decoyfs is a FUSE filesystem that reproduces a tree hierarchy of files and folders stored in SQLite. The files have no content though (hence decoy), this filesystem is only for browsing the hierarchy.
decoy-fs comes with a tool to snapshot an directory tree in such an SQLite database.
Let's say we snapshotted a directory in sample.sqlite. See how it looks:
% sqlite3 sample.sqlite "select * from files"
┌─────────┬───────────────┬───────┬──────────┬───────┬──────┬──────┬──────────┬──────────────────┬──────────────────┬──────────────────┐
│ dirname │ basename │ mode │ ino │ nlink │ uid │ gid │ size │ mtime │ atime │ ctime │
├─────────┼───────────────┼───────┼──────────┼───────┼──────┼──────┼──────────┼──────────────────┼──────────────────┼──────────────────┤
│ │ │ 16877 │ 24698924 │ 4 │ 1000 │ 1000 │ 4096 │ 1672860821.58673 │ 1672860786.2864 │ 1672860821.58673 │
│ │ bar │ 16877 │ 24698926 │ 3 │ 1000 │ 1000 │ 4096 │ 1672861017.35253 │ 1672860821.58673 │ 1672861017.35253 │
│ │ foo │ 16877 │ 24698925 │ 2 │ 1000 │ 1000 │ 4096 │ 1672860903.35948 │ 1672860821.58673 │ 1672860903.35948 │
│ bar │ baz │ 16877 │ 24699120 │ 2 │ 1000 │ 1000 │ 4096 │ 1672861045.02878 │ 1672860821.58673 │ 1672861045.02878 │
│ bar/baz │ archive.tgz │ 33188 │ 24653449 │ 1 │ 1000 │ 1000 │ 11768008 │ 1669434053 │ 1669434053 │ 1672861045.24478 │
│ bar │ script.sh │ 33261 │ 24653447 │ 1 │ 1000 │ 1000 │ 5920 │ 1672860424.38708 │ 1672860424.38708 │ 1672861017.35253 │
│ foo │ document.txt │ 33188 │ 24653307 │ 1 │ 1000 │ 1000 │ 0 │ 1672860857.03905 │ 1672860857.03905 │ 1672860903.35948 │
│ foo │ something.png │ 33188 │ 24653446 │ 1 │ 1000 │ 1000 │ 25296 │ 1672391150.51346 │ 1672391150.46946 │ 1672860894.4594 │
└─────────┴───────────────┴───────┴──────────┴───────┴──────┴──────┴──────────┴──────────────────┴──────────────────┴──────────────────┘Empty dirname column is the root directory.
Now we can run decoyfs to mount that virtual tree to ./mnt/:
% decoyfs -o dbpath=sample.sqlite ./mnt/We can now explore the virtual tree:
% ls mnt
bar foo
% tree -psDFh mnt/
[drwxr-xr-x 4.0K Jan 4 20:33] mnt/
├── [drwxr-xr-x 4.0K Jan 4 20:36] bar/
│ ├── [drwxr-xr-x 4.0K Jan 4 20:37] baz/
│ │ └── [-rw-r--r-- 11M Nov 26 04:40] archive.tgz
│ └── [-rwxr-xr-x 5.8K Jan 4 20:27] script.sh*
└── [drwxr-xr-x 4.0K Jan 4 20:35] foo/
├── [-rw-r--r-- 0 Jan 4 20:34] document.txt
└── [-rw-r--r-- 25K Dec 30 10:05] something.png
4 directories, 4 filesThe import-directory.py tool can generate a SQLite database with the proper schema and fill it with entries from an actual directory:
import-directory.py sample.sqlite real_folder/
The format is pretty simple so it's easy to create another tool to create such a database from another source (tar archive, remote file system, etc.)
The schema is as follows:
CREATE TABLE "files" (
"dirname" TEXT,
"basename" TEXT,
"mode" INTEGER,
"ino" INTEGER,
"nlink" INTEGER,
"uid" INTEGER,
"gid" INTEGER,
"size" INTEGER,
"mtime" INTEGER,
"atime" INTEGER,
"ctime" INTEGER,
PRIMARY KEY("dirname", "basename")
);It's also possible to import catcli databases:
import-catcli.py myfiles.sqlite myfiles.catcli