Add session settings variables like $OutputStyle

[rocky]

It would be nice to be able to set/change setting that can right now only be set at the outset inside a REPL session.

In particular options --style (set Pygments output style) and --full-form (show Full form translation of input)

@mmatera this is something that I could use help on. See https://github.com/Mathics3/mathicsscript/blob/master/mathicsscript/settings.py for what I have so far.

However when I try to set this inside the REPL, I am not able to. I believe the problem lies in Mathics somehwere.

Also I suspect Mathics might be improved somehow to make writing "builtins" that function as settings to be made simpler.

[mmatera]

According to what I see in main.py,

            if full_form:
                print(fmt(query))
            result = evaluation.evaluate(query, timeout=settings.TIMEOUT)
            if result is not None:
                shell.print_result(result, output_style, debug_pygments=pygments_tokens)

pygments are called by the variable fmt, which is defined out of the main loop. One way to make this dependent on the state of a variable would be to put inside the main loop, and look into the definitions object each time to set the behavior.
Regarding FullForm, you could also charge it to the evaluation object, setting the variable $PrePrint
to FullForm from inside mathics. Then, in main, if the parameter is passed, you can do something

evaluation.execute("$PrePrint=FullForm")

and, from inside the command interface, you can deactivate this by calling

$PrePrint=. 

and reactivate it with

$PrePrint=FullForm 

[mmatera]

Regarding the Pygment formatting, I would't do in that way, to do not mix the kernel with the user interface. Instead, you can look at the control variables by calling (from Python)

pygmentsstyle = definitions.get_ownvalue("Settings`$PygmentsStyle").get_string_value()

once a suitable context and variables are defined as builtins/user defined variables. This could be done both by something inside mathics builtins, or just in a initialization file defined in mathicsscript.

[rocky]

@mmatera Thanks! Will do the next time I have some free time to work on this.

[rocky]

Regarding FullForm, you could also charge it to the evaluation object, setting the variable $PrePrint
to FullForm from inside mathics. Then, in main, if the parameter is passed, you can do something

evaluation.execute("$PrePrint=FullForm")

I just tried this and this isn't the same thing as what the option --full-form does: the above formats the output in FullForm, while --full-form shows the input translated to full form before it is fed to evaluation.

Recall that for parsing things like /. (ReduceAll) we want to see the input parsed before it is evaluated.

[mmatera]

Yes, you are right... I tried to get the same behavior with $Pre, but (even in WM) it always evaluates first and then converts the expression to FullForm. In any case, the idea would be then to check the value of a variable in the definitions object and then apply the format according to its value.