Difference between revisions of "SummerOfCodeProposal AI Improvement Crab"

From The Battle for Wesnoth Wiki
m (3. C++ ideas about AI coding phase)
(Hide historical instance of "SVN" from mediawiki search to avoid false positives.)
 
(26 intermediate revisions by 2 users not shown)
Line 48: Line 48:
 
- formula_ai
 
- formula_ai
  
These have two different approaches to implementing ai_inteface::play_turn() and each of them must be preserved and developed. But, a way is needed to develop them without  (A) changing too much and breaking them; (B) changing too little and not achieving anything;
+
These have two different approaches to implementing ai_inteface::play_turn(), recruitment, moves,  and each of them must be preserved and developed. But, a way is needed to develop them without  (A) changing too much and breaking them; (B) changing too little and not achieving anything;
  
To do this, I will start with a refactoring. I will split both default_ai and formula_ai into two (later there may be more) code lines. They can be called 'stable' and 'development', or 'champion' and 'challenger'. There may be more that two - "one champion plus multiple challengers". During this refactoring, pieces which are the same for each AI must be extracted into separate library code to avoid unnecessary duplication. The files will be moved into separate dir 'src/ai', since, IMO, 'ai' is a separate module that should be in it's own directory.
+
There is need for generic master ai, which will have a pluggable flow control, which, in turn, will be configured with specific implementations of specific phases. For example:  There can be a flow control implementation which consists of three phases: ai_component::leader_control, ai_component::candidate_moves, ai_component::moves .  the generic master ai could simply be a vector of ai_components that would be run in turn; it probably doesn't have to be more complicated (plus, later, some exception handling, timeout support, and other wrappers around component invocation). It will be cascaded (ai_component may also have child ai_components), although no dependency loops will be allowed. So if there are N phases of turn sequence, then there'll be master AI A, and ai_component:: implementations B1,B2,...,BN  (B1 may have B1.1,B1.2, B1.2 may have B1.2.1, B1.2.2, .., etc). and the master AI will call the 'turn sequence' components, which will call the rest of the components to make their work...
 +
 
 +
This will also allow us to *fork* the default_ai:: components, and store one copy of them (and don't touch them a lot), and experiment with the other copy, and compare them in-game. or *fork* one component a time. So we won't break anything by trying changes. See the #wesnoth-dev IRC logs of 04.04.2009 for the conversation between me, Dragonking, and silene where we talked about that ( 00:51 - 1:38 )
 +
 
 +
To do this, I will start with a refactoring. I will split both default_ai and formula_ai into components and *fork* them into two (later there may be more) code lines. They can be called 'stable' and 'development', or 'champion' and 'challenger'. There may be more that two - "one champion plus multiple challengers". During this refactoring, pieces which are the same for each AI must be extracted into separate library code to avoid unnecessary duplication. The files will be moved into separate dir 'src/ai', since, IMO, 'ai' is a separate module that should be in it's own directory.
  
 
We'll have the following situation:
 
We'll have the following situation:
Line 75: Line 79:
 
| ai/formula_*.hpp
 
| ai/formula_*.hpp
 
| various support files for formula language, such as parser
 
| various support files for formula language, such as parser
 +
|-
 +
| ai/modular_ai.hpp
 +
| generic master AI which will be constructed using AI components, according to WML configuration
 
|-
 
|-
 
| [[SummerOfCodeProposal_AI_Improvement_Crab_AI_Components|ai/components/*.hpp]]
 
| [[SummerOfCodeProposal_AI_Improvement_Crab_AI_Components|ai/components/*.hpp]]
Line 80: Line 87:
 
|-
 
|-
 
| ai/development/default_ai.hpp
 
| ai/development/default_ai.hpp
| development version of default AI
+
| development version of default AI components
 
|-
 
|-
 
| ai/development/formula_ai.hpp
 
| ai/development/formula_ai.hpp
| development version of formula_ai
+
| development version of formula_ai components
 
|-
 
|-
 
| ai/development/*_ai.hpp
 
| ai/development/*_ai.hpp
Line 89: Line 96:
 
|-
 
|-
 
| ai/stable/default_ai.hpp
 
| ai/stable/default_ai.hpp
| stable version of default AI
+
| stable version of default AI components
 
|-
 
|-
 
| ai/stable/formula_ai.hpp
 
| ai/stable/formula_ai.hpp
| stable_version of formula_ai
+
| stable_version of formula_ai components
 
|-
 
|-
 
| ai/stable/*_ai.hpp
 
| ai/stable/*_ai.hpp
Line 193: Line 200:
  
  
Each unit makes at most 1 move. This means, that position F is firstly optimized by F<sub>i1</sub>, then By F<sub>i2</sub>, ... then by F<sub>iN</sub>. This works in most cases, but there are cases where it doesn't work. It won't work in the situations where it is needed to optimize F by changing a subset of arguments at a time.
+
Each unit makes its moves. This means, that position F is firstly optimized by F<sub>i1</sub>, then By F<sub>i2</sub>, ... then by F<sub>iN</sub>, then by  F<sub>iN+1</sub>,..., (there can be partial moves, so the total number of moves may be not equal to number of units). This works in most cases, but there are cases where it doesn't work. It won't work in the situations where it is needed to optimize F by changing a subset of arguments at a time.
  
 
These situations are cases when units collaborate to achieve a result (backstab, leadership, illuminate, countering power projection by establishing a defensive line, etc). Such cases will not be easily *reachable* using current optimization plan ( for example, a backstab may require: (1) get friendly non-backstabber unit A into position, (2) attack enemy with backstabber B, (3) maybe - attack with unit A.  Note that evaluation of the benefit of doing (1) requires evaluation of all other candidate moves to see 'if any of them are 'backstab attacks' we can help do backstab'. Since current *candidate moves* are not reachable from .fai at the moment, this is very hard to get right. )
 
These situations are cases when units collaborate to achieve a result (backstab, leadership, illuminate, countering power projection by establishing a defensive line, etc). Such cases will not be easily *reachable* using current optimization plan ( for example, a backstab may require: (1) get friendly non-backstabber unit A into position, (2) attack enemy with backstabber B, (3) maybe - attack with unit A.  Note that evaluation of the benefit of doing (1) requires evaluation of all other candidate moves to see 'if any of them are 'backstab attacks' we can help do backstab'. Since current *candidate moves* are not reachable from .fai at the moment, this is very hard to get right. )
Line 238: Line 245:
 
| + ||Mar&nbsp;19&nbsp;-&nbsp;Mar&nbsp;22 || Familiarize myself with important parts of current Wesnoth AI code and Wesnoth way of doing things.
 
| + ||Mar&nbsp;19&nbsp;-&nbsp;Mar&nbsp;22 || Familiarize myself with important parts of current Wesnoth AI code and Wesnoth way of doing things.
 
|-
 
|-
| + || Mar&nbsp;23&nbsp;-&nbsp;Mar&nbsp;27 ||  Discuss implementation ideas and details. Fix some more bugs. Gain SVN commit access.
+
| + || Mar&nbsp;23&nbsp;-&nbsp;Mar&nbsp;27 ||  Discuss implementation ideas and details. Fix some more bugs. Gain S&shy;&shy;V&shy;&shy;N commit access.
 
|-
 
|-
 
| + || Mar&nbsp;28&nbsp;-&nbsp;Mar&nbsp;31 || Improve the AI lifecycle handling to allow AI hot-redeployment. Clean up current mess with AI parameter handling in team.cpp (without breaking anything, that is).  Create a test scenario for interactive testing of the AIs
 
| + || Mar&nbsp;28&nbsp;-&nbsp;Mar&nbsp;31 || Improve the AI lifecycle handling to allow AI hot-redeployment. Clean up current mess with AI parameter handling in team.cpp (without breaking anything, that is).  Create a test scenario for interactive testing of the AIs
 
|-
 
|-
|<b>now</b> || Apr&nbsp;1&nbsp;-&nbsp;Apr&nbsp;3 || Write an demonstration of the AI ideas I am trying to implement.
+
| +|| Apr&nbsp;1&nbsp;-&nbsp;Apr&nbsp;3 || Written an description and pseudocode of the AI idea I am trying to implement (Real implementation needs 'evaluate candidate move in hypothetical situation' part to work, which is hard to do in the current code, although I'll think about hacking it in just to show how it will work.)
 
|-
 
|-
| || April&nbsp;4&nbsp;-&nbsp;April&nbsp;20 || Step 0: Write extended proposed AI redesign description and create the AI testing framework.  
+
| +|| April&nbsp;4&nbsp;-&nbsp;April&nbsp;20 || Step 0: Write extended proposed AI redesign description and begin creating the AI testing framework, begin refactoring of the ai code.
 
|-
 
|-
| || April&nbsp;21&nbsp;-&nbsp;June&nbsp;10 || Step 1: Redesign current AI and FormulaAI code, split them into layers with defined responsibility, make the code cleaner, modular, and simpler, write a library of formula AI functions which will respect basic AI tuning parameters.
+
| +|| April&nbsp;21&nbsp;-&nbsp;July 31&nbsp; || Step 1: Redesign current AI and FormulaAI code, split them into layers with defined responsibility, finish the AI testing framework, make the code cleaner, modular, and simpler.
  
 
|-
 
|-
| || June&nbsp;11&nbsp;-&nbsp;July&nbsp;10 || Step 2: Increase .fai capabilities: namespaces, exception handling, numeric datatype, libraries. Write a good Formula_AI debugger.
+
| + || August&nbsp;1&nbsp;-&nbsp;August&nbsp;14 || Step 2: Refactor fai, increase .fai capabilities.
 
|-
 
|-
| || July&nbsp;11&nbsp;-&nbsp;August&nbsp;10 || Step 3: Implement my AI idea (candidate_move_tree) in C++.
+
| + || August&nbsp;15 || "'pencils down' date. Take a week to scrub code, write tests, improve documentation, etc."
 
|-
 
|-
| || August&nbsp;11 || "'pencils down' date. Take a week to scrub code, write tests, improve documentation, etc." (c) google
+
| now || Afterwards || - continue to improve the AI<br/>
|-
 
| || Afterwards || - continue to improve the AI<br/>
 
 
|}
 
|}
  
Line 262: Line 267:
  
 
I will need this time because AI is quite hard to do right. But I am confident in my abilities and my determination.
 
I will need this time because AI is quite hard to do right. But I am confident in my abilities and my determination.
 +
 +
== Goals ==
 +
{| border="1"
 +
! ID
 +
! PRIORITY
 +
! STEP
 +
! SUBPROJECT
 +
! RESULT
 +
! STATUS
 +
|-
 +
| 0
 +
| <p style="color:red">MUST</p>
 +
| 1.1
 +
| AI Module Redesign
 +
| I will complete the general redesign (file moves, external changes) of the AI module according to the proposed schema.
 +
| DONE
 +
|-
 +
| 1
 +
| <p style="color:red">MUST</p>
 +
| 1.1
 +
| AI Module Redesign
 +
| I will create a new composite_ai which will have pluggable turn sequence and pluggable turn sequence parts.
 +
| DONE
 +
|-
 +
| 2
 +
| <p style="color:red">MUST</p>
 +
| 1.1
 +
| AI Module Redesign
 +
| I will implement a simple turn sequence for my new composite AI which will be a 'vector of configurable phases'
 +
| DONE
 +
|-
 +
| 3
 +
| <p style="color:red">MUST</p>
 +
| 1.1
 +
| AI Module Redesign
 +
| I will split default_ai into several components which will be able to plug into composite_ai.
 +
| DONE
 +
|-
 +
| 4
 +
| <p style="color:red">MUST</p>
 +
| 1.1
 +
| AI Module Redesign
 +
| I will split formula_ai into several components which will be able to plug into composite_ai.
 +
| PARTIALLY
 +
|-
 +
| 5
 +
| <p style="color:red">MUST</p>
 +
| 1.1
 +
| AI Module Redesign
 +
| I will fork default_ai components into 'stable' and development.
 +
| DONE
 +
|-
 +
| 6
 +
| <p style="color:red">MUST</p>
 +
| 1.1
 +
| AI Module Redesign
 +
| I will fork formula_ai components into 'stable' and development.
 +
| DONE
 +
|-
 +
| 7
 +
| <p style="color:orange">SHOULD</p>
 +
| 1.1
 +
| AI Module Redesign
 +
| I will make it possible to 'build'  AI from components using a WML config file (if you are familiar with java, think of 'spring framework' and you'll understand. We don't have reflection in C++, but it still can be done in this case). Marked this as SHOULD because it will work ok even with C++ - based config.
 +
| DONE
 +
|-
 +
| 8
 +
| <p style="color:red">MUST</p>
 +
| 1.1
 +
| AI Module Redesign
 +
| I will make 'AI is an addon' paradigm possible (scan a folder of .cfg files, and construct a list of multiplayer AI's from them; set default AI aliases using a config file)
 +
| DONE
 +
|-
 +
| 9
 +
| <p style="color:red">MUST</p>
 +
| 1.1
 +
| AI Module Redesign
 +
| I will refactor ai_interface::info into a game_interface class, which will hide the internal details of the game state from the AI (I will need this abstraction layer in 1.3, to make my proposed way of AI tuning possible; also, I will need this abstraction layer in stage 3, to make evaluation of 'hypothetical' situations efficient). This class will 'mix-in' the actual game rules and actual game situation and AIs perception of it.
 +
| PARTIALLY
 +
|-
 +
| 10
 +
| <p style="color:orange">SHOULD</p>
 +
| 1.1
 +
| AI Module Redesign
 +
| I will try to enforce the separation of 'AI evaluations' and 'AI game actions' (move a unit, attack a unit, etc), both in c++ and in formula functions. I am marking this with "SHOULD" because I think that this will be met with opposition of other AI developers who prefer the current way of messing-with-game-state-directly-from-the-ai, so I may need to reach an agreement with them on that topic, which will not allow me to reach this goal fully.
 +
| DONE
 +
|-
 +
| 11
 +
| <p style="color:orange">SHOULD</p>
 +
| 1.1
 +
| AI Module Redesign
 +
| I will try to refactor current ai parameter handling to make it simpler and less error-prone. I am marking this as SHOULD because this is highly tied to savegame/scenario reorganization, and backward compatability questions, so I am not sure about how much I will be able to do in this area. AI parameter reorganization is a subject about which we should talk more, to figure out the best way to handle them.
 +
| DONE
 +
|-
 +
| 12
 +
| <p style="color:orange">SHOULD</p>
 +
| 1.2.1
 +
| Unit testing
 +
| There will be unit test for each C++ <i>'AI formula function'</i> present in 1.6. Marking this as SHOULD because if only some of them are covered, that's still good.
 +
|
 +
|-
 +
| 13
 +
| <p style="color:red">MUST</p>
 +
| 1.2.2
 +
| Unit testing
 +
| I will create a unit test for each external public interface function (used by rest of the game) in the AI module. ( This 'external public interface' means, in trunk,  'public interfaces of <i>ai_manager</i> and <i>ai_configuration</i>' )
 +
| NOT DONE
 +
|-
 +
| 14
 +
| <p style="color:red">MUST</p>
 +
| 1.2.3
 +
| Interactive testing
 +
| I will add (to AI Arena) at least 5 situations for each of the test topics: leadership, healing, poisoning, backstab, zoc awareness.
 +
| AI ARENA IMPROVEMENT IN PROGRESS (see patch #1169)
 +
|-
 +
| 15
 +
| <p style="color:red">MUST</p>
 +
| 1.2.4
 +
| Batch testing
 +
| I will organize a test server which will run test scenarios of "ai vs ai" and "ai arena" type and store their results to a small postgresql database with a small web frontend (numbers only). This test server will be able compile latest wesnoth source from s&shy;&shy;v&shy;&shy;n and do the tests, storing the results in the DB and sending the results (or compilation errors, to a pre-defined list of emails.
 +
| DONE
 +
|-
 +
| 16
 +
| <p style="color:orange">SHOULD</p>
 +
| 1.2.4
 +
| Batch testing
 +
| Integrate AI batch testing with stats.wesnoth.org (I'm listing this as 'SHOULD' because this depends not only on me, but also on guys who will do stats.wesnoth.org. On my side, I'll be able to write and host a small script in perl/php/python to pull/push ai batch testing stat data, or code this inside the game in c++ - but I do not know at the moment what method they will prefer.)
 +
| PARTIALLY, WORK IN PROGRESS
 +
|-
 +
| 17
 +
| <p style="color:red">MUST</p>
 +
| 1.3
 +
| Formula function library
 +
| I will refactor all 1.6 formula functions to use the game rule abstraction layer extracted in 1.1, thus they will take 'ai perceptions about game rules' into account.
 +
| PARTIALLY
 +
|-
 +
| 18
 +
| <p style="color:red">MUST</p>
 +
| 1.3
 +
| Formula function library
 +
| I will separate, from formula_ai, all 'formula functions' in formula_ai.cpp which can be made AI-independent.
 +
| DONE BY DK
 +
|-
 +
|
 +
| <p style="color:blue">MILESTONE</p>
 +
| June 10
 +
| Step 1 MILESTONE
 +
| Everything in Step 1 is to be completed by June 10.
 +
|
 +
|-
 +
| 19
 +
| <p style="color:red">MUST</p>
 +
| 2.1.1
 +
| .fai improvements
 +
| I will add library support to .fai, based on my run_file implementation
 +
| NOT DONE
 +
|-
 +
| 20
 +
| <p style="color:red">MUST</p>
 +
| 2.1.2
 +
| .fai improvements
 +
| I will add namespace support to .fai
 +
| NOT DONE
 +
|-
 +
| 21
 +
| <p style="color:orange">SHOULD</p>
 +
| 2.1.3
 +
| .fai improvements
 +
| I will add numeric type support to .fai .  Marking as SHOULD, because It is not that important, since it can be replicated in current code, it will just be harder to write without explicit language support.
 +
| DONE BY DK
 +
|-
 +
| 22
 +
| <p style="color:red">MUST</p>
 +
| 2.1.4
 +
| .fai improvements
 +
| I will allow allow exception handling in .fai - throw and catch exceptions
 +
| WIP
 +
|-
 +
| 23
 +
| <p style="color:orange">SHOULD</p>
 +
| 2.1.4
 +
| .fai improvements
 +
| I will allow to set time limits on .fai execution. Marking this as SHOULD because this requires multithreading (watchdog on a timer) and this may lead to some problems which I cannot foresee at this point.
 +
| WIP
 +
|-
 +
| 24
 +
| <p style="color:red">MUST</p>
 +
| 2.2
 +
| FormulaAI debugger
 +
| I will make a formula AI debugger, with basic stepping capabilities
 +
| DONE
 +
|-
 +
| 25
 +
| <p style="color:red">MUST</p>
 +
| 2.2
 +
| FormulaAI debugger
 +
| I will add breakpoint capabilities to formula AI debugger
 +
| WIP
 +
|-
 +
| 26
 +
| <p style="color:red">MUST</p>
 +
| 2.2
 +
| FormulaAI debugger
 +
| I will add tracing capabilities to formula AI debugger
 +
| DONE
 +
|-
 +
| 27
 +
| <p style="color:red">MUST</p>
 +
| 2.2
 +
| FormulaAI debugger
 +
| I will add time profiling capabilities to formula AI debugger
 +
| WIP
 +
|-
 +
|
 +
| <p style="color:blue">MILESTONE</p>
 +
| July 10
 +
| Step 2 MILESTONE
 +
| Everything in Step 2 is to be completed by July 10.
 +
|
 +
|-
 +
| 28
 +
| <p style="color:red">MUST</p>
 +
| 3
 +
| Candidate Move Evaluation
 +
| I will write a framework for evaluating candidate moves in 'hypothetical' situations (which differ from current game state by a small amount - e.g. differ by expected results of several moves) without doing a full copy of the current game state.
 +
| NOT DONE, POSTPONED
 +
|-
 +
| 29
 +
| <p style="color:red">MUST</p>
 +
| 3
 +
| Candidate Move Evaluation
 +
| I will write a a 'build and Evaluate candidate move tree' turn sequence, based on my 'candidate_move_tree'  search improvement proposal, which will run with 'acceptable' performance for at least max_depth=2 (that is better than todays algorithm, in which it effectively equals 1)
 +
| NOT DONE, POSTPONED
 +
|-
 +
| 30
 +
| <p style="color:orange">SHOULD</p>
 +
| 3
 +
| Candidate Move Evaluation
 +
| I will write an algorithm to aggressively prune unpromising solutions from the candidate move tree, thus allowing to set bigger max_depth in 'candidate_move_tree' implementation. Marking this as 'SHOULD', because even with max-depth=2 my implementation will be better that current in spotting synergy opportunities.
 +
| NOT DONE, POSTPONED
 +
|-
 +
|
 +
| <p style="color:blue">MILESTONE</p>
 +
| August 10
 +
| Step 3 MILESTONE
 +
| Everything in Step 3 is to be completed by August 10.
 +
|
 +
|}
  
 
== Patches and commits ==
 
== Patches and commits ==
I've earned a <b>svn write access</b> to the Wesnoth source repository on Mar 27.<br/>
+
I've earned a <b>s&shy;&shy;v&shy;&shy;n write access</b> to the Wesnoth source repository on Mar 27.<br/>
 
I've already contributed some patches to make debugging easier and to fix some crashes related with FormulaAI. <br/>
 
I've already contributed some patches to make debugging easier and to fix some crashes related with FormulaAI. <br/>
 
[[https://gna.org/bugs/?13218 Bug 13218]] - I've spotted and fixed an infinite loop (which hangs the game) in the AI if it persistently tries to move unit to occupied square, with a hint from <i>Sirp</i> and some help from <i>Dragonking</i><br/>
 
[[https://gna.org/bugs/?13218 Bug 13218]] - I've spotted and fixed an infinite loop (which hangs the game) in the AI if it persistently tries to move unit to occupied square, with a hint from <i>Sirp</i> and some help from <i>Dragonking</i><br/>
Line 275: Line 528:
 
[[http://svn.gna.org/viewcvs/wesnoth?rev=34329&view=rev r34329]] - I've used those new capabilities to create an [[AI_Arena| AI Arena - a framework to easily test the AIs in specified situations]]<br/>
 
[[http://svn.gna.org/viewcvs/wesnoth?rev=34329&view=rev r34329]] - I've used those new capabilities to create an [[AI_Arena| AI Arena - a framework to easily test the AIs in specified situations]]<br/>
 
[[http://svn.gna.org/viewcvs/wesnoth?rev=34378&view=rev r34378]] - I've fixed enemy_units formula to not include incapacitated units, such as stoned units on Caves of the Basilisk<br/>
 
[[http://svn.gna.org/viewcvs/wesnoth?rev=34378&view=rev r34378]] - I've fixed enemy_units formula to not include incapacitated units, such as stoned units on Caves of the Basilisk<br/>
 +
[[https://gna.org/bugs/?13320 FR 13320]] - I've made [modify_side] able to hack the share_view attribute<br/>
 +
[[https://gna.org/bugs/?13345 Bug 13345]] - I've fixed a crash at start of scenario 19 in "The Legend of Wesmere".<br/>
 +
[[https://gna.org/bugs/?13348 FR 13348]] - I've added info about current attack bonus (due to timeofday/alignment) of unit in Formula AI.<br/>
 +
[[https://gna.org/bugs/?13352 Bug 13352]] - I've fixed a bug with allied AI sides behaving stupidly if a leader's initial village capture moves him closer to his ally's keep than his own.<br/>
  
 
<b>My immediate work plans:</b><br/>
 
<b>My immediate work plans:</b><br/>
Show a small working demonstration of my C++ AI Ideas<br/>
 
 
Make the game get multiplayer AI list from the list of AI configuration files in data/ai/ais/ directory to allow 'AI is an addon' paradigm to work <br/>
 
Make the game get multiplayer AI list from the list of AI configuration files in data/ai/ais/ directory to allow 'AI is an addon' paradigm to work <br/>
 +
Continue my refactoring of ai-game interface<br/>
  
 
==Answers to those questions==
 
==Answers to those questions==
Line 291: Line 548:
  
 
<b>1.3) If you have chosen a nick for IRC and Wesnoth forums, what is it?</b><br/>
 
<b>1.3) If you have chosen a nick for IRC and Wesnoth forums, what is it?</b><br/>
IRC: Crab_ <br/>
+
IRC: Crab_ [n=Crab_@wesnoth/developer/crab] <br/>
 
Wesnoth forums: Crab <br/>
 
Wesnoth forums: Crab <br/>
 
Wesnoth wiki: Crab <br/>
 
Wesnoth wiki: Crab <br/>
Line 302: Line 559:
 
Applied Mathematics, 6th year, Faculty of Cybernetics, Kiev University, Ukraine.
 
Applied Mathematics, 6th year, Faculty of Cybernetics, Kiev University, Ukraine.
  
<b>1.6) If you have contributed any patches to Wesnoth, please list them below. You can also list patches that have been submitted but not committed yet and patches that have not been specifically written for Wesnoth. If you have gained commit access to our SVN (during the evaluation period or earlier) please state so.</b><br/>
+
<b>1.6) If you have contributed any patches to Wesnoth, please list them below. You can also list patches that have been submitted but not committed yet and patches that have not been specifically written for Wesnoth. If you have gained commit access to our S&shy;&shy;V&shy;&shy;N (during the evaluation period or earlier) please state so.</b><br/>
Yes, I've contributed some patches and fixed some bugs. I've earned commit access to Wesnoth svn on March 27. See [[SummerOfCodeProposal_AI_Improvement_Crab#Patches_and_commits|My Patches and commits]]
+
Yes, I've contributed some patches and fixed some bugs. I've earned commit access to Wesnoth s&shy;&shy;v&shy;&shy;n on March 27. See [[SummerOfCodeProposal_AI_Improvement_Crab#Patches_and_commits|My Patches and commits]]
  
 
<h3>Experience</h3>
 
<h3>Experience</h3>
Line 385: Line 642:
 
<b>5.1) Are you familiar with any of the following tools or languages?</b><br/>
 
<b>5.1) Are you familiar with any of the following tools or languages?</b><br/>
  
<b>* Subversion (used for all commits)</b><br/>
+
<b>* Subv&shy;&shy;ersion (used for all commits)</b><br/>
 
yes, both with command-line and gui.<br/>
 
yes, both with command-line and gui.<br/>
 
<b>* C++ (language used for all the normal source code)</b><br/>
 
<b>* C++ (language used for all the normal source code)</b><br/>
Line 396: Line 653:
  
 
<b>5.2) Which tools do you normally use for development? Why do you use them?</b><br/>
 
<b>5.2) Which tools do you normally use for development? Why do you use them?</b><br/>
vi, gdb, find, fgrep, small shell scripts, svn, git, Eclipse under GNU/Linux; Visual Studio and Eclipse under Windows. My work machine is a notebook running Debian GNU/Linux.
+
vi, gdb, find, fgrep, small shell scripts, s&shy;&shy;v&shy;&shy;n, git, Eclipse under GNU/Linux; Visual Studio and Eclipse under Windows. My work machine is a notebook running Debian GNU/Linux.
  
 
<b>5.3) What programming languages are you fluent in?</b><br/>
 
<b>5.3) What programming languages are you fluent in?</b><br/>
Line 406: Line 663:
  
 
<b>5.5) At what hours are you awake and when will you be able to be in IRC (please specify in UTC)</b><br/>
 
<b>5.5) At what hours are you awake and when will you be able to be in IRC (please specify in UTC)</b><br/>
I'm in UTC+2 timezone. I am normally awake from +10 UTC to +02 UTC, and most of that time I'm at home, so I'm able to be in IRC.
+
I'm in UTC+3 timezone in summer. I am normally awake from +9 UTC to +01 UTC, and most of that time I'm at home, so I'm able to be in IRC.
  
 
<b>5.6) Would you mind talking with your mentor on telephone / internet phone? We would like to have a backup way for communications for the case that somehow emails and IRC do fail.</b><br/>
 
<b>5.6) Would you mind talking with your mentor on telephone / internet phone? We would like to have a backup way for communications for the case that somehow emails and IRC do fail.</b><br/>

Latest revision as of 02:53, 21 March 2013

Hello

My name is Yurii, My nick on IRC is Crab_, my nick on forum/wiki is Crab, on gna - crab.

terraninfo АТ terraninfo.net

Proposal Summary

I would like to reorganize the current AI subsystem to allow different people collaborate more efficiently on creating new AIs and improving AIs.

Also, I would like to make the Wesnoth AI better.

I propose a three-part project:


1. AI Module design phase

1.1 I will introduce a new AI development process. The two main goals:

- make AI development possible without breaking old AI.

- make AI development convenient.

- make measurement of results possible.

To do so, I propose to fork each ai into 'dev' and 'stable' variants, which will be present in the same game (so we will be able to compare them and gradualy make 'stable' variant better.

Stable versions of each AI files will be changed *only* for bugfixes or when testing shows that the change is good.

This is the only sane way I see of making AI progress without either regularly breaking it (if devs are too confident) or stalling progress (if devs are too cautious) - we need to *preserve* old AI for long enough to see if a patched AI is good.

To do so, I've started with separating AI code from game code, by creating an AI manager class. I will introduce support for getting AI from an algorithm_type via an indirect lookup (from an AI ALIAS)

There will be several important AI ALIASES:

- CHAMPION_MP_AI - AI which is known to be good for multiplayer. It will be default for MP scenarios.

- CHALLENGER_MP_AI - AI which is a candidate to become the "CHAMPION MP AI", and will be promoted to "Champion MP AI" if it proves itself better.

- CHAMPION_SP_AI - AI which is known to be good for singleplayer. It will be default or SP scenarios.

- CHALLENGER_SP_AI - AI which is a candidate to become the "CHAMPION SP AI", and will be promoted to "Champion SP AI" if it proves itself better.

(Of course, naming a concrete AI via ai_algorithm_type will be still possible, nothing will break at this point)

I see two 'active' AIs in current codebase:

- default_ai

- formula_ai

These have two different approaches to implementing ai_inteface::play_turn(), recruitment, moves, and each of them must be preserved and developed. But, a way is needed to develop them without (A) changing too much and breaking them; (B) changing too little and not achieving anything;

There is need for generic master ai, which will have a pluggable flow control, which, in turn, will be configured with specific implementations of specific phases. For example: There can be a flow control implementation which consists of three phases: ai_component::leader_control, ai_component::candidate_moves, ai_component::moves . the generic master ai could simply be a vector of ai_components that would be run in turn; it probably doesn't have to be more complicated (plus, later, some exception handling, timeout support, and other wrappers around component invocation). It will be cascaded (ai_component may also have child ai_components), although no dependency loops will be allowed. So if there are N phases of turn sequence, then there'll be master AI A, and ai_component:: implementations B1,B2,...,BN (B1 may have B1.1,B1.2, B1.2 may have B1.2.1, B1.2.2, .., etc). and the master AI will call the 'turn sequence' components, which will call the rest of the components to make their work...

This will also allow us to *fork* the default_ai:: components, and store one copy of them (and don't touch them a lot), and experiment with the other copy, and compare them in-game. or *fork* one component a time. So we won't break anything by trying changes. See the #wesnoth-dev IRC logs of 04.04.2009 for the conversation between me, Dragonking, and silene where we talked about that ( 00:51 - 1:38 )

To do this, I will start with a refactoring. I will split both default_ai and formula_ai into components and *fork* them into two (later there may be more) code lines. They can be called 'stable' and 'development', or 'champion' and 'challenger'. There may be more that two - "one champion plus multiple challengers". During this refactoring, pieces which are the same for each AI must be extracted into separate library code to avoid unnecessary duplication. The files will be moved into separate dir 'src/ai', since, IMO, 'ai' is a separate module that should be in it's own directory.

We'll have the following situation:

AI Module

ai/manager.hpp Class which is managing the AI lifecycle. This is a single-point-of-interaction between AI module and rest of Wesnoth code. Individual AI implementations will *register* themselves with the ai/manager, and will be available to the rest of the game engine by name or by alias.
ai/configuration.hpp Class which is dealing with AI memory, configuration, effective configuration, global_configuration, etc
ai/ai_interface.hpp an interface for the ai, from the point-of-view of the ai/manager.
ai/game_interface.hpp an interface for the game situation, from the point-of-view of the AI implementation. a class based on current ai_interface::info in order to encapsulate access to everything AI needs.
ai/game_rules.hpp the game rules functions. Game rules functions are those C++ functions which are AI-independent and which evaluate their result based on the AI's perception on the rules of the game (AIs perceptions of the game rules are given to them as a parameter). Therefore, we can have only one copy of these functions, since if we have two such versions which yield different results, then one of them is guaranteed wrong (since they do not make any decisions, they simply answer a question about the game rules (taking AIs perception of the game rules into account))
ai/formula_language.hpp Code that evaluates formula functions. Current formula functions are AI-independent too. So, they should be kept separate from the AIs (to make possible for any AI to use them).
ai/formula_*.hpp various support files for formula language, such as parser
ai/modular_ai.hpp generic master AI which will be constructed using AI components, according to WML configuration
ai/components/*.hpp various components which do *decisions* (they can be reused by various AIs, but it is needed to keep many versions around, because it is not immediately clear whether a decision is good or bad.
ai/development/default_ai.hpp development version of default AI components
ai/development/formula_ai.hpp development version of formula_ai components
ai/development/*_ai.hpp stable version of other AIs
ai/stable/default_ai.hpp stable version of default AI components
ai/stable/formula_ai.hpp stable_version of formula_ai components
ai/stable/*_ai.hpp stable version of other AIs

...

Each of these AI will be backed by its own .cfg/.fai files, containing the AI default configuration and decision support scripts.

NOTE: I have already written ai_manager and ai_configuration, which made me very familiar with current AI lifecycle and integration points between team and AI, and integration points between the game and the AI.

To remove code duplication in stable/development, as much as possible will be extracted into ai-independent part. For example, most of the FormulaAI functions in formula_ai.cpp do not really belong there - they are not related with AI decision-making, and are not related with formula_ai, and are not related to ai_interface (for example, nearest_keep is the same for each ai - so, it should not be a part of formula_ai, it should be part of a common support library). Formula AI functions should be usable by ALL AIs that want to do so.

IMPORTANT NOTE: btw, a question may arise at this point: what is formula_ai then, if 'formula AI functions' are NOT part of it? Answer: formula_ai is code that makes a turn by evaluating some formulas in the specific order. It is that order of evaluation (turn sequence) that makes this AI unique. Even default_ai can evaluate a formula if it wants so. It may be even better to rename formula_ai.?pp to avoid confusion between 'formula ai functions' (which are NOT part of formula AI) and 'formula_ai'. It, also, immediately becomes clear that no AI can be written using just 'formula ai functions' without tuning the C++ part that controls evaluation sequence.

So, I will be able to remove a lot of code from formula_ai before splitting it into two versions. To further remove code duplication in stable/development, modularity in the AIs will be increased. for example, if we have two code paths A,B and A,C, I'll refactor it to A(X) and make both B and C to 'provide X'. This will allow us to keep one copy of code fragment A instead of two. This will have a side benefit of allowing us to plug-in another implementation of X later, if such an idea appears.

The primary candidate for such a refactoring is the formula_ai turn sequence I've spoken earlier. Turn sequence in the critical part of the AI and by allowing it to be pluggable we will allow ourselves to test a low of different ideas and possibilities. Several of the ideas about the turn sequence may include multi-unit evaluations, unit reservations, formula exception handling and flow control interruption, profiling, etc. To be able to test all of them, I will make the turn sequence pluggable. Also, recruitment is a candidate for such a refactoring.

and, as the last step here, I'll point CHAMPION_MP_AI, CHAMPION_SP_AI to "ai/stable/default_ai". CHALLENGER_MP_AI and CHALLENGER_SP_AI to "ai/development/default_ai"

This is, in itself, only a refactoring.

1.2 The next part of the AI Quest is to make any AI changes testable, preferably - with accounting in battle-proven and web-accessible numbers.

We need different kinds of tests.

1.2.1 Firstly, we need unit tests for C++ service functions. This will be made on the base of current unit testing framework (src/tests)

1.2.2 Secondly, we need unit tests for formulas. These, also, will be made on current unit testing framework (src/tests)

1.2.3 Thirdly, we need interactive testing. See AI_Arena - I've created a test scenario which is AI-independent and will allow testing multiple variants of AIs in specific situations. This test scenario also allows shorter formula development cycle (by using AI hot-redeployment functionality I've implemented). I will expand this test framework by creating test situations to cover all basic aspects of gameplay and to cover all unit abilities available in Wesnoth. I will expand this test framework to allow 'test runs' of multiple tests and multiple AI's per one developer command (with storage of the results in the log). I will expand this framework to allow easier 'parallel' testing of 'Champion AIs' and 'Challenger AIs' - to detect regressions easier.

1.2.4 Fourthly, we need non-interactive (batch) testing. In addition to interactive testing, we need to detect degradations and test the ideas without wasting valuable developer time. A good example of such a test is a full-scale battle of CHALLENGER AI vs CHAMPION AI. This also includes unfair battles (with one side having a distinct advantage) and pre-defined scenarios (from AI Arena) This testing has a very nice property - it can run 24x7, run many times, and store the results to the database, which will have a web-accessible frontend with 'interesting numbers' (later, it could be even possible to trick one of those GSoC students which will be working on stats.wesnoth.org into making some nice graphs ). This will act as a limited safety net for the AI developers (if they break the development AI, there is a good chance it will be noticeable during such testing), and this can also be used as an indicator for the replacement of CHAMPION AI with a better one.

NOTE: a question may arise: where to get this 24x7-working machine to make such testing possible ? This will not be a problem, since I've got my own server on a colo, and it will only need to get a small memory upgrade to compile wesnoth quicker, which is easily doable.


1.3 during the aforementioned refactoring, base formula functions will be pushed from formula_ai.?pp into a separate file. This separate file will become a service library of various functions. I will make them aware of the basic AI tune options.

- I will make a library of base FormulaAI functions which will be aware of the basic AI tune options and honour them. I will implement this as a thin layer between 'game rules' and them, called 'perception about game rules'. It will allow the SP scenario designers to fool the AI in a very elegant and declarative (no code to write, only declarations) way.

Example of proposed WML syntax:
[ai]
  [perception]
      [filter]
          [filter_location]
              terrain=Gs^Fp
          [/filter_location]
      [/filter]
      my_defence_bonus=-40%
  [/perception]
[/ai]

and the AI will, in all its decisions (including formula evaluation), *think*, that those forests give its units -40% defence from the standard %value (so, naturally, it will prefer to use non-forested tiles). This will be a very good and easy-to-use-and-easy-to-understand tool for SP designers.


This is a first, and important, milestone for my proposed project. A short overview: refactoring, code split to prevent breakage, testing framework, and tuning-parameter-awareness as a thin layer between game rules and service functions. This milestone will allow me to go forward with improving the formula language and AI ( remember, formula language is NOT formula_ai, and formula language is not a part of a formula_ai. So, after decoupling them we can easily experiment with, for example, adding formula recruitment to development/default_ai - and we will have test numbers about the efficiency of the result, thanks to the constructed test system. )

2. Formula language capabilities phase

2.1 I will increase FormulaAI capabilities by introducing:

2.1.1 .fai library support

2.1.2 .fai namespace support

2.1.3 .fai numeric type (example: 3.14157)

2.1.4 .fai exception handling

2.2 I will also create a FormulaAI debugger\profiler, to allow stepping through formulas, setting breakpoints, tracing and profiling, etc.

This is a second milestone: it is very small compared to the first because all the heavy lifting is supposed to be already done.

3. C++ ideas about AI coding phase

As I expect (1)-(2) to be finished early (till July), I will use the rest of the SoC time to implement my ideas about a new C++ AI.

Let us look at current (trunk) 'candidate move evaluation' ideology from a math-like point of view:

Let F(x1,x2,x3,...,xn,ENEMY_UNITS) be a position evaluation function,

x1,x2,x3,...,xn - our units

This function is unknown to us, but this doesn't matter.

Current candidate move implementation attempts to optimize F, (F -> max), by :


void formula_ai::make_candidate_moves() {
       //evaluate candidate moves
       while( "any candidate moves present" ) {
               // If none of the candidate moves evaluate to > 0, fallback
               // Otherwise, make the best scoring move
               // And re-evaluate candidate moves, get best move.
       }
}


Each unit makes its moves. This means, that position F is firstly optimized by Fi1, then By Fi2, ... then by FiN, then by FiN+1,..., (there can be partial moves, so the total number of moves may be not equal to number of units). This works in most cases, but there are cases where it doesn't work. It won't work in the situations where it is needed to optimize F by changing a subset of arguments at a time.

These situations are cases when units collaborate to achieve a result (backstab, leadership, illuminate, countering power projection by establishing a defensive line, etc). Such cases will not be easily *reachable* using current optimization plan ( for example, a backstab may require: (1) get friendly non-backstabber unit A into position, (2) attack enemy with backstabber B, (3) maybe - attack with unit A. Note that evaluation of the benefit of doing (1) requires evaluation of all other candidate moves to see 'if any of them are 'backstab attacks' we can help do backstab'. Since current *candidate moves* are not reachable from .fai at the moment, this is very hard to get right. )


I propose a better (albeit more cpu-consuming) variant, which will allow to handle those synergistic situations more easily and with easier candidate_move_generation part (this will offset cpu consumption and developer headache to some degree - simpler formulas are simpler to get right )

I name my proposal candidate move tree

void create_candidate_move_tree(candidate_move_tree& cmt, ...)
{
    //(1) for each of leaf moves from the tree
         //(2) create a hypothetical 'game situation after the move'
         //(3) evaluate possible candidate move tree in that hypothetical situation, based on that leaf node.
         //(4) prune bad leaf moves from the tree.
}


(5) To do a candidate move, top-level node that is best, is taken by the AI.

Definition of *best* will be pluggable.

A suitable definition of best: node that yields best expectations is taken.

Another definition of best may be formulated to allow the AI to favor risky choices (nodes that *may* lead to substantially higher result that expected, %-of-a-time), or to allow the AI to favor good defensive position, etc)

NOTES:

(2) as each move/attack only changes a small bit of the battlefield, 'create a hypothetical 'game situation after the move' can be done without 'time+memory-consuming' copying of all the current game situation variables.

(3) should be implemented a recursive call to create_candidate_move_tree. recursion should be limited (by time, by depth, by possible variants)

(4) pruning should become more and more aggressive with depth increase. Pruning is needed to calculate *synergistic* variants in more detail while abandoning non-synergistic ones (for example, 'leader helping several units per turn' is an interesting sequence worth considering in depth, and many moves and attacks can be done in any sequence without synergy benefits), and do not spend much time in situation where benefits of synergy are less powerful than the randomness of Wesnoth. I do not intend to generate a "branch prediction", I need to get only highly synergistic options (leadership and zone of control restrictions on enemy power projection being the most important ones)

P.S: Current situation is "max_depth=1". Even max_depth=2 will allow better AI choices with simpler formulas.

P.P.S. Another solution of the 'synergy' problem is to allow the AI to evaluate formulas for N-units-a-time, but 'N-units-a-time' will scale poorly with N increasing, while my variant allows to combine bigger max_depth with agressive pruning of uninteresting variants, this searching only a subset of C(my_units.size(),N) combination space.

Proposed synergistic-situations search is similar what humans do when they consider turns - they don't think about all the combinations, but they devote time to most promising ones.

Timeline

+ Mar 19 - Mar 22 Familiarize myself with important parts of current Wesnoth AI code and Wesnoth way of doing things.
+ Mar 23 - Mar 27 Discuss implementation ideas and details. Fix some more bugs. Gain S­­V­­N commit access.
+ Mar 28 - Mar 31 Improve the AI lifecycle handling to allow AI hot-redeployment. Clean up current mess with AI parameter handling in team.cpp (without breaking anything, that is). Create a test scenario for interactive testing of the AIs
+ Apr 1 - Apr 3 Written an description and pseudocode of the AI idea I am trying to implement (Real implementation needs 'evaluate candidate move in hypothetical situation' part to work, which is hard to do in the current code, although I'll think about hacking it in just to show how it will work.)
+ April 4 - April 20 Step 0: Write extended proposed AI redesign description and begin creating the AI testing framework, begin refactoring of the ai code.
+ April 21 - July 31  Step 1: Redesign current AI and FormulaAI code, split them into layers with defined responsibility, finish the AI testing framework, make the code cleaner, modular, and simpler.
+ August 1 - August 14 Step 2: Refactor fai, increase .fai capabilities.
+ August 15 "'pencils down' date. Take a week to scrub code, write tests, improve documentation, etc."
now Afterwards - continue to improve the AI


I've got an unique opportunity to improve Wesnoth because I am a last year student and my exams are already over (they were in February-09), and I have no lectures to attend and almost none university work to do - that means I am 90% free to improve Wesnoth (remaining 10% being my teaching practice, coordinating the work of several student developeres, and working towards my diploma). I'm able to work on improving Wesnoth 6 days a week (remaining day being my Sunday D&D session - I've been a D&D DM for 9+ years.), and hacking some code for 12+ hours per day is the thing I love:) So, I can easily devote 48+ hours to Wesnoth each week, even now.

I will need this time because AI is quite hard to do right. But I am confident in my abilities and my determination.

Goals

ID PRIORITY STEP SUBPROJECT RESULT STATUS
0

MUST

1.1 AI Module Redesign I will complete the general redesign (file moves, external changes) of the AI module according to the proposed schema. DONE
1

MUST

1.1 AI Module Redesign I will create a new composite_ai which will have pluggable turn sequence and pluggable turn sequence parts. DONE
2

MUST

1.1 AI Module Redesign I will implement a simple turn sequence for my new composite AI which will be a 'vector of configurable phases' DONE
3

MUST

1.1 AI Module Redesign I will split default_ai into several components which will be able to plug into composite_ai. DONE
4

MUST

1.1 AI Module Redesign I will split formula_ai into several components which will be able to plug into composite_ai. PARTIALLY
5

MUST

1.1 AI Module Redesign I will fork default_ai components into 'stable' and development. DONE
6

MUST

1.1 AI Module Redesign I will fork formula_ai components into 'stable' and development. DONE
7

SHOULD

1.1 AI Module Redesign I will make it possible to 'build' AI from components using a WML config file (if you are familiar with java, think of 'spring framework' and you'll understand. We don't have reflection in C++, but it still can be done in this case). Marked this as SHOULD because it will work ok even with C++ - based config. DONE
8

MUST

1.1 AI Module Redesign I will make 'AI is an addon' paradigm possible (scan a folder of .cfg files, and construct a list of multiplayer AI's from them; set default AI aliases using a config file) DONE
9

MUST

1.1 AI Module Redesign I will refactor ai_interface::info into a game_interface class, which will hide the internal details of the game state from the AI (I will need this abstraction layer in 1.3, to make my proposed way of AI tuning possible; also, I will need this abstraction layer in stage 3, to make evaluation of 'hypothetical' situations efficient). This class will 'mix-in' the actual game rules and actual game situation and AIs perception of it. PARTIALLY
10

SHOULD

1.1 AI Module Redesign I will try to enforce the separation of 'AI evaluations' and 'AI game actions' (move a unit, attack a unit, etc), both in c++ and in formula functions. I am marking this with "SHOULD" because I think that this will be met with opposition of other AI developers who prefer the current way of messing-with-game-state-directly-from-the-ai, so I may need to reach an agreement with them on that topic, which will not allow me to reach this goal fully. DONE
11

SHOULD

1.1 AI Module Redesign I will try to refactor current ai parameter handling to make it simpler and less error-prone. I am marking this as SHOULD because this is highly tied to savegame/scenario reorganization, and backward compatability questions, so I am not sure about how much I will be able to do in this area. AI parameter reorganization is a subject about which we should talk more, to figure out the best way to handle them. DONE
12

SHOULD

1.2.1 Unit testing There will be unit test for each C++ 'AI formula function' present in 1.6. Marking this as SHOULD because if only some of them are covered, that's still good.
13

MUST

1.2.2 Unit testing I will create a unit test for each external public interface function (used by rest of the game) in the AI module. ( This 'external public interface' means, in trunk, 'public interfaces of ai_manager and ai_configuration' ) NOT DONE
14

MUST

1.2.3 Interactive testing I will add (to AI Arena) at least 5 situations for each of the test topics: leadership, healing, poisoning, backstab, zoc awareness. AI ARENA IMPROVEMENT IN PROGRESS (see patch #1169)
15

MUST

1.2.4 Batch testing I will organize a test server which will run test scenarios of "ai vs ai" and "ai arena" type and store their results to a small postgresql database with a small web frontend (numbers only). This test server will be able compile latest wesnoth source from s­­v­­n and do the tests, storing the results in the DB and sending the results (or compilation errors, to a pre-defined list of emails. DONE
16

SHOULD

1.2.4 Batch testing Integrate AI batch testing with stats.wesnoth.org (I'm listing this as 'SHOULD' because this depends not only on me, but also on guys who will do stats.wesnoth.org. On my side, I'll be able to write and host a small script in perl/php/python to pull/push ai batch testing stat data, or code this inside the game in c++ - but I do not know at the moment what method they will prefer.) PARTIALLY, WORK IN PROGRESS
17

MUST

1.3 Formula function library I will refactor all 1.6 formula functions to use the game rule abstraction layer extracted in 1.1, thus they will take 'ai perceptions about game rules' into account. PARTIALLY
18

MUST

1.3 Formula function library I will separate, from formula_ai, all 'formula functions' in formula_ai.cpp which can be made AI-independent. DONE BY DK

MILESTONE

June 10 Step 1 MILESTONE Everything in Step 1 is to be completed by June 10.
19

MUST

2.1.1 .fai improvements I will add library support to .fai, based on my run_file implementation NOT DONE
20

MUST

2.1.2 .fai improvements I will add namespace support to .fai NOT DONE
21

SHOULD

2.1.3 .fai improvements I will add numeric type support to .fai . Marking as SHOULD, because It is not that important, since it can be replicated in current code, it will just be harder to write without explicit language support. DONE BY DK
22

MUST

2.1.4 .fai improvements I will allow allow exception handling in .fai - throw and catch exceptions WIP
23

SHOULD

2.1.4 .fai improvements I will allow to set time limits on .fai execution. Marking this as SHOULD because this requires multithreading (watchdog on a timer) and this may lead to some problems which I cannot foresee at this point. WIP
24

MUST

2.2 FormulaAI debugger I will make a formula AI debugger, with basic stepping capabilities DONE
25

MUST

2.2 FormulaAI debugger I will add breakpoint capabilities to formula AI debugger WIP
26

MUST

2.2 FormulaAI debugger I will add tracing capabilities to formula AI debugger DONE
27

MUST

2.2 FormulaAI debugger I will add time profiling capabilities to formula AI debugger WIP

MILESTONE

July 10 Step 2 MILESTONE Everything in Step 2 is to be completed by July 10.
28

MUST

3 Candidate Move Evaluation I will write a framework for evaluating candidate moves in 'hypothetical' situations (which differ from current game state by a small amount - e.g. differ by expected results of several moves) without doing a full copy of the current game state. NOT DONE, POSTPONED
29

MUST

3 Candidate Move Evaluation I will write a a 'build and Evaluate candidate move tree' turn sequence, based on my 'candidate_move_tree' search improvement proposal, which will run with 'acceptable' performance for at least max_depth=2 (that is better than todays algorithm, in which it effectively equals 1) NOT DONE, POSTPONED
30

SHOULD

3 Candidate Move Evaluation I will write an algorithm to aggressively prune unpromising solutions from the candidate move tree, thus allowing to set bigger max_depth in 'candidate_move_tree' implementation. Marking this as 'SHOULD', because even with max-depth=2 my implementation will be better that current in spotting synergy opportunities. NOT DONE, POSTPONED

MILESTONE

August 10 Step 3 MILESTONE Everything in Step 3 is to be completed by August 10.

Patches and commits

I've earned a s­­v­­n write access to the Wesnoth source repository on Mar 27.
I've already contributed some patches to make debugging easier and to fix some crashes related with FormulaAI.
[Bug 13218] - I've spotted and fixed an infinite loop (which hangs the game) in the AI if it persistently tries to move unit to occupied square, with a hint from Sirp and some help from Dragonking
[Bug 13230] - I've implemented one of the FormulaAI debugging-related suggestions on the EasyCoding page, under the guidance of boucman.
[Bug 13229] - I've fixed two small 'past-end-of-collection-dereference' bug in FormulaAI
[Patch 1136] -I've submitted a patch to fix a segfault in FormulaAI formula parser.
[Patch 1137] I've implemented FormulaAI function run_file - which adds ability to run .fai file directly from in-game console, which allows efficient debugging of the .fai files
[r34200] - I've fixed incorrect handling of poisoning attacks when suggesting best attack in user interface.
[r34298] - I've added basic history and hot-redeployment capabilities to formula console, reworked AI lifecycle management and AI configuration management. That patch alone was more that 2000 lines long.
[r34329] - I've used those new capabilities to create an AI Arena - a framework to easily test the AIs in specified situations
[r34378] - I've fixed enemy_units formula to not include incapacitated units, such as stoned units on Caves of the Basilisk
[FR 13320] - I've made [modify_side] able to hack the share_view attribute
[Bug 13345] - I've fixed a crash at start of scenario 19 in "The Legend of Wesmere".
[FR 13348] - I've added info about current attack bonus (due to timeofday/alignment) of unit in Formula AI.
[Bug 13352] - I've fixed a bug with allied AI sides behaving stupidly if a leader's initial village capture moves him closer to his ally's keep than his own.

My immediate work plans:
Make the game get multiplayer AI list from the list of AI configuration files in data/ai/ais/ directory to allow 'AI is an addon' paradigm to work
Continue my refactoring of ai-game interface

Answers to those questions


Basics

Write a small introduction to yourself.
Hello. My name is Yurii Chernyi. I am 23, and I live in Kiev, Ukraine. I study applied math at the Faculty of Cybernetics, Kiev University.

1.2) State your preferred email address.
terraninfo АТ terraninfo.net

1.3) If you have chosen a nick for IRC and Wesnoth forums, what is it?
IRC: Crab_ [n=Crab_@wesnoth/developer/crab]
Wesnoth forums: Crab
Wesnoth wiki: Crab
GNA: crab

1.4) Why do you want to participate in summer of code?
I would like to make the world better and I've got some free time for this. I think that GSoC is a good opportunity to try to solve interesting problems.

1.5) What are you studying, subject, level and school?
Applied Mathematics, 6th year, Faculty of Cybernetics, Kiev University, Ukraine.

1.6) If you have contributed any patches to Wesnoth, please list them below. You can also list patches that have been submitted but not committed yet and patches that have not been specifically written for Wesnoth. If you have gained commit access to our S­­V­­N (during the evaluation period or earlier) please state so.
Yes, I've contributed some patches and fixed some bugs. I've earned commit access to Wesnoth s­­v­­n on March 27. See My Patches and commits

Experience

good practical knowlegge of C++, Java, computer science concepts, networking, system and database administration.

2.1) What programs/software have you worked on before?
Mostly, I've done website development using Java EE. Also, I've done various C++ projects in regarding numeric computations (for example, in computational fluid dynamics)

2.2) Have you developed software in a team environment before? (As opposed to hacking on something on your own)
Yes, I'm familiar to team development, and, at present, I coordinate a team of student developers for a small university project.

2.3) Have you participated to the Google Summer of Code before? As a mentor or a student? In what project? Were you successful? If not, why?
I haven't participated in the GSoC before. This is my first and last (I graduate this year) chance to participate.

2.4) Open Source

2.4.1) Are you already involved with any open source development projects? If yes, please describe the project and the scope of your involvement.
I use a lot of open source, but Wesnoth is the first open source project I am involved with.

2.5) Gaming experience - Are you a gamer?
Yes, I am a gamer )

2.5.1) What type of gamer are you?
I'm not really sure what this question is about, but I'll try to answer: I prefer to think out a plan, and then execute it. So, according to http://www.quizilla.com/quizzes/result/532090/375850/, I'm a 'strategic gamer'.

2.5.2) What type of games?
I prefer strategy games, both turn-based and realtime.

2.5.3) What type of opponents do you prefer?
Those who play well and are challenging to beat. Those can be both human or computer players.

2.5.4) Are you more interested in story or gameplay?
It depends on gameplay. Some games are worth playing once - to see the story unfold. Some games are interesting to play even without the story :)

2.5.5) Have you played Wesnoth? If so, tell us roughly for how long and whether you lean towards single player or multiplayer. We do not plan to favor Wesnoth players as such, but some particular projects require a good feeling for the game which is hard to get without having played intensively.
yes, I've played Wesnoth, from version 1.4.2. I've finished all single-player campaigns that were in the base install, some of them several times on different difficulty levels. I've played a bit of multiplayer in hot seat mode.

Communication skills

3.1) Though most of our developers are not native English speakers, English is the project's working language. Describe your fluency level in written English.
I think I can read and write English quite well. This wiki page can serve as an example :)

3.2) Are you good at interacting with other players? Our developer community is friendly, but the player community can be a bit rough.
That is yet to be seen, but I expect no problems. Being a player myself, I understand their desires )

3.3) Do you give constructive advice?
Yes, I personally think that I can give constructive advice.

3.4) Do you receive advice well?
Yes, especially if it is constructive :)

3.5) Are you good at sorting useful criticisms from useless ones?
Yes, I am.

Project

4.1) Did you select a project from our list? If that is the case, what project did you select? What do you want to especially concentrate on?
Yes, I looked through that list and I think it will be really hard and interesting to improve the AI (and the AI needs that improvement :) ).

4.2) If you have invented your own project, please describe the project and the scope.
My project is within the scope of 'improving AI and making it easier to tune its behavior' idea.

4.3) Why did you choose this project?
I want to make the AI feel more intelligent and I want it to be more fun to compete with. I would be happy to see Wesnoth with a better AI and I think I could make it better. Also, I want tweaking AI for specific scenarios to be a fun job, not a nightmarish formula debugging experience for scenario developers without programming background.

4.4) Include an estimated timeline for your work on the project. Don't forget to mention special things like "I booked holidays between A and B" and "I got an exam at ABC and won't be doing much then".

See Timeline
4.5) Include as much technical detail about your implementation as you can
See My proposal summary

4.6) What do you expect to gain from this project?
Knowlegge that I have done a good thing, satisfaction playing a better Wesnoth, and a bit of money from Google (I can probably earn more if I would work elsewhere as a full-time developer, but programming for Wesnoth is much more fun :) ).

4.7) What would make you stay in the Wesnoth community after the conclusion of SOC?
I will stay and continue to improve the game.

Practical considerations

5.1) Are you familiar with any of the following tools or languages?

* Subv­­ersion (used for all commits)
yes, both with command-line and gui.
* C++ (language used for all the normal source code)
yes
* Python (optional, mainly used for tools)
no, although I've worked with python code on several occasions and I can read it quite well.. I've had no problems patching wmllint when I needed some extra checks (to check mainline campaigns for usage of undocumented ai parameter syntax )

* build environments (eg cmake/autotools/scons)
'make' - yes, 'scons' - learned a bit while looking at Wesnoth build system. 'autotools' - no, but I know its purpose and I think it will be easy for me to learn to use it, if it becomes necessary. Using scons is all that I needed so far to compile Wesnoth.

5.2) Which tools do you normally use for development? Why do you use them?
vi, gdb, find, fgrep, small shell scripts, s­­v­­n, git, Eclipse under GNU/Linux; Visual Studio and Eclipse under Windows. My work machine is a notebook running Debian GNU/Linux.

5.3) What programming languages are you fluent in?
I am fluent in Java (4+ years of experience) and C/C++ (4+ years of experience) I know a bit of php/perl/sh/sed/awk/javascript/pascal

5.4) What spoken languages are you fluent in?
I can speak in Ukrainian (native), in Russian (native), English.

5.5) At what hours are you awake and when will you be able to be in IRC (please specify in UTC)
I'm in UTC+3 timezone in summer. I am normally awake from +9 UTC to +01 UTC, and most of that time I'm at home, so I'm able to be in IRC.

5.6) Would you mind talking with your mentor on telephone / internet phone? We would like to have a backup way for communications for the case that somehow emails and IRC do fail.
Of course, we can talk by phone.

This page was last edited on 21 March 2013, at 02:53.