Anyone want to write this application?? -- Me love you long time...

[teddyFBI]

It's for dataminers.

Here's the problem: the auto-import feature in PT contains a filtering feature that lets you tell it to ONLY import hands with between x and y players for Party HHs. But the problem is that there's ONLY this feature for Party HHs and not for any of the other clients I mine.

So -- as I am guessing is probably what you've discovered for yourself -- the solution is, for example, to take 1,000 UB HHs, import them into a "dumping" DB, then use the EXPORT feature to, for example, export all UB hands with between 8 and 10 players to a folder, and then RE-import these 'streamlined' HHs into your functional DBs.

So yesterday, I thought of a great program: a stand-alone application that you could quickly dump a whole bunch of HHs into (from UB, Pacific, FTP, Stars, Prima, whatever -- any site other than Party really, it could be used for ALL HHS not just mined hands), and it would automatically spit out 'streamlined' text files with ONLY those HHs that fit the # of players criteria that the user would specify. That would prevent us from doing the whole import/export/RE-import runaround.

This can't be that hard to write...what do you think?

[pokergrader]

Very simple. If somebody doesn't already have it done by the time my flight lands in 12 hours, I will take a stab at it.

[teddyFBI]

Super -- yeah, there's no doubt this isn't all that hard an app to write, but i have 0 programming skills with which to do so.

Here's the type of functionality and interface I had in mind -- of course anyone's free to create their own, but maybe this will be of some guidance.

1) It should have a basic "select files" option from either a dropdown menu or a button on the main GUI, which would bring up a window that lets the user select the files to be filtered.

2) It should also have checkboxes on the main page that correspond to each of the 6 (or however many) sites that this will work with, so you can tell the program which site's HHs it will be looking at, because I'm guessing the reading and parsing of the HHs will have to be slightly different for each -- personally, I'd like to see it be able to handle UB, Pacific, Stars, FTP, and Prima. THis differentiation is needed, because, for example, the format that Gametime+ spits out UB HHs in is one big amalgamated text file with 300 HHs, whereas the Pacific Handgrabber program creates individual text files with one hand each.

3) Lastly, it should have a "desired output folder" field, where the user could specify what folder he wants the 'filtered' HHs to be put in. They could either be separated into single text files with one HH per file, or kept (or converted in Pacific's case) in amalgamated text files with many HHs per file. One quick caveat if it's the latter: I once used a free text amalgamation program that I found on the web to take 500 Stars HHs, and amalgamate them all into a single TXT file with all 500 HHs in it. I then tried to import this combined file into PT, but found that whatever the amalgamator had done in combining the files, it had rendered it unreadable to PT (I can't figure out why, because all it did was insert a line of space, no characters, between each HH in the amalgamated file). Anyway, I just bring that up as a quick caveat that just dumping all the text into a TXT file doesn't mean PT will be able to read it all.

[pokergrader]

Well you have shattered my dreams of a simple command line based solution.

[teddyFBI]

lol - no, just make it as simple as you want. it will be much appreciated either way.

my suggestions were just to make it aesthetically pleasing and user-friendly, but i'll be able to make just as good use of any basic command line thing, as long as the functionality is there. (The sites I care most about are UB, Stars, and Pacific.)

[theRealMacoy]

[ QUOTE ]
Very simple. If somebody doesn't already have it done by the time my flight lands in 12 hours, I will take a stab at it.

[/ QUOTE ]

this makes me very happy!

[theRealMacoy]

[ QUOTE ]
It's for dataminers.


[/ QUOTE ]

you are certainly on to something here....

i too have been taking the long route to keeping my dbs nice and lean.

if mr.pokergrader was able to run with this idea it would be sweet.

cheers,
sean

[pokergrader]

Alright, I am landed. Email me sample hand histories at pokergrader@gmail.com and I can figure out the formats and churn something out.

[SamIAm]

[ QUOTE ]
Alright, I am landed. Email me sample hand histories at pokergrader@gmail.com and I can figure out the formats and churn something out.

[/ QUOTE ]HA! I knew this whole "I'm a nice programmer guy" thing was all an act! He's out to get a couple of your hand-histories!

Just kidding. Thanks, pokergrader.
-Sam

[theRealMacoy]

[ QUOTE ]
Alright, I am landed. Email me sample hand histories at pokergrader@gmail.com and I can figure out the formats and churn something out.

[/ QUOTE ]

i am out all day doing christmas stuff but when i get back to my computer this evening i will send some stars and prima files as well....

thanks,
sean

[pokergrader]

Here is the first version: http://www.pokergrader.com/software/PlayerCount.jar

You need JRE 1.5 (or 5.0, whatever it is called). It should run with a double click on windows, but sometimes you have to go into the command line and type "java -jar PlayerCount.jar".

It is pretty straightforward. Just pick an input directory where files are stored, it will handle figuring out which are stars/ub and pacific. Pick an output directory where you want the new version to be stored, and click convert.

In whatever the output directory is, the program will put all stars histories in a "stars" director, all UB ones in "ub" and all pacific hand histories in "pacific".

It should look something like this:
http://www.pokergrader.com/ss.PNG

[teddyFBI]

Hey, so help me figure out how to run it. I def have the latest Java 5.0 or 1.5 update or whatever it is (cuz my computer just reminded me to get it today).

I double click that link, and it tells me it's a zipped file, so I save it to a folder on my desktop, but all that's in the folder (when i unzip it) are files with .class extensions...and I don't know what to do or click to get it to run...

[teddyFBI]

Ignore my last post - i saved it as "all files" and manually gave it a .jar extension, and now i can just double click it.

I'm putting it thru the paces right now -- just trying pacific to start with. looks awesome.

[teddyFBI]

seems to work like a dream -- here's what i did to test it. i first dumped about 1,000 hand histories from each of those sites through 'playercount' to see how many it 'caught'. then i did my traditional dump/export to see how many of each PT caught. Here were the results.

Pacific: PC identified 695 hands w/ between 8 and 10 players, PT ID'd 684

Stars: PC: 1,534 PT: 1,531

UB: PC: 434 PT: 434

Clearly, the discrepancies are only minor.

I think you just saved me more time than you can imagine!

[illunious]

Why not just write a query to delete observed hands where # of players < X?

[pokergrader]

[ QUOTE ]
Why not just write a query to delete observed hands where # of players < X?

[/ QUOTE ]

Probably a good idea.

[teddyFBI]

i'm confused...can someone explain what that means to me?

Is that some sort of SQL query, meaning that i could just dump EVERYTHING into my SQL DB, and then when i wanted to 'streamline' it, i could just run the query that would purge it of all hands where # of players < X??

[theRealMacoy]

[ QUOTE ]
Here is the first version: http://www.pokergrader.com/software/PlayerCount.jar

You need JRE 1.5 (or 5.0, whatever it is called). It should run with a double click on windows, but sometimes you have to go into the command line and type "java -jar PlayerCount.jar".

It is pretty straightforward. Just pick an input directory where files are stored, it will handle figuring out which are stars/ub and pacific. Pick an output directory where you want the new version to be stored, and click convert.

In whatever the output directory is, the program will put all stars histories in a "stars" director, all UB ones in "ub" and all pacific hand histories in "pacific".

It should look something like this:
http://www.pokergrader.com/ss.PNG

[/ QUOTE ]


holy pickled pinapples!

i get home and intend to send some hh to help with the project and kazaam....you have already put up a working model.

mr.pokergrader i am fully freakin' impressed.

if you still need hh from stars / bet356 / party send me a pm.

i am off to give this little bad boy a whirl.

looks like it will be saving me a whole bunch of trouble and/or space in my databases.

tyvm

sean

[theRealMacoy]

[ QUOTE ]
Why not just write a query to delete observed hands where # of players < X?

[/ QUOTE ]

i echo teddy's question.....

is it possible to run a querry that would select hands that were both below a certain number of players and observed hands (where i was not seated).

could i then delete the resulting hhs permanently from my databases?....could that be built into the query?

i have run queries in the past with pgAdmin....but i have no idea about deleting hhs.

i might try and post this in the pokertracker SQL forum, although it gets much less traffic so i am hoping for some assistance here.

cheers and thanks,
sean

[tizim]

[ QUOTE ]
from UB, Pacific, FTP, Stars, Prima, whatever -- any site other than Party really, it could be used for ALL HHS not just mined hands

[/ QUOTE ]

How do you datamine at Pacific?

[awr000]

I think the issue with a query to delete hands would be the various (maybe only one) summary tables that are kept in pokertracker. The session table is one example. I suppose you could run a query to recreate that(those) tables after running the deletion query.

[jukofyork]

[ QUOTE ]
How do you datamine at Pacific?

[/ QUOTE ]

handgrabber (http://www.pokertracker.com/handgrabber.html) can grab pacific hands using win-pcap library.

Juk /images/graemlins/smile.gif