Topic: [BETA] NEW ninjastic.space search - page 2. (Read 643 times)

This is going to sound controversial but I think that meaningless posts should not be indexed, or at least not returned. The algorithm should be prioritizing information over exact matches.

Open DuckDuckGo or any other search engine and search for it. After the videos, images etc you will start seeing results like this:



The mission of a search engine is to help people find what they are looking for. This is almost always stuff like answers, facts, and other kinds information with a semantic load. There is a lot of spam on Bitcointalk that nobody wants to read on the main site. So as much as possible, search engines must avoid returning spammy posts as results because this will discourage people from using them, as they will not find the kind of information they are looking for.

PS. even merit cannot be used to bias the posts because author merit does not apply if the post in question has no semantic load, and post merit allows for merit farms to manipulate search results by meriting spammy posts.

And what do you expect it to find when you search "I like trains"? Or any actual example, for that matter.

IMO, I think this is going to be a problem.

Exact results are pretty much useless as nobody is going to be searching with entire post contents.

Semantic meaning is not taken into context either, just the lexical word, so we are not only getting results for trains here but about a ton of "training" posts from Gambling Discussion.

I'm sure we would love to have google-like search results but we need to figure out the secret sauce first.

I'm not saying to create filter fields.
What I'm saying is that you create tag buttons, which when you click on them, automatically adds that tag to the search field.


For example, I click on "Topic ID" and the search field shows "topic_id:" and then I just have to write the topic number.
This way, I don't need to keep typing "topic_id:", when I want to search with this tag.

Really? It works to me though, e.g. "scam casino_name" and other crypto related words on the text box with "title:" or not.

Edit: OP's response is more precise for your problem.

Looking at the logs, that's because you are searching:


Notice the space between the author tag and the actual username.

It should be, as per example:

There are very few searches going on, so I noticed that. But please, when reporting an issue, post the details. Otherwise there is no way of debuging what you or I did wrong.

I have tried searching some posts by the title of any related keyword I can remember on the title but it returned nothing. No matter how I tried to include the author according to the placeholder tip, it didn't work for me. I didn't include any user ID though.
But whenever I make a search, I should be able to see some returned results and not an empty return. If I see some returned results, I could manually scan through them or be more confident to search again by changing keywords.
If will be fine that every search, results should be returned even if not precised.

I feel like this is more complex, might be because of using of only single input instead of the per input textbox in ninjastic.space. But with all this search tags for filter, probably this will give more precise results, great job. One to mention that username is case insensitive too.

It's specially focused to people who frequently use searching, be it to investigate other users/scams, find lost content, references, etc... "what is missing when you do your searching?"

And of course, how the UI should look like.

I thought about the single input (what we have now), multiple inputs for each filter like we have on ninjastic.space/search (i.e an input for authors, one for content, another for boards...), complex filters with multiple grouped and organized AND/OR conditions, etc.

First of all, excellent work. Making a good search tool is not easy.

What kind of tests do you want us to do? Research, layout, a little bit of everything!?



I start by making a suggestion, perhaps in line with what has already been said in previous posts.

Creates small tag buttons. When clicking on them, they immediately place the tag in the search field and the person only needs to enter the value of that tag.
I think this makes it easier to use the search, as we may not remember all the possible tags. In addition to helping to avoid errors when writing tags.



I'll look more into the tool and give other tips.
Excellent job!

There is literally a single input and a button on the website.

Over there you write what you want to search and add "author:TryNinja" if you want to search posts where the author is TryNinja.
Or title:"Bitcoin" if you want the title to have "Bitcoin".

I might be wrong, but I think that's dead simple.

I am not after anyone or anything that makes this forum upgrade. What I am after is that you guys as in the software developers, that I am aware of do neglect a simple think, "make it simple & a newbie friendly"

I faced it when I downloaded your bot for Bitcointalk notification. I am assuming that everyone here is good with it and they are good in understanding the notification bot. I am not and that is my reason to question you new update.

I will check it but please understand their are many who are not that smart like you or have brains to understand what you have created. My humble request is to keep it simpler when you as a software enthusiastic individual come up with a new idea.

No offense, but your post seems like a word salad. And looking at the logs, I don't think you even tried the website.

What you have done is amazing and yet it confuses me as a non technical user. I believe that for a few you made this incredible update. I am nobody yet I feel if someone knows how to do Google search in the correct way they as in the user will get more productive outcome. As you have outcome the issue why not create patch for the forum software.

Pardon me but those quoted example doesn't make any sense to me as they are not related to the forum I believe.

Please can you be more specific and detailed with your features. Their might be a lot of arguments by some one who understands it but as a non technical guy it is a waste for me..

Some people have said the searching mechanism on ninjastic.space is bad, so I'm looking for ways to improve it.

Today I'm releasing this beta page:

https://search-beta.ninjastic.space



I'll explore new designs for it and take suggestions (let me know what you think). The plan is for this page to become the new ninjastic.space (at least for the searching part).

- You can search everything with a single query.
- You can use special search tags to better filter your results.
- By default, the search will only consider the content from each post itself (ignoring any user quotes, that is, anything inside quotes that have the "Quoted from: USER" header).
- The content that shows up on the results page is the actual content (no user quotes) and the pagination is now numbered (max of 10k results, so 500 pages).
- Results are ordered by relevance (sorting by date will be added really soon).

This allows more complex searches like this one:

All posts from LoyceV, where he quotes TryNinja or theymos, on topic ID 5255508, made from 2019-01-01 to 2021-01-01, with the exact phrase "easy solution":

P.S: I have messed up separating URLs when indexing the data so that will be fixed in the near future when I reindex everything (takes many hours). Also note that some older posts scraped from Loyce's archive contains no board_id, so they won't be returned on queries where you specify a board_id.

---

SEARCH TAGS:


For title and quotes, it will do an exact match search by default. To do keywords, you can use :* (i.e title:*"title with keywords")
For author, topic_id, board_id, and quoted_user, you can specify multiple options ("OR") by using | (i.e author:TryNinja|theymos|"John Doe" board_id:25|12)
For title, quotes, and the post content, you can negate a keyword using a minus - (i.e "i like bitcoin" -doge)

---

SEARCHING EXAMPLES: