[BUG] Missing D_and_Ds_swap() post-fit correction in 13 out of 22 Python algorithms

[Devguru-codes]

Is there an existing issue for this?

• I have searched the existing issues

Describe the bug

The D_and_Ds_swap() method in OsipiBase corrects mislabelled D and D* (Dp) parameters after fitting. When the optimizer assigns them to the wrong parameter (i.e., D > Dp and Dp < 0.05), it swaps them and adjusts f = 1 - f.

Currently, only 9 out of 22 non-DL/non-MATLAB algorithms call this method after fitting. The remaining 13 algorithms skip this correction, leading to inconsistent parameter labelling across algorithms when the optimizer produces swapped results.

Screenshots [optional]

No response

Steps To Reproduce

import numpy as np
import os, sys
sys.path.insert(0, os.getcwd())

# Check source code for the presence of D_and_Ds_swap
algo_dir = os.path.join("src", "standardized")

algorithms = [
    "IAR_LU_biexp", "IAR_LU_segmented_2step", "IAR_LU_segmented_3step",
    "IAR_LU_subtracted", "IAR_LU_modified_mix", "IAR_LU_modified_topopro",
    "PV_MUMC_biexp",
    "OGC_AmsterdamUMC_biexp", "OGC_AmsterdamUMC_biexp_segmented", "OGC_AmsterdamUMC_Bayesian_biexp",
    "OJ_GU_seg", "PvH_KB_NKI_IVIMfit", "TF_reference_IVIMfit",
    "TCML_TechnionIIT_lsqlm", "TCML_TechnionIIT_lsqtrf", "TCML_TechnionIIT_lsqBOBYQA",
    "TCML_TechnionIIT_SLS", "TCML_TechnionIIT_lsq_sls_lm",
    "TCML_TechnionIIT_lsq_sls_trf", "TCML_TechnionIIT_lsq_sls_BOBYQA",
    "ETP_SRI_LinearFitting", "ASD_MemorialSloanKettering_QAMPER_IVIM",
]

for algo in sorted(algorithms):
    fpath = os.path.join(algo_dir, f"{algo}.py")
    if os.path.exists(fpath):
        with open(fpath) as f:
            content = f.read()
        has_it = "D_and_Ds_swap" in content
        tag = "HAS" if has_it else "MISSING"
        print(f"  [{tag:7s}] {algo}")

Expected behavior

All algorithms should consistently apply D_and_Ds_swap() after fitting to ensure correct D/D* labelling.

Additional context

Output on main branch:

  [HAS    ] ASD_MemorialSloanKettering_QAMPER_IVIM
  [HAS    ] ETP_SRI_LinearFitting
  [HAS    ] IAR_LU_biexp
  [HAS    ] IAR_LU_modified_mix
  [HAS    ] IAR_LU_modified_topopro
  [HAS    ] IAR_LU_segmented_2step
  [MISSING] IAR_LU_segmented_3step
  [MISSING] IAR_LU_subtracted
  [MISSING] OGC_AmsterdamUMC_Bayesian_biexp
  [MISSING] OGC_AmsterdamUMC_biexp
  [MISSING] OGC_AmsterdamUMC_biexp_segmented
  [MISSING] OJ_GU_seg
  [MISSING] PV_MUMC_biexp
  [MISSING] PvH_KB_NKI_IVIMfit
  [HAS    ] TCML_TechnionIIT_lsqBOBYQA
  [HAS    ] TCML_TechnionIIT_lsqlm
  [MISSING] TCML_TechnionIIT_lsq_sls_BOBYQA
  [MISSING] TCML_TechnionIIT_lsq_sls_lm
  [MISSING] TCML_TechnionIIT_lsq_sls_trf
  [HAS    ] TCML_TechnionIIT_lsqtrf
  [MISSING] TCML_TechnionIIT_SLS
  [MISSING] TF_reference_IVIMfit

13 algorithms are MISSING the call.

Suggested Fix

Add results = self.D_and_Ds_swap(results) just before return results in the ivim_fit() method of each of the 13 missing algorithms.

Are you working on this?

Yes

[Devguru-codes]

Please review the fix, I am currently working on this (adding the suggested line). Will make pr soon