Command-line tools

ps_eor installs the pstool command for file-oriented data processing, power-spectrum estimation, ML-GPR, and observation simulations. These tools are independent entry points: using a simulation command does not require following the data-processing or signal-separation commands.

Installation and help

The base installation provides the data, power-spectrum, and simulation commands:

$ pip install ps-eor
$ pstool --help

Install the ML-GPR dependencies before using run_ml_gpr or run_ml_gpr_inj:

$ pip install "ps-eor[ml-gpr]"

Every command provides its exact arguments, current defaults, and available options:

$ pstool COMMAND --help

Commands are grouped here by topic:

• Data-cube commands
  • gen_vis_cube
  • even_odd_to_sum_diff
  • diff_cube
  • run_flagger
  • combine
  • combine_sph
  • vis_to_sph
• Power-spectrum command
  • make_ps
• ML-GPR commands
  • run_ml_gpr
  • run_ml_gpr_inj
• Simulation and sensitivity commands
  • simu_uv
  • simu_noise_img
  • simu_noise_ps
  • simu_noise_ps_zrange

General conventions

Input cubes use the HDF5 formats written by ps_eor. Commands that accept a configuration file use the same flagger, power-spectrum, or ML-GPR configuration objects as the Python interface.

Most output paths are created from an output directory or filename template. Commands that produce several redshift or UV selections create subordinate directories or add the selection to the filename. Existing files at the resolved output paths may be replaced, so use a dedicated output directory for a new analysis.