dicom-tools

Collection of command line utilities to operate on DICOM data.

Setup

The tools are written entirely in Python (version >=3.6)

Download the repository and install required packages:

git clone "https://github.com/hirsch-lab/dicom-tools.git"
cd dicom-tools
python -m pip install -r "requirements.txt"

Build and install the package from source:

# Build package
python setup.py sdist
# Install
python -m pip install "dist/dicom-tools*.tar.gz"

This installs the below command-line tools.

dicom-copy-from-list -h
dicom-copy-headers -h
dicom-inventory -h
dicom-print-info -h

Note: it is possible to run the command-line utilities without installing the package.

# In the root directory of the project
python -m "dicom_tools.scripts.copy_from_list" -h
python -m "dicom_tools.scripts.copy_headers" -h
# ...

Functionality

The following command-line utilities are availble after installing the dicom-tools package.

dicom-copy-from-list

• Implemented in dicom_tools.scripts.copy_from_list
• Copy a subset of DICOM folders as specified in an input list.

dicom-copy-headers

• Implemented in dicom_tools.scripts.copy_headers
• Copy DICOM files without pixel data. This is useful if the image repository is too heavy.

dicom-inventory

• Implemented in dicom_tools.scripts.create_inventory.py
• Create a .csv file with summary information about all DICOM series in folder

dicom-print-info

• Implemented in dicom_tools.scripts.print_info
• Print header info of a DICOM file

Sample calls

# Copy from a list of content
dicom-copy-from-list --in-dir "path/to/dicom/repository" \
                     --out-dir "path/to/output/" \
                     --list-file "path/to/list.csv"

# Copy headers.
dicom-copy-headers --in-dir "path/to/dicom/repository" \
                   --out-dir "path/to/output/"

# Print info about a DICOM file.
dicom-print-info --in-file "path/to/dicom/file-or-folder"
dicom-print-info --in-file "path/to/dicom/file-or-folder" --all

# Create a summary of a DICOM repository.
dicom-inventory --in-dir "path/to/dicom/repository" -v

Development

Run the unitests with pytest. See also the file .coveragerc for the settings of the code coverage analysis

pytest
# To see print statements, use the -s flag
pytest -s 
# To see nicely formatted live logs, use the flag `--log-cli-level`
# Note: Don't confuse failing tests with error-logs. 
pytest --log-cli-level="info"
# To assess code coverage (requires package pytest-cov)
pytest --cov \
       --cov-config="./.coveragerc" \
       --cov-report=html

The package uses type hints. For a static type check, install mypy and run the following command:

mypy "dicom_tools" --ignore-missing-imports --allow-redefinition

Further reading

DICOM standard
DICOM dictionary browser
Pydicom (docs, api docs)
Python type checking