From 3e9a730ec39a3642ca2c97fcf340ad4915877aee Mon Sep 17 00:00:00 2001
From: Joe Corall <jjc223@lehigh.edu>
Date: Fri, 4 Oct 2024 15:22:00 -0400
Subject: [PATCH] Add an integration test for whisper (#45)

---
 ci/docker-compose.yml       | 2 ++
 ci/run.sh                   | 2 +-
 ci/test.sh                  | 9 +++++++++
 examples/whisper/Dockerfile | 2 ++
 examples/whisper/cmd.sh     | 4 ----
 5 files changed, 14 insertions(+), 5 deletions(-)

diff --git a/ci/docker-compose.yml b/ci/docker-compose.yml
index 76d4d39..480e4c1 100644
--- a/ci/docker-compose.yml
+++ b/ci/docker-compose.yml
@@ -11,6 +11,8 @@ services:
         image: ${DOCKER_REPOSITORY_BASE}-imagemagick:${DOCKER_TAG}
     tesseract:
         image: ${DOCKER_REPOSITORY_BASE}-tesseract:${DOCKER_TAG}
+    whisper:
+        image: ${DOCKER_REPOSITORY_BASE}-whisper:${DOCKER_TAG}
     test:
         image: alpine
         command: sleep 300
diff --git a/ci/run.sh b/ci/run.sh
index 1b217a7..79c7018 100755
--- a/ci/run.sh
+++ b/ci/run.sh
@@ -12,4 +12,4 @@ docker exec ci-test-1 /test.sh
 echo $?
 
 docker compose down 2>&1 > /dev/null
-docker compose rm 2>&1 > /dev/null
\ No newline at end of file
+docker compose rm 2>&1 > /dev/null
diff --git a/ci/test.sh b/ci/test.sh
index 335dfd5..0b985b0 100755
--- a/ci/test.sh
+++ b/ci/test.sh
@@ -15,6 +15,7 @@ SERVICES=(
   "imagemagick"
   "crayfits"
   "ffmpeg"
+  "whisper"
 )
 for SERVICE in "${SERVICES[@]}"; do
   URL="http://$SERVICE:8080/"
@@ -58,6 +59,14 @@ for SERVICE in "${SERVICES[@]}"; do
     grep "One time I was ridin' along on the mule" ocr.txt || exit 1
     echo "PDF OCR as expected"
     rm ocr.txt
+  elif [ "$SERVICE" == "whisper" ]; then
+    curl -s -o vtt.txt \
+        --header "Accept: text/plain" \
+        --header "Apix-Ldp-Resource: https://github.com/ggerganov/whisper.cpp/raw/master/samples/jfk.wav" \
+        "$URL"
+    grep "ask not what your country can do for you" vtt.txt || exit 1
+    echo "VTT as expected"
+    rm vtt.txt
   else
     echo "Unknown service"
     exit 1
diff --git a/examples/whisper/Dockerfile b/examples/whisper/Dockerfile
index 1340dfe..4d10d46 100644
--- a/examples/whisper/Dockerfile
+++ b/examples/whisper/Dockerfile
@@ -7,6 +7,8 @@ COPY --from=scyllaridae /app /app
 COPY scyllaridae.yml /app
 COPY cmd.sh /app
 
+RUN bash ./models/download-ggml-model.sh medium.en
+
 RUN apt-get update && apt-get install -y gosu=1.14-1 --no-install-recommends \
   && rm -rf /var/lib/apt/lists/* \
   && groupadd -r nobody \
diff --git a/examples/whisper/cmd.sh b/examples/whisper/cmd.sh
index a4c7c74..e710b63 100755
--- a/examples/whisper/cmd.sh
+++ b/examples/whisper/cmd.sh
@@ -4,10 +4,6 @@
 
 set -eou pipefail
 
-if [ ! -f /app/models/ggml-medium.en.bin ]; then
-  bash ./models/download-ggml-model.sh medium.en > /dev/null 2>&1
-fi
-
 # take stdin and buffer it into a temp file
 input_temp=$(mktemp /tmp/whisper-input-XXXXXX)
 cat > "$input_temp"