PokerStars support can provide a HH archive. The format is similar but not the same as a a standard hh format as it contains an additional line "Hand #X" between each hand.
Patch adds an option -s to GuiBulkImport, which when specified will strip these lines out and continue parsing.
Affected files:
* Configuration.py
* Database.py
* HandHistoryConverter.py
* fpdb_import.py
and setup.py to a lesser extent.
Logging requires a configuration file when initialising the class. If we
are executing from outside the source tree, this file is not present.
Catch config parser's error and try to provide the logging config file
from debian package path.
Update symlink target to match python-support from latest Debian and the
oncoming Ubuntu versions.
modified: FulltiltToFpdb.py
* Modified re_HandInfo regex to take into account Matrix Tournament Hands Histories
* Add Regex for Tourney Summaries files parsing
* New methods : readSummaryInfo, determineTourneyType, getPlayersPositionsAndWinnings
modified: HandHistoryConverter.py
* Add a Tourney object in the attributes to allow storing the summary info retrieved and the ability to hand it over after parsing
* Add a new attribut parsedObjectType (string : "HH" or "Summary")
* In follow = False mode : read the first line to swicth between HH & Summary parsing
* TO DO : Deal with parsing summary files in follow = True mode
* New methods added : isSummary, getParsedObjectType (to be called in fpdb_import), readSummaryInfo : abstract (to be implemented in each specific HHC), getTourney (returns the new attribute)
new file: Tourney.py
* New object
* Lots of attributes regarding the tourney info (buy-in, fee, entries, speed, Rebuy/add-on info, starting chips, KnockOut info, isHeadsUp, isShootout, isMatrix, ...
* List of players with finishing positions (when available) and winnings (when available)
* Methods :
** addPlayer(self, rank, name, winnings)
** incrementPlayerWinnings(self, name, additionnalWinnings): used for KO tourneys when KO occured (for Hero only)
** calculatePayinAmount : unused yet, should be necessary for DB storage
** some Hand methods copied that are still to be done including : assemble and insert
** TO DO : write the insert method for the object to be stored in DB. This will have to deal with the fact that the DB write can occur both before (Bulk Import) or after the HH file info has been stored (tourney might or might not already exist)
modified: fpdb_import.py
* import_file_dict modified :
** after the construction of the hhc, it now calls the new getParsedObjectType method of HHC in order to know what has been parsed by the HHC
** If it's a hand history file (actual expected behaviour) : do as before
** If it's a summary file, gets the tourney object that has been built from the hhc and calls the insert method on it (similar to NEWIMPORT=True for HH)
Better exception handling during parsing
FpdbParseEcxeption takes an optional HID
Stripped pointless __init__s from *ToFpdb subclasses
Logging now has logging.conf config - default is to log only INFO to
file but easily changed to DEBUG for devel work
Make tail work:
-log instead of print inode # change detected,
-make sure splitter regexes have groups around them
-fix a bug
-increase buffering size to something >> a single hand
Conflicts:
pyfpdb/HandHistoryConverter.py
(this is default from cmdline but not from fpdb_import)
Eerleaf gametype tests... added one.
Everleaf action fixed to capture bet amounts, I hope. The 'complete to'
has a problematic . at the end, so we must handle
[ $ 1.40 USD ]
[ $ 1 USD ]
$1.40.
$1. (haven't seen this one yet)
Sorry I've kind of destroyed the sanityCheck thing, some of it belongs
in fpdb_simple where the paths are made. I've retained check for
in_path!=out_path.
Let me know how you get on.
I just remembered autoimport probably broken at this point, I have
changed the way HHC is called and only been testing standalone
filtering..
Just try it and then git checkout ^HEAD
Merge branch 'monkeyfutz'
Conflicts:
pyfpdb/EverleafToFpdb.py
pyfpdb/FulltiltToFpdb.py
pyfpdb/Hand.py
- made it so that Hand.__init__ calls the hhc readX methods in the right
order for that kind of Hand (Stud or Holdem/Omaha)
- hhc.readX methods then callback Hand.addX methods
Why? WHY! I hear you ask. Actually I can see there's very little gain,
sorry about that; but it sort of makes sense. Now you just make the
right kind of Hand, give it the text and the right kind of HHC filter,
and it makes itself.
- apart from that, biggest actual thing done is probably in adding
player cards for stud hands -- revamped and made it really clear (I
think) whats happening (see FullTilt.readStudPlayerCards)
When I run FullTiltToFpdf.py it goes through all the hands.
- Still have some print statements, have been changing them to
logging.debug or logging.info as I come across them.
You may find "tail -f logging.out" useful
Actually I'm not at all convinced about the use of logging except that
you can redirect it to stderr quite easily and then set the threshold
for what gets logged quite easily, so that might be a plus.
- Oh and the subclassing of Hand is getting clearer.
Everleaf converts hands
Moved process hand logic into the hand subclass that knows what it
wants, it calls back to the hhc (which knows about the format) to populate itself
Conflicts:
pyfpdb/EverleafToFpdb.py
pyfpdb/Hand.py
pyfpdb/HandHistoryConverter.py
- command line options
Try the following
./Everleaf --help
cat everleaf.txt | ./Everleaf -i - -o -
Try using -i -, and pasting in a hand. Be aware it needs an EOF
<ctrl>-d (unless you specify --follow, which isn't done yet)
- HHC is a thread
- hand gametypes per hand
- lots of stupid little things just to cause trouble
- added logging
Not yet done:
- tail -f mode
- integration into fpdb_import. Two options for grand plan here:
1) recently modified files get a HHC filter attached
when the HHC runs out of pipe it waits
runUpdated finds modified files and notifies the thread
2) the thread follows independently, when it runs out of input it
sleeps and then tries again at intervals
I like both, 1) involves bigger changes to fpdb_import whilst with 2) we
just point the existing fpdb_import code at the filter output (the way
we currently do filtering has to be altered though)
Comments welcome of course.
About the only place where we are guaranteed to get all the players'
names correctly is in the seating plan in the hand header. We extract
the names and build a regex that matches only the names. We only do this
if the current regex is out of date, to avoid recompiling them too
often.
I also did away with self.rexx because it seemed unnecessary and was
difficult to work with.
