Skip to main content

Documentation Index

Fetch the complete documentation index at: https://mintlify.com/Rubick65/dcemapper/llms.txt

Use this file to discover all available pages before exploring further.

The src.io module provides two submodules for reading MRI data into DCEMapper: nifti_io handles loading NIfTI files directly into NumPy arrays, and bruker_conversion automates batch conversion of raw Bruker scanner output into BIDS-compatible NIfTI files. Together they form the entry point for all data ingestion in the pipeline.

​
nifti_io

​
load_nifti(path)

Loads a NIfTI image from disk and returns its voxel data as a NumPy array alongside the full NiBabel image object. Internally, load_nifti first validates the path via is_valid_nifti, then calls nibabel.load and extracts the floating-point data array with get_fdata().
​
path
string
required
Filesystem path to a NIfTI file. Both uncompressed (.nii) and gzip-compressed (.nii.gz) formats are supported.
Returns
NameTypeDescription
datanumpy.ndarrayVoxel intensity array. Shape is (X, Y, Z) for a 3D volume or (X, Y, Z, T) for a 4D time series.
imgnibabel.Nifti1ImageFull NiBabel image object. Provides .affine (4×4 voxel-to-world matrix) and .header.
from src.io.nifti_io import load_nifti

data, img = load_nifti("path/to/dce_data.nii.gz")
print(data.shape)   # e.g. (128, 128, 20, 50) for a 4D DCE volume
print(img.affine)   # 4x4 affine transformation matrix

​
get_nifti_slices(data, current_t=0)

Extracts all axial slices from a 4D NIfTI array at a specified time frame, returning them as a list of transposed 2D arrays ready for display.
​
data
numpy.ndarray
required
4D array of shape (X, Y, Z, T) as returned by load_nifti.
​
current_t
int
default:"0"
Time frame index to extract. Defaults to 0 (the first frame).
Returns A list of 2D numpy.ndarray objects, one per Z slice. Each element is data[:, :, z, current_t].T, so the orientation matches screen coordinates directly. 
from src.io.nifti_io import load_nifti, get_nifti_slices

data, img = load_nifti("dce_data.nii.gz")
slices = get_nifti_slices(data, current_t=5)
print(len(slices))   # number of slices (Z dimension)

​
bruker_conversion

​
convert_studies_from_bruker(input_dir, output_dir, skip_existing=True)

Scans a directory for Bruker studies (raw subject folders or ZIP archives), identifies each scan’s modality, and converts them to NIfTI format organised in a BIDS-like folder structure under output_dir/sourcedata/. For each study, the function:
  1. Loads the Bruker dataset with brkraw.
  2. Extracts SubjectID and SessionID to build a BIDS-compliant subject/session label (e.g. sub-B001_ses-01).
  3. Iterates over all available scans and reconstructions, skipping derived ISA maps.
  4. Detects the modality (DCE, DWI, T1w, etc.) from sequence parameters.
  5. Saves a .nii.gz file and a companion .json sidecar with BIDS metadata. For DWI scans, .bvec and .bval files are also written.
  6. Writes a conversion_report.log to output_dir if any errors occurred.
​
input_dir
string
required
Path to the directory containing Bruker study folders or ZIP files. The function searches recursively for any path containing a subject file or a valid .zip archive.
​
output_dir
string
required
Destination root directory. Converted files are written under output_dir/sourcedata/sub-XXX_ses-YYY/<category>/.
​
skip_existing
boolean
default:"true"
When True, any subject/session whose target directory already exists is silently skipped. Set to False to force reconversion of previously processed studies.
Returns None. All output is written to disk. A conversion_report.log file is created in output_dir only when one or more scans fail to convert. 
from src.io.bruker_conversion import convert_studies_from_bruker

convert_studies_from_bruker(
    input_dir="/data/bruker_studies",
    output_dir="/data/nifti_output",
    skip_existing=True
)
Time-domain fields (RepetitionTime, EchoTime, InversionTime) are automatically converted from milliseconds (Bruker convention) to seconds (BIDS standard) in the JSON sidecar.

Build docs developers (and LLMs) love

Get started for freeTalk to us