SHELL = /bin/sh #!/usr/bin/make # # This is a makefile to automatise most of the testing. # It is language independent within the framework of the Sámi language # technology project. See further user documentation (and technical as well) # in $CVSROOT/gt/doc/docu-testing.html # VARIABLE DEFINITIONS # ==================== XFST = xfst #-utf8 ifeq (victorio.uit.no, $(shell hostname)) XFST = /opt/sami/xerox/c-fsm/ix86-linux2.6-gcc3.4/bin/fst -utf8 endif # Where all the perl scripts are: BINDIR = ~/../../../gt/script/testing # Directory where all the compiled fst's are: FSTDIR = ../bin # Directory for the source files (which compile into fst's): SRCDIR = ../src # The temp directory: TEMP = ../../tmp # This variable is holding the filename of the save file: # (makes the Makefile truly language independent for # the Sami languages) SAVEFILE = $(wildcard $(FSTDIR)/ipk.save) # Copied from old LS makefile: # Bond, James Bond UMASK=@umask 007 # Empty variable to be overridden by the user when # generating paradigms: WORD = # These variables define the relationship between the source files # and the final output files (the reports), such that one can con- # struct a list of targets for the main target NGenFiles := $(patsubst %.txt,%.greport,$(wildcard n-*.txt)) NAnaFiles := $(patsubst %.txt,%.areport,$(wildcard n-*.txt)) VGenFiles := $(patsubst %.txt,%.greport,$(wildcard v-*.txt)) VAnaFiles := $(patsubst %.txt,%.areport,$(wildcard v-*.txt)) AGenFiles := $(patsubst %.txt,%.greport,$(wildcard a-*.txt)) AAnaFiles := $(patsubst %.txt,%.areport,$(wildcard a-*.txt)) # MAJOR TARGETS, HOUSEKEEPING TARGETS # =================================== # Default target: # Remove the comments in front of the verb and adjective # targets when you have codes and test cases for them .PHONY: all all: n-g.summary \ n-a.summary \ v-g.summary \ v-a.summary \ # a-g.summary \ # a-a.summary # If you only want to test nouns: .PHONY: all-n all-n: n-g.summary n-a.summary # If you only want to test intransitive verbs: .PHONY: all-i all-i: i-g.summary i-a.summary # If you only want to test transitive verbs: .PHONY: all-t all-t: t-g.summary t-a.summary # If you only want to test pronouns: .PHONY: all-p all-p: p-g.summary p-a.summary # If you only want to test verbal generation, use v-g.summary # as target. Etc. for the other cases. And do you only want to # test a specific word or inflection class, type the name of # the file containing the infl. word, but replace the suffix # .txt with .greport for generation, and .areport for analysis. # Clean up all the mess created during a test pass, to prepare for new tests: # *.tgz \ # just to remove archive files used for file transfering # *~ # to remove backup copies that may clutter the directory .PHONY: clean clean: @echo @echo "*** Now deleting ALL the generated files ***" @echo @rm -f *.testbase \ *.gtest \ *.atest \ *.gfacit \ *.afacit \ *.greport \ *.areport \ *.gresult \ *.aresult \ *.summary \ *.ptest \ *paradigm \ *.tgz \ *~ \ $(TEMP)/*test-script .PHONY: almost-clean almost-clean: @echo @echo "*** Now deleting all generated files but the summary & parad. files ***" @echo @rm -f *.testbase \ *.gtest \ *.atest \ *.gfacit \ *.afacit \ *.greport \ *.areport \ *.gresult \ *.aresult \ *.ptest \ *.tgz \ *~ \ $(TEMP)/*test-script # POS SPECIFIC SECTIONS # ===================== # NOUNS # ===== # Create all the noun test file bases: n-%.testbase: n-%.txt noun-codes.txt @echo @echo "*** Creating the test base for n-$*.txt ***" @echo @rm -f $@ $(UMASK) && \ $(BINDIR)/merge-codesNforms.pl noun-codes.txt $< > $@ @chmod 660 $@ # Finally, we concatenate all *.greports n-g.summary: $(NGenFiles) @cat $^ > $@ @echo @echo "*** All noun word form generation testing is done. ***" @echo "*** The concatenated reports are stored in $@. ***" @echo # @less $@ # Finally, we concatenate all *.areports: n-a.summary: $(NAnaFiles) @cat $^ > $@ @echo @echo "*** All noun word form analysis testing is done. ***" @echo "*** The concatenated reports are stored in $@. ***" @echo # @less $@ # Create noun paradigm generation input file: n-para.ptest: noun-codes.txt @echo @echo "*** Creating the noun paradigm input file for '$(WORD)' ***" @echo @rm -f $@ $(UMASK) && \ $(BINDIR)/merge-codesNword.pl $< $(WORD) > $@ @chmod 660 $@ # Paradigm generation: .PHONY: n-paradigm n-paradigm: n-para.ptest savefile @echo @echo "*** Running the paradigm generator for '$(WORD)' ***" @echo @printf "load $(SAVEFILE) \n\ apply down <$< \n\ quit \n" > $(TEMP)/$@-gtest-script $(XFST) < $(TEMP)/$@-gtest-script > $(WORD).paradigm @rm -f $(TEMP)/$@-gtest-script @rm -f $< @less $(WORD).paradigm # INTRANSITIVE VERBS # ===== # Create all the intransitive verb test file bases: i-%.testbase: i-%.txt intrans-codes.txt @echo @echo "*** Creating the test base for i-$*.txt ***" @echo @rm -f $@ $(UMASK) && \ $(BINDIR)/merge-codesNforms.pl intrans-codes.txt $< > $@ @chmod 660 $@ # Finally, we concatenate all *.greports: i-g.summary: $(VGenFiles) @cat $^ > $@ @echo @echo "*** All intrans word form generation testing is done. ***" @echo "*** The concatenated reports are stored in $@. ***" @echo # @less $@ # Finally, we concatenate all *.areports: i-a.summary: $(VAnaFiles) @cat $^ > $@ @echo @echo "*** All intrans word form analysis testing is done. ***" @echo "*** The concatenated reports are stored in $@. ***" @echo # @less $@ # Create intrans paradigm generation input file: i-para.ptest: intrans-codes.txt @echo @echo "*** Creating the intrans paradigm input file for '$(WORD)' ***" @echo @rm -f $@ $(UMASK) && \ $(BINDIR)/merge-codesNword.pl $< $(WORD) > $@ @chmod 660 $@ # Paradigm generation: i-paradigm: i-para.ptest savefile @echo @echo "*** Running the paradigm generator for '$(WORD)' ***" @echo @printf "load $(SAVEFILE) \n\ apply down <$< \n\ quit \n" > $(TEMP)/$@-gtest-script $(XFST) < $(TEMP)/$@-gtest-script > $(WORD).paradigm @rm -f $(TEMP)/$@-gtest-script @rm -f $< @less $(WORD).paradigm # TRANSITIVE VERBS # ===== # Create all the transitive verb test file bases: t-%.testbase: t-%.txt trans-codes.txt @echo @echo "*** Creating the test base for t-$*.txt ***" @echo @rm -f $@ $(UMASK) && \ $(BINDIR)/merge-codesNforms.pl trans-codes.txt $< > $@ @chmod 660 $@ # Finally, we concatenate all *.greports: t-g.summary: $(VGenFiles) @cat $^ > $@ @echo @echo "*** All trans word form generation testing is done. ***" @echo "*** The concatenated reports are stored in $@. ***" @echo # @less $@ # Finally, we concatenate all *.areports: t-a.summary: $(VAnaFiles) @cat $^ > $@ @echo @echo "*** All trans word form analysis testing is done. ***" @echo "*** The concatenated reports are stored in $@. ***" @echo # @less $@ # Create trans paradigm generation input file: t-para.ptest: trans-codes.txt @echo @echo "*** Creating the trans paradigm input file for '$(WORD)' ***" @echo @rm -f $@ $(UMASK) && \ $(BINDIR)/merge-codesNword.pl $< $(WORD) > $@ @chmod 660 $@ # Paradigm generation: t-paradigm: t-para.ptest savefile @echo @echo "*** Running the paradigm generator for '$(WORD)' ***" @echo @printf "load $(SAVEFILE) \n\ apply down <$< \n\ quit \n" > $(TEMP)/$@-gtest-script $(XFST) < $(TEMP)/$@-gtest-script > $(WORD).paradigm @rm -f $(TEMP)/$@-gtest-script @rm -f $< @less $(WORD).paradigm # PRONOUNS # ======== # Create all the pronoun test file bases: p-%.testbase: p-%.txt pron-codes.txt @echo @echo "*** Creating the test base for p-$*.txt ***" @echo @rm -f $@ $(UMASK) && \ $(BINDIR)/merge-codesNforms.pl pron-codes.txt $< > $@ @chmod 660 $@ # Finally, we concatenate all *.greports: p-g.summary: $(PGenFiles) @cat $^ > $@ @echo @echo "*** All pronoun word form generation testing is done. ***" @echo "*** The concatenated reports are stored in $@. ***" @echo # @less $@ # Finally, we concatenate all *.areports: p-a.summary: $(PAnaFiles) @cat $^ > $@ @echo @echo "*** All pronoun word form analysis testing is done. ***" @echo "*** The concatenated reports are stored in $@. ***" @echo # @less $@ # Create pronoun paradigm generation input file: p-para.ptest: pron-codes.txt @echo @echo "*** Creating the pronoun paradigm input file for '$(WORD)' ***" @echo @rm -f $@ $(UMASK) && \ $(BINDIR)/merge-codesNword.pl $< $(WORD) > $@ @chmod 660 $@ # Paradigm generation: p-paradigm: p-para.ptest savefile @echo @echo "*** Running the paradigm generator for '$(WORD)' ***" @echo @printf "load $(SAVEFILE) \n\ apply down <$< \n\ quit \n" > $(TEMP)/$@-gtest-script $(XFST) < $(TEMP)/$@-gtest-script > $(WORD).paradigm @rm -f $(TEMP)/$@-gtest-script @rm -f $< @less $(WORD).paradigm # POS-independent parts: # ====================== # Make sure there is an available smX.save in the parent directory: savefile: @cd $(SRCDIR)/ && $(MAKE) # Word form generation: # --------------------- # Make test cases for word form generation tests, # based on the test file bases: %.gtest: %.testbase @echo @echo "*** Creating test cases for word form generation for $*.txt ***" @echo @rm -f $@ $(UMASK) && \ $(BINDIR)/make-gen-test.pl $< > $@ @chmod 660 $@ # Make facit files containing the expected output from the # word form generation test run: %.gfacit: %.testbase @echo @echo "*** Creating facit file for word form generation for $*.txt ***" @echo @rm -f $@ $(UMASK) && \ $(BINDIR)/make-gen-test-facit.pl $< > $@ @chmod 660 $@ # Here we run the actual generation test: %.gresult: %.gtest savefile @echo @echo "*** Running the word form generator test for $*.txt ***" @echo @printf "load $(SAVEFILE) \n\ apply down <$< \n\ quit \n" > $(TEMP)/$*-gtest-script $(XFST) < $(TEMP)/$*-gtest-script > $@ @rm -f $(TEMP)/$*-gtest-script # Then we compare with the facit, and report any differences: %.greport: %.gresult %.gfacit @echo @echo "*** Diffing the test result and the facit file for ***" @echo "*** word form generation for $*.txt ***" @echo "*** \"Error 1 (ignored)\" means that there were diffs. ***" @echo @rm -f $@ -$(UMASK) && \ diff --side-by-side --width=70 $< $*.gfacit > $@ @chmod 660 $@ # Word form analysis: # ------------------- # First, create the needed test file # NB! The sorting specified does not work completely as intended! # This will lead to small differences between the facit file and # the analysis output. This happens when one input string gives # more than one analysis. The expected output is according to the # order in the inflection lexicon, but the tags in the facit file # are sorted alphabetically (due to the way sorting is performed - # there is no way to tell GNU sort to ONLY sort on a specified key). # For a completely satisfactory result we need a different solution. # The present solution will have to be good enough for now. %.atest: %.testbase @echo @echo "*** Creating test cases for word form analysis for $*.txt ***" @echo @rm -f $@ $(UMASK) && \ $(BINDIR)/make-ana-test.pl $< \ | sort -t '\t' -k 1 | cut -f 1 | uniq > $@ @chmod 660 $@ # Make facit files containing the expected output from the # word form analysis test run: %.afacit: %.testbase @echo @echo "*** Creating facit file for word form analysis for $*.txt ***" @echo @rm -f $@ $(UMASK) && \ $(BINDIR)/make-ana-test.pl $< \ | sort -t '\t' -k 1 | cut -f 2 > $@ @chmod 660 $@ # Here we run the actual analysis test: %.aresult: %.atest savefile @echo @echo "*** Running the word form analysis test for verbs ***" @echo @printf "load $(SAVEFILE) \n\ apply up <$< \n\ quit \n" > $(TEMP)/$*-atest-script $(XFST) < $(TEMP)/$*-atest-script > $@ @rm -f $(TEMP)/$*-atest-script # Finally, we compare with the facit, and report any differences: # Due to the sorting problems mentioned above, there will always (?) # be (small) differences. %.areport: %.aresult %.afacit @echo @echo "*** Diffing the test result and the facit file for ***" @echo "*** word form analysis for $*.txt ***" @echo "*** \"Error 1 (ignored)\" means that there were diffs. ***" @echo -@diff --side-by-side --width=70 $< $*.afacit > $@