Talksearch.io - Advanced Bitcointalk Search Engine

[NotATether]

snip

Hey there NotATether,
I just wanna let you know that you never updated OP with this translation. It was made several days ago. Hope that helps.

Very sorry I didn't add this before, I totally forgot about it. I added it now.

I'm looking for an option here. But it's not working, TalkSearch can't do it.

The thing is that if someone quotes my post and posts it, then if I type the text of my post and search, I will find the links to all the posts in the search results. For example, if I am posting this, now if I copy the text of my post exactly and search, then my post shows: If someone else quotes my post, then all of them will come up on search results.

Why is this necessary? What this will do is ensure that before reporting a post by quoting it, it will show someone has already reported it. This can help when reporting plagiarism posts, AI posts, malware, and to avoid duplicate reports.

This is a known issue and I'm working on fixing this by searching by vector embeddings instead of by plain text, and these embeddings will be made by stripping the quotes from the posts first.

The previous algorithm I designed is (by my own standards) quite garbage.

[beveryu778]

1755945384]

I'm looking for an option here. But it's not working, TalkSearch can't do it.

The thing is that if someone quotes my post and posts it, then if I type the text of my post and search, I will find the links to all the posts in the search results. For example, if I am posting this, now if I copy the text of my post exactly and search, then my post shows: If someone else quotes my post, then all of them will come up on search results.

Why is this necessary? What this will do is ensure that before reporting a post by quoting it, it will show someone has already reported it. This can help when reporting plagiarism posts, AI posts, malware, and to avoid duplicate reports.

This is a known issue and I'm working on fixing this by searching by vector embeddings instead of by plain text, and these embeddings will be made by stripping the quotes from the posts first.

The previous algorithm I designed is (by my own standards) quite garbage.

The sound is good. this option will change the searching experience and will provide many advantage to find information easily.

Good luck, you are doing great.

[DYING_S0UL]

Very sorry I didn't add this before, I totally forgot about it. I added it now.

Sorry for making you do all the extra work, but you did a typo in the language.

It should be "Bangla" not bengala. Opps

[NotATether]

I have spawned a second scraper running in parallel, with a shorter post creation window for indexing.

This will go a long way into making posts get indexed faster. Your posts are now expected to be indexed within five days.

The goal is to eventually get all posts indexed in 5 minutes or less.

PS. Adaptive rate limiting is used in all scrapers, so we always slow down enough to never get any 503 errors.

[GazetaBitcoin]

Hey NotATether, please be aware that a new translation was made by AOBT for your topic:

- Hindi translation, made by M47AK16

I hope this is good news!

Are you so kind, please, to add it to OP?

[NotATether]

Hey NotATether, please be aware that a new translation was made by AOBT for your topic:

- Hindi translation, made by M47AK16

I hope this is good news!

Are you so kind, please, to add it to OP?

It has been added.

[ptrk]

@NotAThether: I'd be interested to know which programming language you used to write the site. Java?

[NotATether]

New app update v1.0.6 and search update v1.1.1 published

You can now report posts that are in the search engine using the "Report Content" feature at the bottom of the page.

This allows you to remove content that exposes sensitive information about yourself or other people.

There is no captcha for now. However if this feature gets abused then I might have to start using reCaptcha.

Note that you can only remove your own posts or topics, and they must have already been deleted from Bitcointalk first. Or they will just be scraped again. The exception to this is posts containing malicious links - please make sure you also use the "Report to Moderator" feature.

I try to delete old reports after 30 days.

@NotAThether: I'd be interested to know which programming language you used to write the site. Java?

The site is written in NextJS. The search algorithms themselves are using Node in Docker.

[NotATether]

For some reason, the scraper has stopped. I am working on getting it back online.

Edit: resolved (since a few hours ago)

[TESY]

I checked it and it seems very useful as a newbie I can find things more easy

I suggest your strongly to add a SSL/TLS encryption so the site is not served as Not Secure

second suggest add  target="_blank"  so when user search and click on a thread it don't leave ur site

[NotATether]

I checked it and it seems very useful as a newbie I can find things more easy

I suggest your strongly to add a SSL/TLS encryption so the site is not served as Not Secure

second suggest add  target="_blank"  so when user search and click on a thread it don't leave ur site

That's strange... Talksearch already has a TLS certificate from Google App Engine. It should not display the "not secure" prompt. Can you try with a different device and browser ans see if you can replicate the issue?

target=_blank sounds like an interesting idea, but it depends on whether people want to leave talksearch after finding the thread they want, or if they are using this as a way to read old posts.

[LoyceV]

target=_blank sounds like an interesting idea, but it depends on whether people want to leave talksearch after finding the thread they want, or if they are using this as a way to read old posts.

Whenever a site does that, I middle-click on the site's tab to close it. It's extra work, and annoying. If I want to open a new tab, I use the same middle-click to open it in a new tab. I don't need any site to decide that for me.

[TryNinja]

Whenever a site does that, I middle-click on the site's tab to close it. It's extra work, and annoying. If I want to open a new tab, I use the same middle-click to open it in a new tab. I don't need any site to decide that for me.

FYI, I used to use the same target=_blank for all external links in my projects/sites. I think it was an old post of yours that changed my mind, and now I'm changing all my links so that the user has the option to decide where they want it to open.

[joker_josue]

Whenever a site does that, I middle-click on the site's tab to close it. It's extra work, and annoying. If I want to open a new tab, I use the same middle-click to open it in a new tab. I don't need any site to decide that for me.

FYI, I used to use the same target=_blank for all external links in my projects/sites. I think it was an old post of yours that changed my mind, and now I'm changing all my links so that the user has the option to decide where they want it to open.

You're absolutely right. I'd never thought of it from that perspective. I'll soon be adjusting this situation at TalkImg and applying this "philosophy" to other projects. It's simpler, safer, and gives the website visitor more control!

[Fivestar4everMVP]

I checked it and it seems very useful as a newbie I can find things more easy

I suggest your strongly to add a SSL/TLS encryption so the site is not served as Not Secure

second suggest add  target="_blank"  so when user search and click on a thread it don't leave ur site

That's strange... Talksearch already has a TLS certificate from Google App Engine. It should not display the "not secure" prompt. Can you try with a different device and browser ans see if you can replicate the issue?

I just checked, the site is secure from my end, I am visiting it for the first time and I did not see any "Not secure" prompt, I am assuming the issue is from the user's device, possibly his device needs a software update or upgrade, either of the two.

target=_blank sounds like an interesting idea, but it depends on whether people want to leave talksearch after finding the thread they want, or if they are using this as a way to read old posts.

The target=_blank is a good idea, I don't know about other users but I personally like the search engine to remain active/open even when I had click on a search result, as there is always the possibility that the results I click on may not be exactly what I am looking for but I still may like to leave that open and return to the search engine and click on another result.
So, instead of opening a clicked link on the same tab, I think it would make more sense if the link opens in a new tab while the search engine maintains its own tab.

Congratulations bud, this is a very useful tool for this forum for sure.

[NotATether]

The target=_blank is a good idea, I don't know about other users but I personally like the search engine to remain active/open even when I had click on a search result, as there is always the possibility that the results I click on may not be exactly what I am looking for but I still may like to leave that open and return to the search engine and click on another result.
So, instead of opening a clicked link on the same tab, I think it would make more sense if the link opens in a new tab while the search engine maintains its own tab.

Google and DuckDuckGo don't open search result links in a new tab, to take an example. So it's one of those things where you have to go by precedent and do what all the other services are doing. So I don't think I will be adding target=_blank to the links. Besides, I feel like the search results page could use some polishing if people actually stay on it for more than a few seconds.

Congratulations bud, this is a very useful tool for this forum for sure.

Thanks. I've meant to add vector embeddings to each of the posts in order to make ML search possible, but I've never had the time to do that yet. I have the hardware and everything ready (48 CPU cores but no GPUs), it's just waiting for my action.

[GazetaBitcoin]

Hey NotATether, please be aware that a new translation was made by AOBT for your topic:

- Serbian translation, made by katanic97

Are you so kind, please, to add it to OP as well?

[NotATether]

Hey NotATether, please be aware that a new translation was made by AOBT for your topic:

- Serbian translation, made by katanic97

Are you so kind, please, to add it to OP as well?

It has been done.

[NotATether]

Add Talksearch to Google Chrome search engine!



1. Go to chrome://settings/search
2. Click on "Manage search engines and site search"
3. Scroll down to Site search and click on "Add"
4. Fill in the values as follows:
  - Name: Talksearch
  - Shortcut: @talk
  - URL: https://talksearch.io/search?q=%s

Now you can directly search Bitcointalk posts from the address bar. For example, you can type "@talk bitcointalk memes" to find memes on Bitcointalk.

[NotATether]

Bump

Natural language search coming soon...