lvmio

Typed reader and writer for LabVIEW Measurement (.lvm) files, standard library only.

.lvm is the text format LabVIEW's Write To Measurement File express VI produces. lvmio reads those files into frozen dataclasses and writes files in the same layout, byte for byte: the files LabVIEW 2026 wrote in the test captures come back unchanged after a read and a write (the guide explains the few exceptions), and LabVIEW's own reader reads lvmio's files with the values that were written (to the last bit, except for about one cell in a few thousand that LabVIEW's reader rounds one unit off).

Install

pip install lvmio

Python 3.10 or newer. No runtime dependencies.

Quick start

>>> from lvmio import Channel, LvmFile, Segment, format_lvm, parse_lvm
>>> channels = [
...     Channel("Voltage", [1.0, 2.5, -3.25], unit="V", delta_x=0.1),
...     Channel("Current", [0.5, 0.25, 1e-9], unit="A", delta_x=0.1),
... ]
>>> text = format_lvm(LvmFile([Segment(channels, comment="run 1")]), operator="lab")
>>> lvm = parse_lvm(text)
>>> segment = lvm.segments[0]
>>> segment.names, segment["Current"].data, segment.comment
(['Voltage', 'Current'], [0.5, 0.25, 1e-09], 'run 1')
>>> segment["Voltage"].time_axis()
[0.0, 0.1, 0.2]

write_lvm(path, data) and read_lvm(path) do the same with files. data can be an LvmFile, one Segment, a list of Channel objects or a mapping of name to samples (any iterable of numbers, NumPy arrays included):

>>> import pathlib, tempfile
>>> from lvmio import read_lvm, write_lvm
>>> path = pathlib.Path(tempfile.mkdtemp(), "quick.lvm")
>>> write_lvm(path, {"x": [1.0, 2.0], "y": [3.0, 4.0]}, decimal_separator=",")
>>> read_lvm(path).segments[0]["y"].data
[3.0, 4.0]

Keyword arguments set the file header: operator, date, time, separator ("Tab" or "Comma"), decimal_separator ("." or ","), x_columns ("No", "One" or "Multi"), precision (decimals per sample, 6 by default as in LabVIEW) and, for write_lvm, encoding.

What it handles

• The three X column layouts LabVIEW writes: No (an empty X column), One (one shared X column) and Multi (an X column before every channel).
• Tab and comma separators, the decimal comma, several segments with their comments, and a file header repeated in the middle of a file (LabVIEW writes one when a later session appends to the file).
• LabVIEW's number format, including how it rounds exact ties (2.500000, 1.000000E-9, NaN, Inf, -Inf); empty cells read as NaN.
• Comments with LabVIEW's escapes (\09 for a tab, \\ for a backslash).
• Text in UTF-8 or in Windows-1252, the code page LabVIEW writes on a Western-European Windows system; read_lvm tries UTF-8 first and then Windows-1252. For a file LabVIEW will read, write with encoding="cp1252": LabVIEW reads files in its code page.

Where LabVIEW reads part of a cell (2.5V as 2.5), lvmio refuses the cell; where both read a value they mostly agree, and the guide lists every case where they do not (for example LabVIEW's reader gives a file without units the unit V and rounds a few cells one unit in the last place off). The guide lists every difference.

Command line

lvmio header run.lvm                 # file header fields, --json for JSON
lvmio channels run.lvm               # channels and sample counts per segment
lvmio dump run.lvm --channel Voltage --time   # one channel as CSV with its X axis
lvmio dump run.lvm --format json     # a segment as JSON
lvmio header old.lvm --encoding cp1252
cat run.lvm | lvmio header -

python -m lvmio runs the same command. Reading errors print one line to standard error and exit with status 1; invalid arguments exit with status 2.

Why

LabVIEW reads and writes .lvm itself. On the Python side, the lvm_read package (1.26) reads the format into NumPy arrays and has no writer. lvmio writes files in the layout LabVIEW's own writer produces, so a Python program can hand data to a LabVIEW VI, and it reads them back into typed objects without NumPy.

Documentation

• Guide: the format as LabVIEW 2026 writes it, reading rules, errors, and upgrading from 0.2.
• Changelog

License

MIT. See LICENSE.

Links

• Website: https://lianes8studio.com
• GitHub: https://github.com/JaviLianes8
• LinkedIn: https://www.linkedin.com/in/jlianes/