14.4. Using Larch from Python

Although Larch contains its own scripting language, this is not Python, and it is perfectly reasonable to expect that the Larch analysis functionality be available in Python without using the Larch scripting language at all. That is, you may want to consider Larch to be a set of Python modules for the analysis of X-ray spectroscopic and related data. There are plenty of good reasons to want to do this, and it is certainly possible. However, because Larch is intended for use independent of an installed Python system, there are two points to keep in mind when using Larch from Python.

First, Larch keeps essentially all its functionality in plugins, which are not installed into the standard Python tree of installed modules, but into a folder specific to Larch. This means that, while Larch can be used from Python, Python will need to be told about where Larch is installed in order for the import statements to work.


add the named Plugin folder to Python’s sys.path, allowing python functions to be imported from the modules in the specified plugin folder.

The argument is the subfolder for each plugin, relative to the installed Larch plugins (typically $HOME/.larch/plugins for Unix or Mac OSX or $USER\larch\plugins on Windows).

Thus to get the _xafs.autobk() function into a Python module, you could do either:

import larch
from larch_plugins.xafs import autobk

The second consideration is that many of the functions in the Larch plugins will only work if they are passed an instance of the Larch interpreter. This interpreter instance is primarily used inside Larch plugins to create Groups and place them in the current symbol table, or to access data from the builtin _sys module or from data resouces loaded into the _xray module. Though you won’t need to write Larch scripts inside python (you can, but if you’re reading this section, you probably want to use Python instead of Larch), you will need an instance of the interpreter. This is easily created, and can then be passed to any of the plugin functions with the _larch keyword argument:

from larch import Interpreter
from larch_plugins.xafs import autobk
from larch_plugins.io import read_xdi

mylarch = Interpreter()
dat = read_xdi('../xafsdata/fe3c_rt.xdi', _larch=mylarch)
dat.mu = dat.mutrans
autobk(dat, rbkg=1.0, kweight=2, _larch=mylarch)

That is, the _larch=mylarch argument is vital to having _io.read_xdi() properly create a Larch group, and for allowing _xafs.autobk() to do the actual fit, and organize the results.

Note that you can create the interpreter without loading all the plugins using with_plugins=False. When running from python, this may be a reasonable default. If you want to add some of the plugins for a particular interpreter session, you can:

>>> import larch
>>> session = larch.Interpreter()
>>> session.run("cu_ka = xray_line('Cu', 'Ka1')")
>>> session.symtable.cu_ka
(8046.3, 0.577108, u'K', u'L3')

This would be nearly the same as doing:

>>> import larch
>>> from larch_plugins.xray import xray_line
>>> session = larch.Interpreter()
>>> xray_line('Cu', 'Ka1', _larch=session)
(8046.3, 0.577108, u'K', u'L3')

except that in the former, the Larch session retains the value in cu_ka.