Use regex recursively taking indentation level into account - regex

Question

I am trying to parse a custom input file for a simulation code I am writting. It consist of nested "objects" with properties, values (see the link).
Here is an example file and the regex I am using currently.
([^:#\n]*):?([^#\n]*)#?.*\n
It is made such that each match is a line, with two capture group, one for the property and one for its value. It also excludes "#" and ":" from the character set, as they correspond to a comment delimiter and a property:value delimiter respectively.
How can I modify my regex so as to match the structure recursively? That is if line n+1 has an identation level higher than line n, it should be matched as a subgroup of line n's match.
I am working on Octave, which uses PCRE regex format.

Answer 1

I asked if you have control over the data format because as it is, the data is very easy to parse with YAML instead of regex.
The only problem is that the object is not well formed:
1) Take the regions object for example, it has many attributes called layer all of them. I think your intention is to build a list of layers instead of a lot of properties with the same name.
2) Consider now each layer property that has a corresponding value. Following each layer are orphan attributes that I presume belong to each layer.
With these ideas in mind. If you form your object following YAML rules, it would be a breeze to parse it.
I know that you are working in Octave, but consider the modifications I made to your data, and how easy it is to parse it, in this case with python.
DATA AS YOU HAVE IT NOW
case :
name : tandem solar cell
options :
verbose : true
t_stamp : system
units :
energy : eV
length : nm
time : s
tension : V
temperature: K
mqty : mole
light : cd
regions :
layer : Glass
geometry:
thick : 80 nm
npoints : 10
optical :
nk_file : vacuum.txt
layer : FTO
geometry:
thick : 10 nm
npoints : 10
optical :
nk_file : vacuum.txt
MODIFIED DATA TO COMPLY WITH YAML SYNTAX
case :
name : tandem solar cell
options :
verbose : true
t_stamp : system # a sample comment
units :
energy : eV
length : nm
time : s
tension : V
temperature: K
mqty : mole
light : cd
regions :
- layer : Glass # ADDED THE - TO MAKE IT A LIST OF LAYERS
geometry : # AND KEEP INDENTATION PROPERLY
thick : 80 nm
npoints : 10
optical :
nk_file : vacuum.txt
- layer : FTO
geometry:
thick : 10 nm
npoints : 10
optical :
nk_file : vacuum.txt
With only these instruction you get your object parsed:
import yaml
data = yaml.load(text)
""" your data would be parsed as:
{'case': {'name': 'tandem solar cell',
'options': {'t_stamp': 'system', 'verbose': True},
'regions': [{'geometry': {'npoints': 10, 'thick': '80 nm'},
'layer': 'Glass',
'optical': {'nk_file': 'vacuum.txt'}},
{'geometry': {'npoints': 10, 'thick': '10 nm'},
'layer': 'FTO',
'optical': {'nk_file': 'vacuum.txt'}}],
'units': {'energy': 'eV',
'length': 'nm',
'light': 'cd',
'mqty': 'mole',
'temperature': 'K',
'tension': 'V',
'time': 's'}}}
"""