solved selecting a result while rerunning a script filter

[giovanni]

Hello, I was hoping to double check something about the rerun feature in script filters.

 

If the script filter is returning more than one result while being rerun, and the user attempts to select an option beyond the first using arrow-down on the keyboard, the selection will be lost upon refreshing. 

 

So, I understand one can select one of the results only via mouse (if you are quick) or typing matching text, is that correct?

Or, is there a way to store the result currently selected/highlighted via keyboard and serve it back upon refreshing?

thanks in advance!

 

 

 

[vitor]

If you set a uid, your selection will be maintained (and jump around) as the order of results changes.

[giovanni]

AHA! awesome, thank you so much!

[giovanni]

@vitor this works but Alfred will learn the selection and change the order of results (undesired here). I read that a random UID is recommended to obviate this issue, however this random number should be generated outside the script filter, correct? (otherwise a new one gets generated for each rerun)... I used a Random utility and passed the output to the python script and that works, however now the script filter needs to be triggered either by a hotkey or a keyword (so an extra ↩️ is needed). Am I missing anything? In other words, is there any other way to pass to a script filter a random number that stays the same across multiple reruns of that script filter? thanks!

[deanishe]

12 minutes ago, giovanni said:

read that a random UID is recommended to obviate this issue, however this random number should be generated outside the script filter, correct?

If you want random UIDs that persist till the workflow is closed, use a pseudo-random function, but store the seed in a top-level workflow variable (alongside rerun).

 

When you seed the generator with the same number, the same “random” results come out.

 

Most languages have a pseudo-random number generator. It’s often the standard random function (which is why you’re not supposed to use it for generating passwords).

[giovanni]

thanks, but wouldn't I need to pass a new (random) seed, each time the workflow is triggered? 

[deanishe]

16 minutes ago, giovanni said:

thanks, but wouldn't I need to pass a new (random) seed, each time the workflow is triggered? 

 

Yes. That's the point, isn't it? Just use the current time.

[giovanni]

I must be missing something... If I do that, the random generator gets reset at each rerun... here is my simplified code: 

 

import sys
import time
from workflow import Workflow3
import random

def main(wf):
    
    random.seed()
    # setting a value, e.g. random.seed(4) will produce the same number every time    
    myRandString = str(random.randint(0, 1000))
    wf.rerun = 1

    goals=get_goals()

    for x in goals:
        
        wf.add_item(title=x['blabla1'],
            subtitle="blabla",
            icon = myIcon,
            valid='TRUE',
            uid=x['blabla2'] + myRandString,
            arg=x['myArg'])

    wf.send_feedback()


 

if __name__ == u"__main__":
    wf = Workflow3()
    log = wf.logger
    sys.exit(wf.run(main))

 

[deanishe]

49 minutes ago, giovanni said:

If I do that, the random generator gets reset at each rerun...

 

Not if you set a workflow variable to the seed. Top-level variables get passed back to your script on rerun.

 

import os
import random
import time

# get seed from environment variable or current time
seed = float(os.getenv('seed') or time.time())
# initialise RNG with seed to get deterministic values
random.seed(seed)

def main():
    # set top-level variable that will be returned on re-run
    wf.setvar('seed', str(seed))
    wf.rerun = 1

 

[giovanni]

brilliant, that works! thanks as always 🙏