How to replace/remove "🖤"?

[EpicLPer]

Heya!

As the title says, I have one filename where DropBox really doesn't want to let me sync the file despite "path-restrict": "windows" already being in the config file, it always comes up with a "🖤" in the filename.
Is there any other way to get rid of specific symbols and replace/remove them automatically?

Thanks!

[mikf]

"path-restrict": "ascii" should get rid of any non-ASCII character like 🖤.

Alternatively, you could use a custom mapping like here and include 🖤.

gallery-dl/docs/gallery-dl-example.conf

Lines 16 to 27 in d50c312

	"#": "replace invalid path characters with unicode alternatives",
	"path-restrict": {
	"\\": "⧹",
	"/" : "⧸",
	"|" : "￨",
	":" : "꞉",
	"*" : "∗",
	"?" : "？",
	"\"": "″",
	"<" : "﹤",
	">" : "﹥"
	},

[EpicLPer]

Thanks for the answer, but I'm not sure how to implement this into the config. I tried putting "path-remove": "\\u1f5a4" instead of restrict now but it simply seems to remove all "4" from the filename afterwards...

[mikf]

The JSON representation of 🖤 is apparently \\ud83d\\udda4

>>> json.dumps("🖤")
'"\\ud83d\\udda4"'

edit: It has to be \ud83d\udda4 when putting it in a JSON file.

[Hrxn]

@mikf
According to configuration.rst, the "ascii" setting for "path-restrict" is basically "ascii": "^0-9A-Za-z_."
I assume this is like a regex range expression?
Isn't this a bit too strict then, because it would also replace valid ASCII characters like parentheses and the square brackets?

[mikf]

@Hrxn Yeah, it is a bit too strict. I can't remember exactly, but I might have used something close to youtube-dl's --restrict-filenames as a reference and that replaces basically everything that is not a letter or number with _.

[Hrxn]

Ah gotcha, makes sense. Yeah, you're right about this --restrict-filename option, it is pretty stringent, I mean whitespace is also valid ASCII, obv., but this is also always replaced by this option.

I would not suggest a change anyway, keep the option working as it is..
Maybe it can be extended by another special value, something like "ascii+" or whatever, as a preset that is somewhat less "strict".

Can I ask you another thing here?

If you use the object mapping variant of "path-restrict", is the replacement character enforced as a single character, i.e. limited in length?
Or would something like _-_ as replacement be possible, e.g.

        "path-restrict": {
            " " : "_",
            "\\": "_-_",
            "/" : "_-_",
            "|" : "_-_",
            ":" : "_-_"
            }

?

[mikf]

It is possible to use more than a single character as replacement.
That's why it says "replacement character (s) " in the docs.

[EpicLPer]

The JSON representation of 🖤 is apparently \\ud83d\\udda4

>>> json.dumps("🖤")
'"\\ud83d\\udda4"'

Awesome, this seems to have worked! I tried to find how to generate that specific string but couldn't find anything, now I even know how to do so :) Thanks!

[EpicLPer]

Actually @mikf this didn't work, I put "path-remove": "\\ud83d\\udda4" in my config and it still created a file with the black heart emoji in the path.

[mikf]

I got it to work by using only 1 backslash: "path-remove": "\ud83d\udda4"

And it seems to also be possible to use 🖤 directly, i.e. "path-remove": "🖤".

{
    "filename": "<🖤>",
    "path-remove": "🖤"
}

$ gallery-dl -c cfg.json imgur.com/asdqwe
/tmp/_/imgur/<>