Error reported with the following hand hostory from FTP France:
Full Tilt Poker.fr Game #22632637058: Daily Euro (174599326), Table 3 - 10/20 - No Limit Hold'em - 20:15:39 CET - 2010/07/27
The existing FTP parser crashed on CET times.
1) Fixed parser to accept CET
2) Added timezone offset into HHC
Not happy with TZ1/TZ2 solution.
Decided to knock up a quick proof of concept for fetching and parsing tournament results from a website.
Specifically uses:
- Pythons Beautiful Soup library
- P5s results pages - replace playername with the desired p5s username (eg. taypaur)
Of note; no tournament ids are displayed on any of the sites, so for this to work we may need to add a field like 'p5sid'
- updated copyright notices
- added copyright notices that were missing
- changed python to python2
- added gpl-2/3 license texts
- removed THANKS.txt file as the wiki version is more up to date
- removed empty fpdb_db.py
- added GPL2 to debian license file
Conflicts:
pyfpdb/Database.py
pyfpdb/DerivedStats.py
pyfpdb/SQL.py
pyfpdb/SummaryEverleaf.py
pyfpdb/TournamentTracker.py
pyfpdb/Tourney.py
run_fpdb.py
Stars HHC, HHC itself and Hand.
Should not get the first 100 characters of an failing hand in the log, which contains the handid for later reference.
Played around with the number of characters a while ago - 100 chars is about the sweet spot.
Kangaderoo has a patch which potentially fixes storage and display issues for users who do not have their database text storage as utf8.
Functions and variables added to Charset to disect the patch
- Remove added whitespace
- Reformatted debug strings in preferred printf style (and starting with DEBUG:)
- Swapped len(aggrers)>0 and i>0 for short circuit evaluation, and reformatted spacing for consistent coding style
-- (NOTE: for NLHE this will definitely be faster as less hands go to flop - may be slower for limit games)
Also added a couple of comments near enumerate() calls about 2.5 to 2.6 syntax
Fix a couple of crashers
- Make sure last session in list is displayed correctly
- Actually calculate hands/hour (Thanks Socratic)
- Make graph display the correct number of sessions
Patch from bbtgaf@googlemail.com aka gimick
DerivedStats.betStreet() was only functioning if the player was the first person to act on a street.
If the player was checked to the function would exit as False before ever finding the player
If a player sits out a sb in a cash game they can post it after the button
There is still an accounting error in the way we calculate the returned bet.
The bet action could become multiline greedy if there was a playername starting with 'card' in the next line.
"""
danny purse: bets $0.50
cardjunkie25: calls $0.50
"""
Would match "danny purse: bets $0.50\ncard", and the rest of that line would be ignored.
Added '$' to the end of the regex to make sure that it is limited to 1 line
The bet action could become multiline greedy if there was a playername starting with 'card' in the next line.
"""
danny purse: bets $0.50
cardjunkie25: calls $0.50
"""
Would match "danny purse: bets $0.50\ncard", and the rest of that line would be ignored.
Added '$' to the end of the regex to make sure that it is limited to 1 line, and wrap 'cards' as a group.
From Gerko on mailing list:
> I just did a trial. Use bulkimport for a single Pokerstars file
> switch setting and bulkimport a single Full tilt file.
> When screening the db all players in the database show the same siteId
> (2 for pokerstars)
> Or the other way if the first site was Full Tilt.
The Lambda function wasn't initialising the lambda function correctly.
Occasionally Stars throws up a cash table that looks like
PokerStars Game #22920951206: Triple Draw 2-7 Lowball Limit ($0.10/$0.20) - 2008/12/14 17:44:00 ET
Table 'T #252691988' 6-max Seat #2 is the button
a # was previously excluded from the regex.
Worros