QUICK_START.md

PyTIA Quick Start

1) Install

pip install pytia

For local development:

pip install -e ".[dev]"

2) Create a minimal config

inputs:
  images:
    - t0.nii.gz
    - t1.nii.gz
    - t2.nii.gz
  times: [1.0, 24.0, 72.0]

time:
  unit: hours
  sort_timepoints: true

io:
  output_dir: ./output
  prefix: patient_01

physics:
  half_life_seconds: 21600.0

3) Validate, then run

pytia validate --config config.yaml
pytia run --config config.yaml

Optional:

pytia info --config config.yaml
pytia --version

4) Expected outputs

By default, io.output_dir contains:

• tia.nii.gz
• r2.nii.gz
• sigma_tia.nii.gz
• model_id.nii.gz
• status_id.nii.gz
• pytia_summary.yaml

If io.prefix is set, filenames are prefixed (for example patient_01_tia.nii.gz).

5) Python API quick example

from pytia import run_tia

result = run_tia(
    images=["t0.nii.gz", "t1.nii.gz", "t2.nii.gz"],
    times=[1.0, 24.0, 72.0],
    config="config.yaml",
)

print(result.summary["status_counts"])

6) Single-timepoint quick example

inputs:
  images: [activity_single.nii.gz]
  times: [0.0]

io:
  output_dir: ./output_stp

physics:
  half_life_seconds: 21600.0

single_time:
  enabled: true
  method: phys

Run:

pytia run --config config_stp.yaml

7) Common issues

Unknown config key(s)...

PyTIA enforces strict key validation. Check spelling and section names against CONFIG.md.

Number of times must match number of images

Ensure len(inputs.times) == len(inputs.images).

mask.provided_path is required

If mask.mode: provided, set mask.provided_path.

Unexpected TIA scale

Verify:

• time.unit matches inputs.times
• physics.half_life_seconds is in seconds
• noise_floor settings are intentional

Next

• Full usage: USER_GUIDE.md
• Complete config reference: CONFIG.md
• STP-specific details: STP_USER_GUIDE.md
• Release flow: RELEASE.md