Write script to run optimization tests in VM.

[kenorb]

For each test.ini we need to run file with .rules.

The goal is to run optimization tests based on YAML configuration, then propose changes with better optimized settings by sending a pull request.

• [ ] write Python script to parse YAML file (e.g. optimizer.yml) passed from 2st argument

  ./run_optimizer.py -d deposit -s spread -b dukascopy -y `optimizer.yml` -f `EA31337-foo.set` -D dest/

  • [ ] default values: -d 2000 -s 30, mandatory parameters: -y, -f
  • [ ] given YAML file, for each main mapping section prepare the test data which is:
  • [ ] load and read SET file from the 3nd argument
  • [ ] write function to read and change settings in-place in SET file based on YAML rules which is:
    • [ ] disable given EA options in Settings based on values in Disable array
    • [ ] enable given EA options in Settings based on values in Enable array
    • [ ] if entry key is not Disable or Enable, set value for specific option name (e.g. LotSize)
    • [ ] if wildcard is specified, change value of all matching options (e.g. *_TrailingStopMethod: 0)
    • [ ] make sure it's executed in order
  • [ ] write function to change settings in-place in SET file based on YAML which is:
    • [ ] for each entry key starting with Test-, load entries into optimization SET file
    • [ ] given start, step, end, set appropriate values in SET file
    • [ ] after optimization test is run, go to the next test (within the same section) if present
  • [ ] after main section is finished, go to the next one (keeping the previous values)
  • [ ] load pair, deposit, spread values from for parameters and set into appropriate ini file
  • [ ] given SET file with prepared test data execute (e.g. run_optimizer.sh)
  • [ ] script should be configured to run from VM
• [ ] write run_optimizer.sh (or reuse run_backtest.sh if possible) to run platform with given SET file
  • [ ] configure terminal.ini with appropriate settings for given optimization test
  • [ ] load year, deposit, spread and source of backtest data from the command line
  • [ ] run optimization test using platform under VM with given settings
  • [ ] Use: Genetic algorithm
  • [ ] Optimized parameter: Balance
  • [ ] Make placeholder for Limitation rules
  • [ ] save results into pre-configured destination folder
• [ ] pass back the result into Python script then analyse and parse the data
  • [ ] save results into pre-configured destination folder
  • [ ] read results and update in-place SET file with optimized settings

• [ ] after each test is completed, commit changes into local git repository with given new optimized settings:

  • [ ] each commit should have the following dynamic message:

  $TestExpert: $YML-SectionName: $YYYY-$mm-$dd - $YYYY-$mm-$dd, $TestSymbol $TestPeriod, Deposit: \$$dep, Spread: $spread (Time: $TimeInMin min)

• [ ] after each test is completed, push change into remote branch with given RSA key
  • [ ] branch name: $TestExpert-$pair-$deposit-$spread
• [ ] after all tests are completed, generate pull request for that branch

[kenorb]

Implemented as part of run_backtest.sh (-o/-i).