Fuse-bgzip is a read-only overlay filesystem to transparently uncompress indexed GZIP files. This allows access the content of large GZIP files without the need to first uncompress them.

(The index file is to make it possible to seek() in the compressed file without taking a huge performance hit.)

BGZIP

BGZIP is a utility to create a special type of GZIP files that are indexed.

Compiling

gcc -o fuse-bgzip fuse-bgzip.c -lfuse -ltdb -lhts

Create an index file

bgzip -i foo.bin.ecm

Which will create the two files foo.bin.ecm.gz and foo.bin.ecm.gz.gzi

Mounting an overlay

fuse-bgzip -m

Unmouning the filesystem

fusermount -u

Example

Assume we have a directory that contains a n indexed file:

ls -l test total 485548 -rw-rw-r-- 1 sahlberg sahlberg 382199471 Oct 24 2007 foo.bin.ecm.gz -rw-rw-r-- 1 sahlberg sahlberg 91337 Oct 24 2007 foo.bin.ecm.gz.gzi -rw-rw-r-- 1 sahlberg sahlberg 132267 Oct 24 2007 foo.bin.ecm.edi -rw-rw-r-- 1 sahlberg sahlberg 114 Oct 24 2007 foo.cue

Then mount fuse-unecm ontop of this directory:

fuse-bgzip -m test

And the result becomes:

ls -l test total 485548 -rw-rw-r-- 1 sahlberg sahlberg 497190410 Oct 24 2007 foo.bin.ecm -rw-rw-r-- 1 sahlberg sahlberg 132267 Oct 24 2007 foo.bin.ecm.edi -rw-rw-r-- 1 sahlberg sahlberg 114 Oct 24 2007 foo.cue