Documentation Center

Content Importer changes

Content Importer underwent major changes.

Overview

For Content Importer, we:
  • Entirely re-designed the user interface
  • Developed the translation import feature
  • Improved the resume behavior
  • Improved the performance
  • Improved the progress information
  • Distributed the metadata temporary storage among several files
  • Regrouped all metadata operations within the conversion phase

These changes are making Batch Import obsolete. Therefore Batch Import is removed from Content Manager in this release.

The user interface

Content Importer's user interface now lets you choose between three types of import, depending on your requirements: with conversion, without conversion, and translation only. You then access to the settings corresponding to your choice.

During the import, the user interface provides detailed progress information. After the import, you can sort and locate files that returned errors, if any.

The guidance have been considerably improved. Moreover, additional information is provided such as links to folders and the ability to monitor last imports.

Translation import

Importing translations for objects already in the Repository with Content Importer is now simple and reliable. It can be done with multiple languages in one go.

The only metadata operation for this import is a change of status for the imported translation.

Resume

If an import is interrupted and then resumed, the already converted files are not re-run, and the imported files are not imported again. However, the conversion is run again on files that have been changed after the interruption, and added files are converted as well. The files that are no longer in the source content folder are removed from the conversion folder.

The resume feature does not apply if a filemap.xml file is provided by the user in the source folder, or if the import settings are changed after the interruption.

Performance

The performance of several steps has been improved. Those steps include:
  • The loading of the filemap.
  • The resolving of resolutions.
  • The resolving of mixed content in maps.
  • The resolving of content types for folders.
  • The retrieval of titles.

Progress information

You now have a more precise view of the current progress of an import. Also, you can now easily detect at which stage the interruption occurred in case of interruption:
  • Percentage information for steps is provided in addition to the amount of objects.
  • The Apply XSL step is now reported.
  • Only one line is printed for each object.
  • The UI is refreshed less often.

Metadata storage

During the conversion and import processes, the metadata is now stored in separate files having a .3sish extension, created in the same folder as the XML. This replaces the single filemap.xml file that could cause out of memory errors and decreased Content Importer throughput.

Metadata operations

All metadata operations are now performed during the conversion phase. Therefore, by examining the content of the metadata files produced after conversion, you will have a precise view of what will be imported in the Repository.

This metadata grouping also makes future metadata management improvement easier to implement.

Parallel use

You can now run multiple conversions and imports in parallel. The welcome screen shows the progression of the ongoing imports.

Other changes

  • The interruption process is now more friendly: it doesn't kill the process but instead lets it end gracefully. This prevents from possible imported data corruption.
  • When absolute URLs (starting with HTTP, HTTPS, etc.) are detected in <href> elements, Content Importer now adds a scope=external attribute for that link, and produces a warning. It is however possible to revert to the previous behavior, where Content Importer returned one error for each link, by changing the value of a new configuration item: addScopeWhenExternalReference (located in Trisoft.ContentImporter.config), by default set to true.

Deprecated filemap.xml

Since the previous version of Content Manager, we no longer copy the metadata information, created during the conversion phase, in the filemap.xml file. This was changed for performance reasons. However, it is still possible to use a filemap.xml as input file, which perpetuates the old process, and removes the performance gain.

With the Content Importer changes introduced in this version, there are additional reasons why the use of a filemap.xml, even though still available, should be abandoned. First it is not compatible with the resume feature. Second it typically contains custom changes that cannot be controlled by the application and it makes future automation and improvement of the process impossible.

Therefore, we strongly recommend that you no longer use a filemap.xml file as input file.