YZF_Sudoku

Programs which generate, solve, and analyze Sudoku puzzles

Re: YZF_Sudoku

Postby yzfwsf » Wed Jun 10, 2020 11:00 pm

1. Fix bug in finding sk-loops.Examples are as follows.
Code: Select all
9876.....65....7...........4...3..2...67..8.......1..4..98..5......5..1......2..3

2. Add the function of opening Andrew Stuart's weely unsolvable puzzle through the Internet
3. Add the function of "copy with cell state" (compatible with Hodoku)

Please download the two files at the same time and put them in the same directory, remove ". 7z" at the end of each file name, and then decompress them.
YZF_Sudoku610V9.7z.001.7z
(256 KiB) Downloaded 45 times
YZF_Sudoku610V9.7z.002.7z
(246.34 KiB) Downloaded 59 times
yzfwsf
 
Posts: 247
Joined: 16 April 2019

Re: YZF_Sudoku

Postby ghfick » Mon Jun 29, 2020 1:20 pm

Here are a few more minor edits:
Als should be ALS in : Als-XZ, Als-XY-Wing, Als-W-Wing
extra cands should be extra candidates
Broken Wing step appears twice in 'All Possible Steps
After a manual entry, one sees 'Game unique solution'. This should be 'Game has a unique solution'.

May I suggest that next you should add :
Chains with ALS nodes [ as in HoDoKu ] ...then chains with other more elaborate nodes?
Chains with memory [ called 'Complex Chains' in PhilsFolly ]
Steps that are intermediate to Dynamic Chains [ like Kraken Unit {row, column, box}, Kraken Fish {as in HoDoKu} ]
ghfick
 
Posts: 102
Joined: 06 April 2016

Re: YZF_Sudoku

Postby yzfwsf » Mon Jun 29, 2020 2:00 pm

Hi ghfick:
Thank you for your suggestion. I will consider it carefully. At the beginning, the implementation of the
chain function was that the data structure did not reserve the interface of complex nodes, so if you want
to support complex nodes, you need to change a lot, which requires a lot of time to conceive and write code.

Recently, I modified ALS w wing's code to support group nodes, which can solve some problems that used to
require dynamic forcing chain.Next, I want to enhance the senior Exocet technique to support situations
where the two target pair contains a locked non-base digit or target triple contains two locked non-base digits.
Next, I want to enhance the senior Exocet technique to support situations where the target pair contains a locked
non-base digit or the target triple contains two locked non-base digits.

The following example used to need dynamic chain, now it is not needed.
Code: Select all
1..2.34...5.....1....6.....6.37.2.......3....7..4.62..8....46...1......5.......9.

ALS W Wing.png
ALS W Wing.png (80.12 KiB) Viewed 376 times
yzfwsf
 
Posts: 247
Joined: 16 April 2019

Re: YZF_Sudoku

Postby ghfick » Mon Jun 29, 2020 6:43 pm

610 gives me an MSLS with almost the same set of exclusions. No need for a dynamic chain here?

MSLS:27 Cells r1346c25789+r3c13,r7c89,r8c78,r9c79,r1c3,27 Links 6789r1,234789r3,489r4,389r6,15c5,15c7,5c8,1c9,23478b9
11 Eliminations:r5c79,r79c5<>1,r5c7,r7c5<>5,r3c6<>7,r3c6,r6c3<>8,r3c6,r6c3<>9
ghfick
 
Posts: 102
Joined: 06 April 2016

Re: YZF_Sudoku

Postby SpAce » Tue Jun 30, 2020 2:05 am

ghfick wrote:Chains with memory [ called 'Complex Chains' in PhilsFolly ]

Funny that you should mention that. Have they improved at all since we last discussed the topic? Looks like the help page has at least improved tremendously, since it has copied my memory chain verbatim:

1) philsfolly wrote:(5)r1c4 = r3c5 - (5=3)r3c2 - (3=7)r5c2 - r56c1 = r8c1 - (57=3)r8c5 - r9c4 = r9c9 - (3=5)r7c78 => -5 r7c4

2) SpAce wrote:(5)r1c4 = r3c5* - (5=3^)r3c2 - (3=7)r5c2 - r9c2 = r8c1 - (7|*5=3)r8c5 - r9c4^2 = r9c9 - (3=15)r7c78 => -5 r7c4

3) philsfolly wrote:(5)r1c4 = r3c5* - (5=3^)r3c2 - (3=7)r5c2 - r9c2 = r8c1 - (7|*5=3)r8c5 - r9c4^2 = r9c9 - (3=15)r7c78 => -5 r7c4

What a quantum leap! Well, imitation is the sincerest form of flattery, so I'll take that as a compliment 8-) Thanks, Phil!

Of course it would be nice to credit the original author, but I guess that's too much to ask. (Better that I don't say what I really think :evil: or I get banned again. Lol.)
User avatar
SpAce
 
Posts: 2559
Joined: 22 May 2017

Re: YZF_Sudoku

Postby yzfwsf » Sun Jul 05, 2020 1:41 pm

1. Enhance als-w-wing to support group nodes.
2. SE is enhanced to support target cell groups with locked members of non-base candidates.
3. Fixed the crash problem when refreshing the grid panel when using the search all possible
steps function when there are hint step on the grid panel.
4. As the next period of time to busy work, so the enhancement of chaining technique plan can only be delayed.

Please download the two files at the same time and put them in the same directory, remove ". 7z" at the end of each file name, and then decompress them.
Release version 611
YZF_Sudoku611.7z.001.7z
(256 KiB) Downloaded 47 times
YZF_Sudoku611.7z.002.7z
(245.02 KiB) Downloaded 57 times
yzfwsf
 
Posts: 247
Joined: 16 April 2019

Re: YZF_Sudoku SE+

Postby yzfwsf » Sun Jul 05, 2020 1:50 pm

Senior Exocet Plus example
Code: Select all
.......6.2....4..1..9.8.42.4....561...3.......9..7.........824.8....21...2.5....8
..2..1....7..8....4.83....9..1..7.8..8..2....3.......5......49....9..5....6..3..8

SE2+.png
SE2+.png (34.61 KiB) Viewed 306 times
SE+.png
SE+.png (33.21 KiB) Viewed 306 times
yzfwsf
 
Posts: 247
Joined: 16 April 2019

Re: YZF_Sudoku

Postby ghfick » Sun Jul 05, 2020 5:49 pm

Hi yzfwsf,
Very interesting extension to Senior Exocet. 611 does not accept paste from copy from the forum. Manual entry works fine. I can copy/paste from 610 to 611.
Please post an example to show the extension to ALS W-Wing.
I am unable to help with the programming. I do understand that some advances may be harder than others.
Gordon
ghfick
 
Posts: 102
Joined: 06 April 2016

Re: YZF_Sudoku

Postby yzfwsf » Sun Jul 05, 2020 11:02 pm

ghfick wrote:Please post an example to show the extension to ALS W-Wing.
Gordon

Code: Select all
1...6.7...5.4...9...7..3..8.92.1..7.4...9.8...7...2..4....3.1...3......92....8.6.

You can press "Ctrl + Shift + B" to make YZF_ Sudoku generates the puzzle when the grid panel has focus, ALS-W-Wing = 41.
ghfick wrote:Very interesting extension to Senior Exocet. 611 does not accept paste from copy from the forum. Manual entry works fine. I can copy/paste from 610 to 611.
Gordon

I have no problem pasting here, please give the page link which can't use paste function.
yzfwsf
 
Posts: 247
Joined: 16 April 2019

Re: YZF_Sudoku

Postby ghfick » Mon Jul 06, 2020 7:41 pm

I will try to determine the occasions when copy/paste does not work [with Wine] with 611. There appears to have been a change since 610.
Grouped ALS W-Wing is really something. I cannot imagine a human solver ever finding it but it is certainly possible.
I am unclear how to determine when the grid panel has focus. Is that a key stroke?
ghfick
 
Posts: 102
Joined: 06 April 2016

Re: YZF_Sudoku

Postby yzfwsf » Mon Jul 06, 2020 11:15 pm

If you are not sure if the grid has focus, you can click anywhere on the grid to get the focus. In fact, whether the grid has focus is marked. When the grid has focus, the active cell has a light yellow outline. When the grid is operated with the keyboard, the input focus is in the active cell.
yzfwsf
 
Posts: 247
Joined: 16 April 2019

Re: YZF_Sudoku

Postby ghfick » Tue Jul 28, 2020 5:00 pm

UL_NT_2.jpg
UL_NT_2.jpg (48.06 KiB) Viewed 188 times

Here is a step described by the solver as a Uniqueness Loop Type 3 with a Naked Triple from the patterns Game 391. I am still trying to understand it. Help would be appreciated.
The solver appears to miss the 'easier' uniqueness step : Hidden Unique Rectangle of 58 at r69c23 : r6c2<>5 [found by PhilsFolly]
ghfick
 
Posts: 102
Joined: 06 April 2016

Re: YZF_Sudoku

Postby tarek » Tue Jul 28, 2020 7:27 pm

This is a uniqueness loop of size 8 (8 cells) compared to the unique rectangle which has size 4. The naked triple would mean looking at 3 cells. I wouldn't call this easy by any stretch of the imagination. I'm not sure how other solvers classify & order their difficulty ratings but in SE the Unique Rectangles and loops regardless of difficulty will be investigated prior to the naked quad. SE would actually report the above Unique loop type 3 with naked triple before the naked quad!

Tarek
User avatar
tarek
 
Posts: 3745
Joined: 05 January 2006

Re: YZF_Sudoku

Postby mith » Tue Aug 11, 2020 11:30 pm

yzf, do you have a google drive link to the latest release? I haven't been able to get the split 7z files to extract for some reason I haven't figured out yet.
mith
 
Posts: 291
Joined: 14 July 2020

Re: YZF_Sudoku

Postby ghfick » Fri Aug 14, 2020 5:51 pm

After downloading the two files, you must rename them as :
YZF_Sudoku611.7z.001 and YZF_Sudoku611.7z.002
Then you need the archiving software 7-Zip. There are several ways to set up 7-Zip and each way depends on your operating system [OS]. You may need to associate 7z files with 7-Zip. The two files are indeed 7z files. Your OS may object to the renaming but you must force the renaming.
Then, you should be able to click on the file that ends in 001 and the unarchiving should happen. I suggest saving the three files [ a .exe file and 2 .dll files] in a new folder.
If your OS is MSWindows, you should be able to run the .exe file. YZF_Sudoku should create a file with suffix .ini [which is needed each time you launch YZF_Sudoku]
If your OS is Mac or Linux, you will need software called wine. There are, again, many ways to set up wine. Version 611 of YZF_Sudoku will run quite well with wine however there are still issues as with any software in development. Again, you will need to keep the .ini file. If you change any preferences in YZF_Sudoku, the changes are kept in the .ini file.
As YZF_Sudoku will continue to be revised, you may wish to copy and move the .ini file to newly created folders containing the revised software.
I am not a programming person so there may be better ways to set up and/or understand all of this.
ghfick
 
Posts: 102
Joined: 06 April 2016

PreviousNext

Return to Software