Where communities thrive

Join over 1.5M+ people
Join over 100K+ communities
Free without limits
Create your own community

Explore more communities

siddhartha-gadgil/Superficial

People

Repo info

Activity

Aug 10 2021 05:06

siddhartha-gadgil opened #99
Aug 10 2021 04:50

siddhartha-gadgil closed #94
Aug 10 2021 04:49

siddhartha-gadgil closed #96
Aug 07 2021 03:43

siddhartha-gadgil on master

helpers for timing (compare)
Aug 06 2021 05:19

siddhartha-gadgil on master

"require" correction slight refactoring (compare)
Aug 05 2021 05:22

siddhartha-gadgil on master

helpers for curves being pants … (compare)
Aug 05 2021 02:47

siddhartha-gadgil on master

figures for pants length 2 (compare)
Aug 04 2021 10:33

siddhartha-gadgil on master

reduced laziness to run in REPL (compare)
Aug 04 2021 09:50

siddhartha-gadgil on master

segment lengths 1 and 2 bound (compare)
Aug 04 2021 05:26

siddhartha-gadgil on master

parallel normal path enumeration (compare)
Aug 03 2021 10:41

siddhartha-gadgil on master

margulis constant corrected (compare)
Jul 16 2021 06:40

siddhartha-gadgil on master

Updates to function shortPathsD… Merge pull request #98 from sha… (compare)
Jul 16 2021 06:40

siddhartha-gadgil closed #98
Jul 16 2021 06:38

shabarishch opened #98
Jul 16 2021 05:21

siddhartha-gadgil closed #97
Jul 16 2021 05:21

siddhartha-gadgil on master

introducing compact types More compact types Merge pull request #97 from sha… (compare)
Jul 16 2021 05:16

shabarishch opened #97
Jul 05 2021 10:55

siddhartha-gadgil commented #96
Jul 05 2021 10:52

siddhartha-gadgil on master

speedup by memoizing hash (compare)
Jul 05 2021 10:16

siddhartha-gadgil commented #94

Siddhartha Gadgil

@siddhartha-gadgil

It is the region between them that needs to be determined.

Arka Ghosh

@anotherArka

Arka Ghosh

@anotherArka

I have started working on crossings. Is anyon else working on it?

Arka Ghosh

@anotherArka

Got this warning

[warn] /media/arkaghosh/Important5/ARKA/10th sem/Computer assisted topology and geometry/Superficial/superficial/src/main/scala/superficial/EdgePath.scala:725:34: method - in trait SetOps is deprecated (since 2.13.0): Use &- with an explicit collection argument instead of - with varargs
[warn]       val edgeVect = (quad.edges - (e1, e1.flip, e2, e2.flip)).toVector
[warn]                                  ^
[warn] one warning found
[info] Done compiling.

Is it important?

Siddhartha Gadgil

@siddhartha-gadgil

Not important, but it would be good if you can switch to a non-deprecated method. Or I can do this if it is not clear to you what you do.

Arka Ghosh

@anotherArka

I dont know what deprecated means in this context. It's better if you do this.

Siddhartha Gadgil

@siddhartha-gadgil

Sure.

Deprecated means that it is recommended that you do not use this, and it will be removed in a future version.

Siddhartha Gadgil

@siddhartha-gadgil

Fixed.

Arka Ghosh

@anotherArka

Oh ok. Thanks

Arka Ghosh

@anotherArka

I have the following thought. Is it right?
The region

\Delta

is the region bounded by

c_L

and

c_R

which are the leftmost and rightmost canonical geodesic. If we start right canonisation starting from

c_L

we will reach

c_R

. Also, the edges traversd during the canonisation is exactly the edges in

\Delta

.
If it is true then I think it gives us a possible way to calculate

\Delta

and the Spokes

Siddhartha Gadgil

@siddhartha-gadgil

The region

\Delta

is the region bounded by

c_R

and

c_L

. I am not sure of the traversed path statement, but in any case it is cleaner to just deduce it from the two boundary curves.

Namely, the paths

c_L

and

c_R

can be decomposed into coinciding segments and segments that are boundaries of stairs.

One can make a structure for a stair, and such a decomposition. Then define a recursively defined function computing this.

Siddhartha Gadgil

@siddhartha-gadgil

If the curves

c_L

and

c_R

start with a non-trivial coinciding segment, we find a maximal one and then recurse.

Otherwise we find the first staircase, which is bounded by the segments of the curves

c_L

and

c_R

between the initial point and the second point where they agree. Again we recurse.

Arka Ghosh

@anotherArka

I think the traversed path statement should follow as the region

\Delta

is an annulus and if any of the spokes are skipped then the outer and inter boundaries, which are

c_L

and

c_R

can not be homotopic. But, maybe I am oversimplifying.

Siddhartha Gadgil

@siddhartha-gadgil

But is there anything gained by that?

Arka Ghosh

@anotherArka

It can give a possible method to calculate

\Delta

. But your algorithm is simpler I think.

Arka Ghosh

@anotherArka

While pattern matching in scala, is there a way to assert that certain patterns are unreachable.

Siddhartha Gadgil

@siddhartha-gadgil

For what reason?

I mean, why do you want to show this. Is it to avoid compiler warnings?

Arka Ghosh

@anotherArka

Compiler warnings is an issue but I was curious in general.

Arka Ghosh

@anotherArka

Also, currently I am writing a method to separate into matching and non-matching edges. This will help in building the staircases. After that it remains to add faces to get the actual staircases. For that I have the following thought, tell me if this is ok -

Say we have a non-matching pair $(a,b)$ from two paths, where $a$ is part of the leftmost geodesic $c_L$ and $b$ is part of the rightmost geodesic $c_R$ then $b.a^{-1}$ is the boundary of a staircase $S$ . Also each face of the staircase should contain at least two edges from $b.a^{-1}$ .

Siddhartha Gadgil

@siddhartha-gadgil

Yes, that is correct.

You have to decide which kind of staircase it is, and then the spokes just join vertices with index differinb by 1

Arka Ghosh

@anotherArka

Compiler warnings is an issue but I was curious in general.

Anything about this?

Siddhartha Gadgil

@siddhartha-gadgil

The compiler is reasonably smart about this, but can only work with type information. If the missed case is based on object level, there is no way you can tell the compiler.

The practice I follow is have a match for the missed case and throw an exception with information of why that match should never have been reached (helps in debugging, but also serves as documentation)

Chinmaya Kausik

@Chinmaya-Kausik

Okay, to get to examples - would writing out the ones you mentioned in the email the first time we worked on GIN be good, Prof. Gadgil?

Arka Ghosh

@anotherArka

Do note that I have only written the code for primitive curves.

Siddhartha Gadgil

@siddhartha-gadgil

I do not remember what I wote. But some things to try are:

Various pairs of intersections of generators $\alpha_i$ and $\beta_j$ and also their slef-intersections.
Self-intersections of $\alpha_1\alpha_2$ and $\alpha_1\bar{\alpha_2}$ to check one is 0 and the other is not.
Generate families such as $\alpha_1\beta_1^n$ and look for intersection numbers of pairs in these.
Intersection number of the commutator $[alpha_1, \beta_1]$ with $\alpha_1\alpha_2$.

By the way, @Chinmaya-Kausik please address me here by my Gitter handle to maximize the chance I am alerted.

Copying the above with double dollars.
I do not remember what I wote. But some things to try are:

Various pairs of intersections of generators $\alpha_i$ and $\beta_j$ and also their slef-intersections.
Self-intersections of $\alpha_1\alpha_2$ and $\alpha_1\bar{\alpha_2}$ to check one is 0 and the other is not.
Generate families such as $\alpha_1\beta_1^n$ and look for intersection numbers of pairs in these.
Intersection number of the commutator $[alpha_1, \beta_1]$ with $\alpha_1\alpha_2$ .

By the way, @Chinmaya-Kausik please address me here by my Gitter handle to maximize the chance I am alerted.

Chinmaya Kausik

@Chinmaya-Kausik

In that case @anotherArka will you have the time to implement the code for non-primitive loops too? I'll write some examples but the tests obviously won't compile without that.

I'll do some primitive ones right now, of course.

Arka Ghosh

@anotherArka

I will try to do it. But I have a presentation tomorrow so I can't gurantee.

Chinmaya Kausik

@Chinmaya-Kausik

Oh definitely, no problem. I'll do the primitive ones for now.

Arka Ghosh

@anotherArka

I have uploaded a code to break a loop into primitive loops. I think the logic is correct but the code can be optimized slightly.

Unfortunately my laptop is out of charge and it's stormy outside so I can't charge it.

Chinmaya Kausik

@Chinmaya-Kausik

I have uploaded a code to break a loop into primitive loops. I think the logic is correct but the code can be optimized slightly.

Ah thanks a lot!

Chinmaya Kausik

@Chinmaya-Kausik

Running the tests I wrote, it seems like there's a flaw in the canonicization code, and it's very souped up and I can't quite figure out what it is.

Had anyone independently checked the general reasoning behind the code (the use of turn indices and rightward shifts)?

Chinmaya Kausik

@Chinmaya-Kausik

I do remember explaining my method (since there was no explicit algorithm in the paper and the algorithm in the paper referenced seemed a bit involved), but I'm not sure if its correctness was discussed at all.

Chinmaya Kausik

@Chinmaya-Kausik

I have made a pull request with the code I added for tests so far, which has been commented out though since the tests crashed due to some problem with canonicization.

I do remember explaining my method (since there was no explicit algorithm in the paper and the algorithm in the paper referenced seemed a bit involved), but I'm not sure if its correctness was discussed at all.

the algorithm in the other paper referenced*

Chinmaya Kausik

@Chinmaya-Kausik

Had anyone independently checked the general reasoning behind the code (the use of turn indices and rightward shifts)?

Though I guess it's mostly what Prof. Gadgil told us to do, so it should ideally be correct.

Siddhartha Gadgil

@siddhartha-gadgil

@Chinmaya-Kausik can you clarify and narrow the error - do you mean the problem is hanging, presumably an infinite loop, or giving the wrong answer, and what minimal triggers this, and make an issue with this.

Chinmaya Kausik

@Chinmaya-Kausik

Oh okay, yes I can specify what is going wrong. The "why" is what I'm having trouble with.

Chat via Matrix