[Bug] Using 'archive' with 'parent-metadata' options breaks Imgur Album downloads

[harvii]

When downloading a reddit post linking to an imgur album with both 'archive' and 'parent-metadata' options enabled, only the first image in the album is saved and the rest are skipped.

I'm guessing this is because the imgur 'id' is overwritten by the reddit 'id':

gallery-dl/gallery_dl/extractor/imgur.py

Line 146 in befe635

archive_fmt = "{album[id]}_{id}"

Would using archive_fmt = "{album[id]}_{filename}" be a better option?

[mikf]

I'm guessing this is because the imgur 'id' is overwritten by the reddit 'id':

Yep, that's exactly it. Using {filename} should work since the metadata from Reddit doesn't have such a field in case it's an Imgur URL.

I've also extended the parent-metadata option a bit (e95f998) so you can tell it to not overwrite child metadata but to put the parent metadata entries into their own field. This also makes it possible to distinguish between child extractor and default usage with conditional directories and/or filenames:

{
    "extractor": {
        "reddit": {
            "parent-metadata": "parent",
            "parent-directory": true
        },
        "imgur": {
            "directory": {
                "'parent' in locals()": ["child extractor", "{parent[author]}"],
                "": ["default", "{album[id]}"]
            }
        }
    }
}