Winboard Forum

In our project "chess960athome" we are now able to create the first test units.

As the result from a work unit we 'll get the uci protocol output from a engine.
We are using a strong free chess960 engine (Glaurung) but we are able to use each uci engine later in the project.

The problem is the following:

We need help to "translate" the uci engine output into the pgn move format because we want to publish the pgn's in the database later.
We need a program to generate the moves in pgn format from the uci output file of the engine.

I believe such parts of code are in each GUI and/or engine.
We are not familar with engine / GUI programming therefore I ask here.
The project focus is not to generate a new engine or improving of existing engines, we only want to use the engines.

If u are able and willing to help please contact me (joerg.nowak at gmail.com) for details. We can only pay a small amount for your efforts because it is a non commercial project.


Thanks.

Joerg Nowak wrote:In our project "chess960athome" we are now able to create the first test units.

As the result from a work unit we 'll get the uci protocol output from a engine.
We are using a strong free chess960 engine (Glaurung) but we are able to use each uci engine later in the project.

If you are interested in a more recent version stronger than the public one, I can send it to you.

The problem is the following:

We need help to "translate" the uci engine output into the pgn move format because we want to publish the pgn's in the database later.
We need a program to generate the moves in pgn format from the uci output file of the engine.

Could you please be a little bit more specific? Can you give some example input, and show us how you want the output to look? I don't have much time for programming right now (or, to be more precise, programming has very low priority for me this summer), but if it's just a couple of hours of work I may be able to do it some rainy weekend.

We can only pay a small amount for your efforts because it is a non commercial project.

I don't want money. The problem is always time, not money, and contrary to popular belief, the two are not the same.

Tord

If you are interested in a more recent version stronger than the public one, I can send it to you

Of course I'm interested in .

Could you please be a little bit more specific?

Of course I can.

After some workunits (one move is one workunit) we 'll get for instance the following result (only 2 moves played).

position fen qbbnnrkr/pppppppp/8/8/8/8/PPPPPPPP/QBBNNRKR w KQkq - 0 1 moves d2d4 d7d5 ........

For your understanding, what we do:
Each workunit will be played on one of the computers of the project participants, the boinc software will handle this. We plan to play with fixed depths at the beginning, may be with fixed numbers of nodes later (if supported from the engine side).

We will parse the output file from the engine to generate the input for the next workunit (next move). At the end we have a fen string plus the moves in the notation as it comes from uci engine output.

From this format we want to have the moves in pgn file format (with a dummy header which contains the initial position at least).

For instance:

[Event "Chess960athome game"]
[Site "Chess960athome.org"]
[Date "2006.06.14"]
[Round "?"]
[White "Glaurung"]
[Black "Glaurung"]
[Result "?"]
[Time "08:11:20"]
[TimeControl "fix depth"]
[SetUp "1"]
[FEN "qbbnnrkr/pppppppp/8/8/8/8/PPPPPPPP/QBBNNRKR w KQkq - 0 1"]
[Termination "normal"]
[PlyCount "2"]
[WhiteType "program"]
[BlackType "program"]
1. d4 d5 ...

The header is not important now but the FEN string and the moves are essential.

Is this enough information ?

Joerg Nowak wrote:

If you are interested in a more recent version stronger than the public one, I can send it to you

Of course I'm interested in .

OK, a couple of days ago I put a new, semi-public beta version here. You can also download the source code if you need it.

I think this version is clearly stronger than Glaurung 1.1, but because it hasn't received much independent testing yet I cannot be sure.

Could you please be a little bit more specific?

Of course I can.

(detailed explanation snipped)

Is this enough information ?

Yes, I hope so. I can't promise anything right now, but it is possible that I will find some time for this later this summer.

Tord

Joerg Nowak wrote:In our project "chess960athome" we are now able to create the first test units.

As the result from a work unit we 'll get the uci protocol output from a engine.
We are using a strong free chess960 engine (Glaurung) but we are able to use each uci engine later in the project.

The problem is the following:

We need help to "translate" the uci engine output into the pgn move format because we want to publish the pgn's in the database later.
We need a program to generate the moves in pgn format from the uci output file of the engine.

I believe such parts of code are in each GUI and/or engine.
We are not familar with engine / GUI programming therefore I ask here.
The project focus is not to generate a new engine or improving of existing engines, we only want to use the engines.

If u are able and willing to help please contact me (joerg.nowak at gmail.com) for details. We can only pay a small amount for your efforts because it is a non commercial project.


Thanks.

If you can send me a raw log of the UCI->engine i/o I can see what I can do. CAn try and do a perl translator then maybe a C/C++ version.

I've been doing some ICS to PGN stuff lately so this kind of stuff is fresh in my mind.

Email me in priv jshriver <at> gmail.com

-Josh

Curious but what is the goal of the project? To create a pgn database of game for a FRC opening book? If so I did work on this last year. Here is a kinda simple way to do it.

Tools:
* Glaurug *excuse spelling* engine
* xboard or windboard interface
* polyglot for UCI UCI->WB bridge
* FRC db (listing of all FRC openings combinations)

Take whatever opening you want to work on and put the FEN line in a file called frc.fen or any filename you want. Then do this

xboard -size small -fcp glaurug -scp glaurug-mg 1000 -sgf test1.pgn -lpf frc.fen

Voila let it burn away the games. When it's done you'll have a 1000 game pgn database of games that used the opening you chosen.
Keep in mind if you want a strong opening book I'd recommend setting the -tc (timecontrol) to higher levels.. 30 0 perhaps?

I believe this is all have to do IF you want I can gather up all the pgns I generated last year (if I still have them) and put them online.

-Josh