Caching for remote files during extraction

[awagner-mainz]

When resources are fetched from remote servers, download them once and use the cached version instead of re-downloading them every time.

(Code is operational at my place (TM), move it to git branch and test it!)

Think about:

• invalidate/clean cache

[metacontext]

To keep cache pruning simple I think it would suffice to add a slim cleanup function that gets called last on return of the main query. All resources fetched for a specific extraction could be saved to a subfolder in /temp named with a temporary identifier. The cleanup function could then delete the whole cache folder with a xmldb:remove during return.

[metacontext]

Just to keep all thoughts we had about this feature so far here is a related snippet from my TODO.txt:

Only load external resources once => retrieve and cache them for all following statements
maybe by checking the configuration first for all resource attributes, load them and temp. put them in DB with URI <=> hashname map
file: cache/extraction_1235612rauafd.xml

<cachedResources>
<resource key="http://my.gnd.resource/12345" value="temp/13717dhgqf.xml" />
</cachedResources>