Approach 1
While working in the fix of CM-44 I found that the lastest version (0.6.1) may include a default option that could fix the problem.
The thing is that upgrading this library will imply including/upgrading its dependencies (see the section Transitive Dependencies):
any23-061-deps.xls (33.5 KB)
NOTE: The site of Any23 shows an updated and shorter list of dependencies.
To test that this approach actually works:
[ul]I created a user-library from Eclipse to include the jars (I got them using Maven).[/ul]
[ul]Included in the build path of the Metadata-indexer project.[/ul]
[ul]Tweaked some code and run all the Unit Test for the project.[/ul]
Approach 2
Another approach or workaround would imply modifying PSTripleHandler to check every property’s name:
if it is “dcters:title” and its value contains more than 2 whitespaces anywhere
ignore it,
else
continue with normal workflow, adding to a “properties” collection, etc.
Approach 3
Restrict the input in CM1 so it does not accept consecutives white spaces.
So, what do you think about these approaches?