Classification Toolkit

The main documentation for the classification-toolkit can be found at the fw-classification Documentation

Classification toolkit is a library that can act on arbitrary metadata in the form of JSON, with a declarative set of rules. The main use case for this library is for classification of medical imaging files based on their metadata, for example, the SeriesDescription tag in a DICOM file.

Install

poetry add fw-classification, or with optional dependencies poetry add fw-classification -E all.

Run classification via cli

~/ ➜  classify run ./fw_classification/profiles/main.yml ~/Downloads/example.dicom.zip -a dicom
fw_classification:      Block mr_file result: True
fw_classification:      Block set_modality result: True
fw_classification:      Block set_classification_from_acquisition_label_or_series_description result: True
fw_classification:      Block set_contrast_from_acquisition_label_or_series_description result: False
fw_classification:      Block add_features_from_acquisition_label_or_series_description result: False
fw_classification:      Block add_measurement_from_acquisition_label_or_series_description result: False
fw_classification:      Block add_intent_from_acquisition_label_or_series_description result: False
{
  "file": {
    "info": {
      "header": {
        "dicom": {
          "FileMetaInformationGroupLength": 216,
          "WindowWidth": 199.0,
          ... # Many more tags
          "dBdt": 0.0
        }
      }
    },
    "type": "dicom",
    "modality": "MR",
    "classification": {
      "Intent": [
        "Structural"
      ],
      "Measurement": [
        "T1"
      ]
    }
  }
}

Run via python

from fw_classification.classify import run_classification
import json

input_data = {}
with open('/path/to/input','r') as fp:
   input_data = json.load(fp)

result, out_data = run_classification('MR_classifier', input_data)

Out:

~/ ➜   python test.py
{
  "file": {
    "info": {
      "header": {
        "dicom": {
          "FileMetaInformationGroupLength": 216,
          ... # Many more tags
          "WindowWidth": 199.0,
          "dBdt": 0.0
        }
      }
    },
    "type": "dicom",
    "modality": "MR",
    "classification": {
      "Intent": [
        "Structural"
      ],
      "Measurement": [
        "T1"
      ]
    }
  }
}

Development

Install the project using poetry (requires poetry > 1.2.0b1), enable plugin and pre-commit:

poetry install --extras "all"
pre-commit install

Release files for fw-classification 0.6.3

For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.

Built distribution (wheel)

Table of built distributions (wheels) for fw-classification 0.6.3
File	Interpreter	ABI	Platform	Reset
fw_classification-0.6.3-py3-none-any.whl 31.0 kB Sep 29, 2026	Python 3	none	any	Details

Release files / fw_classification-0.6.3-py3-none-any.whl

Download URL	fw_classification-0.6.3-py3-none-any.whl
Size	31.0 kB
Tags	Python 3
SHA-256 checksum How to use checksums	`a8d5f9fe3909339c0d9355dc8f8914d4970536466fe4efdd31c3a74bbc68868b`
BLAKE2b-256 checksum How to use checksums	`90eeb9977a82b3a4769f1fb7a48999b665dc56bfaa4408b1354f6447d1396359`
Upload date	Sep 29, 2026
Uploaded using Trusted Publishing? What is trusted publishing?	No
Uploaded via	`poetry/2.4.3 CPython/3.12.14 Linux/5.15.154+`