Question

I have an LV2 plugin and I want to use Python to extract its metadata – plugin name, description, list of control and audio ports and specification of each port.

With LADSPA the instructions were pretty clear, although a bit difficult to implement in Python: I just needed to call ladspa_descriptor() function. Now with LV2 there’s a .ttl file, simples to access but more complicated to parse.

Is there any python library that will make this job simple?

Answer 1

The LV2 documentation generation tools use RDFLib. It is probably the most popular RDF interface for Python, though does much more than just parse Turtle. It is a good choice if performance is not an issue, but is unfortunately really slow.

If you need to actually instantiate and use plugins, you probably want to use an existing LV2 implementation. As Steve mentioned, Lilv is for this. It is not limited to any static default location, but will look in all the locations in LV2_PATH. You can set this environment variable to whatever you want before calling Lilv and it will only look in those locations. Alternatively, if you want to specifically load just one bundle at a time, there is a function for that: lilv_world_load_bundle().

There are SWIG-based Python bindings included with Lilv, but they stop short of actually allowing you to process data. However there is a project to wrap Lilv that allows processing of audio using scipy arrays: http://pyslv2.sourceforge.net/ (despite the name they are indeed Lilv bindings and not bindings for its predecessor SLV2)

That said, if you only need to get static information from the Turtle files, involving C libraries is probably more trouble than it is worth. One of the big advantages of using standard data files is ease of use with existing tools. To get the number of ports on a plugin, you simply need to count the number of triples that match the pattern (plugin, lv2:port, *). Here is an example Python script that prints the number of ports of a plugin, given the file to read and the plugin URI as command line arguments:

#!/usr/bin/env python
import rdflib
import sys

lv2 = rdflib.Namespace('http://lv2plug.in/ns/lv2core#')

path = sys.argv[1]
plugin = rdflib.URIRef(sys.argv[2])

model = rdflib.ConjunctiveGraph()
model.parse(path, format='n3')

num_ports = 0
for i in model.triples(plugin, lv2.port, None]):
    num_ports += 1

print('%s has %u ports' % (plugin, num_ports))