{"payload":{"feedbackUrl":"https://github.com/orgs/community/discussions/53140","repo":{"id":620995167,"defaultBranch":"master","name":"zzt-puzzle","ownerLogin":"kristomu","currentUserCanPush":false,"isFork":false,"isEmpty":false,"createdAt":"2023-03-29T19:30:30.000Z","ownerAvatar":"https://avatars.githubusercontent.com/u/19836467?v=4","public":true,"private":false,"isOrgOwned":false},"refInfo":{"name":"","listCacheKey":"v0:1680118231.570487","currentOid":""},"activityList":{"items":[{"before":"40c5de80d4045fcdf227f44169652cf4f7fde6db","after":"e1c6a4f9a2374678a5b2fe50941117a168220a7a","ref":"refs/heads/master","pushedAt":"2023-09-21T17:07:18.000Z","pushType":"push","commitsCount":1,"pusher":{"login":"kristomu","name":"Kristofer Munsterhjelm","path":"/kristomu","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/19836467?s=80&v=4"},"commit":{"message":"Factorial design: add progressive order output.\n\nThis change lists the puzzles in an \"anytime\" or progressive order so that\nif we stop early instead of solving every puzzle, the puzzles done so far\nwill still give decent data to do regression on.","shortMessageHtmlLink":"Factorial design: add progressive order output."}},{"before":"81d13c4c415ec5a42b3e9fe807b3986ee43f465e","after":"40c5de80d4045fcdf227f44169652cf4f7fde6db","ref":"refs/heads/master","pushedAt":"2023-09-20T21:17:14.000Z","pushType":"push","commitsCount":1,"pusher":{"login":"kristomu","name":"Kristofer Munsterhjelm","path":"/kristomu","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/19836467?s=80&v=4"},"commit":{"message":"Add Python code for doing factorial experimental design to pick puzzles.\n\nThe purpose is to pick a large number of puzzles with different parameters\nor characteristics, to make it easier to find a function that predicts how\nhard any given puzzle is to solve.","shortMessageHtmlLink":"Add Python code for doing factorial experimental design to pick puzzles."}},{"before":"a16a2951cb3e58176f5780ed9da5e44c8f66819f","after":"81d13c4c415ec5a42b3e9fe807b3986ee43f465e","ref":"refs/heads/master","pushedAt":"2023-05-06T10:24:01.000Z","pushType":"push","commitsCount":2,"pusher":{"login":"kristomu","name":"Kristofer Munsterhjelm","path":"/kristomu","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/19836467?s=80&v=4"},"commit":{"message":"Clean up retry logic in generator.cc (extract inner functionality).\n\nRelabel the board indices to start in N as well, to distinguish them\nfrom old (MT19937-based) indices.","shortMessageHtmlLink":"Clean up retry logic in generator.cc (extract inner functionality)."}},{"before":"18454016532d7e473d71d5af74e252b14e65436c","after":"a16a2951cb3e58176f5780ed9da5e44c8f66819f","ref":"refs/heads/master","pushedAt":"2023-05-01T23:21:34.000Z","pushType":"push","commitsCount":1,"pusher":{"login":"kristomu","name":"Kristofer Munsterhjelm","path":"/kristomu","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/19836467?s=80&v=4"},"commit":{"message":"Add weighted tile generation and retries to generator.cc's grow_board.\n\nWeighted tile generation lets the frequency of the different\nobstacles be random per board; retries lets the solver try skipping\npast a number of tiles if they make the board unsolvable, thus\nincreasing the chance that a solvable complex puzzle is found.\n\nThe latter may still be a bit too slow due to how the retries\nmess up the IDDFS. Better code is needed (it would be better to\njust pretend that tile didn't happen, but that would require some\nextraction of functionality that I don't have the time to do right\nnow).","shortMessageHtmlLink":"Add weighted tile generation and retries to generator.cc's grow_board."}},{"before":"df350f5afc76d9a08f63e58d9fd915d08be27941","after":"18454016532d7e473d71d5af74e252b14e65436c","ref":"refs/heads/master","pushedAt":"2023-05-01T22:18:23.000Z","pushType":"push","commitsCount":1,"pusher":{"login":"kristomu","name":"Kristofer Munsterhjelm","path":"/kristomu","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/19836467?s=80&v=4"},"commit":{"message":"Replace std::mt19937 RNG with the xorshift generator from quadelect.\n\nThis will make interoperability/porting easier, though it *does*\ninvalidate prior indices.","shortMessageHtmlLink":"Replace std::mt19937 RNG with the xorshift generator from quadelect."}},{"before":"49f46934d844ba18f8879fddb19e0520e6772af5","after":"df350f5afc76d9a08f63e58d9fd915d08be27941","ref":"refs/heads/master","pushedAt":"2023-04-11T18:36:44.000Z","pushType":"push","commitsCount":1,"pusher":{"login":"kristomu","name":"Kristofer Munsterhjelm","path":"/kristomu","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/19836467?s=80&v=4"},"commit":{"message":"Add multi-board and multitasking functionality to writer.","shortMessageHtmlLink":"Add multi-board and multitasking functionality to writer."}},{"before":"e2c01a7dc7b7bdef80940b8f9a859c460fcb1561","after":"49f46934d844ba18f8879fddb19e0520e6772af5","ref":"refs/heads/master","pushedAt":"2023-04-10T21:17:38.000Z","pushType":"push","commitsCount":1,"pusher":{"login":"kristomu","name":"Kristofer Munsterhjelm","path":"/kristomu","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/19836467?s=80&v=4"},"commit":{"message":"Build out the proof-of-concept writer so it writes puzzles.\n\nI had to add a lot of stuff to linux-reconstruction, but it should work\nnow.","shortMessageHtmlLink":"Build out the proof-of-concept writer so it writes puzzles."}},{"before":"074775c462e570db98a58487d375f6e1aef69014","after":"e2c01a7dc7b7bdef80940b8f9a859c460fcb1561","ref":"refs/heads/master","pushedAt":"2023-04-09T23:20:55.000Z","pushType":"push","commitsCount":3,"pusher":{"login":"kristomu","name":"Kristofer Munsterhjelm","path":"/kristomu","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/19836467?s=80&v=4"},"commit":{"message":"Add empirical heuristic for refutations in grow_board.\n\nIt appears that if a tile is added, most of the time if a solution\nisn't found at (xsize + ysize) steps deeper in the tree than the current\nsolution, there is no solution to be had. Sometimes this fails, but the\nspeedup we get in proving puzzles unsolvable is worth it IMHO.\n\nAlso boost the max depth to take advantage of this.","shortMessageHtmlLink":"Add empirical heuristic for refutations in grow_board."}},{"before":"51d865729d91ed0f278e0a30748a40f2ac556c07","after":"074775c462e570db98a58487d375f6e1aef69014","ref":"refs/heads/master","pushedAt":"2023-04-07T09:57:07.000Z","pushType":"push","commitsCount":1,"pusher":{"login":"kristomu","name":"Kristofer Munsterhjelm","path":"/kristomu","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/19836467?s=80&v=4"},"commit":{"message":"Add linux-reconstruction submodule for writing to ZZT files.\n\nI haven't implemented the zzt-puzzle code that calls to World yet.","shortMessageHtmlLink":"Add linux-reconstruction submodule for writing to ZZT files."}},{"before":"544c1fb70f0e9dafed6e4e99d28ee69d9033e78f","after":"51d865729d91ed0f278e0a30748a40f2ac556c07","ref":"refs/heads/master","pushedAt":"2023-04-03T22:02:22.000Z","pushType":"push","commitsCount":1,"pusher":{"login":"kristomu","name":"Kristofer Munsterhjelm","path":"/kristomu","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/19836467?s=80&v=4"},"commit":{"message":"Add negative result of IDDFS move experiments.\n\n(The actual code hasn't been added because I couldn't justify its\ncomplexity.)\n\nDo a few comment adjustments too.","shortMessageHtmlLink":"Add negative result of IDDFS move experiments."}},{"before":"258223a9b08ca3d87b80a9558c8912a39b25057a","after":"544c1fb70f0e9dafed6e4e99d28ee69d9033e78f","ref":"refs/heads/master","pushedAt":"2023-04-02T21:41:43.000Z","pushType":"push","commitsCount":2,"pusher":{"login":"kristomu","name":"Kristofer Munsterhjelm","path":"/kristomu","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/19836467?s=80&v=4"},"commit":{"message":"Add explanatory comments and ideas of what to do next.","shortMessageHtmlLink":"Add explanatory comments and ideas of what to do next."}},{"before":"2cfe6eadd7a25b2ac9b987526b3a95cabfb35508","after":"258223a9b08ca3d87b80a9558c8912a39b25057a","ref":"refs/heads/master","pushedAt":"2023-04-02T18:05:03.000Z","pushType":"push","commitsCount":3,"pusher":{"login":"kristomu","name":"Kristofer Munsterhjelm","path":"/kristomu","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/19836467?s=80&v=4"},"commit":{"message":"Clean up DFS now that it works.\n\nUnfortunately, it is now slower than it was before I found the bug.","shortMessageHtmlLink":"Clean up DFS now that it works."}},{"before":"88c8c373044f85843b975a58393d140b1223dcf9","after":"2cfe6eadd7a25b2ac9b987526b3a95cabfb35508","ref":"refs/heads/master","pushedAt":"2023-04-02T14:47:27.000Z","pushType":"push","commitsCount":1,"pusher":{"login":"kristomu","name":"Kristofer Munsterhjelm","path":"/kristomu","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/19836467?s=80&v=4"},"commit":{"message":"Remove flawed transposition table code, try to make it work this time.\n\nIt's not fully implemented yet. Fix silly bug in solver.h where\na result with a longer solution would be considered \"better\" (greater)\nthan one with a shorter.","shortMessageHtmlLink":"Remove flawed transposition table code, try to make it work this time."}},{"before":"cfe898dd45d9368916053bce46d9e963b53b989e","after":"88c8c373044f85843b975a58393d140b1223dcf9","ref":"refs/heads/master","pushedAt":"2023-04-01T21:21:08.000Z","pushType":"push","commitsCount":4,"pusher":{"login":"kristomu","name":"Kristofer Munsterhjelm","path":"/kristomu","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/19836467?s=80&v=4"},"commit":{"message":"Try to fix DFS, then find more failing test cases.\n\nThe principal value reconstruction is a pain and full of subtle\nedge cases...","shortMessageHtmlLink":"Try to fix DFS, then find more failing test cases."}},{"before":"6a5b1dd050eecd8b8784bee7fbb0743f13757b22","after":"cfe898dd45d9368916053bce46d9e963b53b989e","ref":"refs/heads/master","pushedAt":"2023-04-01T14:29:27.000Z","pushType":"push","commitsCount":1,"pusher":{"login":"kristomu","name":"Kristofer Munsterhjelm","path":"/kristomu","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/19836467?s=80&v=4"},"commit":{"message":"Add simple move ordering and horizon shortening to minmax.\n\nThe former improves performance considerably while the latter\nmay be more marginal, but it's free, so why not?\n\nIt's also pretty unsophisticated; I'd like to use the prior depth's\nPV or transposition table to guide the next depth in IDDFS. But if\nI do that, I either have to pull IDDFS into minmax or inelegantly\nexport the PV or transposition table from the latter to the former.\n\nNeither choice is very appealing.","shortMessageHtmlLink":"Add simple move ordering and horizon shortening to minmax."}},{"before":"89517dc8cdceb40fde0c2c40c7b322ed2272c5ff","after":"6a5b1dd050eecd8b8784bee7fbb0743f13757b22","ref":"refs/heads/master","pushedAt":"2023-03-31T21:07:36.000Z","pushType":"push","commitsCount":3,"pusher":{"login":"kristomu","name":"Kristofer Munsterhjelm","path":"/kristomu","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/19836467?s=80&v=4"},"commit":{"message":"Split up the solvers (interface, minmax, and iterative deepening).","shortMessageHtmlLink":"Split up the solvers (interface, minmax, and iterative deepening)."}},{"before":"7adc2f8dd7baccad8c4fbebd88945ce790d6f084","after":"89517dc8cdceb40fde0c2c40c7b322ed2272c5ff","ref":"refs/heads/master","pushedAt":"2023-03-30T21:13:20.817Z","pushType":"push","commitsCount":3,"pusher":{"login":"kristomu","name":"Kristofer Munsterhjelm","path":"/kristomu","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/19836467?s=80&v=4"},"commit":{"message":"Add grow_board board generator.\n\nIt uses a quick and dirty IDDFS approximation for now.","shortMessageHtmlLink":"Add grow_board board generator."}},{"before":"13892dbdebe8cfe779bb1d7906ea4e38907f9f02","after":"7adc2f8dd7baccad8c4fbebd88945ce790d6f084","ref":"refs/heads/master","pushedAt":"2023-03-30T14:22:06.377Z","pushType":"push","commitsCount":1,"pusher":{"login":"kristomu","name":"Kristofer Munsterhjelm","path":"/kristomu","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/19836467?s=80&v=4"},"commit":{"message":"Refactor: Split components into separate files, and add CMake.\n\nAlso update README with references for the computational complexity\nclaims.","shortMessageHtmlLink":"Refactor: Split components into separate files, and add CMake."}},{"before":"7c9db70cc57f5edbc644364722223cc29ca4e6e7","after":"13892dbdebe8cfe779bb1d7906ea4e38907f9f02","ref":"refs/heads/master","pushedAt":"2023-03-29T20:52:43.246Z","pushType":"push","commitsCount":1,"pusher":{"login":"kristomu","name":"Kristofer Munsterhjelm","path":"/kristomu","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/19836467?s=80&v=4"},"commit":{"message":"Initial commit.","shortMessageHtmlLink":"Initial commit."}},{"before":null,"after":"7c9db70cc57f5edbc644364722223cc29ca4e6e7","ref":"refs/heads/master","pushedAt":"2023-03-29T19:30:31.570Z","pushType":"branch_creation","commitsCount":0,"pusher":{"login":"kristomu","name":"Kristofer Munsterhjelm","path":"/kristomu","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/19836467?s=80&v=4"},"commit":{"message":"Initial commit","shortMessageHtmlLink":"Initial commit"}}],"hasNextPage":false,"hasPreviousPage":false,"activityType":"all","actor":null,"timePeriod":"all","sort":"DESC","perPage":30,"cursor":"djE6ks8AAAADhfNVTwA","startCursor":null,"endCursor":null}},"title":"Activity · kristomu/zzt-puzzle"}