lvm_read

LabView Measurement file reader

Reads the text .lvm files written by National Instruments LabVIEW into plain Python dictionaries and NumPy arrays. Tab- and comma-separated files, the comma decimal separator, several segments per file, missing fields and comments are all handled.

Installation

pip install lvm_read

Requires Python 3.10 or newer and NumPy.

Usage

import lvm_read

lvm = lvm_read.read('short.lvm')

lvm['Segments']              # 1
lvm[0]['Channel names']      # ['Excitation (Trigger)', 'Response (Trigger)', 'Comment']
lvm[0]['data']               # array of shape (10, 2), one column per channel
lvm[0]['data'][0, 0]         # 0.914018

The time axis is not stored column by column unless the file says so; it is reconstructed from X0 and Delta_X:

import numpy as np

segment = lvm[0]
time = segment['X0'][0] + np.arange(len(segment['data'])) * segment['Delta_X'][0]

What read returns

A dictionary with the header fields of the file (Date, Operator, Separator, Decimal_Separator, X_Columns, ...), the number of segments under Segments, and one entry per segment under the integer keys 0, 1, ...:

data	2-D NumPy array of floats, one row per sample
Channel names	names from the X_Value line
Channels	number of channels
Samples	samples per channel
X0, Delta_X	start and step of the X axis, per channel
comments	text of the comment of each row, '' where there is none; present only if the last channel is Comment

Anything that is not a number - a comment, an empty field - becomes nan in data, so the array stays numeric. Rows of unequal length, which occur when only some rows carry a comment, are padded with nan.

Reading a string instead of a file

lvm = lvm_read.read_str(open('short.lvm', encoding='utf8').read())

Both read and read_str take the column separator from the Separator field of the header. Pass separator explicitly for a file whose header does not have it. A file no segment can be read from raises ValueError.

Caching

read writes a <filename>.pkl next to the file and reads it back on the next call, which is much faster than parsing the text again. Turn it off with read_from_pickle=False and dump_file=False. Note that the cache is a pickle: do not point read at .lvm files from an untrusted source while it is enabled.

More

See the Showcase lvm_read.ipynb notebook for a walk through a few sample files, and the file format specification for what the fields mean.