Aline supports associating some data with each line in a file (example: profiling data indicating hot path / cold path), and building this into a compressed map that can later be reapplied to a new version of the file, preserving line associations across insertions/deletions.
This works using the patience diff algorithm described by Bram Cohen, and this code was derived from watt/ruby_patience_diff.
(see example/run-example)
Given an original source file file.txt:
line1
line2
line3
line4
And a data.json:
[0.8969537475692057,0.293077462817166,0.5673016517223046,0.24424795082483]
We run aline build-map...
bin/aline build-map file.txt data.json > file.alinemap
Each line and each data element is represented by a single byte so this map will be 8 bytes long.
Later, we have a new version of the file (new-file.txt):
line1
lineFOOBAR
line4
line5
You can see there's been a change, an addition, and a deletion. If we run:
bin/aline remap file.alinemap new-file.txt
You'll see...
line1 # (in bright red)
lineFOOBAR
line4 # (in dark red)
line5
This is representing that line1 and line4 are present in the original file, and correspond to relatively high and low values repectively; while lineFOOBAR and line5 are not present in the original file.
The line association works pretty much the same as git's better diff algorithms.