Back to General discussions forum
Hello, Rodion!
I am sorry, there was indeed a problem with one of my scripts.
I do not mind if you keep account disabled - there is a lot of work with analyzing results anyway.
Sorry again, and a nice day for you all!
No worry, just notify when it is ok to re-enable :)
It was disabled just for technical reason - currently there are again some interesting cases of cheating from some countries/universities we know - and I couldn't understand perhaps they hacked you or something like this :)
No, it is probably not like I'm hacked...
But I'm playing around some interesting codes like CodeAbbeySaver and CodeAbbeySubmitter recently.
My current code is based mostly on these two scripts, but there are also other interesting work.
P. S. The links above are already posted on the forum, in the previous topics.
Thank you for sharing these :)
There is no objection on my side about automatic collection of solutions or submitting without using web-interface - even I vaguely remember there were requests to make such API, but I don't remember if anything was done in this direction. Perhaps I should return to implement the idea.
If not a secret I'm curious about the purpose of your current scripts. My weak understanding was that people may want obviously to backup their sources - and a kind of restore backup in case of necessity to reconstruct/clone an account, or saved sources were inadvertently damaged etc... But I don't know for sure.
If not a secret I'm curious about the purpose of your current scripts.
I have two scripts for now: one to backup the solutions, and second to automatically resubmit these solutions if needed. So yes, a sort of backup feature.
Rodion, please reenable submissions for a moment...
I suspect that checker for task 34 is incorrect, but probably need more submissions to investigate.
Example. For input data:
a = 2.52990223
b = 0.95498318
c = 38.48936504
d = 1106.58265989
Your checker expects the following answer:
x = 94.260370512149
Which yields:
y = 0.00000382549
Correct answer is:
x = 94.26034028106201
With:
y = 0.00000000000
And the difference between checker answer and correct answer is:
dx = 0.00000023108699
Which is bigger than tolerance stated in problem statement (1e-7
).
But there are so many successful submissions... I wonder if actual tolerance is 1e-6
or so, but cannot check due to current account restrictions...
Hi Friend! sure, the block is relieved now, you are welcome!
There is some confusion about tolerance, I'm sorry to say. All these problems which want 1e-7
tolerance actually mean
relative error, e.g. it should be less than expected_value * 1e-7
. It's here in the code,
if it clarifies better. Perhaps I should improve wording in the corresponding tasks, but I generally felt that if the
submission achieves 1e-7
absolute tolerance it will fit anyway :)
Everything seems ok, sorry for bothering you!
Made 40 submissions, 20 with one solution, 20 with other, both solutions have 100% success rate.
So it is just a problem with me, my offline validator is using absolute tolerance...
I rewrote my offline validator using the code of verifyFloatAnswer
function from link above, and offline tests are passing as well. Great, thank you, Rodion!
Are there any particular limits on collecting test data for tasks? Or it is not allowed at all?