Re: [Tutor] problem with creating paths
On 17/10/18 18:18, Mats Wichmann wrote: On 10/17/2018 10:07 AM, Shall, Sydney via Tutor wrote: Firstly, I would like to thank Steven for reminding me of the assert statement. I should have remembered this. It allowed me to isolate the problem, which predictably (for me) was very elementary. I am too embarrassed to say how simple the error was. However, my original problem was not solved by correcting this error. So, I will now try and narrow down the location of the problem and then if I cannot solve it, I shall return for more good advice. Many thanks to Steven and to Peter. I'll weigh in with a mini- (and unasked-for-) lecture here: this is often the point at which someone says "boy, I wish Python were strongly typed, so things didn't change types in flight". This is where I say Python *IS* strongly, although dynamically typed. Why do people have such a problem with this? -- My fellow Pythonistas, ask not what our language can do for you, ask what you can do for our language. Mark Lawrence ___ Tutor maillist - Tutor@python.org To unsubscribe or change subscription options: https://mail.python.org/mailman/listinfo/tutor
Re: [Tutor] problem with creating paths
On 17/10/2018 18:18, Mats Wichmann wrote: On 10/17/2018 10:07 AM, Shall, Sydney via Tutor wrote: Firstly, I would like to thank Steven for reminding me of the assert statement. I should have remembered this. It allowed me to isolate the problem, which predictably (for me) was very elementary. I am too embarrassed to say how simple the error was. However, my original problem was not solved by correcting this error. So, I will now try and narrow down the location of the problem and then if I cannot solve it, I shall return for more good advice. Many thanks to Steven and to Peter. I'll weigh in with a mini- (and unasked-for-) lecture here: this is often the point at which someone says "boy, I wish Python were strongly typed, so things didn't change types in flight". But in fact, the list didn't change types, it's still a list. In fact we even know where that list is: it's the first element of that tuple you ended up with. The _name_ you gave to that list carries no typing meaning, however (although you can give it type hints that an external tool could use to warn you that you are changing something). So you have somewhere given that name to a completely different object, a tuple which contains your list and another element. So clearly what you're looking for is the place that happens. So here's a sketch of how you might use type hinting to find this, to bring it back to something practical: === types.py: from typing import List, Tuple a: List[str] = [ '/a/path', '/b/path', ] print(type(a)) print(a) a = (a, '/c/path') print(type(a)) print(a) === this works just fine: $ python3 types.py ['/a/path', '/b/path'] (['/a/path', '/b/path'], '/c/path') === but a hinting tool can see a possible issue: $ mypy types.py types.py:10: error: Incompatible types in assignment (expression has type "Tuple[List[str], str]", variable has type "List[str]") Thanks for this. Most helpful. Sydney ___ Tutor maillist - Tutor@python.org To unsubscribe or change subscription options: https://emea01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fmail.python.org%2Fmailman%2Flistinfo%2Ftutor&data=01%7C01%7Csydney.shall%40kcl.ac.uk%7Ca58896d72c374cda3c1508d63454b783%7C8370cf1416f34c16b83c724071654356%7C0&sdata=OHZtedYdy0UHKDagLO1TI%2BUIjEJuzRjZjD4HRdSQmNI%3D&reserved=0 -- _ Professor Sydney Shall Department of Haematology/Oncology Phone: +(0)2078489200 E-Mail: sydney.shall [Correspondents outside the College should add @kcl.ac.uk] ___ Tutor maillist - Tutor@python.org To unsubscribe or change subscription options: https://mail.python.org/mailman/listinfo/tutor
Re: [Tutor] problem with creating paths
On 10/17/2018 10:07 AM, Shall, Sydney via Tutor wrote: > Firstly, I would like to thank Steven for reminding me of the assert > statement. I should have remembered this. It allowed me to isolate the > problem, which predictably (for me) was very elementary. I am too > embarrassed to say how simple the error was. > > However, my original problem was not solved by correcting this error. > > So, I will now try and narrow down the location of the problem and then > if I cannot solve it, I shall return for more good advice. > > Many thanks to Steven and to Peter. I'll weigh in with a mini- (and unasked-for-) lecture here: this is often the point at which someone says "boy, I wish Python were strongly typed, so things didn't change types in flight". But in fact, the list didn't change types, it's still a list. In fact we even know where that list is: it's the first element of that tuple you ended up with. The _name_ you gave to that list carries no typing meaning, however (although you can give it type hints that an external tool could use to warn you that you are changing something). So you have somewhere given that name to a completely different object, a tuple which contains your list and another element. So clearly what you're looking for is the place that happens. So here's a sketch of how you might use type hinting to find this, to bring it back to something practical: === types.py: from typing import List, Tuple a: List[str] = [ '/a/path', '/b/path', ] print(type(a)) print(a) a = (a, '/c/path') print(type(a)) print(a) === this works just fine: $ python3 types.py ['/a/path', '/b/path'] (['/a/path', '/b/path'], '/c/path') === but a hinting tool can see a possible issue: $ mypy types.py types.py:10: error: Incompatible types in assignment (expression has type "Tuple[List[str], str]", variable has type "List[str]") ___ Tutor maillist - Tutor@python.org To unsubscribe or change subscription options: https://mail.python.org/mailman/listinfo/tutor
Re: [Tutor] problem with creating paths
Firstly, I would like to thank Steven for reminding me of the assert statement. I should have remembered this. It allowed me to isolate the problem, which predictably (for me) was very elementary. I am too embarrassed to say how simple the error was. However, my original problem was not solved by correcting this error. So, I will now try and narrow down the location of the problem and then if I cannot solve it, I shall return for more good advice. Many thanks to Steven and to Peter. Sydney ___ Tutor maillist - Tutor@python.org To unsubscribe or change subscription options: https://mail.python.org/mailman/listinfo/tutor
Re: [Tutor] problem with creating paths
Shall, Sydney via Tutor wrote:
> There are two items that are 'wrong' in this output.
>
> 1. The property 'paths' is defined in the program as a list and the
> items are added using paths.append(), yet the test says that when tested
> it is a tuple.
>>> paths = ["foo", "bar"],
>>> paths += "baz", # ("baz",) leads to the same result
>>> paths
(['foo', 'bar'], 'baz')
This might happen if you have accidental trailing commas in two places, but
that seems rather unlikely. Maybe you have a function with a star
def f(*args):
... # args is a tuple, even if you pass one argument or no arguments
# to f().
?
> 2. The tuple arises by the addition of the last entry in the file, AFTER
> the closing bracket of the list which is the first item in the tuple.
Sorry, I don't understand that sentence.
> When I test the length of 'paths' I get a value of 2!
That's because you have a tuple with two entries, one list and one string.
> I apologise for the lengthy explanation, but I am at a loss.
>
> I have looked for an error that might have added an item as a + and I
> find nothing.
> The character of the final item is also puzzling to me.
>
> I would much appreciate any guidance as to how I should search for the
> fault or error.
Look at the assignments. If you find the pattern I mentioned above remove
the commas.
If you don't see anything suspicious make a test script containing only the
manipulation of the paths variable. Make sure it replicates the behaviour of
the bigger script; then show it to us.
___
Tutor maillist - Tutor@python.org
To unsubscribe or change subscription options:
https://mail.python.org/mailman/listinfo/tutor
[Tutor] problem with creating paths
I can now add to my previous email the following observation. If I do not delete the output file and redo the test I get the following as the 'extra' entry in paths: '/Users/sydney/AnacondaProjects/capital_reproduction/Current_Version/Results/20181017D/B_Cycle_Zero/Text_Files') If however, I delete the output file and then redo the test I get the following as the 'extra' entry in paths: '/Users/sydney/.Trash/20181017D/B_Cycle_Zero/Text_Files') This seems to be consistent. The upper incorrect entry is item path19 in the list part of paths. I have studied every example of 'path19' in the program and I cannot find an explanation. help! Sydney _ Professor Sydney Shall Department of Haematology/Oncology Phone: +(0)2078489200 E-Mail: sydney.shall [Correspondents outside the College should add @kcl.ac.uk] ___ Tutor maillist - Tutor@python.org To unsubscribe or change subscription options: https://mail.python.org/mailman/listinfo/tutor
Re: [Tutor] problem with creating paths
On Wed, Oct 17, 2018 at 03:18:00PM +0100, Shall, Sydney via Tutor wrote: [...] > After searching I have found this unexpected output illustrated in the > copy-paste below. > > > test > > > The type of the paths is: Believe Python when it tells you something is a tuple. Trust me, it knows! > The values of the paths are : > > > ([ > '/Users/sydney/AnacondaProjects/reproduction/Current_Version/Results/', > '/Users/sydney/AnacondaProjects/reproduction/Current_Version/Results/20181017D', > '/Users/sydney/AnacondaProjects/reproduction/Current_Version/Results/20181017D/A_POCI_Input_Data', [...] > > '/Users/sydney/AnacondaProjects/reproduction/Current_Version/Results/20181017D/B_Cycle_Zero/Text_Files', [...] > '/Users/sydney/AnacondaProjects/reproduction/Current_Version/Results/20181017D/C_Final_Results/Plots/Population_Data/Ratios'], > '/Users/sydney/.Trash/20181017D/B_Cycle_Zero/Text_Files') Here is a good trick for debugging. Start by simplifying your data, to make it easier to see the needle in the haystack. Long paths like the above present you with a wall of text, so (temporarily) replace each path with a single character to cut down on the visual complexity. That makes it easier to see what is going on. Replacing each path with a single letter gives me: (['a', 'b', 'c', [...] 'x', 'y'], 'z') Notice that your value is a tuple of two items: Item One is a list, ['a', 'b', 'c', [...] 'x', 'y'] Item Two is a string, 'z' > There are two items that are 'wrong' in this output. > > 1. The property 'paths' is defined in the program as a list and the > items are added using paths.append(), yet the test says that when tested > it is a tuple. > 2. The tuple arises by the addition of the last entry in the file, AFTER > the closing bracket of the list which is the first item in the tuple. > > When I test the length of 'paths' I get a value of 2! That's because it is a tuple of two items. > I apologise for the lengthy explanation, but I am at a loss. > > I have looked for an error that might have added an item as a + and I > find nothing. Without seeing your code, there's no way of telling how you constructed this value. You intended a list, and built a list (somehow!), but then you did *something* to replace it with a tuple. Perhaps you did: paths = [] for some_path in something_or_rather: paths.append(some_path) then later on: paths = (paths, another_path) but there's a million ways you could have got the same result. And of course you could have used any variable name... I'm assuming it is called "paths", but you should substitute whatever name (or names!) you actually used. > I would much appreciate any guidance as to how I should search for the > fault or error. Start by looking for any line of code that starts with: paths = and see if and where you replaced the list with a tuple. If that gets you nowhere, start looking for *every* reference to "paths" and see what they do. If *that* gets you nowhere, start adding debugging code to your program. Put assertions like this: assert isinstance(paths, list) in various parts of the code, then run the program and see where it fails. That tells you that *at that point* paths is no longer a list. E.g. something like this: # build the paths... paths = [] for blah blah blah blah: paths.append(whatever) assert isinstance(paths, list) # line 50 (say) do_this() do_that() assert isinstance(paths, list) # line 53 do_something_else() and_another_thing() assert isinstance(paths, list) # line 56 If the first two assertions on line 50 and 53 pass, but the third at line 56 fails, you know that the bug is introduced somewhere between line 53 and 56. -- Steve ___ Tutor maillist - Tutor@python.org To unsubscribe or change subscription options: https://mail.python.org/mailman/listinfo/tutor
[Tutor] problem with creating paths
I am a novice (at programming). I use MAC OS 10.13.6 Anaconda. Python 3.5.4 Spyder 3.5.6 I have just re-written a moderately complex program (a Class) on the advice of Alan and Steven. The rewriting proved to be very useful. The working program uses instances of the Class with User chosen parameters. The output data seems correct to me. So I then began redoing all the tests. The first Methods tested gave OK. But I have just started testing a new Method and I get a Universal error in my tests. It says that the output file is already present. These files (paths) are correctly deleted by the 'teardown' Method, when only the earlier portion of the program is tested. But are not deleted with the last method tested. After searching I have found this unexpected output illustrated in the copy-paste below. test The type of the paths is: The values of the paths are : ( [ '/Users/sydney/AnacondaProjects/reproduction/Current_Version/Results/', '/Users/sydney/AnacondaProjects/reproduction/Current_Version/Results/20181017D', '/Users/sydney/AnacondaProjects/reproduction/Current_Version/Results/20181017D/A_POCI_Input_Data', . . . . '/Users/sydney/AnacondaProjects/reproduction/Current_Version /Results/20181017D/B_Cycle_Zero/Text_Files', . . . '/Users/sydney/AnacondaProjects/reproduction/Current_Version/Results/20181017D/C_Final_Results/Plots/Population_Data/Ratios'], '/Users/sydney/.Trash/20181017D/B_Cycle_Zero/Text_Files') There are two items that are 'wrong' in this output. 1. The property 'paths' is defined in the program as a list and the items are added using paths.append(), yet the test says that when tested it is a tuple. 2. The tuple arises by the addition of the last entry in the file, AFTER the closing bracket of the list which is the first item in the tuple. When I test the length of 'paths' I get a value of 2! I apologise for the lengthy explanation, but I am at a loss. I have looked for an error that might have added an item as a + and I find nothing. The character of the final item is also puzzling to me. I would much appreciate any guidance as to how I should search for the fault or error. Sydney _ ___ Tutor maillist - Tutor@python.org To unsubscribe or change subscription options: https://mail.python.org/mailman/listinfo/tutor
