Fuzzy search helper for Script Filters

[deanishe]

Add fuzzy search to your Script Filters

This is a simple script you can add to your Script Filters to replace "Alfred filters results" with a fuzzy search algorithm.

 

https://github.com/deanishe/alfred-fuzzy

 

 

 

How it works

Instead of calling your script directly, you call it via fuzzy.py, which caches your script's output for the duration of the user session (as long as the user is using your workflow), and filters the titles of the items emitted by your script against the user's query using a fuzzy algorithm.

Example usage

fuzzy.py only works in Script Filters, and you should run it as a bash/zsh script (i.e. with Language = /bin/bash or Language = /bin/zsh).
Instead of running your own script directly, place ./fuzzy.py in front of it.
For example, if your Script Filter script looks like this:

/usr/bin/python myscript.py

You would replace it with:

# export user query to `query` environment variable, so `fuzzy.py` can read it
export query="$1"
# or if you're using "with input as {query}"
# export query="{query}"
# call your original script via `fuzzy.py`
./fuzzy.py /usr/bin/python myscript.py

Note: Don't forget to turn off "Alfred filters results"!

 

Caveats

As the script is written in Python and uses a more complex matching algorithm, it can only handle a few thousands items at most before it becomes irritatingly sluggish (Alfred can handle many tens of thousands).

If there's interest in the script, I will rewrite it in a compiled language. My Go library uses the same algorithm, and it can comfortably handle 20K+ items.
 

You can grab a demo workflow from GitHub to see it in action.

See the GitHub repo for more information.
 

[GuiB]

Thanks @deanishe, I'll have a look and try this! Thanks for your work

[dfay]

Thanks again for this.  

 

In case others need this info -- you can keep the fuzzy searching without sorting the results by deleting (or commenting out) the line 

items.sort(reverse=True)

 

[dfay]

I've run into a hitch.  I want to use this to search a list of categories, then again within the same workflow to search elements within a category.  But the second run is always displaying the cached results of the first.  I tried duplicating the script and running each separately but it seems to still use the same session_id and hence display the cached results instead of the results of the second run.

 

Any quick & easy ways to reset the session ID?  I suppose I could duplicate it & change the os.getpid() to something else. 

 

UPDATE:

 

I ungraciously and crudely solved the problem by

 

1. duplicating the file,

2. editing the duplicate copy

a. changing line 45 to 

SID = 'fuzzy_session_id_2'

b. changing the end of line 311 to 

str(os.getpid())+"A"

I call the original in the first run and the modified duplicate in the second.  I'm sure it could be done more gracefully/elegantly/pythonically but that approach works for me.

Edited January 13, 2018 by dfay

[deanishe]

8 minutes ago, dfay said:

Any quick & easy ways to reset the session ID?

 

8 minutes ago, dfay said:

I suppose I could duplicate it & change the os.getpid() to something else

 

Wouldn't work. That's only used to set the session ID when none exists. The problem is that both Script Filters are reading the session ID from the same variable (and therefore using the same cache).

 

I've added a new session_var option. Put export session_var=filter2 (or some such) in your downstream Script Filter, so it stores its session ID in a different variable to the first one (and therefore uses a different cache).

 

1 hour ago, dfay said:

In case others need this info -- you can keep the fuzzy searching without sorting the results by deleting (or commenting out) the line 

items.sort(reverse=True)

 

 

Hmm. If you only want match/no match, without the ranking, you can do that 10x more quickly with re:

pat =  '.*' + '.*'.join(query) + '.*'
match = re.compile(pat, re.I).match

 

[giovanni]

Is this still current? I tried it on a simple script returning a JSON list of names and got the following error:

  File "/System/Library/Frameworks/Python.framework/Versions/2.7/lib/python2.7/json/decoder.py", line 382, in raw_decode
    raise ValueError("No JSON object could be decoded")
ValueError: No JSON object could be decoded

my input is 'as {query}' and I have tried both 

export query="$1"
./fuzzy.py python myScript.py

and 

export query="{query}"
./fuzzy.py /usr/bin/python myScript.py

🙏 thanks!

[deanishe]

It doesn't like the output of your script, so I can't possibly say what's wrong without seeing your script and its output…

[giovanni]

very simple, it's one of yours modified

some of the items have emojis, perhaps that's the problem? 

It works well with the 'Alfred filters results' option, but I wanted to add fuzzy filtering. 

 

#!/usr/bin/env python
# encoding: utf-8
#
# Copyright (c) 2014 deanishe@deanishe.net
#
# MIT Licence. See http://opensource.org/licenses/MIT
#
# Created on 2014-07-03
# Modified from books, to show folders for filtering

"""Workflow Script Filter to show search results in Alfred."""

from __future__ import print_function, unicode_literals

import sys
import csv
import os
import struct
from time import time

from workflow import Workflow, ICON_INFO, ICON_WARNING
from workflow.background import run_in_background, is_running

from config import INDEX_DB

log = None



def main(wf):

    mylabels=[]
    with open("FOLDERS_ALL.tsv", "r") as fp:
        reader = csv.reader(fp, delimiter=b'\t')
        for row in reader:
            mylabel, myCount = [v.decode('utf-8') for v in row]
            toShow = mylabel
            #log.info (mylabel)
            wf.add_item(toShow, myCount + " papers", valid=True, arg=mylabel, icon='icon_folder.png')
        wf.send_feedback()


if __name__ == '__main__':
    wf = Workflow()
    log = wf.logger
    sys.exit(wf.run(main))

 

[deanishe]

I can't run that script without the rest of the files. Please upload your workflow somewhere and post a link.

[giovanni]

will do!

just to clarify, fuzzy.py is the only script I am supposed to copy into my own workflow's folder, or are there others? 

[deanishe]

fuzzy.py is standalone. It's a question of the other files your script requires.

[giovanni]

here you go!

keyword: zzl

thank you!!

 

https://github.com/giovannicoppola/paperpAlfred/blob/main/paperpAlfred troubleshooting.alfredworkflow

 

[deanishe]

It isn’t working because your script doesn’t output JSON. If you want to use Alfred-Workflow (which is probably unnecessary here) with fuzzy.py, you need to use the Workflow3 class to generate JSON. You’re using Workflow, which outputs XML.

 

[giovanni]

awesome, thank you so much! I will try to figure it out. 

[giovanni]

thanks again for your feedback @deanishe. I saved the dictionary as JSON instead of CSV and then read it in, which allowed me to ✅ avoid using Alfred-Workflow and ✅  be able to use fuzzy.pl.

thanks!!

 

[marck]

Hey @deanishe,

 

When I carbon-copy your Fuzzy demo, it won't work some reason, outputting
 

Quote

Code 1: [fuzzy] .

[fuzzy] cmd=['cat', 'books.json'], query='', session_id=None

Traceback (most recent call last):

  File "./fuzzy.py", line 377, in <module>

    main()

  File "./fuzzy.py", line 363, in main

    cache = Cache(cmd)

  File "./fuzzy.py", line 279, in __init__

    self.cache_dir = os.path.join(CACHEDIR, '_fuzzy')

  File "/System/Library/Frameworks/Python.framework/Versions/2.7/lib/python2.7/posixpath.py", line 70, in join

    elif path == '' or path.endswith('/'):

AttributeError: 'NoneType' object has no attribute 'endswith'

 

Am I missing something here? Do I need to install a local package or something?

 

Thanks for your help,

 

Marc

[marck]

Answering my own Q: The bundle ID must be set in order for the cache to work. Duplicating a workflow – obviously? – nullifies the bundle id.

[andy4222]

@deanishe The fuzzy script does not seem to be working after upgrading to 12.3. This is a really important script since it powers multiple workflows for me. Anything I can do to make it work? I tried changing /usr/bin/python to /usr/bin/python3 but that didn't help. Other errors came up which I could not fix:

 

error:

raceback (most recent call last):
  File "./fuzzy.py", line 377, in <module>
    main()
  File "./fuzzy.py", line 364, in main
    fb = cache.load()
  File "./fuzzy.py", line 312, in load
    json.dump(fb, fp)
  File "/Applications/Xcode.app/Contents/Developer/Library/Frameworks/Python3.framework/Versions/3.8/lib/python3.8/json/__init__.py", line 180, in dump
    fp.write(chunk)
TypeError: a bytes-like object is required, not 'str'

[andy4222]

@deanishe buzzing again if migrating this script to python 3 would be possible?

[dfay]

I've got a Python 3 version working & just posted it in @deanishe's github repo at https://github.com/deanishe/alfred-fuzzy/issues/3