GetDist

GetDist can read parameter sample/chain files in a specific plain text format. A set of chains is identified by a root name xxx, requiring the following files in the same directory:

1. xxx_1.txt, xxx_2.txt, etc.: The actual chain files. Each row follows the format: weight like param1 param2 param3 ...
   • weight: Number of samples or importance weight.
   • like: -log(posterior).
   • paramN: Parameter values.
2. xxx.paramnames: A file listing parameter names (one per line).
   • Supports LaTeX labels: x1 x_1.
   • Derived parameters are indicated by a trailing asterisk: xy* x_1+y_1.
3. xxx.ranges: Defines hard bounds for parameters: param_name lower_limit upper_limit. Use N for unbounded limits (e.g., x2 0 N).
4. xxx.properties.ini (Optional): Allows setting burn_removed=T or ignore_rows=x (where x is a fraction 0 < x < 1 or a fixed number of rows if x > 1).

The GUI allows you to visualize cosmological parameter chains and grids (including paramgrid and Cobaya grids).

Workflow

1. Open Directory: Open a folder containing chain files. If you downloaded a grid from the Planck Legacy Archive, unzip it and open the main directory.
2. Select Chains: Select the chain root names you want to plot. They will appear in a list box. You can drag and drop them to change the plotting order.
3. Select Parameters: Use the checkboxes to select parameter names (these correspond to names available in the first selected chain).
4. Plotting Types: The GUI supports 1D, 2D (line and filled), 3D (using 'color by' for a second parameter), triangle, and rectangle plots.

Exporting Plots

• Quick Export: Use the File menu to export a plot as-is to PDF or other image formats.
• High Quality/Customization: Use the Script Preview tab to view the Python script generated for the current plot. You can export this script, edit it (e.g., to add custom legend text), and run it separately for fine control and better resolution.

To generate plots from samples, you must first create a plotter instance using either get_single_plotter or get_subplot_plotter. Once the instance is created, you can use its methods to generate and export specific plots.

Plotter functions accept a roots argument, which can be:

• A root name for chain files.
• An in-memory MCSamples instance.
• A list of either of the above (used for making comparison plots).

Parameters are referenced by their names (as defined in the .paramnames file or the MCSamples instance). For functions requiring lists of parameters, you can provide lists of names or use glob patterns to match subsets (e.g., x* to match all parameters starting with 'x').

# Example conceptual usage
from getdist.plots import get_single_plotter

# Create a plotter for a single set of samples
plotter = get_single_plotter(root_name_or_mcsamples)

# Use the plotter to make plots (specific methods depend on the plotter instance)
# plotter.plot_contour(...)