Automatic disc ripper

[DanForever]

I've been working on a little c# program that will read a disc with makemkv and then query TheDiscDb (using graphql) to find any matching discs. It does this by matching the runtimes of the various titles (which is how I believe cd rippers did it when they checked cddb)

It's a little rough and ready, it needs a bunch of UX polish to smooth out the flow though the program, but if the disc is in TheDiscDb, it makes backing up your own media in a plex-friendly way super easy.

I would like to be able to add in the ability for the program to duplicate some of what ImportBuddy does. Either using ImportBuddy via the command line (if that's possible) or recreating the functionality with c#, but this way there's a nice GUI to help

[lfoust]

I like this idea. Looks like a great start.

The most reliable way to lookup a disc would be if you can calculate the hash (it is based on the files on the disc as shown here:

data/tools/ImportBuddy/source/ImportBuddy/TheDiscDb.Core/DiscHash/HashingExtensions.cs

Line 7 in 23cae72

public static string CalculateHash(this IEnumerable<FileHashInfo> files)

) and then you could query using the hash (here is a sample query)

As for using ImportBuddy - what functionality would you need? ImportBuddy does have a couple of command line arguments it can take to do specific tasks (like finalize a release or fill in the hash for a disc.json file) so we could add more if it makes sense.

[DanForever]

Regarding the hash, I'm hesitant because (unless I'm mistaken) you'd need to rip all the files first before they could be identified - the current method will present the "best match" based on times, but then also let you select any other matches (in theory, in practise it's untested because I don't actually know of any two films with the exact same runtime - down to the second - even my 1080p blu-ray UK copy of the collectors edition of avatar doesn't match up with the US 4K blu-ray). The graphql query just uses the runtime of the longest title on the track, but then it will match up every title on the disc against all the titles on all the releases returned by the query and rank them.

But if it becomes unreliable then switching to the hash is definitely a possibility.

Regarding ImportBuddy - The easiest would be to provide all the information it asks for in the form of command line parameters, so that this program could just present that in a nice gui. I like to imagine you could fill out all the information in the gui and hit a single button and then everything happens automagically under the hood.
With something like this, anything you can do to reduce friction for the user would improve the rate at which new items are submitted (assuming that's something you want ;) )

[lfoust]

you'd need to rip all the files first before they could be identified

No, you don't need to rip anything. The hash is calculated directly from the file system on the disc itself

With something like this, anything you can do to reduce friction for the user would improve the rate at which new items are submitted (assuming that's something you want ;) )

Yes! Honestly, the friction of submitting data is thediscdb's biggest problem. I have been working on a design to address this issue but just haven't had the time to build it.

I'll have to give some more thought to what it would look like to expose all the possible ImportBuddy inputs to the command line. It sounds like you want to essentially create an alternate GUI for ImportBuddy which isn't how the code is currently built - so it would likely be a big re-design of ImportBuddy

[DanForever]

OK so knowing that then yeah the hash is certainly a possibility.

Is ImportBuddy also open source on github? Maybe I could help there as well...

[lfoust]

Yes, the code is here: https://github.com/TheDiscDb/data/tree/main/tools/ImportBuddy/source/ImportBuddy/ImportBuddy

[DanForever]

I've cleaned up the code a bit and submitted it here: https://github.com/DanForever/DiscRipper
I'd love it if you could grab it and take a look and suggest next steps. I'm also gonna put a bunch of tasks/"todo"s into the github "project" thingy

[DanForever]

@lfoust Sneak preview for you:

I actually found I could link the ImportBuddy assemblies and reuse a lot of functionality there (I also had to copy a bunch from the source code too, where it was hardcoded to work off console input, but it works!)

The image above is very WIP, but should hopefully simplify a lot of stuff

[lfoust]

This looks very useful! I am actually working on an online submission process (no ImportBuddy needed!) which incorporates some of these same ideas

[DanForever]

Do you think it will be possible to integrate my app such that it can submit directly to your website (rather than via a PR request)?

[lfoust]

Not likely - even the online submission system I am building will end up in github. The data in github is ultimately the source of truth of the data for the site

[DanForever]

ok, so with that in mind, how are you tackling that particular issue? Are you hooking up your process so that people sign in via github and then you generate the PR request automatically? or will the process deliver you a zip file that you'll integrate into the repository manually?

[DanForever]

(or some other approach that I haven't thought of? 😁 )

[DanForever]

Check this out @lfoust: DanForever@45a3543

This is data generated using my DiscRipper app (which incorporates some of the ImportBuddy assemblies). Most of the data is roughly the same, except for the metadata.json which I think is actually because the imdb functions in ImportBuddy aren't working currently? (Either that or I haven't seen how to enable them, but I think I get the same values just running ImportBuddy directly).

Things of note:

• I haven't bothered to add the year to the summary text file, but I can do that if you'd prefer
• The filenames are different because I didn't bother to modify them as a user (and that data's not available/useful for DiscRipper in general)
• The locale property now has proper capitalisation
• SortTitle appears to have the release/edition year rather than a "1" at the front, not sure why that is (I haven't investigated it, but either I changed something accidentally or some data was present that wasn't before)

[DanForever]

You can try it for yourself now, if you want: https://github.com/DanForever/DiscRipper/releases/tag/0.1.0

[DanForever]

@lfoust I've recently submitted some changes that should make using the app a whole lot more user friendly, so I would love it if you would be able to find the time to run the app and try ripping a disc or adding a new release to the database though it (as opposed to import buddy).

I would love to know what you think, and what issues you encounter.

I want to avoid posting any screenshots or instructions as I want to get your first impressions (I haven't even written a readme yet)