How to shorten image sequence output in tree command?

Question

I'm in a directory where running tree command produces something like this:

├── directory1
│ └── image_sequence
│ ├── image.0001.jpg
│ ├── image.0002.jpg
│ ├── image.0003.jpg
│ ├── image.0004.jpg
│ ├── image.0005.jpg
│ └── image.0006.jpg
│ 
└── directory2
 ├── somefile.ext
 └── someanotherfile.ext2

The image sequence inside image_sequence produces a large listing that I want to trim. My desired output is something like below:

├── directory1
│ └── image_sequence
│ └── image.####.jpg
│ 
└── directory2
 ├── somefile.ext
 └── someanotherfile.ext2

Can the output of tree command somehow be modified?

• Jeff Schaller: I'm confused about whether you want to use the tree command or not.
• Jeff Schaller: unix.stackexchange.com/q/334438/117549 may be the answer to this question; see also unix.stackexchange.com/q/368884/117549 and unix.stackexchange.com/q/47805/117549
• Santosh Kumar: Tree command can be used, except that the whole sequence listing needs to converted to image.####.jpg format. Then use Python to modify image sequence output. The links you gave suggested the flags which will completely omit the output of image sequence. There's also an option called --filelimit which is close to what I want, but again it completely omits the image sequence output. I fear that using Python completely will not produce the identical output (the symbols and all).

Answer 1

Try this:

tree | sed '/\xe2\x94\x9c\xe2\x94\x80\xe2\x94\x80 image\.[0-9]\+\.jpg/d; s/\(\xe2\x94\x94\xe2\x94\x80\xe2\x94\x80 image\.\)[0-9]\+\(\.jpg\)/1円####2円/'

• The first /.../d; deletes all lines containing ├── image.[0-9]+.jpg (pseudo-pattern) entries
• The second s/.../1円####2円/ replaces the last line └── image.[0-9]+.jpg

Output:

$ tree | sed '/\xe2\x94\x9c\xe2\x94\x80\xe2\x94\x80 image\.[0-9]\+\.jpg/d; s/\(\xe2\x94\x94\xe2\x94\x80\xe2\x94\x80 ima
ge\.\)[0-9]\+\(\.jpg\)/1円####2円/'
.
├── directory1
│  └── image_sequence
│  └── image.####.jpg
└── directory2
 ├── someanotherfile.ext
 └── somefile.ext
3 directories, 8 files

This will of course only work if all files in image_sequence match the image pattern and will modify filenames in other directories matching the patterns. If the last file in image_sequence for example is readme.txt, then you will remove all image entries instead.

• Santosh Kumar: What are those hex values?
• Freddy: ├── and └──. This is why it only works if all files match those patterns. So I wouldn't use this solution if you have other files inside image_sequence or other image.xxxx.jpg files in other directories.
• Santosh Kumar: One last doubt. I want to replace ├── and └── to something simpler. Say |- and L. Because my shell is not printing it well. How can I do that?
• Santosh Kumar: Well, I see there's an option for --charset which take ascii or unicode. Can you please modify your answer with that?
• Freddy: With the ascii option it's tree --charset=ascii | sed '/|-- image\.[0-9]\+\.jpg/d; s/\(`-- image\.\)[0-9]\+\(\.jpg\)/1円####2円/'

Answer 2

You could replace the numerical parts of the sequential filenames with #s, using a sed expression (similar to the second one in Freddy's answer). uniq can then remove the duplicate lines:

tree | sed 's/\.[0-9]\+\.jpg/.####.jpg/g' | uniq

This will still leave two entries for the images (because the final line uses a different symbol in the tree-drawing part), but it has still trimmed the list down to a manageable length:

.
|-- directory1
| `-- image_sequence
| |-- image.####.jpg
| `-- image.####.jpg
`-- directory2
 |-- someanotherfile.ext2
 `-- somefile.ext

• Santosh Kumar: bash: s/\.[0-9]\+\.jpg/.####.jpg/g: No such file or directory
• JoostM: It's important to quote the s/.../.../g part with single quotes ('), or else bash will try to expand the [0-9] part (and fail).
• Santosh Kumar: I just copy pasted your snippet.
• JigglyNaga: @SantoshKumar I accidentally chopped the word sed out while editing. Could you try again?