Fix dropped tei2korapxml flags for standard TEI P5 sources
The build/tei.zip recipe spliced the unquoted TEI2KORAPXML_FLAGS value
(-s -tk) into a single-quoted `sh -c` script string. The embedded space
split the shell word, so sh -c received only `tei2korapxml -l warn -s`
as the script and treated -tk, TEI_FLAGS and the input glob as ignored
positional parameters. tei2korapxml then ran without a tokenizer and
without inputs ("No tokenizer chosen").
Drop the sh -c indirection: it existed only to glob /input/*.xml, but
Make already knows the file list, so compute the container paths
(TEI_INPUTS) and run tei2korapxml directly. The recipe's bash now does a
single round of quote processing, so -s -tk and TEI_FLAGS reach the tool
intact.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Change-Id: I8a14de8e4c8b3c1d0721ee2f541446f47f895b9f
diff --git a/Makefile b/Makefile
index 7cd5022..d4cca60 100644
--- a/Makefile
+++ b/Makefile
@@ -9,9 +9,10 @@
# Standard TEI P5 support
TEI_DIR ?= TEI
TEI_FILES := $(wildcard $(TEI_DIR)/*.xml)
+# Same files as seen from inside the container (the dir is mounted at /input).
+TEI_INPUTS := $(patsubst $(TEI_DIR)/%,/input/%,$(TEI_FILES))
TEI_ZIP_NAME ?= tei
TEI_FLAGS ?= --auto-textsigle 'TEI/XYZ.00001' # --xmlid-to-textsigle '([A-Z]+)\.(.*)\.([0-9]+)\.*([0-9])@$$1/$$2/$$3$$4'
-export TEI_FLAGS
# If set to a foundry name (e.g. gingko), tei2korapxml will use inline annotations and tokens from the source XML
USE_INLINE_ANNOTATIONS_AS ?=
@@ -60,7 +61,7 @@
$(BUILD_DIR)/$(TEI_ZIP_NAME).zip: $(TEI_FILES)
mkdir -p $(BUILD_DIR)
- docker run --rm --entrypoint /bin/sh $(if $(DOCKER_CPU_SHARES),--cpu-shares $(DOCKER_CPU_SHARES)) -v "$(abspath $(TEI_DIR)):/input:ro" korap/tei2korapxml:latest -c 'tei2korapxml -l warn '$(TEI2KORAPXML_FLAGS)' '"$$TEI_FLAGS"' /input/*.xml' > $@ 2> >(tee $(@:.zip=.log) >&2)
+ docker run --rm $(if $(DOCKER_CPU_SHARES),--cpu-shares $(DOCKER_CPU_SHARES)) -v "$(abspath $(TEI_DIR)):/input:ro" korap/tei2korapxml:latest -l warn $(TEI2KORAPXML_FLAGS) $(TEI_FLAGS) $(TEI_INPUTS) > $@ 2> >(tee $(@:.zip=.log) >&2)
printf "%s\t%s\n" "$$(cat $^ | grep -c '<teiHeader')" "$$(unzip -l $@ | grep data.xml | wc -l)"