code

Discussions about Red scripts from red/code github repository.

Respectech

@Respectech

random/only - Learning so much new stuff!

ralfwenske

@ralfwenske

I have a printF function which I find very usefull for creating formatted tabular reports. I am interested in getting feedback regarding concept and code. Perhaps others might want to play with it too. Is this the place to show it?

Boleslav Březovský

@rebolek

@ralfwenske certainly

ralfwenske

@ralfwenske

this syntax :

Red []

do load %printF.red

printF/underlined ["Some examples" [20] "This" [-14] "That" [-10] "A Rest" [-8 ] "When" [-12]] #"_"
printF ["none shown" [20] 345.2345 [-14 2 "$ " 'blank] 123.4567 [-10 3] none [-8 ] 2000-1-1 [-12]]
printF ["various 'blank" [20] 98765.7321 [-14 2 "$ " 'blank] -1.87654 [-10 3] 0 [-8 'blank] 2019-3-6 [-12]]
printF ["zero blank / not blank" [20] 0.0 [-14 2 "$ " 'blank] 0 [-10 3] -98 [-8 0 "" 'blank]]
printF ["a block shown" [20] [0 "xyz"] [-14] 0 [-10 3] -98 [-8 0 "" 'blank]]
printF/underlined ["too big float" [20] 98765444444.7321 [-14 2 "$ " 'blank] 1.87654 [-10 3] 54321 [-8 'blank] none [-12]] #"-"

str: printF/asString/underlined ["more underlined" [20] 345.2345 [-14 2 "$ " 'blank] 123.4567 [-10 3] 12345678 [-8 'blank] now/date [-12]] #"="
print [str]

produces this report:

Some examples                  This       That   A Rest         When 
____________________________________________________________________
none shown                 $ 345.23    123.456            1-Jan-2000 
various 'blank           $ 98765.73     -1.876            6-Mar-2019 
blank / not blank                        0.000      -98 
a block shown             [0 "xyz"]      0.000      -98 
too big float        $ 98765444444.73      1.876    54321              
--------------------------------------------------------------------
more underlined            $ 345.23    123.456 12345678  20-Feb-2020 
====================================================================

Code for inspection:

Boleslav Březovský

@rebolek

That's nice!

Gregg Irwin

@greggirwin

Nice work @ralfwenske. Thanks for posting.

ralfwenske

@ralfwenske

Thanks for encouragement @rebolek @greggirwin . I have made a refinement by replacing the refinement /underline:
alt

There are two odd things though (on Mac)

still strange behaviour when resizing window (already reported)
In the demo1-area some linefeeds are missing from mould :demo1
as I have verified this with probe also this seems OS independant. Is this a bug?

Gregg Irwin

@greggirwin

Is it this? red/red#3964

If not, can you post a small reproducible case?

ralfwenske

@ralfwenske

Yes, that's it @greggirwin . Can someone confirm the lost linefeeds related to mold? Can be produced when running 'printf-test.red'.

GaryMiller

@GaryMiller

I noticed in the large numbers such as 98765.73 that there were no comma separators such as 98,765.73

I wasn't sure if that was because it was just an American standard formatting feature or because of the comma being more complicated to include in a number due to a RED dual meaning.

Boleslav Březovský

@rebolek

@GaryMiller Yes, because the comma is used as decimal separator in other languages, it can't be used as thousands separator.:

>> 123.45
== 123.45
>> 123,45
== 123.45

However, you can use an apostrophe as a thousands separator:

>> 12'345.67
== 12345.67
>> 12'345,67
== 12345.67

Adam Sherwood

@theSherwood

Are there any implementations of iterators or generators floating around in the community?

Dave Andersen

@dander

@theSherwood I did some experimentation on that - https://gist.github.com/dander/66ad65555b30286115cee675b16bf1d7
It's a little messy, but once you get past the different partial implementations, I think the sample generator functions down below should make sense. I've been meaning to get back to it to clean it up. I've learned a lot since then too, but life is crazy right now

Adam Sherwood

@theSherwood

@dander Great! Thanks!

Dave Andersen

@dander

Sure thing! I'd be interested to hear any thoughts you have about it. I think it would also be interesting to also add some sort of yield method

Adam Sherwood

@theSherwood

Yeah. I'm trying to conceive of a way to yield without a bunch of slow parsing. But I haven't thought of a good approach yet.

Dave Andersen

@dander

I was thinking maybe something with do/next saving the current position in the block

Toomas Vooglaid

@toomasv

I was recently playing with a special syntax for these.

Adam Sherwood

@theSherwood

@toomasv I remember seeing that. Really cool stuff

Toomas Vooglaid

@toomasv

:smile: Thanks! If you will play with it, any remarks/ideas are welcome.

Adam Sherwood

@theSherwood

The examples you show demonstrate just how flexible those functions are! I think the lazy sequence function you have there might be the one I focus in on.

R cqls

@rcqls

Really cool stuff @toomasv. I love the syntax when range (and I guess sequence) is followed by block as function body.

Toomas Vooglaid

@toomasv

@rcqls I didn't remember if it worked for sequences too, but seems so. E.g. this generates endless Fibonacci sequence:

>> fib: ...| [seed: [0 1] also seed/1 seed/2: sum reverse seed] ()
>> loop 10 [prin [fib ""]]
0 1 1 2 3 5 8 13 21 34

Petr Krenzelok

@pekr

Eeeh, you took a pipe char, which I dreamed of as an expression separator when writing one liners 😀

Toomas Vooglaid

@toomasv

No problem, pipe char in this syntax is special after ... and .. only.

Toomas Vooglaid

@toomasv

Hmm, looking at it... this does too:

fib: does [seed: [0 1] also seed/1 seed/2: sum reverse seed]

No need for fancy syntax here :flushed:

Vladimir Vasilyev

@9214

Reminded me of streams implementation.

Toomas Vooglaid

@toomasv

Quite an inception. Finally wrapped my head around it.

Adam Sherwood

@theSherwood

@9214 I've been probeing that thing for a little while. I still can't figure out how/where the growth function is being called in thaw. What am I missing?

OOOOOh. nevermind. It just clicked.

Vladimir Vasilyev

@9214

stream/-1 stream.

Adam Sherwood

@theSherwood

Gotcha.

That thing is some dark magic.

Very cool stuff.

Vladimir Vasilyev

@9214

Here's the kicker.

Red [
    Title:  "Bubble sort"
    Date:   17-Apr-2018
    Author: 9214
]

bubble: func [list [series!]][
    also list until [
        tail? next list: skip list do reduce [
            pick [do last] list/1 > list/2
            [swap list next list negate 1]
        ]
    ]
]

Toomas Vooglaid

@toomasv

It is nice demonstration of dynamic construction, but consumes lot of memory due to reducing. Following stripped down version takes ~300 times less memory:

bubble2: func [list [series!]][
    until [
        tail? next list: skip list either list/1 > list/2 [
            swap list next list -1
        ][1]
    ] head list
]

Adam Sherwood

@theSherwood

Interesting. @toomasv how do you do the memory profiling?

Vladimir Vasilyev

@9214

>> recycle x: stats bubble [3 7 9 9 5 0 2] stats - x
== 3844
>> recycle x: stats bubble2 [3 7 9 9 5 0 2] stats - x
== 156

GiuseppeChillemi

@GiuseppeChillemi

I didn't know about this group. What is it for?

Vladimir Vasilyev

@9214

Have you read this room's description?

Adam Sherwood

@theSherwood

@9214 Thanks!

Follow up: why does reducing result in a lot of memory consumption?

Vladimir Vasilyev

@9214

Because it allocates a new block. reduce/into avoids that and reuses the one that you give it.

Adam Sherwood

@theSherwood

Good to know. Thanks

GiuseppeChillemi

@GiuseppeChillemi

Vladimir, I am on mobile, there is no description other than the name. The android client lacks of many things like this. I see just red/red/code 😔.

Toomas Vooglaid

@toomasv

Ah, yes, of course, memory consumption grows with growing length of sorted series:

>> recycle/off
>> string: "sdlkjflkjlsdmcvLKJFawjrfaoiajfliuh.ghlskls" ()
>> profile/show [[bubble copy string][bubble2 copy string]]
Count: 1
Time         | Time (Per)   |      Memory | Code
0:00:00.001  | 0:00:00.001  |         360 | [bubble2 copy string]
0:00:00.002  | 0:00:00.002  |      104428 | [bubble copy string]
>> 104428 / 360
== 290.0777777777778

@theSherwood I used the profile func by @greggirwin.

Toomas Vooglaid

@toomasv

With reduce/into it is almost as good as stripped down version:

bubble: func [list [series!] /local code][
    code: clear []
    also list until [
        reduce/into [
            pick [do last] list/1 > list/2
            [swap list next list negate 1]
        ] clear code
        tail? next list: skip list do code
    ]
]

>> profile/show [[bubble copy string][bubble2 copy string]]
Count: 1
Time         | Time (Per)   |      Memory | Code
0:00:00.001  | 0:00:00.001  |         360 | [bubble2 copy string]
0:00:00.003  | 0:00:00.003  |         516 | [bubble copy string]

Adam Sherwood

@theSherwood

@toomasv Thanks! I'll check out the profile func

Toomas Vooglaid

@toomasv

@theSherwood You are welcome!

Where communities thrive

People

Repo info

Activity