(2015 December) 0.9 First version distributed via the Git system
(2015 02 11) 0.91 Eliminated unused file ObserverSwarm.py in folder '6 objectSwarmObserverAgents_AESOP_turtleLib_NetworkX/basic'
(2015 03 01)
0.911 Improved the presentation with a quick introduction
(2015 04 10)
0.92 Introduced in folder 6, file start.py, a debug common variable; if set to True, a large part of the try/except structures will be bypassed, so the errors will be managed directly by the Python interpreter; this choice can be useful when you build a new project and as an expert user you want to check the errors in a basic way
(2015 04 10)
0.921 Set, in start.py, common.debug=False as default
(2015 04 21)
0.922 Modified ObserverSwarm.py correcting cycle => common.cycle to facilitate the access to this information in any part of the applications
(2015 04 24)
0.93 Added the project 'debug' to check debug capabilities. Have a look to the file debug.txt in the folder 'debug' of folder '6 objectSwarmObserverAgents_AESOP_turtleLib_NetworkX'
Modified a few error message of ModelSwarm.py if the required method does not exist
Improved the Tools.py output in case of error on a method
(2015 06 07)
0.94 'project' is by default both the name of the application and of the subfolder that contains its code; the subfolder is supposed to be placed within the SLAPP tree
With this version, the folder can be placed outside the SLAPP tree if we place a file project.txt in the folder "6 objectSwarmObserverAgents_AESOP_turtleLib_NetworkX"
The file has to contain the path and the name of the folder of the project
(2015 06 10)
0.95 simplified the reference to agentList in Agent.py in production project
simplified Agent.py in projects using a superAgent class as a container of unspecific methods
(2015 08 08)
0.951 fixes a bug in the version control for the library NetworkX, in production/parameters.py
(2015 08 08)
0.952 fixes a residual bug in the version control for the library NetworkX, in production/parameters.py
(2015 08 20)
0.953 adds the capability of recognizing if running in IPython or in Python
(2015 08 26)
0.96 clarified scheduling in Observer.py about 'visualizeNet'; corrected also oActions.py in production project; no changes in the results
(2015 08 31)
1.0 Having now a Handbook, this is version 1.0
(2015 09 02)
1.01 Reorganized start.py in the folder '6 objectSwarmObserverAgents_AESOP_turtleLib_NetworkX' considering the possibility running SLAPP also in IPython
(2015 09 10)
1.1 New starting point in the main folder (runShell.py)
The file project.txt (look at the Reference Handbook) can be also placed in the main folder
Better management of 'end' condition for graphical projects
(2015 09 29)
1.11 Allowing a wider use of WorldState. This is only a temporary patch, the interface to WorldState in ModelSwarm.py will be quite soon re-engineered, also having as optional the presence of WorldState.py in the folder of a project
(2016 02 15)
1.2 Introducing the extension .txtx (extended txt) to be used for the definition of the files describing the agents
Have a look to the Reference Handbook, section The agents and their sets, sub section Files .txtx in defining the agents
(2016 04 08)
1.21 Modifications facing v. 1.5.1 of matplotlib
New independent folder matplotlib_aQuestForAFewGraphicCapabilities to explore matplotlib in Python & IPython environments
From now on, SLAPP runs only via a terminal or in IPython (jupyter notebook), using runShell.py or iRunShell.ipynb (look at the Reference Handbook, section How to run SLAPP); it is also possible to run SLAPP in IPython via Spyder and in iptyhon in a terminal
In IPython, the magic command '%matplotlib inline' is internally added if missing; if '%matplotlib' is the explicit choice, the 'inline' option is internally stated
In the main folder now we have runShell.py to start the shell in Python and iRunShell.ipynb to start it in IPython (using "jupyter notebook"); have a look to the Reference Handbook, again section How to run SLAPP
The patch for matplotlib 1.3.4 is no more necessary and we have eliminated it; you can anyway retrieve it in the branch v.1.2 in the github repository of SLAPP
matplotlib produces an annoying warning about creating fonts; to avoid it, several hints online suggest to delete in ~.cache the folder fontconfig or matplotlib Instead, in MacOS go to the folder .matplolib in your home and delete the file fontList.cache The annoying warning will appear only one more time
(2016 04 15)
1.22 maintenance
corrected ModelSwarm.py in
class superAgent in agTools.py in
range # a b in schedule.xls now repeats (b - a + 1) times and not (b - a) times
more detailed graphic reactions to the graphical environment: Now SLAPP works
-
from a plain terminal, with: python runShell.py OR ipython AND internally %run runShell.py (with the same graphic results of python in terminal)
-
from jupiter notebook (with inline graphics)
-
from Jupiter QtConsole - e.g. Anaconda launcher - (with inline graphics)
-
from Spyder ipython (with inline graphics)
(look at the Reference Handbook, section How to run SLAPP)
(2016 07 25)
1.3 significant collection of maintenance modification (in GitHub history)
last change: fixed an error with Tkinter, using IPython 5.0 or greater
(2016 08 02)
1.31 new feature: the report in output of the probabilities of the methods, if any
(2016 08 17)
1.32 debug option can now be set from the project; clarified also in the Reference
(2016 08 21)
1.33 Reengineering WorldState, as anticipated in version 1.11, to have a clean structure, now working both as calculation tool and a repository of values, have a look to the Reference Handbook
In mAction.py modify any occurrence of
worldStateList[0]
to
worldState
Unfortunately you have to modify a bit your projects
If you do not use the WorldState feature, simply delete the file WorldState.py
If you use it, please modify
def __init__(self,number):
to
def __init__(self):
and eliminate any reference to number or to self.number, maybe also in the initial print, always in WorldState.py
(2016 08 22)
1.34 Eliminated everywhere the dictionary of the action groups
actionDictionary, unnecessary, but kept in Tools.py of
(2016 08 25)
1.35 Introduced the capability of the agents of adding or eliminating tasks in the schedule (at the level of the schedule reported in the schedule.xls file)
Have a look both to the example in 'basic' project, file scheduleAddElim.xls (to use it, rename it as schedule.xls) and to the section 'Agents adding and eliminating tasks into the detailed schedule' of the Reference Handbook
(2016 08 31)
1.36 Big step, now SLAPP uses agent of different classes, as in the example basic2classes
(2016 09 01)
1.4 Correction of minor problems and alignment of the Reference Handbook to the novelty of the multi-class capability, look at the Section "SLAPP multi-class: the basic2classes example"
(2016 09 11)
1.41 Improving the introductory files and the Reference Handbook
(2016 09 17)
1.5 Creating http://slapp-online.net:6789 to simplify the initial knowledge of SLAPP
(2016 09 25)
1.6 A step ahead in SLAPP presentation
(2016 11 17)
1.6.1 Fixed an error in production project: in addAFactory method - and in removeItself one, when applied to an added agents - the list of the agents was not updated because it was not set in the creation phase for the added agents
Also the lists of the other agents and that of the ModelSwarm, via the address now generated by ModelSwarm in commonVar.py, are updated with a new specific code, within the two method reported above
(2017 04 01)
1.6.2 Multi-class structure is now more general (look at Chapter 3 of the Reference Handbook contained in the distribution of SLAPP)
(2017 04 18)
1.6.3 Fixed an error in recognizing library versions and updated the Reference Handbook: the description of the basic example was not aligned to the modifications of version 1.6.2 about agent creation (Section 2.3)
(2017 04 24)
1.6.4 New feature 'specialAction' acting in specific time steps, look at the example 'basicSpecialAction' and to the paragraph pointed by index voice 'specialAction'
Put in common space the variable 'project' containing the path to the project
(2017 05 11)
2.0 Last version in Python 2, we will only report corrections for bugs, no improvements
The next version, with 3.0 number, will be running in Python 3
(2017 05 11)
3.0 New version in Python 3
Folder 7 (toBeDeveloped_aFewHints) is no more useful and has not reported (but you can find it until v.2.0)
Transition notes into the file from2to3.md with a technical part dedicated to a digression on methods and functions in Python 3
Fixing also .txtx file translation
(2017 06 07)
3.0.1
Applying PEP8, the Style Guide for Python Code
We used the program autopep8 via terminal, with:
autopep8 --in-place -a -a <filename>
and a few handmade corrections
(2017 07 22)
3.0.2
adapting to work also without TkInter, in case with limitations, to run also online in beta.mybinder.org, in the future simply mybinder.org
(2017 07 24)
3.0.3
adding the oligopoly project as a new example (version 5bPy3)
modified the .ipynb files to adapt them to the use with Binder or without it
(2017 07 28)
3.0.4
fixed the specialAction feature (with the basicSpecialAction example), broken with the conversion from Python 2 to Python 3
(2017 09 05)
3.0.5
in Tools.py the functions askEachAgentInCollection, askEachAgentInCollectionAndExecLocalCode, askAgent, and askAgentAndExecLocalCode were canceling the effect of an os.sys.exit(1) operation raised in the method under execution (if debug==False)
fixed duplicating the except condition, with except SystemExit in first position to repeat the os.sys.exit(1) action
the same in ModelSwarm.py when applying methods via the functions of Tools.py
(2017 11 13)
3.0.6
adapted production project to the version 2.0 of NetworkX
(2017 11 28)
3.0.7
adapted production project again to the version 2.0 of NetworkX
updated requirements.txt, addying statsmodels and patsy for readingCsvOutput.ipynb in oligopoly model
(2018 01 28)
3.0.8
in SLAPP3 a 0 probability of the execution of a method is now working
3.0.9
improved the report of method probabilities starting the execution
(2018 02 09)
3.1.0
new version of the 'basic' project to introduce 2D visualization
improved contents of iRunShellOnline.ipynb file
(2018 02 12)
3.1.1
fixed a misleading calculation about the default space limits in
consistently, we have corrected the init function in Agent.py of basic, basic2D, basic3D, basic2classes (also in otherAgent.py), basicSpecialAction, debug, and production, to eliminate numerical specific values, leaving 0 values as unused default initializations
new version of the 'basic' project to introduce 3D visualization
(2019 04 16)
3.2
new oligopoly folder (version 6book)
(2019 05 14)
3.3
improved oligopoly folder (version 6book, with stable results in Mac and Linux)
(2019 05 16)
3.3.1
eliminating the heavy files from the folder exampleGauss in oligopoly folder (version 6book, with stable results in Mac and Linux); they are now at https://github.com/terna/oligopolyBookCasesGaussianValues
(2019 08 07)
3.3.2
improvements in random number generation, look at Section "A note on random number generation" in the SLAPP_Reference_Handbook.pdf file
(2019 08 11)
3.3.3
a new way to input the parameters, look at Section "An option: adopting the INI file format for parameter input" in the SLAPP_Reference_Handbook.pdf file
(2019 08 16)
3.3.4
introducing the capability of managing subfolders (within the project folders, with path of any length), containing the files specifying the agents' classes; look at Section "Having agent classes in subfolder: the project basic2classesPathsToTypes"
(2020 03 26)
3.3.5
fixed an error in the choice of the project if, in presence of a predefined path, the reply was different from "n" or "y" or "Y" or "" (enter)
fixed an error in production project due to the new release of NetworkX
(2020 04 03)
3.3.6
changes in oligopoly project with: better myGauss functioning; nicer first time series graphic; more robust initial input cell reactions in case of wrong replies, such as empty reply or blanks when a content is mandatory