Workflow volume

Dockerfile

@@ -70,6 +70,7 @@ EXPOSE 22
 
 WORKDIR /data
 VOLUME /data
+VOLUME /workflows
 
 # simulate a virtual env for the makefile,
 # coinciding with the Python system prefix

Makefile

@@ -23,6 +23,8 @@ Variables:
 	  currently: "$(PRIVATE)"
 	- DATA		host directory to mount into `/data`
 	  currently: "$(DATA)"
+	- WORKFLOWS	host directory to mount into `/workflows`
+	  currently: "$(WORKFLOWS)"
 	- UID		user id to use in logins
 	  currently: $(UID)
 	- GID		group id to use in logins
@@ -47,6 +49,7 @@ help: ; @eval "$$HELP"
 KEYS ?= $(firstword $(wildcard $(HOME)/.ssh/authorized_keys* $(HOME)/.ssh/id_*.pub))
 PRIVATE ?= $(firstword $(filter-out %.pub,$(wildcard $(HOME)/.ssh/id_*)))
 DATA ?= $(CURDIR)
+WORKFLOWS ?= $(CURDIR)/workflows
 UID ?= $(shell id -u)
 GID ?= $(shell id -g)
 UMASK ?= 0002
@@ -62,6 +65,7 @@ run: $(DATA)
 	--name ocrd_manager \
 	--network=$(NETWORK) \
 	-v $(DATA):/data \
+	-v $(WORKFLOWS):/workflows \
 	--mount type=bind,source=$(KEYS),target=/authorized_keys \
 	--mount type=bind,source=$(PRIVATE),target=/id_rsa \
 	-e UID=$(UID) -e GID=$(GID) -e UMASK=$(UMASK) \
@@ -88,7 +92,8 @@ test-production: $(DATA)/testdata-production
 ifeq ($(NETWORK),bridge)
 	ssh -i $(PRIVATE) -Tn -p $(PORT) ocrd@localhost $(SCRIPT) $(<F)
 else
-	docker exec -t -u ocrd `docker container ls -qf name=ocrd-manager` $(SCRIPT) $(<F)
+	if test -t 0 -a -t 1; then TTY=-i; fi; \
+	docker exec $$TTY -t -u ocrd `docker container ls -n1 -qf name=ocrd-manager` $(SCRIPT) $(<F)
 endif
 	test -d $</ocr/alto
 	test -s $</ocr/alto/00000009.tif.original.xml
@@ -100,7 +105,8 @@ test-presentation:
 ifeq ($(NETWORK),bridge)
 	ssh -i $(PRIVATE) -Tn -p $(PORT) ocrd@localhost $(SCRIPT) $(<F)/mets.xml
 else
-	docker exec -t -u ocrd `docker container ls -qf name=ocrd-manager` $(SCRIPT) $(<F)/mets.xml
+	if test -t 0 -a -t 1; then TTY=-i; fi; \
+	docker exec $$TTY -t -u ocrd `docker container ls -n1 -qf name=ocrd-manager` $(SCRIPT) $(<F)/mets.xml
 endif
 	diff -u <(docker run --rm -v $(DATA):/data $(TAGNAME) ocrd workspace -d $(<F) find -G FULLTEXT -g PHYS_0017..PHYS_0021) <(for file in FULLTEXT/FULLTEXT_PHYS_00{17..21}.xml; do echo $(PREFIX)/$$file; done)
 

README.md

@@ -44,10 +44,12 @@ Build or pull the Docker image:
 
 ### Starting and mounting
 
-Then run the container – providing a **host-side directory** for the volume …
+Then run the container – providing a **host-side directory** for the volumes …
 
  * `DATA`: directory for data processing (including images or existing workspaces),  
    defaults to current working directory
+ * `WORKFLOWS`: directory for scripts (preconfigured workflows),  
+   defaults to `./workflows` in current working directory
 
 … but also files …
 
@@ -73,7 +75,7 @@ Then run the container – providing a **host-side directory** for the volume 
 
 … thus, for **example**:
 
-    make run DATA=/mnt/workspaces MODELS=~/.local/share KEYS=~/.ssh/id_rsa.pub PORT=9022 PRIVATE=~/.ssh/id_rsa
+    make run DATA=/mnt/workspaces WORKFLOWS=/mnt/workflows KEYS=~/.ssh/id_rsa.pub PORT=9022 PRIVATE=~/.ssh/id_rsa
 
 (You can also run the service via `docker-compose` manually – just `cp .env.example .env` and edit to your needs.)
 
@@ -140,7 +142,7 @@ which contains a trivial workflow:
 - preprocessing, layout analysis and text recognition with a single Tesseract processor call
 - format conversion of the result from PAGE-XML to ALTO-XML
 
-It can be replaced with the (path) name of any workflow script mounted under `/data`.
+It can be replaced with the (path) name of any workflow script mounted under `/workflows` or `/data`.
 
 For example (assuming `testdata` is a directory with image files mounted under `/data`):
 
@@ -180,7 +182,7 @@ ENVIRONMENT VARIABLES:
  CONTROLLER: host name and port of OCR-D Controller for processing
 ```
 
-The same goes here for the `workflow parameter`.
+For the `workflow` parameter, the same goes here as [above](#from-image-to-alto-files).
 
 For example (assuming `testdata` is a directory with image files mounted under `/data`):
 

docker-compose.yml

@@ -33,6 +33,7 @@ services:
         source: ${MANAGER_KEY} # Manager private key (for access to Controller)
         target: /id_rsa
       - ${MANAGER_DATA}:/data # metadata directory
+      - ${MANAGER_WORKFLOWS}:/workflows
       - shared:/run/lock/ocrd.jobs
 
     tty: true # docker run -t

ocrd_lib.sh

@@ -10,10 +10,17 @@ logerr() {
   logger -p user.info -t $TASK "terminating with error \$?=$? from ${BASH_COMMAND} on line $(caller)"
 }
 
+stopbg() {
+  logger -p user.crit -t $TASK "passing SIGKILL to child $!"
+  # pass signal on to children
+  kill -KILL $!
+}
+
 # initialize variables, create ord-d work directory and exit if something is missing
 init() {
   trap logerr ERR
-
+  trap stopbg INT TERM KILL
+
   PID=$$
 
   cd /data
@@ -29,13 +36,25 @@ init() {
     exit 2
   fi
 
+  WORKDIR=ocr-d/"$PROCESS_DIR" # use subdirectory of same volume so --reflink CoW still possible
+  if ! mkdir -p "$WORKDIR"; then
+    logger -p user.error -t $TASK "insufficient permissions on /data volume"
+    exit 5
+  fi
+  # try to be unique here (to avoid clashes)
+  REMOTEDIR="KitodoJob_${PID}_$(basename $PROCESS_DIR)"
+
   WORKFLOW=$(command -v "$WORKFLOW" || realpath "$WORKFLOW")
   if ! test -f "$WORKFLOW"; then
     logger -p user.error -t $TASK "invalid workflow '$WORKFLOW'"
     exit 3
   fi
   logger -p user.notice -t $TASK "using workflow '$WORKFLOW':"
   ocrd_format_workflow | logger -p user.notice -t $TASK
+  if test "${WORKFLOW#/workflows/}" = "$WORKFLOW"; then
+      cp -p "$WORKFLOW" "$WORKDIR/workflow.sh"
+      WORKFLOW="$WORKDIR/workflow.sh"
+  fi
 
   if test -z "$CONTROLLER" -o "$CONTROLLER" = "${CONTROLLER#*:}"; then
     logger -p user.error -t $TASK "envvar CONTROLLER='$CONTROLLER' must contain host:port"
@@ -44,18 +63,11 @@ init() {
   CONTROLLERHOST=${CONTROLLER%:*}
   CONTROLLERPORT=${CONTROLLER#*:}
 
-  WORKDIR=ocr-d/"$PROCESS_DIR" # will use other mount-point than /data soon
-  if ! mkdir -p "$WORKDIR"; then
-    logger -p user.error -t $TASK "insufficient permissions on /data volume"
-    exit 5
-  fi
-  # try to be unique here (to avoid clashes)
-  REMOTEDIR="KitodoJob_${PID}_$(basename $PROCESS_DIR)"
-
   # create stats for monitor
   mkdir -p /run/lock/ocrd.jobs/
   {
     echo PID=$PID
+    echo TIME_CREATED=$(date --rfc-3339=seconds)
     echo PROCESS_ID=$PROCESS_ID
     echo TASK_ID=$TASK_ID
     echo PROCESS_DIR=$PROCESS_DIR
@@ -68,7 +80,8 @@ init() {
 }
 
 logret() {
-    sed -i 1s/.*/RETVAL=$?/ /run/lock/ocrd.jobs/$REMOTEDIR
+    sed -i "1s/PID=.*/RETVAL=$?/" /run/lock/ocrd.jobs/$REMOTEDIR
+    sed -i "2a TIME_TERMINATED=$(date --rfc-3339=seconds)" /run/lock/ocrd.jobs/$REMOTEDIR
 }
 
 init_task() {
@@ -83,15 +96,15 @@ ocrd_format_workflow() {
 
 # ocrd import from workdir
 ocrd_import_workdir() {
+  echo "echo \$\$ > $REMOTEDIR/ocrd.pid"
   echo "if test -f '$REMOTEDIR/mets.xml'; then OV=--overwrite; else OV=; ocrd-import -i '$REMOTEDIR'; fi"
   echo "cd '$REMOTEDIR'"
-  echo 'echo $$ > ocrd.pid'
 }
 
 ocrd_enter_workdir() {
+  echo "echo \$\$ > $REMOTEDIR/ocrd.pid"
   echo "if test -f '$REMOTEDIR/mets.xml'; then OV=--overwrite; else OV=; fi"
   echo "cd '$REMOTEDIR'"
-  echo 'echo $$ > ocrd.pid'
 }
 
 ocrd_process_workflow() {

process_images.sh

@@ -16,7 +16,7 @@ parse_args() {
   SCRIPT=
   PROCESS_ID=
   TASK_ID=
-  WORKFLOW=ocr-workflow-default.sh
+  WORKFLOW=/workflows/ocr-workflow-default.sh
   IMAGES_SUBDIR=images
   RESULT_SUBDIR=ocr/alto
   while (($#)); do

process_mets.sh

@@ -15,7 +15,7 @@ parse_args() {
   SCRIPT=
   PROCESS_ID=
   TASK_ID=
-  WORKFLOW=ocr-workflow-default.sh
+  WORKFLOW=/workflows/ocr-workflow-default.sh
   PAGES=
   IMAGES_GRP=DEFAULT
   RESULT_GRP=FULLTEXT