Option to omit printing Ipopt header information

[juliohm]

Even though I've been using IpoptSolver(print_level=0), I still get this header printed on the first run:

******************************************************************************
This program contains Ipopt, a library for large-scale nonlinear optimization.
 Ipopt is released as open source code under the Eclipse Public License (EPL).
         For more information visit http://projects.coin-or.org/Ipopt
******************************************************************************

Is there any way to omit this information?

[odow]

Not to my knowledge. The list of options that can be passed to Ipopt is here.

You could use redirect_stdout as suggested here, but if you are distributing the code, I'm not sure of the licence implications.

[mlubin]

No, it's baked into the code (https://github.com/coin-or/Ipopt/blob/27f645c0ce7207fc589116c2b489df23f0196a40/Ipopt/src/Algorithm/IpIpoptAlg.cpp#L956). You can modify the source if you like.

[juliohm]

Wow, baked in the code, unbelievable. I would never do that to my users.

Thank you for pointing to the source code where this annoying message lies.

[mlubin]

Actually, a bit more digging reveals there's an undocumented option to suppress this warning. IpoptSolver(sb="yes") does the trick for me. TIL.

[juliohm]

Good catch! Works like a charm, maybe add to a FAQ? I am sure other people would feel happy about it.

[mlubin]

Not sure about that. It could be perceived as disregarding the authors' intentions. I'd first want to see a discussion of this on the Ipopt mailing list.

[odow]

Just to follow up: this came up in a discussion I had with Andreas. He intentionally made the option obscure and undocumented. We won't be adding this to the FAQ, or advertising it too widely.

[juliohm]

This is not necessarily a smart move. You will just get more annoying issues asking for this feature.

[mlubin]

@juliohm Could you clarify what you mean? Respecting the intentions of the author of Ipopt is not a smart move?

[juliohm]

I mean that many people will open issues here in the future asking for such an option. And that it is more work on your side to repeat it over and over the same answer.

Alternatively, a better option in my view, is to be explicit about the author's license, given them credit in many places in the README, the docs, etc; contact the author of Ipopt, ask if he agrees that printing messages in the console can be annoying for users that do not need interactivity, but just run optimization in a batch.

[odow]

contact the author of Ipopt

Andreas is the author of Ipopt.

[juliohm]

And he prefers to enforce the message on users? Ok then.