[Alfred 3] Google Inline Search (open URL + copy to clipboard)

[aviaryan]

This workflow allows you to do inline Google searches right from the Alfred bar. You can either open the result URL or copy it to clipboard. Please go to GitHub for documentation. 

 

GitHub: https://github.com/aviaryan/alfred-google-search

Download: https://github.com/aviaryan/alfred-google-search/raw/master/Google Search.alfredworkflow

 

Screenshots

 

 

 

Edited September 7, 2017 by aviaryan

[deanishe]

The whole "$ to copy URL" is unnecessary: just hit ⌘C on a result to copy the URL.

[aviaryan]

1 hour ago, deanishe said:

The whole "$ to copy URL" is unnecessary: just hit ⌘C on a result to copy the URL.

Thanks. Didn't know about that. 

I have updated the workflow to incorporate this change.

 

[deanishe]

Yeah. If nothing is selected in the query box, and you haven't set a "copy text" for an item, ⌘C copies the selected item's arg.

 

BTW: You might want to add a copy of the .alfredworkflow file to your releases on GitHub. People generally expect to find the installable files in releases.

Edited September 7, 2017 by deanishe

[aviaryan]

2 hours ago, deanishe said:

Yeah. If nothing is selected in the query box, and you haven't set a "copy text" for an item, ⌘C copies the selected item's arg.

1

Cool. I didn't know we had a copy text parameter. Any place where I can find all this documented?

 

2 hours ago, deanishe said:Yeah. If nothing is selected in the query box, and you haven't set a "copy text" for an item, ⌘C copies the selected item's arg.

BTW: You might want to add a copy of the .alfredworkflow file to your releases on GitHub. People generally expect to find the installable files in releases.

Yes. I had been including the binary in the releases in my previous 2 workflows. Just got a bit lazy this time. I have added it now. 

[deanishe]

9 minutes ago, aviaryan said:

Any place where I can find all this documented?

 

https://www.alfredapp.com/help/workflows/inputs/script-filter/json/

 

See the "text" section. Also the example workflow you can add via Alfred Preferences > Workflows > + > Getting Started > Script Filter shows most of what you can do (it doesn't demonstrate rerun or variables).

[multicolor]

Love it! So useful to look up quick things, like "gs 720p def" and then use shift to preview the first page, way faster than using a browser. Way to go!

[Skone]

Anyone using this successfully with Alfred 4? Seems to have the same problem as ethan-funny's workflow: 

 

[xilopaint]

18 hours ago, bazerman said:

Anyone using this successfully with Alfred 4?

 

It works for me, but with this important caveat:

 

Quote

Overusing this workflow might lead to your IP being blocked by Google Search servers. Searches through Chrome or another browser might still work but this workflow will stop working. I recommend keeping a 10-15 seconds gap after each usage of this workflow. In most cases, much lower gaps or even continuous use of the workflow will still work but still this is something to be kept in mind. If you see a 'rate limit' or a 503 error, it's best to stop using the workflow and try back after some time (~1 minute).

 

The rate limit makes the workflow unfeasible for my routine.

[deanishe]

19 hours ago, bazerman said:

Anyone using this successfully with Alfred 4? Seems to have the same problem as ethan-funny's workflow:

 

It has exactly the same problem. Google isn't returning the HTML the workflow expects. And this one's a real mess to fix. One simply does not use regular expressions to parse HTML.

[xilopaint]

1 hour ago, deanishe said:

 

It has exactly the same problem. Google isn't returning the HTML the workflow expects. And this one's a real mess to fix. One simply does not use regular expressions to parse HTML.

 

Really? This workflow worked for me.

[deanishe]

1 hour ago, xilopaint said:

Really? This workflow worked for me.

 

For me, Google isn't sending the HTML markup the workflow expects. This parser is looking for the same tags as the other workflow's.

 

The problem is that Google doesn't give everybody the same HTML, or the same person the same HTML at different times. Changing the user agent can get you a completely different page.

 

I suspect such a workflow might have more luck with a user agent like Lynx/2.8.5dev.16 libwww-FM/2.14 SSL-MM/1.4.1 OpenSSL/0.9.7a. That gets you a really old-school version of the page that hasn't changed in 15 years (and probably won't change).

 

[joeynotjoe]

I used to use a few different modified versions of this workflow, to return results for specific sites (prefix site:site.com). It was indispensable, but now, I am definitely rate-limited / blocked, and I can't seem to get it to work anymore, even from different IP addresses. 

 

Does anyone have a recommendation for where to begin on replacing such functionality? I tried modifying @deanishe's Searchio! workflow, but I think the suggest URL doesn't work after adding in my site: prefix. 

 

 

[deanishe]

10 minutes ago, joeynotjoe said:

but I think the suggest URL doesn't work after adding in my site: prefix

 

In theory, you could edit the search URL for the results page to add your site:xyz.com, but you can't have suggestions for a specific site. And anyway, it only provides suggestions, not results like this workflow does.

 

13 minutes ago, joeynotjoe said:

I can't seem to get it to work anymore, even from different IP addresses

 

That probably won't help. Google seems to have changed the HTML of the results page (for some users at least).

[joeynotjoe]

Ah you are right, thanks for pointing that out. I missed the fundamental distinction between results and suggestions. So essentially, this workflow will need to be refactored to support the new HTML results if that's possible, otherwise, go back to the dark ages of opening the results page before going to what I want :). 

Thanks for the quick response!  

Edited June 1, 2019 by joeynotjoe
typo

[deanishe]

3 hours ago, joeynotjoe said:

So essentially, this workflow will need to be refactored to support the new HTML results if that's possible

 

Yup. It’s pretty normal to have to fix HTML parsers often. Webpages change a lot.

[Skone]

23 hours ago, joeynotjoe said:

this workflow will need to be refactored to support the new HTML results if that's possible, otherwise, go back to the dark ages of opening the results page before going to what I want :). 

 

🤞️ - was one of my favourites! Huge timesaver

[Skone]

I am still so gutted about this - anyone found any alternatives? As a non developer, what would be the learning curve like if I was to take this on myself and try to fix the HTML parsers?

Edited July 5, 2019 by bazerman

[Briscoe]

Curious if anyone was able to find an alternative to this that works in Alfred 4 🙂

[vitor]

Doesn’t @deanishe’s Searchio! do the same (and more) as this Workflow?

[joeynotjoe]

@vitor - Searchio! is great, but the main difference was, as far as I can tell, actioning the searchio 'results' takes you to the google results page, whereas actioning this workflow's results actually took you to the result's URL.

 

aviaryan's workflow's results:

Searchio's results:

So, the original workflow from this thread is 'better' if you know what you're looking for - it saves you the step of going to the google page.

[deanishe]

11 hours ago, vitor said:

Doesn’t @deanishe’s Searchio! do the same (and more) as this Workflow?

 

Mine shows search suggestions. This one scrapes the actual search results from the webpage. That's the reason it breaks quite often.

[Skone]

@aviaryan do you have any plans to update this workflow? Is it still possible to achieve inline search results or do you feel google changes this stuff too often?

[deanishe]

On 7/4/2019 at 5:39 PM, bazerman said:

As a non developer, what would be the learning curve like if I was to take this on myself and try to fix the HTML parsers?

 

Scraping HTML ranges from pretty simple to really complex, depending on what you're trying to scrape. Scraping Google results isn't technically difficult. The problem is that they change the website frequently, which breaks the workflow. They also have a lot of defences in place to prevent scraping.

 

2 hours ago, bazerman said:

do you feel google changes this stuff too often?

 

I've had some luck scraping Google by sending a really old user agent. They did change the HTML last year, but it was stable for a long time before that. The page looked to be straight out of 2005.

 

[Skone]

Thanks Deanishe, makes sense.

 

I wonder if there's an alternative search engine that's more stable to build an in-line search workflow from. I'm honestly surprised this isn't at the top of everyone's lists, it seems like one of the most obvious and effective uses of alfred!