dicomweb-dump

CLI tool to dump DICOMweb WADO-RS responses to disk

Status

Beta (as of Nov 16, 2021), target is to do first release by Nov 24, 2021

Features

• Dump study metadata
• Dump series metadata
• Dump instance metadata
• TODO: Dump instance
• Dump instance frames
• Records HTTP headers, time to get response and multi-part mime headers to disk $.dump.json
• Concurrent requests for faster downloads (and stress testing servers)
• Supports requests over http2 (on by default - can be disabled)
• Stream based to minimize memory overhead and maximize concurrency
• Support for using http2 (off by default)

Prerequisites

• NodeJS (v16.13.0 or later)
• NPM

Install dependencies

npm install

Install globally

npm install -g .

Uninstall

npm uninstall -g dicomweb-dump

Usage

Usage: index.mjs [options]

Options:
      --version                Show version number                     [boolean]
  -w, --wadoRsRootUrl          WADO-RS Root Url              [string] [required]
  -s, --studyUid               The Study Instance UID to dump[string] [required]
  -o, --outputFolder           path                          [string] [required]
  -q, --quiet                  suppresses status messages to stdout    [boolean]
  -i, --include full instance  adds the full instance to the dump (DICOM P10 ins
                               tance)                                  [boolean]
  -c, --concurrency            controls maximum concurrent request count (defaul
                               t is 1)                                  [number]
  -a, --abort                  abort processing on any errors (default is false/
                               off)                                    [boolean]
  -r, --retry                  request failure retry count (default 3)  [number]
  -z, --authorization          HTTP Authorization header value          [string]
      --h2, --http2            Use HTTP2 (default is HTTP1)            [boolean]
  -h, --help                   Show help                               [boolean]

Output Structure

$outputFolder/$studyUid/metadata - study metadata
$outputFolder/$studyUid/series/<seriesuid>/metadata - series metadata
$outputFolder/$studyUid/series/<seriesuid>/instances/<sopinstanceuid>/metadata - instance metadata
$outputFolder/$studyUid/series/<seriesuid>/instances/<sopinstanceuid>/frames/<frames 1..n> - Image Frame
$outputFolder/$studyUid/series/<seriesuid>/instances/<sopinstanceuid>/_/<sopinstanceuid> - DICOM P10 instance (if -i option is passed)

Examples

Single Image CR

dicomweb-dump -w https://server.dcmjs.org/dcm4chee-arc/aets/DCM4CHEE/rs/studies -s 1.3.6.1.4.1.25403.345050719074.3824.20170126085406.1 -o output -i -m

Large CT (Note - some images in this study return errors)

dicomweb-dump -w https://server.dcmjs.org/dcm4chee-arc/aets/DCM4CHEE/rs/studies -s 1.3.6.1.4.1.14519.5.2.1.7009.2403.129940714907926843330943219641 -o output -s

Each output file will have a corresponding file with the prefix .dump.json which includes details such as time to complete request, HTTP headers returned and multi-part mime header

TODO

Features

• Dump bulkdata
• Add option to just dump instance metadata/frames/bulk data
• Enhance getAttachments() to return more info from the multi-part mime header
• Add option to use qido-rs study series and series instances to get UIDs
• Add option to dump qido-rs study series and series instance responses
• Report performance/timing information
• Add option to pretty print JSON?
• Write to disk with multi-part mime headers removed in $.bin

Reuse

• Consider separating the fetch logic into a separate library so it can be re-used?

Optimizations

• See if getAttachments() can be optimized - just use Buffer instead of Uint8Array?