Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Extend PCS format #83

Closed
mlindauer opened this issue Oct 13, 2016 · 9 comments
Closed

Extend PCS format #83

mlindauer opened this issue Oct 13, 2016 · 9 comments
Assignees

Comments

@mlindauer
Copy link
Contributor

Hi,

we (i.e., Frank, Holger, Kevin, Thomas, Manuel and I) extensively discussed at the Dagstuhl seminar, that
the new pcs format (see AClib 2.0 format spec or the latest SMAC pcs format) will be the standard in AClib in the near future.
Therefore, I think, it will be also crucial to support it in SMAC3.
Irace does already support it.

Best,
Marius

@mfeurer
Copy link
Contributor

mfeurer commented Oct 13, 2016

@sfalkner @Krxsy you mentioned that you're working on this for the fANOVA, right?

@Krxsy
Copy link

Krxsy commented Oct 13, 2016

Yes, I'm working on it. The ordinal parameters are the only ones missing in the new version right now . I'm going to talk to Stefan about that today and send a pull request afterwards.

@mfeurer
Copy link
Contributor

mfeurer commented Oct 13, 2016

I'm not sure if Marius only means the ordinal parameters, but also the new conditions etc.

@mlindauer I thought the ACLIB standard is not yet final/official, that's at least what I understand from your comment on PR #45. Since the SMAC and AClib2.0 specification are different, it would be good to know what exactly you mean.

@mlindauer
Copy link
Contributor Author

Here a summary of the format:

parameter_name categorical {value_1, ..., value_N} [default value]
parameter_name ordinal {value_1, ..., value_N} [default value]
parameter_name integer [min_value, max_value] [default value]
parameter_name integer [min_value, max_value] [default value] log
parameter_name real [min_value, max_value] [default value]
parameter_name real [min_value, max_value] [default value] log

Conditionals:
  child_name | condition [&&,||] condition ...

Condition Operators: 
  parent_x [<, >] parent_x_value (if parameter type is ordinal, integer or real)
  parent_x [==,!=] parent_x_value (if parameter type is categorical, ordinal or integer)
  parent_x in {parent_x_value1, parent_x_value2,...}

Forbiddens:
  {parameter_name_1=value_1, ..., parameter_name_N=value_N}

@mfeurer
Copy link
Contributor

mfeurer commented Oct 17, 2016

Some more questions?

  • Are child_name in conditionals still allowed to appear in multiple lines and have to be merged?
  • How does the file itself look like? Are conditionals preceded by the term 'Conditionals' (and forbiddens) as well.
  • Are there forbidden characters as in the SMAC2 pcs files?

@mlindauer
Copy link
Contributor Author

Are child_name in conditionals still allowed to appear in multiple lines and have to be merged?

No

Are conditionals preceded by the term 'Conditionals' (and forbiddens) as well.

No key words such as Conditionals or Forbiddens, but the file should be in the following order: 1. parameters, 2. conditionals, 3. forbiddens

Are there forbidden characters as in the SMAC2 pcs files?

I don't know.

@mfeurer
Copy link
Contributor

mfeurer commented Mar 30, 2017

Just merged #132, this should by done now.

@mfeurer mfeurer closed this as completed Mar 30, 2017
@HYDesmondLiu
Copy link

Hi experts,
I tried to put " x1 integer [0, 10] [7]" in my PCS file, however error is reported. How do I limit the solution of optimization to be integer?

@mlindauer
Copy link
Contributor Author

Hi,

First of all, please don't re-open old issues where you were not involved in. Furthermore, closed issues are not shown in the issue overview and thus, we will loose track for these.
Furthermore, your problems does not provide sufficient information. Which error was thrown? Which version have you used? Can you please provide a mini-example to reproduce the error?

So, please open a new issue ticket and provide all required details.

Best,
Marius

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

4 participants