Skip to content

Releases: harvard-lts/fits

Release 1.6.0

16 Jan 18:41
Compare
Choose a tag to compare

This FITS release contains a lot of changes. Almost all of the tools FITS uses were upgraded, some, such as MediaInfo, significantly so. Because of this, some of the metadata coming out of this version of FITS will be different from previous versions, and hopefully it's more accurate. Please file an issue if you encounter any problems.

Please review the full change log, but there are a few points that merit special mention:

  1. FITS now requires Java 11 or later to run.
  2. The linux version of MediaInfo that's bundled with FITS is the amd64 Ubuntu build. If this build does not work on your system, please review these notes on how to install a version that works for you.
  3. FITS now packages a Dockerfile that can be used to build a FITS image, if you'd like to run FITS through Docker. You can find this file in the root of the FITS release zip, and usage instructions are available here.
  4. Two new tools were added, embARC for DPX files and jpylyzer for validating jp2 files.

Change Summary

  • Java 11 or later is now required to run FITS
  • Add embARC tool integration for processing DPX files.
  • Add jpylyzer 2.1.0 for processing jp2 files.
  • Tool upgrades
    • DROID 6.5.2
    • ExifTool 12.50
    • File 5.43
    • Jhove 1.26.1
    • MediaInfo 22.09
    • Tika 2.6.0
  • Add FITS config option consolidate-first-identity, defaulted to false. This means that metadata from
    all tools that identify a file is included, even if the identified format is not the highest ranked
    identity. This is a behavior change from previous versions. If you prefer the old behavior, set the value
    to true in your fits.xml.
  • Add CLI option -t for including raw tool output in the report.
  • Add CLI option -d for enabling debug output.
  • Fix DROID container format identification.
  • exiftool raw output now only outputs XML.
  • Fix bug where errors from one run were being incorrectly reported on subsequent runs.
  • Improve ePub file identification.
  • Improve TEI and KML file identification.
  • Fix miscategorization of some tiffs as TIFF EXIF.
  • Remove improper MIX namespace from Jhove output.
  • Improve MIX generation by cycling through all available metadata elements looking for valid values.
  • Add support for processing plaintext documents with exiftool.
  • Do not infer a timezone for exif timestamps.
  • Preserve whitespace within text elements. This ensures that significant spaces aren't trimmed from codec codes,
    but could also result in unwanted whitespace in other elements.
  • No longer run NZME on image files by default.
  • General internal dependency upgrades.

Pre-Release 1.6.0-SNAPSHOT

30 Nov 00:57
Compare
Choose a tag to compare
Pre-release

This is an interim pre-release of the upcoming FITS 1.6.0.

Release 1.5.5

10 May 21:38
Compare
Choose a tag to compare

This release does not contain any work from the Community Sprint. Its primary change from the previous release is that it resolves a log4j (v.1) security vulnerability by replacing log4j with reload4j. There has also been the upgrade to the latest version of OTS 1.0.58.

Release 1.5.4

03 May 15:49
Compare
Choose a tag to compare

This release is built from the 1.5.1 code with the substitution of the reload4j dependencies replacing the security-vulnerable log4j.

Release 1.5.3

02 May 14:58
Compare
Choose a tag to compare

Updates from log4j to log4j2 and slf4j.

Release 1.5.1

03 Jan 19:33
19ff874
Compare
Choose a tag to compare

There are no functional changes in this release.
A few minor updates of dependencies. More will follow in subsequent releases.
Due to building on MacOS 11.6, there is now more output from system-level commands resulting in increased output for the test suite. Expected output in the tests have been updated to reflect this.

Release 1.5.0

10 Sep 19:49
Compare
Choose a tag to compare

This FITS release contains several changes.

FITS code is now built with Java 1.8 and is expected to be run with a minimum of Java 1.8.

The identification of various video formats and mime types have been normalized (.mov, .mp4, .meg-4, .avi, .rm, .rmvb, .wmv) whereas previously there had been conflicts in the stanza of the FITS output.

There is a new command line flag related to recursively processing files in nested directories. Currently, when the -r flag is used and the -o flag points to a directory with one or more sub-directories, all files will be processed and the output placed into the single output directory as designated by the -o flag. With the new -n flag (in conjunction with the -r flag), the output will be placed in the configured -o flag output directory yet the output file will have the same sub-directory structure as the input directory. Any subdirectories in the output directory will be created if they do not already exist.

The embedded Tika tool has been updated to v.1.22 to address a security vulnerability. The embedded Exiftooll has been updated to v.11.54. The fast-md5 dependency has also been upgraded.

Change Summary

  • Format and mime type output for some video formats have now been normalized: .mov, .mp4, .meg-4, .avi, .rm, .rmvb, .wmv
  • Add command line flag (-n) so that, if the input (-i) is a nested directory structure, the output (-o) will be similarly nested.
  • Update Tika to v.1.22 due to security vulnerability.
  • Update Exiftool to v.11.54.
  • Improve JHOVE identification of TIFF EXIF files.
  • Improve capture and reporting of tool errors.
  • Update to latest OTS JAR file.
  • Command line errors will no longer result in Java stack traces. Only a simple error message will be output.
  • All command line errors will result in an exit code of 1. (As before, success will result with an exit code of 0.)
  • Improve MediaInfo access to OS-specific native libraries; Remove the element <fits_home>.</fits_home> from fits.xml since it is no longer necessary for MediaInfo.
  • Update fast-md5 dependency to latest version.

Release 1.4.1

10 Apr 15:34
Compare
Choose a tag to compare
  • Update Tika to v.1.19.1 due to security vulnerability.
  • Add m4a file extension to the file utility tool exclusion list to avoid a format conflict with Exiftool thus resulting in no metadata output.
  • Add m4a file and test.
  • Bug fixed: Identification of JP2000 files that are not well-formed did not go into generated metadata.

Release 1.4.0

07 Nov 17:30
4197754
Compare
Choose a tag to compare

We are pleased to announce a new release of both FITS and FITS Service web application.

The FITS release contains update to the following tools contained within FITS: DROID 6.4, Tika 1.19.1, and Exiftool 11.14. There have been updates to XSL transformations corresponding to these tools' output into FITS metadata to accommodate some (but not all) of the additional metadata provided by these tools. MP3 files are now excluded from processing by JHOVE by default. This can be modified in the fits.xml configuration file. There have also been other minor tweaks and bug fixes.

The FITS Service project -- a Java web application -- contains a few bug fixes and a new, optional request parameter to turn off returning standard metadata schema output contained within the FITS metadata element. See the FITS Service project on GitHub for more details. The significant change in this project is that it is now built using Maven rather than Ant. However, this should have no bearing on end users.

Change Summary

  • Added mp3 to Jhove file exclusion list in fits.xml; Add unit tests for mp3 files.
  • DROID updated to version 6.4; updated signature file to V94.
    • There have been two minor modifications to this signature file. See the README.txt file within the same tools/droid folder where the signature file is located.
  • Tika updated to version 1.19.1.
    • There are now Tika outputs warnings -- which have been turned off via a new Tika configuration file -- about not processing JPEG2000 files. This is due to a missing (intentionally omitted) dependency that is not compatible with the Apache 2.0 license (for both Tika and FITS). This can be added independently. See: JAI Image I/O
    • This new Tika configuration file, tika-config.xml, has been added to the directory xml/tika/. See Apache Tika Configuration to turn these warning messages back on.
  • Exiftool updated to version 11.14.
    • This version tends to identify PDF/A documents as only PDF whereas earlier versions identified as PDF/A. However, other tools properly identify PDF/A documents. As a result, in rare cases metadata identified only by Exiftool is dropped due its less specific file identification.
  • Other bug fixes.

Release 1.3.0

07 Jun 17:52
a438c51
Compare
Choose a tag to compare

This release contains essentially the same functionality as release 1.2.1. However, from a code development perspective, the build tool used to create the FITS downloadable zip file has been changed from Ant to Maven. Also, due to some licensing issues, some of the Open Source dependencies contained within FITS have been updated. It is now also possible to change the Java heap size in the scripts that start FITS. This can be configured in either fits-env.sh (for *nix) or fits.bat (for Windows). Previous versions of FITS have been removed from the downloads page and have been archived.

Note: Undocumented from a previous release, it’s possible to limit the number of bytes read by Droid tool when analyzing input file -- this may be useful for very large video and audio file. This can be configuring in the fits.xml file by uncommenting the following line, editing the file extension (separated by commas if more than one), and editing the number of bytes (in kilobytes) to be read.

<!-- <droid_read_limit include-exts="mov,mxf" read-limit-kb="64" /> -->

Change Summary

  • Allow for Java heap size to be set in the Unix (fits-env.sh) and Windows (fits.bat) shell scripts.
  • The ZIP file artifact is now built using Maven instead of Ant. The output of this file is in the target directory.