"--max-depth" support - allow for summary of size per directory

[sforsythe-wiley]

I was hoping the output -o json would do this, but that feature appears more to save the state, to be fed back into gdu at a later time interactively.

I was hoping for the equiv of du --max-depth flag
with possible improvement of output in json format,

eg:

du -h --max-depth=0
1.2G    .

du -h --max-depth=1
1.2G    .
 92M    ./2024
863M    ./2023
305M    ./2022

du -h -max-depth=2
1.2G    .
141M    ./2022/07
31M     ./2022/08
25M     ./2022/09
41M     ./2022/10
48M     ./2022/11
20M     ./2022/12
305M    ./2022
113M    ./2023/01
355M    ./2023/02
121M    ./2023/03
49M     ./2023/04
53M     ./2023/05
18M     ./2023/06
13M     ./2023/07
6.2M    ./2023/08
20M     ./2023/09
30M     ./2023/10
55M     ./2023/11
35M     ./2023/12
863M    ./2023
14M     ./2024/01
78M     ./2024/02
92M     ./2024

[rwahyudi]

What would be the drawback of including the directory usage and file count in the json file for directory?
I think this would allow us to run jq filter and produce the above output

Example :

[
  1,
  2,
  {
    "progname": "gdu",
    "progver": "v5.29.0",
    "timestamp": 1724120998
  },
  [
    {
      "name": "/etc/audisp",
      "mtime": 1701018116,
	  "asize" : 9313,	<-----------
	  "dsize" : 20480,	<-----------
	  "items" : 4		<-----------
    },
    [
      {
        "name": "plugins.d",
        "mtime": 1701018116,
		"asize": 4971,	<-----------
		"dsize": 12288,	<-----------
		"items": 3	<-----------
      },
      {
        "name": "af_unix.conf",
        "asize": 358,
        "dsize": 4096,
        "mtime": 1557149431
      },
      {
        "name": "syslog.conf",
        "asize": 517,
        "dsize": 4096,
        "mtime": 1557149431
      }
    ],
    {
      "name": "audispd.conf",
      "asize": 246,
      "dsize": 4096,
      "mtime": 1689530538
    }
  ]
]

[sforsythe-wiley]

Yes, apologies, I didn't mean to imply the output should look the same as the du example, just the desired equivalent outcome.
Putting into the json output and using jq to filter/sort was what I was envisioning as well, sorry if didn't make that clear.

Possible caveat that maybe need to indicate a directory vs a file.
I guess you could filter on elements that contain an "items" count value.