NBIA Data Retriever version 4.4 provides the option to save the error report and then retry.

1. Starting from version 4.3, any version which is below a minimum version or has a unknown version will be prompted for upgrade 2. Resume downloading feature is added to the command line interface (CLI) of the app 3. The policy window is set to read-only

1. Added a new option under the File menu allowing the MD5 checksum of the downloaded files to be validated against the server’s checksum information.  2. Automatically compares the size of downloaded file with the server’s size information.  3. Fixed the bug that shows incorrect information in the Data Description URI column of metadata.csv. 4. Fixed the bug that shows the Data Usage Policy in the LICENSE file twice.

1. You can now resume a download if you closed the NBIA Data Retriever in the middle of a download, as long as you choose the same download directory as you originally did before the interruption. You can specify whether you want to download just the missing series or download all of the series in the original manifest file. 2. You must now accept the TCIA Data Usage Policy before proceeding to download data. This policy is now also included in the license file. 3. The NBIA Data Retriever now warns you if there are series in your download that have commercial use restrictions. 4. More technical features in this release include replacing Oracle Java 8 with Amazon-Corretto as the embedded Java runtime environment and several bug fixes.

1. An overall progress bar is at the top of the Downloads table. 2. Under the directory the user selects for downloaded files, a new directory has been added and it has the same name as the manifest file. 3. A metadata file in CSV format, which provides an overview of the downloaded files, is now under the new directory mentioned above. This file includes the following columns: Series Instance UID, Collection, 3rd Party Analysis, Data Description URI, Subject ID, Study UID, Study Description, Study Date, Series Description, Modality, Manufacturer, SOP Class Name, SOP Class UID, Number of Images, File size, File Location, and Download Time Stamp. 4. If the license information for a collection is present, the license file has been added to the collection level of the directory. 5. Limitation for restricted series not to exceed 10,000 series is removed. Please see details on https://wiki.cancerimagingarchive.net/display/NBIA/What%27s+New+in+NBIA+Data+Retriever+4.0

The file naming system used by the NBIA Data Retriever has been revised to provide a more natural ordering of the files within each series. This has been accomplished by first ordering the DICOM files by ordinal position of acquisition number and then by ordinal position of instance number. The files are then assigned numbers with the lowest acquisition being 1 and the lowest instance number within that acquisition being 1, separated by a dash. The numbers are incremented by 1 as the next values are encountered. All values are left-padded with zeros to provide for the ordering within the file system. As an example, a series with 2 acquisition numbers with each acquisition having 42 instance numbers would start with 1-01.dcm for the lowest acquisition number and the lowest instance number in that acquisition. The last file in that acquisition would be 1-42.dcm. Then the second acquisition would start with 2-01.dcm, ending with 2-42.dcm, in that acquisition. For the purposes of ordering, an empty value in either acquisition number or instance number is lower than an file having a value for those number.

Removed download restriction of 9999 series in a manifest file per run.

* Use Long instead of Integer to hold file size internally. * Add a video tutorial link to the Error Dialog. * Upgrade Java run time environment to 1.8.0_211. * Add HttpRequestRetryHandler for automatically reconnect to download server for maximum five times. * Fix unknown node issue.

* Use Long instead of Integer to hold file size internally. * Add a video tutorial link to the Error Dialog. * Upgraded Java run time environment to 1.8.0_211. * Fix unknown node issue.

* Use Long instead of Integer to hold file size internally. * Add a video tutorial link to the Error Dialog. * Fix unknown node issue.

* Use Long instead of Integer to hold file size internally. * Add a video tutorial link to the Error Dialog. * Fix unknown node issue.

* Use Long instead of Integer to hold file size internally. * Add a video tutorial link to the Error Dialog. * Fix unknown node issue.

* Use Long instead of Integer to hold file size internally. * Add a video tutorial link to the Error Dialog. * Fix unknown node issue.

* Use Long instead of Integer to hold file size internally. * Add a video tutorial link to the Error Dialog.

* Add an informative dialog to remind the user the app needs to work with a manifest file. * Load the help desk info and the user documentation link from the server. So one version of NBIA Data Retriever can be used with different NBIA instances, TCIA included.

The latest version of the app provides additional options for the handling downloads with errors. When a series does not download completely after multiple automatic retries (four, by default), whether due to data or network connectivity issues, the app reports an error. You can now either create an error report that provides information you can send to the help desk, or download only the failed series again.