Final steps for first release

[Aklakan]

• Add build-graph.sh script
  • Extend script to include download URLs for all available datasets. If we used dcat:downloadURLs then we have to go for a DCAT model.
• Publish the patch-finder script (the client to the output of build-graph.sh)
  • Update the patch-finder script (implements dijkstra over the patch graph)

Later:

• Switch to a model where each relevant file X can be annotated with a single X.meta.json file. This way, we can collect arbitrary metadata uniformly. Example metadata to include: byte size, creation date. Date and byte size could be leveraged by the patch-finder.
• Check whether there are specialized path search algos that improve complexity. Probably more of an academic exercise because patch graphs are usually quite small and linear, so generic dijkstra should work sufficiently well.