This package provides utilities for loading and analyzing MITgcm data assimilation files, primarily costfunction*, adxx, and xx files. xmitgcm doens't handle these automatically. this uses the MITgcm readmds Python funcitonality. This code depends on xarray, numpy, MITgcmutils, and xmitgcm. If you provide the ctrls functions a grid directory from an MITgcm model run, then xmitgcm will load the grid and use that to put the adxx and xx files on the grid in a DataArray. If you don't give it a grid directory then the coords will be kept as time, i, j, k.

Install the package in development mode:

cd mitgcm-assim
pip install -e .

import mitgcm_assim.cost as cost

# Read all costfunction* files in current directory
ds = cost.read_costfunction('.')

# Access total cost for all iterations
fc_cost = ds.cost.sel(dataset='fc')
print(fc_cost.values)

# Access specific dataset costs
tao_u_cost = ds.cost.sel(dataset='TAO_WO_2012_ADCP_prof_U')
print(f"U-velocity cost: {tao_u_cost.values}")

import mitgcm_assim.ctrls as ctrls

# Load all default variables
ds = ctrls.read_controls_and_sensitivities('./data', './grid', 0)

# Load specific variables only
ds = ctrls.read_controls_and_sensitivities(
    './data', './grid', 0,
    control_vars=['xx_theta', 'xx_salt'],
    sensitivity_vars=['adxx_theta', 'adxx_salt']
)

# Load a single control variable
da = ctrls.read_single_control('./data', 'xx_theta', 0)

# Check what variables are available
available = ctrls.get_available_controls('./data', 0)
print("Available controls:", available['controls'])

Functions for reading MITgcm cost function files into xarray Datasets:

• read_costfunction(directory): Read all costfunction* files in a directory
• parse_costfunction_file(filepath): Parse a single costfunction file

Functions for reading MITgcm control variables and sensitivities:

• read_controls_and_sensitivities(): Load control variables and sensitivities with grid coordinates
• read_single_control(): Load a single control variable
• get_available_controls(): List available control files in a directory

• numpy
• xarray
• MITgcmutils
• xmitgcm