script the generation of the examples/queries

.github/workflows/publish.yml

@@ -43,17 +43,8 @@ jobs:
       - name: Sleep to give server time to start
         run: sleep 90
 
-      - name: Load add-condition-transaction-example-1
-        run: curl -X POST -H "Content-type:application/fhir+json;fhirVersion=4.0" -d @fsh-generated/resources/Bundle-add-condition-transaction-example-1.json http://localhost:8080/fhir; 
-
-      - name: Query add-condition-transaction-example-1
-        run: sleep 10; curl $(cat input/queries/get-flags-associated-with-patient.md) | jq '.id = "QUERY-OUTPUT--add-condition-transaction-example-1"' > input/resources/QUERY-OUTPUT--add-condition-transaction-example-1.json
-
-      - name: Load add-condition-transaction-example-2
-        run: curl -X POST -H "Content-type:application/fhir+json;fhirVersion=4.0" -d @fsh-generated/resources/Bundle-add-condition-transaction-example-2.json http://localhost:8080/fhir; \
-
-      - name: Query add-condition-transaction-example-2
-        run: sleep 10; curl $(cat input/queries/get-flags-associated-with-patient.md) | jq '.id = "QUERY-OUTPUT--add-condition-transaction-example-2"' > input/resources/QUERY-OUTPUT--add-condition-transaction-example-2.json
+      - name: Load examples and run queries to be added to IG
+        run: ./scripts/run-examples.sh
 
       - name: Build IG
         run: ./_genonce.sh

input/queries/add-condition-transaction-example-1.md

@@ -0,0 +1 @@
+Patient?identifier=9912003888&_revinclude=Consent:patient&_revinclude=Flag:patient&_revinclude=Condition:patient

input/queries/add-condition-transaction-example-2.md

@@ -0,0 +1 @@
+Patient?identifier=9912003888&_revinclude=Consent:patient&_revinclude=Flag:patient&_revinclude=Condition:patient

scripts/run-examples.sh

@@ -0,0 +1,40 @@
+#!/bin/bash
+
+# TODO The order of the examples may be important depending on the queries.  Would there be a tidy way of chaining them and makes the depedancies
+# clear, explicit and useful in the guidance?
+# Probably makes more sense just to describe the queries and add output in statically, rather trying to combine integrations testing
+# into the build of the IG.  See if it serves as a useful example... 
+
+SERVER_BASE="http://localhost:8080/fhir/"
+HEADERS="Content-type:application/fhir+json;fhirVersion=4.0"
+
+# TODO - Make the script runnable from any path
+SCRIPT_ROOT=$( cd "$(dirname "${BASH_SOURCE[0]}")" ; pwd -P );
+
+for EXAMPLE in ./input/fsh/examples/*; do
+
+  # Get the filename from the path
+  EXAMPLE=$(echo $EXAMPLE | awk -F '/' '{print $5}');
+
+  # This relies on SUSHI having been ran to generate JSON representations of the resources.
+  GENERATED=$(echo ./fsh-generated/resources/*${EXAMPLE%.*}*);
+
+  # Sushi will prepend the name of the Resource, which will be used in the server calls.
+  RESOURCE_NAME=$(echo $GENERATED | awk -F '/' '{print $4}' | awk -F '-' '{print $1}');
+
+  # Assumption here that Bundle examples will be transactions, guessing it'll fail if not...
+  # Set RESOURCE_NAME to empty string so post is made to the base server url.
+  if [ "$RESOURCE_NAME" == "Bundle" ]; then
+    RESOURCE_NAME=""
+  fi;
+
+  # Post the resource
+  curl -X POST -H $HEADERS -d @${GENERATED} $SERVER_BASE$RESOURCE_NAME;
+
+  # Filename of query (minus extension) must match the filename of the example posted.
+  QUERY=$(cat ./input/queries/${EXAMPLE%.*}*);
+
+  # Run corresponding query, change the ID of the bundle to something readable and output to resources in input to be included in IG build
+  curl $SERVER_BASE$QUERY | jq '.id = "QUERY-OUTPUT--'${EXAMPLE%.*}'"' > ./input/resources/QUERY-OUTPUT--${EXAMPLE%.*}.json
+
+done;