parseUV

A Python package and PyQt6 GUI application for reading, visualizing, and exporting proprietary binary files from:

• Varian / Agilent Cary UV-Vis-NIR spectrophotometers (.DSW, .BSW)
• Shimadzu UV-Vis-NIR spectrophotometers (.SPC)

Features

• Direct Binary Parsing: Reads proprietary .DSW, .BSW, and .SPC files natively in Python without requiring proprietary software.
• Drag-and-Drop PyQt6 GUI: Interactive desktop GUI app with Matplotlib plotting widgets, sample/background spectrum categorization, CSV export, and instant reset.
• Batch Processing Tool: Command-line and programmatic batch processor to convert all spectrometer files in a directory to CSV and high-resolution plots (.png, .pdf, .svg).
• Publication-Ready Styling: Includes custom .mplstyle profile styled with TeX Gyre Heros, Helvetica, and Arial fonts.
• Font Installer: Automated helper script (parseuv-install-fonts) to discover and register TeX Gyre Heros fonts from MiKTeX or TeX Live into Matplotlib.
• Pandas Integration: Converts spectra into clean pandas.DataFrame objects aligned by wavelength.

Supported Formats

Installation

The package is available on PyPI.

Using pip

pip install parseuv

Using uv (Recommended)

uv add parseuv

Development Install (from source)

git clone https://github.com/RJFernandezTeran/ParseUV.git
cd ParseUV
uv pip install -e .

Or with standard pip:

pip install -e .

Font & Style Configuration

1. Install TeX Gyre Heros Fonts (Optional, Recommended)

To install TeX Gyre Heros fonts into Matplotlib's font manager from a local MiKTeX or TeX Live installation:

parseuv-install-fonts

Or from Python:

from parseuv.fonts import install_fonts

install_fonts()

2. Apply Custom Plotting Style

from parseuv import apply_style, parse_uv

# Apply the regular style profile (TeX Gyre Heros / Helvetica / Arial)
apply_style("regular")

data = parse_uv("path/to/file.DSW")
data.plot()

Available style profile:

• 'regular' (HLV_plt.mplstyle): Regular publication profile (18pt bold labels, sans-serif fonts).

Usage

1. PyQt6 Drag-and-Drop GUI

Launch the interactive desktop GUI application:

parseuv-gui
# or
python run_gui.py
# or
parseuv --gui

Features:

• Drag and drop .BSW, .DSW, or .SPC files into the drop zone.
• File selection dialog filter: Varian/Agilent Cary (*.DSW, *.BSW) or Shimadzu UV-Vis-NIR (*.SPC).
• 2-row subplot layout with 3:1 height ratio (main absorption spectra vs baselines).
• Export Spectra as CSV: Exports sample spectra to a CSV/ subfolder.
• Export Background as CSV: Exports background/baseline recordings.
• Reset: Clears all data and returns to the drop zone.

2. Python API

from parseuv import parse_uv, apply_style

apply_style("regular")

# Read a single Varian Cary spectrum file (.DSW)
cary_file = parse_uv("path/to/file.DSW")
print(cary_file)  # <CaryFile 'file.DSW' (DSW) | 1 spectra>

# Read a Shimadzu spectrum file (.SPC)
spc_file = parse_uv("path/to/file.spc")
print(spc_file)  # <CaryFile 'file.spc' (SPC) | 1 spectra>

spectrum = spc_file[0]
print(spectrum.title)  # 'Sample Title'
print(spectrum.num_points)  # 601
print(spectrum.start_wavelength)  # 800.0
print(spectrum.end_wavelength)  # 200.0

# Export to CSV
spc_file.to_csv("spectrum.csv")

# Read a batch spectrum file (.BSW)
cary_bsw = parse_uv("path/to/file.BSW")
df = cary_bsw.to_dataframe()
print(df.head())

# Plot all spectra automatically
cary_bsw.plot(save_path="spectra.png")

Manually Plotting a Specific Spectrum from a Multi-Spectrum File

To manually extract the X (wavelengths) and Y (absorbances) coordinates, title, and metadata for custom plotting:

import matplotlib.pyplot as plt
from parseuv import apply_style, parse_uv

# Apply regular publication plot style
apply_style("regular")

# Read a multi-spectrum file (.BSW or batch .SPC)
cary_bsw = parse_uv("path/to/file.BSW")

# Select a specific spectrum by index (e.g. 0) or by title
spectrum = cary_bsw[0]  # or cary_bsw["Spectrum Title"]

# Get X (wavelengths in nm) and Y (absorbance) arrays directly
x_wavelengths = spectrum.wavelengths  # 1D numpy array
y_absorbances = spectrum.absorbances  # 1D numpy array
title = spectrum.title
metadata = spectrum.metadata

print(f"Plotting '{title}' with {spectrum.num_points} data points.")

# Custom plot using Matplotlib
fig, ax = plt.subplots(figsize=(8, 5))
ax.plot(x_wavelengths, y_absorbances, label=title, color="#1f77b4", linewidth=2.0)
ax.set_xlabel("Wavelength (nm)", fontweight="bold")
ax.set_ylabel("Absorbance", fontweight="bold")
ax.set_title(f"Manual Plot: {title}", fontweight="bold")
ax.legend(loc="best")
plt.tight_layout()
plt.savefig("manual_spectrum_plot.png", dpi=300)
plt.show()

3. Batch Processing Tool

Process an entire folder of .BSW, .DSW, and .SPC files:

python batch_process.py path/to/folder -f png

Or interactively select format (png, pdf, svg):

python batch_process.py

Outputs are automatically organized into format-specific subfolders:

• CSV/
• PNG/ (or PDF/, SVG/)

4. Command Line Interface (CLI)

# Convert a single binary file to CSV and PNG plot
parseuv path/to/file.spc -o output.csv -p plot.png

# Run batch mode on a directory
parseuv --batch path/to/folder -f pdf

Technical Specifications

1. Varian / Agilent Cary Binary Format (.DSW, .BSW)

Supports both standard fixed-step Cary files and newer Cary WinUV version 3.00+ .DSW (single spectrum) and .BSW (batch spectra) binary files:

• Header Magic: Starts at offset 0x00 with the Pascal string Varian UV-VIS-NIR (length byte 0x11 = 17 followed by ASCII text Varian UV-VIS-NIR).
• Global Header: Offset 0x5D..0x71 contains legacy initial start wavelength (float32), end wavelength (float32), and total point count (int32).
• Text Metadata Blocks:
  • Located 256 bytes prior to each spectral data stream (data_offset - 256).
  • Contains null-terminated ASCII parameters: Sample Title, Collection Time / Date/Time stamp, Scan Software Version (e.g. Scan Software Version: 3.00(182)), Instrument (e.g. Cary 50), Start (nm) and Stop (nm), UV-Vis Scan Rate (nm/min), UV-Vis Data Interval (nm), and Baseline Correction parameters.
• Spectral Data Points Stream:
  • Sequential 8-byte little-endian IEEE float32 pairs: (wavelength_nm, absorbance).
  • Wavelength Encoding Variants:
    • Arithmetic Fixed Step: Exact arithmetic integer or fractional steps (e.g., -1.0 nm, -0.5 nm, +1.0 nm).
    • Empirical Monochromator Encoder Values (WinUV v3.00+): Hardware monochromator encoder readings per point (e.g., 799.998, 799.026, 798.053, ..., 199.968 nm) accounting for physical motor positioning tolerances.
  • Variable Sampling Intervals: Supports standard step sizes (0.05, 0.1, 0.5, 1.0, 2.0 nm) as well as fast custom sampling intervals (up to 25.0 nm interval, e.g., 5.0 nm step yielding 121 points from 800 to 200 nm).
  • Sweep Direction & Bounds: Supports decreasing (start > stop) and increasing (start < stop) wavelength sweeps between 190.0 and 1100.0 nm.

2. Shimadzu Binary Format (.SPC)

• OLE2 Compound Document Container: Microsoft OLE2 Compound Container (0xD0CF11E0A1B11AE1) generated by Shimadzu UVProbe software.
• Data Streams: Extracts streams DataSpectrumStorage/Data/X Data.N (wavelengths) and Y Data.N (absorbances) stored as 64-bit IEEE float64 (double) arrays.
• Galactic GRAMS SPC Fallback: Direct parsing of Galactic SPC headers (version 0x4B / 0x4D) reading fnpts, ffirst, flast, and float32 absorbance arrays.

Code Quality & Development

Format and lint code using ruff:

uv run ruff format .
uv run ruff check .

Run test suite:

uv run pytest

Author & License

• Author: Dr. Ricardo J. Fernández-Terán
• Contact: ricardo.fernandezteran[at]unige.ch
• License: Distributed under the BSD 3-Clause License. See LICENSE for details.

Acknowledgements

Special thanks and acknowledgement to SpectraGryph (optical spectroscopy software developed by Dr. Friedrich Menges, effemm2.de/spectragryph/) for serving as an explicit inspiration for format discovery, conversion workflows, and spectroscopy tooling design.

Copyright (c) 2026, Dr. Ricardo J. Fernández-Terán.