>> isinstance(a, str)
True
You won't need to use exception handling with this route. You can check for instance type, then exit, if necessary.
I tried to write a code that can distinguish the following four different errors.
TypeError: The first parameter is not an integer;
TypeError: The second parameter is not a string;
ValueError: The value of the first parameter is not in the range of 1 to 13; or
ValueError: The value of the second parameter is not one of the strings in the set {'s', 'h', 'c', 'd'}.
However, I only can get the first one to work but not the other three errors. I tried different ways to make it work, but still can't figure out what's wrong.
class Card: # One object of class Card represents a playing card
rank = ['','Ace','Two','Three','Four','Five','Six','Seven','Eight','Nine','Ten','Jack','Queen','King']
suit = {'d':'Diamonds', 'c':'Clubs', 'h':'Hearts', 's':'Spades'}
def __init__(self, rank=2, suit=0): # Card constructor, executed every time a new Card object is created
if type(rank) != int:
raise TypeError()
if type(suit) != str:
raise TypeError()
if rank != self.rank:
raise ValueError()
if suit != 'd' or 'c' or 'h' or 's':
raise ValueError()
self.rank = rank
self.suit = suit
def getRank(self): # Obtain the rank of the card
return self.rank
def getSuit(self): # Obtain the suit of the card
return Card.suit[self.suit]
def bjValue(self): # Obtain the Blackjack value of a card
return min(self.rank, 10)
def __str__(self): # Generate the name of a card in a string
return "%s of %s" % (Card.rank[int(self.rank)], Card.suit[self.suit])
if __name__ == "__main__": # Test the class Card above and will be skipped if it is imported into separate file to test
try:
c1 = Card(19,13)
except TypeError:
print ("The first parameter is not an integer")
except TypeError:
print ("The second parameter is not a string")
except ValueError:
print ("The value of first parameter is not in the range of 1 to 13")
except ValueError:
print ("The value of second parameter is not one of the strings in the set {'s','h','c','d'}")
print(c1)
I know maybe it is due to that I have same TypeError and ValueError. Therefore, Python can't distinguish the second TypeError which I hit c1 = Card(13,13) is different from the first TypeError. So, I only get the message that "The first parameter is not an integer" when I have c1 = Card(13,13).
You are trying to distinguish between the error sources in completely the wrong place. By the time the error gets out of Card.__init__, there is no way to tell why e.g. a TypeError was thrown. For each error class (TypeError, ValueError) only the first except will ever be triggered:
try:
...
except TypeError:
# all TypeErrors end up here
except TypeError:
# this is *never* reached
except ValueError:
# all ValueErrors end up here
except ValueError:
# this is *never* reached
Instead, you should provide the specific error messages inside Card.__init__, when you actually raise the error and already know what the reason is:
if not isinstance(rank, int): # better than comparing to type
raise TypeError("The first parameter is not an integer")
Then you can handle them much more simply:
try:
c1 = Card(19,13)
except (TypeError, ValueError) as err: # assign the error to the name 'err'
print(err) # whichever we catch, show the user the message
else:
print(c1) # only print the Card if there were no errors
If you have a particular need to distinguish between different errors of the same class, you can either explicitly check the message:
except TypeError as err:
if err.args[0] == "The first parameter is not an integer":
# do whatever you need to
or create your own, more specific Exception sub-classes, so you can have separate except blocks:
class FirstParamNotIntError(Exception):
pass
(this is just an example, they are too specific in your particular case).
It's probably worth having a read through the documentation on exceptions and the tutorial on using them.