|
Thread Rules
1. This is not a "do my homework for me" thread. If you have specific questions, ask, but don't post an assignment or homework problem and expect an exact solution.
2. No recruiting for your cockamamie projects (you won't replace facebook with 3 dudes you found on the internet and $20)
3. If you can't articulate why a language is bad, don't start slinging shit about it. Just remember that nothing is worse than making CSS IE6 compatible.
4. Use [code] tags to format code blocks. |
On October 21 2017 01:43 travis wrote:zigg, like your idea, that makes sense hanh, i took a glance. I'll look at it a bit closer at some of the algorithms for that and see if I can apply them. Show nested quote +On October 20 2017 21:15 Manit0u wrote:
Just remove all 1's that are adjacent to 2's and then remove the 2's?
Edit:
Simpler yet:
1. Find all 2's
2. Make an adjacency list for 2's (ignoring duplicates) which is dead simple for 2d array
3. Profit! I don't understand your first suggestion. For the 2nd suggestion, you are saying to 1.) put each 2_location[x][y] into a list 2.) then for each element of the list create a sublist of each adjacent_1_location[x][y] 3.) and then for the next iteration I can just repeat for the elements in the sublist? 4.) repeat 2 and 3 until I am done? Is this correct? If so, I really like this. And you'd probably mean for me to avoid adding the same adjacent_1_location[x][y] into 2 different sublists, but actually, because of the nature of what I am trying to do with this problem it is actually solving a future problem for me to just put them in there anyways. Long term I was going to need to do math based on the relationship of every square of ring 2 that touches any given square of ring 1. So this is PERFECT for that.
Here's how to do it step by step:
1. Note the location of all 2's (x, y).
2. Note the location of all elements adjacent to those 2's (we don't care if it's 0, 1, 2).
3. Set all those locations to 0.
Example:
1 1 1 1
1 1 2 1
1 2 2 1
1 1 1 1
grid_size = 4
loc_2 = [ [1, 2], [2, 1], [2, 2] ]
def find_adjacent_to(coords)
adj = []
adj << [ coords[0], coords[1] - 1] if coords[1] > 0 # left
adj << [ coords[0], coords[1] + 1] if coords[1] < grid_size # right
adj << [ coords[0 - 1], coords[1]] if coords[0] > 0 # top
adj << [ coords[0] + 1, coords[1]] if coords[0] < grid_size # bottom
adj
end
# all cells of interest are loc_2 + find_adjacent for each loc_2
This is pretty naive example but it gives you every single cell you need to work with by traversing the grid just once (to find the 2's).
All you're working with later is a specific set of coordinates being just a subset of the grid, which makes your work that much easier.
On October 21 2017 04:55 spinesheath wrote:
I'm reading your solution as
^$|^b*(ab)*b*(a?)$
Is that right? Not sure what those dots below the stars mean here.
Anyways, my counterexample would be:
abbbabbbab
Reason: you only allow a single occurrence of the a - many b's - a pattern. The NFA allows an arbitrary number of those by going 100010001000...
My solution:
^(b*(ab)*)*(a?)$
Also you shouldn't need to explicitly have the epsilon case when everything in your regex is optional.
This is not in line with what he stated earlier in the question. Travis marked "ababb" as acceptable and this will fail on such expression (you can't start with "a").
Here's a proper regexp for that:
^(?:(?!aa)[ab])+$
|
I am not familiar with the "?:" "?:". Is this language specific ?
But I think spinesheath's solution actually does work
for ababb he sees
(b*(ab*))
which he uses to construct abab
but outside of (b*(ab)*) we have another * (you may have missed this)
so we run through again (b*(ab)*) again to get our final b
and then have ababb
|
|
|
yeah, I think that is the most concise version.
|
On October 21 2017 22:29 travis wrote:
I am not familiar with the "?:" "?:". Is this language specific ?
But I think spinesheath's solution actually does work
for ababb he sees
(b*(ab*))
which he uses to construct abab
but outside of (b*(ab)*) we have another * (you may have missed this)
so we run through again (b*(ab)*) again to get our final b
and then have ababb
"?:" is a non-capturing group in the regex (when you want to know if it matches but don't care about the result). "?!" is negative look ahead (in the regex I posted it allows any combination of letters "a" and "b" as long as there are no 2+ letters "a" adjacent to each other).
|
No offence but ^(?:(?!aa)[ab])+$ ...
1. it is not obvious why it should be equivalent to the NFA,
2. in fact it is not equivalent, the empty string is acceptable in the NFA but not in the regex,
3. it uses uncommon regex features, the ?! part
4. it does things that are not required: nothing about capturing group was mentioned in the question
5. the ?! has performance issue because of the lookahead backtrack
|
On October 21 2017 19:59 Manit0u wrote:Show nested quote +On October 21 2017 04:55 spinesheath wrote:
I'm reading your solution as
^$|^b*(ab)*b*(a?)$
Is that right? Not sure what those dots below the stars mean here.
Anyways, my counterexample would be:
abbbabbbab
Reason: you only allow a single occurrence of the a - many b's - a pattern. The NFA allows an arbitrary number of those by going 100010001000...
My solution:
^(b*(ab)*)*(a?)$
Also you shouldn't need to explicitly have the epsilon case when everything in your regex is optional. This is not in line with what he stated earlier in the question. Travis marked "ababb" as acceptable and this will fail on such expression (you can't start with "a"). Here's a proper regexp for that:
^(?:(?!aa)[ab])+$
I fail to see why my regex wouldn't match ababb. Especially since an online regex tester confirms it does.
I'm fairly confident that stuff like negative lookahead is not really a topic in university at that level (or any level at all). As far as I know it is not part of the formal definition of regex.
Also I need to play more regex golf. (b|ab)*a? is so clean and honestly, it should have been obvious.
|
On October 22 2017 11:25 Hanh wrote:No offence but ^(?:(?!aa)[ab])+$ ... 1. it is not obvious why it should be equivalent to the NFA, 2. in fact it is not equivalent, the empty string is acceptable in the NFA but not in the regex, 3. it uses uncommon regex features, the ?! part 4. it does things that are not required: nothing about capturing group was mentioned in the question 5. the ?! has performance issue because of the lookahead backtrack
Haha, yeah. I'm way over my head with regexes lately. Still trying to figure out the best way to validate allowed path and filenames on Windows with it without much success, there's so many conditions that I keep running into catastrophic backtrace with strings that are > 12 characters long and have failing condition at the end.
On October 22 2017 16:14 spinesheath wrote:Show nested quote +On October 21 2017 19:59 Manit0u wrote:On October 21 2017 04:55 spinesheath wrote:
I'm reading your solution as
^$|^b*(ab)*b*(a?)$
Is that right? Not sure what those dots below the stars mean here.
Anyways, my counterexample would be:
abbbabbbab
Reason: you only allow a single occurrence of the a - many b's - a pattern. The NFA allows an arbitrary number of those by going 100010001000...
My solution:
^(b*(ab)*)*(a?)$
Also you shouldn't need to explicitly have the epsilon case when everything in your regex is optional. This is not in line with what he stated earlier in the question. Travis marked "ababb" as acceptable and this will fail on such expression (you can't start with "a"). Here's a proper regexp for that:
^(?:(?!aa)[ab])+$
I fail to see why my regex wouldn't match ababb. Especially since an online regex tester confirms it does. I'm fairly confident that stuff like negative lookahead is not really a topic in university at that level (or any level at all). As far as I know it is not part of the formal definition of regex. Also I need to play more regex golf. (b|ab)*a? is so clean and honestly, it should have been obvious.
The problem with this regex is that it allows multiples of "a" in succession (and I thought the goal was not to allow them).
|
Is it possible with cmake to have multiple independant executables with different names? I have a folder of very small files where I like to try things out (think less then 1k loc) and I'd really like to not have to organise this disaster.
|
On October 23 2017 15:06 Manit0u wrote:
The problem with this regex is that it allows multiples of "a" in succession (and I thought the goal was not to allow them).
Hmm ... no it doesn't... Besides, the goal is not to avoid multiple "a" in succession but to translate the NFA.
|
On October 23 2017 17:40 Hanh wrote:Show nested quote +On October 23 2017 15:06 Manit0u wrote:
The problem with this regex is that it allows multiples of "a" in succession (and I thought the goal was not to allow them). Hmm ... no it doesn't... Besides, the goal is not to avoid multiple "a" in succession but to translate the NFA.
Well, the NFA doesn't allow multiple "a"s in succession... 
Anyway, spinesheath got by far the most elegant solution (and it was really obvious). I was wondering where the hell you guys found the problem, but I just figured out that travis had an image link that I had clicked on, but didn't open properly on my phone.
|
On October 24 2017 02:17 Acrofales wrote:Show nested quote +On October 23 2017 17:40 Hanh wrote:On October 23 2017 15:06 Manit0u wrote:
The problem with this regex is that it allows multiples of "a" in succession (and I thought the goal was not to allow them). Hmm ... no it doesn't... Besides, the goal is not to avoid multiple "a" in succession but to translate the NFA. Well, the NFA doesn't allow multiple "a"s in succession... Anyway, spinesheath got by far the most elegant solution (and it was really obvious). I was wondering where the hell you guys found the problem, but I just figured out that travis had an image link that I had clicked on, but didn't open properly on my phone.
Whoa hold it. That's Hanh's solution.
|
On October 23 2017 17:36 bo1b wrote:
Is it possible with cmake to have multiple independant executables with different names? I have a folder of very small files where I like to try things out (think less then 1k loc) and I'd really like to not have to organise this disaster.
add_executable(exec_a a1.c a2.c)
add_executable(exec_b b1.c b2.c)
?
|
After reading the intro, still not sure what that framework is like. The intro is basically a docker-compose + doctrine tutorial.
|
On October 24 2017 03:02 Blisse wrote:Show nested quote +On October 23 2017 17:36 bo1b wrote:
Is it possible with cmake to have multiple independant executables with different names? I have a folder of very small files where I like to try things out (think less then 1k loc) and I'd really like to not have to organise this disaster.
add_executable(exec_a a1.c a2.c)
add_executable(exec_b b1.c b2.c)
?
I tried doing something similar and had an error come up. Tried doing it again and it works. :\ I feel a little silly now lol.
|
Does anyone have experience with PyQt exit codes? I am running PyQt5 and python 3.6.2 and my app crashes with some exit codes like: "Process finished with exit code 3" not usual python traceback info. This makes it really hard to debug since i cant find any explantion for those codes.
I think this is PyQt thing and the only solution i found on net is overwriting exeption handler.
Any ideas?
|
|
|
On October 24 2017 08:52 LG)Sabbath wrote:After reading the intro, still not sure what that framework is like. The intro is basically a docker-compose + doctrine tutorial.
It's pretty much Symfony in API-only mode 
|
On October 25 2017 04:16 Acrofales wrote:Don't use PyQT, but seems this is relevant: https://stackoverflow.com/questions/33736819/pyqt-no-error-msg-traceback-on-exitHave you tried just putting your entire script between try catch and printing the exception and callback? Reading this seems like PyQT treats this as a last resource if the exception isn't being handled, and exits with an information-less error code.
Yeah i found this also. I think i will try to run line by line and if that doesnt help try override sys.execpthook. Still there is number provided in that exit process message (its 0 for example if application finish without errors), so i wondered if anyone knows what 3 stands for in my example, thats probably some group of errors. Knowning that could probably help me a lot.
|
Completely unrelated:
module Api::V1::Media::JobProcessingService
def self.find_all_processes_for_job(job)
subquery = Api::V1::Media::Process.select(:id).where(
process_step_id: Api::V1::Media::ProcessStep.select(:id).where(
job_id: job.id
)
)
query = lambda { |model|
model
.joins(process: :process_step)
.select("#{model.table_name}.*, api_v1_media_process_steps.position")
.where(process_id: subquery)
.order('api_v1_media_process_steps.position ASC')
}
{
automated_processes: query.call(Api::V1::Media::AutomatedProcess),
manual_processes: query.call(Api::V1::Media::ManualProcess)
}
end
end
Resulting query (code returns objects):
SELECT api_v1_media_automated_processes.*, api_v1_media_process_steps.position
FROM "api_v1_media_automated_processes"
INNER JOIN "api_v1_media_processes" ON "api_v1_media_processes"."id" = "api_v1_media_automated_processes"."process_id"
INNER JOIN "api_v1_media_process_steps" ON "api_v1_media_process_steps"."id" = "api_v1_media_processes"."process_step_id"
WHERE "api_v1_media_automated_processes"."process_id" IN (
SELECT "api_v1_media_processes"."id"
FROM "api_v1_media_processes"
WHERE "api_v1_media_processes"."process_step_id" IN (
SELECT "api_v1_media_process_steps"."id"
FROM "api_v1_media_process_steps"
WHERE "api_v1_media_process_steps"."job_id" = $1
)
)
ORDER BY api_v1_media_process_steps.position ASC
Working with modern languages and frameworks can be really rewarding at times.
|
|
|
|
|
|
EPT
