Please add support for logging in with an account on TorrentGalaxy

[NoX1De]

Is there an existing issue for this?

• I have searched the existing issues

Is your feature request related to a problem? Please describe

TorrentGalaxy supports logging in with a user account, but the indexer definition does not support this for some reason, being authenticated with an account on the website is beneficial for a number of reasons that I could explain if need be.

Describe the solution you'd like

Please add a login block to https://github.com/Prowlarr/Indexers/blob/b7798267fab07d44b1d73a1e2a21ae552df78eaa/definitions/v7/torrentgalaxy.yml that allows configurable username and password for login in addition to the captcha settings.

Describe alternatives you've considered

N/A

Anything else?

Thank you !

[bakerboy448]

@ilike2burnthing @mynameisbogdan

[ilike2burnthing]

From a quick look, the only thing I could see as potentially being useful for a minority of users would be disabling adult content. Is there anything else I'm missing? Even signed in they're still just using public trackers.

There's also seemingly an API for those ranked verified uploaders or higher, but that would need a whole new indexer and an account with access to the API, and would ultimately benefit next to no one.

[mynameisbogdan]

actually I was debating to propose we remove this indexer since the captcha is already making it unusable. 😬

[ilike2burnthing]

Really? I haven't been having any major issues with it. The odd time I'll see a healthcheck warning for it in the *arrs, but it either resolves itself in the next RSS sync or I reconfigure the indexer with the new captcha.

[NoX1De]

Adding support for logging in appears to reduce captcha frequency in cursory manual testing...I'm hoping this makes the indexer more useful.

Edit: In other words, while logged in manually I get prompted far less frequently for a captcha...I'm hoping this translates to the same behavior once logged in via Prowlarr. I think the change is worth making/testing, this is one of the best sources IMO.

[mynameisbogdan]

Really? I haven't been having any major issues with it. The odd time I'll see a healthcheck warning for it in the *arrs, but it either resolves itself in the next RSS sync or I reconfigure the indexer with the new captcha.

Well, many complaints of it not working and considering there's a captcha doesn't make it very automation friendly. 🤷‍♂️

[NoX1De]

Well, many complaints of it not working and considering there's a captcha doesn't make it very automation friendly. 🤷‍♂️

See my comment above, this is a top notch source from a content quality perspective, if logging in with an account via Prowlerr also greatly reduces the captcha as it does manually then this change would be a significant improvement. Being logged in manually seems to mostly functionally eliminate it/revert back to the timing before they made recent changes to greatly increase the frequency of the captcha...at least making it far more usable and less annoying, based on my testing today.

[bakerboy448]

TorrentGalaxy and this is a top notch source from a content quality perspective is an oxymoron of a sentence

Shitty public tracker that inflates and lies about its seed count & is anti-automation given it's increased captcha is rather comical

but anyway. maybe someone with enough drive to implement the log in version will come around, one day... maybe

[mynameisbogdan]

The issue is that adding auth to the existing definition is gonna piss of a lot of people.

[ilike2burnthing]

We could have a second semi-private login indexer, but it'd need to be able to handle both the login and galaxy fence pages, which Cardigann can't currently do.

Either the second indexer would need to be written in C# or Cardigann adapted to:

1. fetch captcha from galaxyfence.php
2. submit user's captcha response
3. post login inputs to account-login.php
4. load test path and check for selector

[mynameisbogdan]

wait a second, even if we add auth the captcha won't go away? 🤔 seems a lot of work for 0 return. 😬

[NoX1De]

TorrentGalaxy and this is a top notch source from a content quality perspective is an oxymoron of a sentence

Shitty public tracker that inflates and lies about its seed count & is anti-automation given it's increased captcha is rather comical

but anyway. maybe someone with enough drive to implement the log in version will come around, one day... maybe

I agree they are a pain but I'm sure they have their own expenses they have to modulate/minimize, who knows their motivation for being such a pain but reducing operating expenses would be my first guess...agree to disagree on available content there, lots of quality 4K/UHD not found elsewhere in public tracker land, at least not that I'm currently aware of. Never really cared to pay attention to any kind of supposed seed count fraud. lol.

We could have a second semi-private login indexer, but it'd need to be able to handle both the login and galaxy fence pages, which Cardigann can't currently do.

Either the second indexer would need to be written in C# or Cardigann adapted to:

1. fetch captcha from galaxyfence.php
2. submit user's captcha response
3. post login inputs to account-login.php
4. load test path and check for selector

This is along the lines of what I've been pondering, was also thinking this should be semi-private and a second indexer definition for that makes a ton of sense. Basically, had been wondering if this can currently be done in a definition... sounds like that's a no.

Any idea how much effort adapting Cardigann would be to enable this functionality? Or which option would be the least effort, I'm not currently familiar with Cardigann's codebase or how it's being leveraged wrt to Prowlarr indexers. At least it's written in golang...heh. Just taking a super quick glance, it looks like the following file would need to be adapated, perhaps?

https://github.com/cardigann/cardigann/blob/4ad57c5e08e96f86bf08e5f97db40f458a29e119/indexer/parser.go

[mynameisbogdan]

Both Jackett and Prowlarr have their own implementation of Cardigann in C#.

[ilike2burnthing]

wait a second, even if we add auth the captcha won't go away? 🤔 seems a lot of work for 0 return. 😬

I haven't tested @NoX1De's claim that being logged in reduces the frequency with which challenges occur, but there are forum posts about it being an issue, and I can purposefully trigger a challenge.

We have a few indexers (e.g. CRT, PixelCove, BJ-Share) for which cookies are used rather than credentials just because 2FA appears on a second page, so adapting Cardigann to handle multi-page logins wouldn't be solely for this indexer.

Unfortunately, we can't use cookies here, as you'd need a new one every time there's a captcha challenge.

[NoX1De]

@ilike2burnthing I'm pretty confident that being logged in makes a substantial difference in the frequency of being prompted for a captcha after testing this for a week or so.

However, I cannot currently account for if they are using something like the user agent to trigger captcha frequency in a variable fashion depending on what the user agent is. I suppose we could test that by logging into TGx via normal browser and spoofing the user agent to the Prowlarr user agent string and seeing if the captcha occurrences stay the same.

I don't know/have the Prowlarr user agent string handy but that seems like a plausible way to test with the least amount of effort:

1.) Create TorrentGalaxy account
2.) Launch web browser and spoof user agent to Prowlarr user agent string
3.) Login via web browser
4.) Browse away and observe in comparison to browsing while logged in with normal browser user agent

[ilike2burnthing]

Prowlarr/{Prowlarr version} ({platform} {platform version})
e.g.:
Prowlarr/1.10.5.4116 (windows 10.0.19041.0)

No, it doesn't trigger a captcha challenge (or at least not immediately).

[soredake]

I'm pretty confident that being logged in makes a substantial difference in the frequency of being prompted for a captcha after testing this for a week or so.

For me, it changed nothing (with normal browser or prowlarr user-agent), I registered on tgx, after 3 days testing I got like 10 captchas, no difference.

[NoX1De]

I'm pretty confident that being logged in makes a substantial difference in the frequency of being prompted for a captcha after testing this for a week or so.

For me, it changed nothing (with normal browser or prowlarr user-agent), I registered on tgx, after 3 days testing I got like 10 captchas, no difference.

Hmmm well that's discouraging. I never got around to testing this with the Prowlarr user-agent but when I did test with normal browser it seemed to reduce the CAPTCHA frequency significantly. I wonder if they got even more aggressive with logged in accounts since then....shrug, really sucks too because I think they are one of the best in terms of content breadth, clearly not with implementing access policy for their site.

[mikekuzak]

drive

Any public alternative? Rarbg style?

[ilike2burnthing]

TheRARBG