nguyenpham/oobs

Open Opening Book Standard (OOBS)

oobs's Issues

Learning or feature question

I am fond of the intent in both your standard proposals repositories. Notable, the care to not lose information.

I may not understand certain aspects, but my question is about the relations that might exist between the 2 repositories.

I understand that this is about the data structure here, but what would I need to be able to use any game set database in the other SQL compatible proposal, and using appropriate tools (? Would your banskia GUI has a bookmaker that would create in this repository format?), end up with a pair of full information databases.

My understanding is that there is always some loss of information when grinding a book db from a game database, some about the games instance themselves (beyond some bookmaking max depth parameter, for "duplicate" last branching point the bookmaker would be set to consider that way, sorry, for expansion, just want to make sure my words complement each others, in case one is ambiguous, 2 or more, might either not make sense, or they do, more chances I transmit my intent, ok. not sure if it works, as who wants a wall of text when they get the shorter version....).

Anyway. Are those 2 standards easy to keep tied with each other. (I mean the win draw loss triplets, would be crunched from the game database given a position, for all the games visiting such "position" (as EPD without op codes? Extra points question.), but would there be a way to create a no loss bridge SQL db or patch or whatever (my ignorance betrays me), so that at least we could keep backtracking querying traversal. I guess that might be the bookmaker job, to output residuals db for its full games maxbookdepth rejects, but what about this.

If trouble with user case: I really like not just the no loss direction, but liberating chess databases in general from chess software programs specific previous choices, when as you notices computer resources had to be catered to, with lots of deference, and human needs might have suffered from it. I am not saying those tools are not doing what they were design (and limited) to do. But I would like the freedom to use position databases (maybe that is more the book) and still keep the full graph of relations from the game mother database.

I would like then to have the bigger ecosystem from SQL, and likely all the libraries that might connect from there to tools of my own gardening experience in the past, and explore the other ML that have been developed since then. Exploratory data analysis in chess, and no loss. For the interactivity needed in such a research context (about chess, yes. Research), relying on processing per instance of position or game or databases of each, would not be human brain timescale reaction time.

The point of storage space being bigger, is acceptable if it allows human time scales of reasoning while doing exploratory data analysis. And other fun stuff from there. All expansions of some human-knowledge embedded in all sort of existing databases, would have to be made first, and I find that your 2 things are a wheel I might not have to reinvent, if I were ever to go tangible on above rambling. or ideas. I hope I made sense, despite ramblings.

Which of ocgdb versus oobs for position database (before book creation)

Somewhere between a full game database (PGN or ocgdb) and the final tradition book digestion, there are positions being visited.

I have been eying ocgdb principles to be able to eventually work with other database tools than dedicated chess ones, and being able to go back and forth, with emphasis on having also the full relation database of the games being queryable for their full position information field, before any zobrification and other book making reductions.

I am interested in being able to make position graphs representation of the games, not decision trees only, which I think is the usual expectation from a book format.

But I am illiterate in chess databases and software traditions for them. I am also not sure I have much clues about mainstream databases software. I would be more of a user and data analysis user type.. able to handle the querying logic.. I might have more clues than stated but I think my question is best stated that way, for others near that user type. Chess insight targeting, I should add.

convert obs.db3 book to bin and pgn

Is it possible to convert obs.db3 books to bin and to pgn format?

How to create a book from pgn?

Recency: Newer games contribute more to weights than older ones.

Can you please add an option for Recency for calculating the book move weights.
If a move is out of fashion in the input pgn file, it most of the time has a good reason.

What compiler will be used for windows binary?

Could you just release a windows binary?

How can I browse the books to see the created books.

Can I navigate the created books in obs.db3 format somehow?
Like you can with polyglot bin books in Scid

Can only standard chess pgn format moves be supported?

Can only standard chess pgn format moves be supported?

For example, can you support Asian traditional chess pgn like janggi?

How to make Stockfish create and read this type of opening book?

Hi Nguyen,
I hope you are doing very well
How to make Stockfish create and read this type of opening book? or at least read