User pleiade67 in the LifeIn19x19 forum packaged Kombilo for Mac OS X to allow for a very easy installation: Just download and unpack the zip file below, and start the program.
The package uses Wine (and WineBottler) to package the Windows version of Kombilo and the required libraries into a package that can be used on Mac OS X.
When I tried it (on a MacBook Pro around 5 years old) the program ran noticeably slower than the native version, but I might have been doing something wrong. In any case, it's a great way to try what Kombilo can do before taking the slightly more involved path of installing the native Mac version.
Download: kombilo-084.app.zip (80MB)
Unpacking the zip file, you get a folder kombilo-084.app which serves as an executable program.
Using the SGF tree feature on a large database of more than 1.6 million games, I produced a large corner dictionary. The search region was an 11x11 square in the lower right corner of the board. Only moves which were played in at least 200 games were considered, in each position only the 30 most frequent continuations were considered (of course, usually there are much less), and at most 35 moves were considered.
After unpacking, you obtain a 14.3 MB SGF file.
Another bug fix release which corrects some issues in the pattern search in rare cases (handling of continuations in patterns with special symmetries, and searching for move sequences in patterns with certain kinds of symmetry), and made the sgf tree building more user friendly. You can reuse the databases from any 0.8.* version.
Yet another release, mostly in order to fix some issues on Mac OS X. Also contains a number of other small improvements, so everybody is encouraged to update. If you installed a previous 0.8.* version using pip, upgrading is as eays as:
pip install -U kombilo
With the Windows installer, it is recommended to deinstall the old version and then use the installer for the new version.
Some improvements over 0.8.1 (most notably better memory management and more flexible handling of large SGF collections).
On Linux upgrade (or install) via pip, e.g. on Ubuntu:
sudo apt install python-pip python-tk libboost-all-dev libsqlite3-dev pip install --upgrade kombilo
On Windows, you have the following two options:
The clean way is to install Python 2.7 from <a href="http://www.python.org/">python.org</a> yourself, and then install the Kombilo package by typing, on a command prompt:
c:\Python27\Scripts\pip.exe install --upgrade kombilo
This upgrades/installs an executable kombilo.exe in c:\Python27\Scripts\ and allows you to update later by downloading only Kombilo itself, rather than Kombilo+Python.</p>
Alternatively, use the installer. It will install install Kombilo as a stand-alone package (but with Python integrated inside, so you do not save much space compared with the previous option), and install a desktop icon and/or a start menu entry, at your choice: Link removed -- use version 0.8.3 instead Windows installer
To upgrade from 0.8.1, it is probably safest to deinstall the old version, although you can probably just as well install the new one directly.
On Linux install via pip, e.g. on Ubuntu:
sudo apt install python-pip python-tk libboost-all-dev libsqlite3-dev pip install kombilo
To update, use pip install -U kombilo instead.
On Mac OS X, I would expect that this works as well, in principle, but might need some more fiddling. Update: Version 0.8.3 should run OK on Macs.
On Windows, you have the following two options:
The clean way is to install Python 2.7 from python.org yourself, and then install the Kombilo package by typing, on a command prompt:
c:\Python27\Scripts\pip.exe install kombilo
This installs an executable kombilo.exe in c:\Python27\Scripts\ and allows you to update later by downloading only Kombilo itself, rather than Kombilo+Python.
Alternatively, use the installer. It will install install Kombilo as a stand-alone package (but with Python integrated inside, so you do not save much space compared with the previous option), and install a desktop icon and/or a start menu entry, at your choice.
Link removed - use Kombilo 0.8.2 instead Windows installer. (Update on Nov 2: The installer has been "signed" to ensure its authenticity.) Tested on Windows 7 SP1, Windows 8.1, Windows 10. Since development is done on Linux, feedback is particularly welcome.
After a long time, I can finally publish Kombilo 0.8 with a couple of improvements compared with the 0.7.* versions:
I am working on a Windows installer and hope to add it within a couple of days ... (Update: Now available in version 0.8.1)
A small bug fix release to cure a few issues in version 0.7.5.
The new version has a few small bug fixes (thanks to everybody, in particular Bram Vandenbom, Gilles Arcas, Claude Brisson for pointing out bugs). Notably, the search for the empty board was broken in Kombilo 0.7.4 and now works again.
The source code is now hosted on Github, since I now have most of my projects there: Kombilo repository.
For this version, there is (so far) no Windows installer. Since I do not work with Windows and do not currently have access to a Windows machine, building a Windows installer is too much trouble for me at the moment. If somebody wants to have a go at this, I will be glad to give some hints (also see the file "fabfile.py" in the code repository). You will need a C++ compiler to build the C++ extension, a tool (such as py2exe) to package things together with Python into a stand-alone exe file, and a packaging tool such as InnoSetup to create an installer.
Update: This version is really outdated. Using version 0.8.* is highly recommended.
A new version which fixes a bug in the search algorithm: in some cases, partial board patterns where the stone completing the pattern causes a capture outside the search region were not found. Thanks to John Fairbairn for pointing this out!
After a long hiatus in Kombilo development, the new version Kombilo 0.7 is finished.
In comparison to previous versions, it is overall much faster. The pattern search is 'parallelized' and makes use of all processor cores that are available. You can search in variations and search for move sequences.