Use OFN as the intermediate format more widely.

[gouttegd]

OFN is already the default intermediate format for import modules and components. This commit makes it the format for more files, including:

• the $(EDIT_PREPROCESSED) file;
• the $(SRCMERGED) file (strictly speaking this is not necessary: the $(SRCMERGED) variable points to a file with a .ofn extension, so ROBOT already writes that file in OFN format;
• remotely sourced components;
• the merged mirror file.

Part of INCATools/ontology-development-kit#1299

[matentzn]

Here is one that makes sense to change:

$(TRANSLATIONSDIR)/%.babelon.owl: $(TRANSLATIONSDIR)/%-preprocessed.babelon.tsv
	$(BABELONPY) convert $< --output-format owl -o $@.tmp
	$(ROBOT) merge -i $@.tmp \
		annotate \
			--ontology-iri $(ONTBASE)/translations/$*.babelon.owl \
			-V $(ONTBASE)/releases/$(VERSION)/translations/$*.babelon.owl \
			--annotation owl:versionInfo $(VERSION) \
		convert -f owl --output $@

Silly and not necessary in this PR (only cosmetic if you want to fix and boost performance a bit):

reason_test: $(EDIT_PREPROCESSED)
	$(ROBOT) reason --input $< --reasoner $(REASONER) --equivalent-classes-allowed {{ project.allow_equivalents }} \
		--exclude-tautologies {{ project.exclude_tautologies }} --output test.owl && rm test.owl

We can drop --output test.owl && rm test.owl to not even generate this file at all.

[gouttegd]

Here is one that makes sense to change:

$(TRANSLATIONSDIR)/%.babelon.owl: $(TRANSLATIONSDIR)/%-preprocessed.babelon.tsv

Along with its siblings $(TRANSLATIONSDIR)/%.synonyms.owl, I suppose?

Silly and not necessary in this PR (only cosmetic if you want to fix and boost performance a bit):
[…]
We can drop --output test.owl && rm test.owl to not even generate this file at all.

I plan to deal with this type of silliness in a dedicated, “Uberon-Makefile-cleanup-style” PR later. :)

[matentzn]

wonderful PR and much needed