v1.0.0 - First Public Release

The sigima library is part of the DataLab open-source platform.
See DataLab roadmap page for future and past milestones.

sigima 1.0.0

💥 New features and enhancements:

• Signals to image conversion: New feature to combine multiple signals into a 2D image

  • New computation function signals_to_image() in sigima.proc.signal.arithmetic
  • Takes a list of signals and combines them into an image by stacking Y-arrays
  • Two orientation modes:
    • Rows: Each signal becomes a row in the image (default)
    • Columns: Each signal becomes a column in the image
  • Optional normalization:
    • Supports multiple normalization methods (Z-score, Min-Max, Maximum)
    • Normalizes each signal independently before stacking
    • Useful for visualizing signals with different amplitude ranges
  • Validates that all signals have the same size before combining
  • New parameter class SignalsToImageParam with orientation and normalization settings
  • New enum SignalsToImageOrientation for specifying row/column orientation
  • Comprehensive validation tests for all combinations of parameters
  • Ideal for creating spectrograms, heatmaps, or waterfall displays from signal collections

• Non-uniform coordinate support for images: Added comprehensive support for non-uniform pixel coordinates

  • ImageObj now supports both uniform and non-uniform coordinate systems:
    • Uniform coordinates: defined by origin (x0, y0) and pixel spacing (dx, dy)
    • Non-uniform coordinates: defined by coordinate arrays (xcoords, ycoords)
  • New methods for coordinate manipulation:
    • set_coords(): Set non-uniform X and Y coordinate arrays
    • is_uniform_coords: Property to check if coordinates are uniform
  • New computation function set_uniform_coords(): Convert non-uniform to uniform coordinates
    • Automatically extracts uniform spacing from non-uniform arrays
    • Handles numerical precision issues from linspace-generated arrays
    • Preserves image data while transforming coordinate system
  • Enhanced calibration() function with polynomial support:
    • Now supports polynomial calibration up to cubic order: dst = a0 + a1*src + a2*src² + a3*src³
    • Parameter class changed from a, b (linear) to a0, a1, a2, a3 (polynomial)
    • Works on X-axis, Y-axis (creating non-uniform coordinates), and Z-axis (data values)
    • Linear calibration is a special case with a2=0, a3=0
    • Automatically handles conversion between uniform and non-uniform coordinate systems
  • Enhanced I/O support:
    • HDF5 format now serializes/deserializes non-uniform coordinates
    • Coordinated text files support non-uniform coordinate arrays
  • All geometric operations updated to handle both coordinate types:
    • Coordinate transformations preserve or create appropriate coordinate system
    • ROI operations work seamlessly with both uniform and non-uniform coordinates

• DateTime support for signal data: Added comprehensive datetime handling for signal X-axis data

  • Automatic detection and conversion of datetime columns when reading CSV files
    • Detects datetime values in the first or second column (handling index columns)
    • Validates datetime format and ensures reasonable date ranges (post-1900)
    • Converts datetime strings to float timestamps for efficient computation
    • Preserves datetime metadata for proper display and export
  • New SignalObj methods for datetime manipulation:
    • set_x_from_datetime(): Convert datetime objects/strings to signal X data with configurable time units (s, ms, μs, ns, min, h)
    • get_x_as_datetime(): Retrieve X values as datetime objects for display or export
    • is_x_datetime(): Check if signal contains datetime data
  • Enhanced CSV export to preserve datetime format when writing signals with datetime X-axis
  • New constants module (sigima.objects.signal.constants) defining datetime metadata keys and time unit conversion factors
  • Comprehensive unit tests covering datetime conversion, I/O roundtrip, and edge cases
  • Example test data file with real-world temperature/humidity logger data (datetime.txt)

• New client subpackage: Migrated DataLab client functionality to sigima.client

  • Added sigima.client.remote.SimpleRemoteProxy for XML-RPC communication with DataLab
  • Added sigima.client.base.SimpleBaseProxy as abstract base class for DataLab proxies
  • Included comprehensive unit tests and API documentation
  • Maintains headless design principle (GUI components excluded)
  • Enables remote control of DataLab application from Python scripts and Jupyter notebooks
  • Client functionality is now directly accessible: from sigima import SimpleRemoteProxy

• New image ROI feature: Added inverse ROI functionality for image ROIs

  • Added inside parameter to BaseSingleImageROI base class, inherited by all image ROI types (PolygonalROI, RectangularROI, CircularROI)
  • When inside=True, ROI represents the region inside the shape (inverted behavior)
  • When inside=False (default), ROI represents the region outside the shape (original behavior)
  • Fully integrated with serialization (to_dict/from_dict) and parameter conversion (to_param/from_param)
  • Signal ROIs (SegmentROI) are unaffected as the concept doesn't apply to 1D intervals
  • Optimal architecture with zero code duplication - all inside functionality implemented once in the base class
  • Individual ROI classes no longer need custom constructors, inheriting directly from base class

• New image operation:

  • Convolution.

• New image format support:

  • Coordinated text image files: Added support for reading coordinated text files (.txt extension), similar to the Matris image format.
    • Supports both real and complex-valued image data with optional error images.
    • Automatically handles NaN values in the data.
    • Reads metadata including units (X, Y, Z) and labels from file headers.

• New image analysis features:

  • Horizontal and vertical projections
    • Compute the horizontal projection profile by summing values along the y-axis (sigima.proc.image.measurement.horizontal_projection).
    • Compute the vertical projection profile by summing values along the x-axis (sigima.proc.image.measurement.vertical_projection).

• New curve fitting algorithms: Complete curve fitting framework with sigima.tools.signal.fitting module:

  • Core fitting functions: Comprehensive set of curve fitting algorithms for scientific data analysis:
    • linear_fit: Linear regression fitting
    • polynomial_fit: Polynomial fitting with configurable degree
    • gaussian_fit: Gaussian profile fitting for peak analysis
    • lorentzian_fit: Lorentzian profile fitting for spectroscopy
    • voigt_fit: Voigt profile fitting (convolution of Gaussian and Lorentzian profiles)
    • exponential_fit: Single exponential fitting with overflow protection
    • piecewiseexponential_fit: Piecewise exponential (raise-decay) fitting with advanced parameter estimation
    • planckian_fit: Planckian (blackbody radiation) fitting with correct physics implementation
    • twohalfgaussian_fit: Asymmetric peak fitting with separate left/right parameters
    • multilorentzian_fit: Multi-peak Lorentzian fitting for complex spectra
    • sinusoidal_fit: Sinusoidal fitting with FFT-based frequency estimation
    • cdf_fit: Cumulative Distribution Function fitting using error function
    • sigmoid_fit: Sigmoid (logistic) function fitting for S-shaped curves
  • Advanced piecewise exponential (raise-decay) fitting: Enhanced algorithm with:
    • Standard piecewise exponential model: y = a_left*exp(b_left*x) + a_right*exp(b_right*x) + y0
    • Multi-start optimization strategy for robust convergence to global minimum
    • Support for both positive and negative exponential rates (growth and decay components)
    • Comprehensive parameter bounds validation to prevent optimization errors
  • Enhanced asymmetric peak fitting: Advanced twohalfgaussian_fit with:
    • Separate baseline offsets for left and right sides (y0_left, y0_right)
    • Independent amplitude parameters (amp_left, amp_right) for better asymmetric modeling
    • Robust baseline estimation using percentile-based methods
  • Technical features: All fitting functions include:
    • Automatic initial parameter estimation from data characteristics
    • Proper bounds enforcement ensuring optimization stability
    • Comprehensive error handling and parameter validation
    • Consistent dataclass-based parameter structures
    • Full test coverage with synthetic and experimental data validation

• New common signal/image feature:

  • Added phase (argument) feature to extract the phase information from complex signals or images.
  • Added operation to create complex-valued signal/image from real and imaginary parts.
  • Added operation to create complex-valued signal/image from magnitude and phase.
  • Standard deviation of the selected signals or images (this complements the "Average" feature).
  • Generate new signal or image: Poisson noise.
  • Add noise to the selected signals or images.
    • Gaussian, Poisson or uniform noise can be added.
  • New utility functions to generate file basenames.
  • Deconvolution in the frequency domain.

• New ROI features:

  • Improved single ROI title handling, using default title based on the index of the ROI when no title is provided.
  • Added combine_with method to ROI objects (SignalROI and ImageROI) to return a new ROI that combines the current ROI with another one (union) and handling duplicate ROIs.
  • Image ROI transformations:
    • Before this change, image...

v0.2.0 - Preliminary release

Bump version to 0.2.0 and update development status to Beta

v0.1.0 - Preliminary release

sigima 0.1.0

This first version of the library is the result of the externalization of the signal and image processing features from the DataLab main repository.