This guide, designed for a Web Curator Tool developer and system administrator, covers the release notes from version 1.5. Versions are in reverse chronological order, with the most recent version first. While the Web Curator Tool System Administrator Guide, Web Curator Tool Developer Guide, Web Curator Tool Quick Start Guide and Web Curator Tool User Manual are accurate for the current release, the Release Notes can give some idea of how things have changed since the last major release.
Contents of this document¶
Following this introduction, the Web Curator Tool Release Notes includes the following sections:
- Changes since 3.1.0 - Changes since the last official release 3.0.3.
- 3.1.0 - Release 3.1.0.
- 3.0.3 - Release 3.0.3.
- 3.0.2 - Release 3.0.2.
- 3.0.1 - Release 3.0.1.
- 3.0.0 - Release 3.0.0.
- 2.0.2 - Release 2.0.2.
- 2.0.1 - Release 2.0.1.
- 2.0.0 - Release 2.0.0.
- 1.6.2 - Release 1.6.2.
- 1.6.1 - Release 1.6.1.
- 1.6.0 - Release 1.6.0.
- 1.5.2 - Release 1.5.2.
- 1.5.1 - Release 1.5.1
- 1.5 - Release 1.5.
- Previous versions - Versions prior to release 1.5.
Changes since 3.1.0¶
This is a placeholder for changes since the official 3.1.0 release. Please add notes here for changes and fixes as they are released into the master branch.
Released Sept 2021, v3.1.0 includes the following changes:
New Harvest Visualization and Bulk Patching
A new network visualization of the harvested domains in a crawl, combined with improved tree views for deeper analysis of the captured URLs.
Basic harvest patching improvements, as well as the addition of bulk importing and pruning.
Removal of previous Tree View area.
Removal of large redundant database tables,
Retiring of H1 Harvest Agent.
Removal of redundant bandwidth functionality.
Released May 2021, v3.0.3 includes the following bug fixes and changes:
- Documentation typos fixed.
- H3 Harvest Agent heartbeats fail when started before Webapp
- ‘Too many files open’ error when Store is transferring files
Released April 2021, v3.0.2 includes the following bug fixes:
- Duplicate crawl running on multiple harvest agents
- Occasional failing to archive using Submit-to-Rosetta plugin
- Escape special characters in Target Annotation field
- Old bandwidth restrictions skipped for H3
- Better handling of heartbeat failed issues
- Extra profile validation for blocked and included URL regexs
- Hard coding of version number to login page
- Fix for slow deletion of harvest resources in MySQL/MariaDB
Released November 2020, v3.0.0 includes a major technical uplift of WCT. This version addresses technical debt and uplifting of several core frameworks, libraries and dependencies used by WCT. Other significant changes include the migration to Spring Boot, and the restructure of some WCT components.
- Spring 1.2.7 -> Spring 5.1.x
- Acegi Security -> Spring Security
- Hibernate 3.1.3 -> Hibernate 5.x.x with JPA
- Quart scheduler 1.5.2 -> Quartz scheduler 2.3.1
- Struts 1.2.8 / Tiles 1 -> Spring MVC 5.1.x / Tiles 3
- Apache Tomcat 8.5/9 -> Spring Boot 2.1
- Maven 3.5 -> Gradle 5.x
Major changes include upgrading to the latest Spring framework version, migrating to Spring Security, and from Struts to Spring MVC. All the web application controller classes were upgraded in some capacity to align with modern Spring coding conventions.
Migration to Spring Boot. WCT components are now run individually with their own embedded Tomcat server. Note, this has implications for networking and firewall rules, as WCT is no longer running behind one Apache Tomcat instance, and one port.
WCT configration has shifted to an application.properties file within each component. Local Spring profiles can also be utilized for overriding default configurations and providing more flexibilty for environment specific setups.
WCT build technology has been switched to Gradle. Note, Maven is still reqiuired for installing legacy dependencies.
The WCT codebase has been tested and refined to be OpenJDK 8.x compliant. Note, while WCT has been tested successfully with Java 11, it is not yet officially supported.
The WCT codebase has been refactored to separate the web application and the shared core logic into different modules. Previously, WCT-Core was the web application, and also included as a dependency within the other modules (WCT-Store, Harvest Agents) due to it’s common WCT data models and interfaces. This has now been separated into webcurator-webapp and webcurator-core.
- The SOAP implementation has changed. As part of that change, the ex-libris Rosetta SDK dependency has moved from 2.2.0 to 5.5.0. This means that the dps-sdk-5.5.0.jar must be installed in a local Maven repository for the maven build to work. This jar is now sourced from the github project rosetta-dps-sdk-projects-maven-lib, found at (https://github.com/NLNZDigitalPreservation/rosetta-dps-sdk-projects-maven-lib). The installation of this jar includes a pom with its maven dependencies so the wct-store and wct-submit-to-rosetta subprojects no longer need to explicitly include or track the dps-sdk dependencies in their project poms. It has a different groupId and artifactId from the dps-sdk-5.5.0.jar downloaded from Rosetta.dps-sdk-projects (https://github.com/ExLibrisGroup/Rosetta.dps-sdk-projects). This dependnecy is installed into the local maven repository by running the script install_maven_dependencies.[sh|bat].
- Because of some classpath issues, harvest-agent-h1 now uses a modified version of heritrix that has been created with the github project https://github.com/WebCuratorTool/heritrix-1-14-adjust. This version of heritrix and its necessary dependencies are installed into the local maven repository by running the script install_maven_dependencies.[sh|bat]. Note that this script now requires that the program git works from the command line.
Released December 2018, this version builds on release 1.7.0, which was a proof-of-concept integrating Heritrix 3 with WCT. Version 2.0.0 completes that integration.
Heritrix 3 profile management¶
- The configuration options available for Heritrix 3 are different from the old Heritrix 1 profiles, but management of them stays the same.
- Heritrix 3 profile options are contained within a single simplified ‘scope’ tab. This relies on a correctly formatted set of fields within the background profile xml. Due to this, imported Heritrix 3 profiles cannot be edited through the same screen, and are only editable via an in-screen xml editor.
- Validation of Heritrix 3 profiles is achieved using an available H3 Harvest Agent. The profile is used to build a special one-off job within the agent, which in essence validates the integrity of the Heritrix 3 profile. The job is then destroyed and any unsuccessful outcome is fed back to the WCT user interface.
- Heritrix 3 Targets can now be scheduled, and will be assigned to an available H3 Harvest Agent when due to run.
- Running Heritrix 3 Target Instances have an H3 script console available to use. This console can be used to run scripts against the Target Instance job in Heritrix 3, similar to the scripting console available in H3’s own UI.
- Heritrix 1 integration has been preserved for now, allowing for Targets to transition to using Heritrix 3. A period of experimentation is expected when replacing the old Heritrix 1 profiles.
- The sql scripts for setting up the WCT database have been consolidated and brought up to date. The folder structure has been refactored and legacy scripts separated to reduce confusion. Any script changes have been reflected in the documentation.
- An additional parent script has been added to simplify the setup process, enabling the setup to be completed through running a single script.
- The documentation has been migrated from PDF to the reStructedText format, and now hosted on the readthedocs.io platform. This increases the accessibility of the documentation and makes it simpler to maintain and update.
- All documentation has been brought up-to-date to reflect v2.0.0 changes.
- The old Harvest Agent module has been separated into a Heritrix 1 and Heritrix 3 version. This has been done with a view to using the core Harvest Agent component to interface with other crawlers in the future.
- Usage of the old heritrix-1.14 dependency, aheritrix-1.14.1.jar, has been upgraded where possible to use the webarchive-commons library.
Things to be aware of¶
- The Bandwidth restriction functionality is not currently applicable to the new Heritrix 3 crawling. The Bandwidth feature has been underused in recent years and was not compatible out-of-the-box with Heritrix 3. A decision on whether to develop the feature to be compatible or remove it entirely will be made in the future.
- The existing prune and import functionality within the QA tool is not currently compatible with Target Instances harvested using Heritrix 3. These components of QA functionality are no longer fit-for-purpose in version 2.0.0, and will be re-developed as part of the WCT development road-map.
- The Groups feature is not currently compatible with Heritrix 3 profiles. This is intended to be resolved in the near future with a minor release.
This is the WCT 1.6.3 GA version.
Released July 2017, this version contains minor changes to the Submit-to-Rosetta module.
Alma compatibility upgrades for Submit to Rosetta module¶
Changes required by the National Library of New Zealand to be compatible with archiving to a Rosetta DPS integrated with Alma (library cataloguing and workflow management system from Ex Libris). All changes have been implemented as backward compatible as possible. The exposure of these changes and their configuration are through the files wct-das.properties, wct-das.xml inside WCT-Store.
This is the WCT 1.6.2 GA version.
Obtaining the source files¶
The WCT code is now stored in a GIT repository on sourceforge - available from the code link on the main WCT sourceforge project page.
The previous versions of WCT are available via the Legacy Code link, if needed. This is still a CVS repository.
Please ensure the user that WCT uses to login to your database has the correct permissions to create temporary tables. Failure to grant this will result in problems during the purge process.
WCT new features and changes for v1.6.2¶
UI new features and improvements¶
- Import https urls
- Configurable Rosetta access rights
The Rosetta access codes that are used in the Submit-to-Rosetta module are now configurable via the wct-das.properties file in the wct-store app. These codes are used in the mets.xml when a harvest is archived to Rosetta.
OMS Codes (Rosetta)
dpsArchive.dnx_open_access=xxxx dpsArchive.dnx_published_restricted=xxxx dpsArchive.dnx_unpublished_restricted_location=xxxx dpsArchive.dnx_unpublished_restricted_person=xxxx
- Submit-to-Rosetta compatibility with newer Rosetta versions
Later versions of Rosetta system complained when performing xsd validation on the mets.xml file submitted by WCT when archiving a harvest. The structure map schema used by WCT was old. As Rosetta auto generates structure maps for deposits that are missing them, structure map generation was removed from the WCT process.
Allowing the version of Rosetta you are archiving to to generate the appropriate structure map.
- Quality Review tool uses original seed url
The harvest quality review tools were not available previously if the original target seed URL was modified.
Now the target seed URL can be changed, and the QR tool will always look for the original URL of the Target Instance instead.
- Pruning and importing for warc files fixed
- Pruning and importing on warcs in the Tree View screen was encountering a bug. When parsing a warc, the input stream was over-reading the number of bytes in the warc-info header, causing unexpected characters to be read when trying to access the next record. This was mainly visible when trying to import and prune.
- Indexing breaking for compressed warcs
- Harvesting as compressed warc was breaking the indexing of a harvest. The Heritrix class handling the reading of the compressed warc was missing the functionality to move to the next record. The Heritrix library included has been recompiled to include a fix.
- Duplicate schedules when saving annotations
When creating/editing a Target - if a schedule is created/edited without saving the Target, and then the Target is saved whilst adding an annotation, WCT creates target instances for that schedule but the Target remains in a state where it contains a cache of new a schedule(s). So if the Target is then saved via the bottom save button, another group of target instances will be generated for the new schedule(s).
This bug has now been fixed. If a schedule already has target instances generated (at Annotations tab), then WCT will flag this to prevent any duplicates from being generated.
- No strippedcrawl.log generated on non-windows os
- WCT was hard-coded to use a Windows file path separator when saving this log file. Now system specific file path separator is used.
UI new features and improvements¶
- Date pickers for date fields
- All date fields in WCT now have associated date pickers to aid in selection.
- Edit button for view screens
- All possible view screens now have buttons to enable edit mode where the user has authority to edit the associated record.
- Harvest optimization incl. global option
There is now the option to specify harvest optimization on any target. This allows the harvesters to perform harvesting of the associated target instances earlier than the schedule otherwise permits. The window for this look-ahead is configurable, and defaults to 12 hours.
This feature can also be disabled on a global basis, temporarily, from the Management->Harvester Configuration->General screen. Upon restart this setting is enabled.
- Harvester queue pause
The queue for harvesters can now be paused on a per-harvester basis. This pause only affects harvests which have not yet started - it is still possible to pause harvests using the traditional mechanism. To activate/deactivate this feature, click the pause/play icon in the Accept tasks column on the Management->Harvester Configuration->General screen.
The intent of this is to be able to pause a specific harvester in order to stop it and perform maintenance once harvests are finished.
- Scheduling heat map
A heat map is now available on the target scheduling screen. This allows a user to see a rough overview of when jobs are scheduled in the next few months or so, in order to choose a day where harvesting is least intensive.
To view the heat map, visit the Target->Schedule->edit/new schedule page. Click the calendar icon labelled heat map - the days will be colored based on how many harvests are scheduled on those days.
The color of the heat map, and the thresholds used to display the colors, are configurable on the Management->Harvester Configuration->Bandwidth page. This allows organizations of any size to customize the heat map to the capabilities of their harveters.
- Import profile to any agency
- The profile import page now has the ability to select any agency to import a profile into. This option is only available when the logged in user has the authority to manager the profiles for all agencies. When this authority is not present, that user’s agency is used for the import.
- Ability to delete all intray tasks
- There is now a button to allow the deletion of all intray tasks, intended mainly for organizations that do not make use of the tasks.
- Ability to hide intray tasks
- Similar to the ability to delete all tasks, the tasks can also be hidden from view on a per-agency basis. The configuration for this feature is on the edit agency page.
- Target URL edit
- It is now possible to edit Target URLs once they have been created. Note that this will affect all existing and future scheduled target instances!
- Target description search
- It is now possible to search inside the description of targets on the target listing screen.
- Reply-to email address in permissions
- There is now a configurable reply-to email address on the permission template edit screen. This will show in most modern email clients as Reply-to: and should be used as the default reply to address in clients which support it.
- Annotations prompt to save
- When clicking the add button for annotations, a prompt now asks whether the user wants to save the associated target, target instance, or group.
- Indicator flag color picker improvement
- The indicator flag color picker now updates when colors are selected, rather than having to click the color wheel icon in the bottom right.
- Completed harvests can be harvested now where user has authority, state is reset to Approved
- Where a user has authority to reinstate and approve a target, they do not need to manually change the state to approved when adding a new schedule or using harvest now.
- Completed harvests can have schedules added where user has authority, state is reset to Approved
- Where a user has authority to reinstate and approve a target, they do not need to manually change the state to approved when adding a new schedule or using harvest now.
- Groups with sub-groups can now be styled using CSS
- The text for groups with sub-groups in the group listing screen can now be styled using CSS.
- Rejection reason is shown against rejected harvest results
- The rejection reason was not visible in any UI element for a rejected harvest result. This has been added to the harvest result listing screen.
- Non-english character support for all WCT screens (providing database is configured correctly)
When the database is configured to support UTF-8 characters, the user interface now supports non-english characters on all screens, including permissions emails.
If you are experiencing problems with UTF-8 characters after this release, ensure that the database tables explicitly support UTF-8.
- Non-existant scheduling alert
When attempting to create a Target schedule which falls on non-existant dates, an alert will be displayed. This is show for custom schedules as well as any schedule with a frequency of monthly or less.
For example, a monthly schedule on the 30th day of the month will not fire in February, and a monthly schedule on the 31st day of the month will only fire seven months a year, as February, April etc have less than 31 days.
- Profile null pointers fixed
- Null pointer exceptions caused by the absence of a default profile have been fixed. This was especially a problem when users were creating targets using the bootstrap user, and was generally experienced by new users of WCT.
- Various other null pointers fixed
- A variety of other NullPointerException errors have been fixed.
- Permissions orphan records
- The database was amended so that permissions records were not duplicated then orphaned when any change to permissions was made. In organizations where a lot of permissions changes were made, this could result in a large number of orphaned records.
- Indicator flags can now only be applied to targets for the same agency
Previously if a user had the “manage flags” authority they could assign any indicator flag to any target instance. This can result in users without that privilege from being able to find those target instances during a TI search by indicator flag.
Updated Target Instance edit screen to only allow indicator flags for the same agency as the owner of the target.
- Viewing other TIs in harvest history changes the TI being reviewed
When reviewing a target instance, clicking on any other target instance in the harvest history screen caused a change in the target instance originally being reviewed. In some cases users were endorsing the wrong target instance, believing that they were still reviewing the one they originally chose to review.
The target instance being reviewed now does not change unless the user decides to review the one selected in the target history, and a warning is displayed indicating this fact.
- Target instances are now completely created for targets with repeating schedules
- A bug was introduced in WCT 1.6 that meant target instances were not created when adding a schedule to a target and saving. Any subsequent saves would create one target instance, but it could result in missing target instances. This has been fixed.
- Max width of target, QA indicator screens has been limited to prevent scroll bars
- When using particularly long seeds or target names, a scroll bar on the target listing screen was necessary, similarly for the QA indicator listing. The table contents are now wrapped and sized appropriately.
- The eSerial next function (used by NLNZ) has been included on the QA Target Instance Summary page
- Previously, the archive button would not show the custom deposit form for Rosetta. A next button now allows this function as per the Harvest Result screen.
- Deletion of harvest resources fixed (requires that WCT database can create temporary tables)
- A potential problem with the deletion of harvest resources was fixed - a null pointer exception was possible, which meant that only one harvest was deleted per execution of the purge process. Additionally, the WCT database user needs authority to create temporary tables (e.g. for Oracle GRANT CREATE TEMPORARY TABLE to usr_wct)
Release 1.6.0 greatly enhances the automated quality assurance (QA) features by providing a preview of each harvest and an automated recommendation. It contains a large number of updates summarised in the list below. Further details can be found in the release notes in the download and on the website.
- Added config parameter to enable new QA module
- Added new target instance summary screen (QA control and streamlines access to other WCT functions)
- Added the QA Recommendation Service
- Added website preview to target instances screen
- Extended target instance flags (enables adhoc grouping)
- Enhanced target instance search screen (sortable columns, filters and annotations as tooltips)
- Integrated existing schedule service into new summary screen
- Added ‘Auto-prune’ service
- New Report: Heritrix Status Code Summary
- New Report: Crawl differential comparison (New URIs + Matching URIs + Missing URIs)
- New Report: URL count by Domain Summary
- New Report: Off-scope URIs
- New Report: Long URIs
- New Report: Unknown MIME Types
- New Report: robots.txt entries disallowed
- New Report: Repeating patterns in URIs
SourceForge bug fixes¶
In addition, the following SourceForge bug fixes have been applied:
- Warc write process with prune tool
- Group schedule target to harvest agent errors
Community and internal testing bug fixes¶
The following bugs have also been fixed as a result of user community and internal testing:
- Memory leak caused by target instances being pinned into memory by tag-pooling (also see ‘WCT Configuration and Deployment Guide (WCT 1.6).pdf’)
- Removed target instance from session after exiting edit mode
- Malformed initial <select> HTML tag within the <wct:list> tag caused options to disappear
Release 1.5.2 is primarily a maintenance update of the Web Curator Tool. It contains a number of bugfixes and enhancements. These are summarised in the list below:
Further details for each item can be found in the SourceForge Enhancement Tracker where relevant.
- SourceForge Ref: 3162584 - Fix bug where Targets with open schedules were erroneously going to completed status
- SourceForge Ref: 3162582 - Fix problem with Illegal group reference error on review
- SourceForge Ref: 3169679 - Add Rejection Reason functionality
- SourceForge Ref: 3162580 - Fix bug where errors thrown when Re-start indexing used
- SourceForge Ref: 3072266 - Implemented batch re-assignment of Target profiles to fix issues such as ‘bad effect on Approved targets when profile is Disabled’.
- SourceForge Ref: 2941648 - Add facility to reject harvests stuck in ‘restart indexing’
- SourceForge Ref: 2952587 - Approved targets should stay approved after profile edits - enhanced logic regarding Target state changes when editing profiles
- SourceForge Ref: 2940542 - Seed URL too long for database column - column width increased
- SourceForge Ref: 3162604 - By default checkProcessor bean will be disabled in harvest agent
- SourceForge Ref: 3162649 - Property file update - to fix typographical error in das.properties file
- SourceForge Ref: 3162600 - Date locale - bandwith settings bug
- SourceForge Ref: 3025576 - Fix bug introduced by Endorse/unendorse actions in WCT version 1.5
- SourceForge Ref: 3162610 - fix absolute paths coded into certain jsp and css files
- SourceForge Ref: 2943743 - Fix bug causing error after approving a target in some circumstances
- SourceForge Ref: 3006785 - Log IP address of harvested files to the crawl.log
- SourceForge Ref: 3162609 - Disk check bean not checking correct partition
- SourceForge Ref: 3162581 - Fix bug where URIListRegExpFilter not working in some profiles
- SourceForge Ref: 2983692 - Correct the permission checking logic where users are allowed to create permission request templates
- SourceForge Ref: 3162597 - Add ‘view target instances’ link to Target Action Column
- SourceForge Ref: 2983693 - Add new field to Harvest Auths permissions tab to allow saving of permission responses
- SourceForge Ref: 3016176 - Crawler Activity Report modifications - add filters
- SourceForge Ref: 2970877 - Link to associated target instance records from Harvest History table and navigate back again
- SourceForge Ref: 3097070 - Fix profile issues regarding redundant fields in DecideRules when setting scope
- SourceForge Ref: <none, BL specific enhancement> - Switch Automated QA on/off on a per target basis
- SourceForge Ref: <none, BL specific enhancement> - Allow the importing of missing content into a harvest via the wctaqa report
- SourceForge Ref: <none, BL specific enhancement> - SOAP API call enhancements which allow automation of certain setup tasks from external applications
Release 1.5.1 is primarily a maintenance update of the Web Curator Tool. It contains a number of bugfixes, enhancements and performance improvements. These are summarised in the lists below:
Further details for each item can be found in the SourceForge Tracker.
- SourceForge 2935731: Ability to add missing files to a harvest before submitting to archive
- SourceForge 1828045: Ability to set harvest alerts, triggered from Target record via alertable annotations
- SourceForge 2892942: Ability to export and import profiles to xml files
- SourceForge 2892356: Ability to sort the views of targets, target instances and harvest authorisations by name and date
- SourceForge 2934308: Ability to view hidden targets, target instances and groups (where display flag is false)
- SourceForge 2892361: Highlight the primary seed URL on target records
- SourceForge 2892965: Set Targets to Completed status when appropriate
- SourceForge 2932069: Ability to create Group sub-categories
- SourceForge 1828045: Display alert icon against scheduled TIs when associated Target has alertable annotations
- SourceForge 2932065: Add a ‘Submit to Archive’ icon to action column of endorsed Target Instances
- SourceForge 2931964: Facility to add Annotations on Access tab of Targets/Groups and the Display tab of TIs
- SourceForge 2892358: Capture WCT and Heritrix version numbers used when harvesting, display on Target Instance
- SourceForge 2892367: Distinguish between first time harvests and repeat harvests for a given Target
- SourceForge 2617632: Implement Tree Tool display enhancements
- SourceForge 2511377: Add ability to display crawl path (hop path) in tree tool view
- SourceForge 2892363: Display the date that permissions letter/email was sent on Harvest Authorisations view
- SourceForge 1774427: Selection Note and Evaluation Note fields on Target record annotations tab were too short
- Fixed issue 2932075: Allow pre v1.5 harvests to be reviewed using wayback
- Fixed issue 2892945: Harvest profile description field length bug
- Fixed issue 2156378: Two default active profiles causes crash on new target screen
- Fixed issue 2931967: Submitting Target instance to archive now returns user to instances list screen
- Fixed issue ???????: WCT timeout was occuring when reviewing large harvests (raised directly via BL no Sourceforge ref)
- Fixed issue 3004090: Slow performance when accessing WCT targets/instances with group schedules
- Fixed issue 2931964: Annotations on Access tab of Targets and the Display tab of TIs
- Fixed issue 2928219: System Activity Report slow or unresponsive
- Fixed issue 1557611: Name duplication conflict between Target and Group names
Release 1.5 is the fourth major update of the Web Curator Tool. This release is mainly concerned with the optional integration of Wayback as an additional quality review tool, and the simplification of system configuration using properties files; but also contains a small number of additional enhancements and bugfixes summarised in the list below. Further details for each item can be found in the SourceForge Tracker.
- Quality Review Update to use wayback (2807159)
- Properties file for spring config (2807161)
- Autopopulate dublin core title field from target title (2815658)
- Target section multiple action on seeds (2811357)
- Add Harvested link to list of quick links (SourceForge# 2811364)
- Ability to submit to a Rosetta based archive
- Fixed issue 2815654: Reindexing fails
- Fixed issue 2807164: MYSQL install file update
- Fixed issue 2810210: Sub-group deletion exception
- Fixed issue 2775423: Browse tool throwing exceptions with bad URI’s
This docuemnt does not include the Release notes for versions before 1.5.