Is it possible to add option to keep user search query as first item always? (or make it configurable) about alfred-searchio HOT 8 CLOSED

It’s in there, but only in the Python version.

The Go version of the search is currently a quick proof-of-concept hacked together to see if it’s meaningfully faster (between 3 and 20x FWIW).

I’ll put the query back in as the last option, so UP ARROW will take you straight to it.

I don’t see much point in making it the first option. If you want to enter the entire query yourself, you should be using Alfred’s built-in Web Searches, which are massively faster.

It would also require turning off Alfred’s knowledge, as that’s the only way to ensure the first item you emit is also the first result.

from alfred-searchio.

I don’t see much point in making it the first option.

It's mostly that I want to keep Searchio searches as the main search option for the searches it provides. So I don't want to switch to Alfred web searches or my web searches workflow since I can just use Searchio.

And more often than not, I want to search for a specific thing and not look at suggestions or search for a thing and see if suggestions are any good, if they are, pick them if not, just use my query.

I really like how Chrome and Safari and all other browsers do it where they prioritise the User query first and provide suggestions below to DOWN ARROW to.

If I attempt at making it a configuration where you can mark it in Searchio (where the User query goes), will you accept the PR? I will really appreciate it as this is something that would really be helpful for me. Thank you.

from alfred-searchio.

It would also require turning off Alfred’s knowledge, as that’s the only way to ensure the first item you emit is also the first result.

I am not too familiar with this. Previous Searchio didn't use Alfred knowledge I suppose because this worked there.

from alfred-searchio.

Previous Searchio didn't use Alfred knowledge

Yeah, it did.

from alfred-searchio.

Searchio v1 users could have the first Alfred item be User query

That's true. But as stated in #20, v2 will add the query to the end of the results, not the beginning.

If other people weigh in and say, "No! It's better at the top!", then I'll reconsider, but to my mind the end is a better choice. Especially given the added speed of v2.

The purpose of the workflow is to provide suggestions from the search engine API. If searching for the query you entered were top priority, I would assume you'd use Alfred's built-in Web Search feature, as it's much faster.

you say it would require turning off Alfred's knowledge for this to work

It requires turning of Alfred's knowledge to ensure the query is at the top. It may be there most of the time (as in v1), but the only way to guarantee it is to disable Alfred's knowledge, so it doesn't re-order the results.

Again, if other people chime in with an opinion, I will reconsider.

But at the moment, it's your opinion vs mine, and mine wins because it's my workflow…

Personally, I far prefer not having the query at the top, as the extra speed of v2 means I can often TAB the first result to expand the query and keep on typing.

from alfred-searchio.

Personally, I far prefer not having the query at the top, as the extra speed of v2 means I can often TAB the first result to expand the query and keep on typing.

That's a really valid point.

Thanks for the clarification too. I really appreciate it.

from alfred-searchio.

I haven't noticed a speed gain with V2 and I have pretty fast internet speed

What does "fast" mean? Latency is what counts here, not bandwidth: the amount of data transferred is tiny.

You may also have a much faster computer than me. My Mac Pro is 10 years old now, and my MacBook Air is a MacBook Air (and also 5 years old)…

The speed advantage of the Go program is all in the time it takes to load and start running.

On my machine, the Python version takes ~0.3s to load, while the Go version takes ~0.03s. If your machine loads the Python version more quickly, the speed difference will shrink (they both load results from the network and cache at more or less the same speed).

If you have jupyter installed, you can try running the Benchmark.ipynb notebook in the extras directory.

from alfred-searchio.

It is really fast, don't get me wrong.

I thought it'd be as fast as something like AnyComplete or close to it. But I remember you said why that is impossible to do in Alfred.

But I really appreciate the new update, really.

Specifically the generation of script filters is really handy as I didn't use many of the script filters in Searchio v1.

from alfred-searchio.