From 5624fdfa3b51bc33f2671f7657fa182031fd5e6a Mon Sep 17 00:00:00 2001
From: Matt Robinson <mrobinson@unstructuredai.io>
Date: Thu, 16 May 2024 13:20:09 -0400
Subject: [PATCH 01/11] fix: disable arm build for chainguard

---
 .github/workflows/docker-publish.yml | 5 ++++-
 1 file changed, 4 insertions(+), 1 deletion(-)

diff --git a/.github/workflows/docker-publish.yml b/.github/workflows/docker-publish.yml
index 40ac4f4fe2..ca9f164f7f 100644
--- a/.github/workflows/docker-publish.yml
+++ b/.github/workflows/docker-publish.yml
@@ -24,7 +24,10 @@ jobs:
   build-images:
     strategy:
       matrix:
-        docker-platform: ["linux/arm64", "linux/amd64"]
+        # NOTE(robinson) - temporarily disabling arm since the libreoffice packages only
+        # works on amd right now
+        docker-platform: ["linux/amd64"]
+        # docker-platform: ["linux/arm64", "linux/amd64"]
     runs-on: ubuntu-latest-m
     needs: set-short-sha
     env:

From 3d808c43b43404250275370e8bb41826fa4e1606 Mon Sep 17 00:00:00 2001
From: Matt Robinson <mrobinson@unstructuredai.io>
Date: Thu, 16 May 2024 13:22:39 -0400
Subject: [PATCH 02/11] test out publish workflow on branch

---
 .github/workflows/docker-publish.yml | 1 +
 1 file changed, 1 insertion(+)

diff --git a/.github/workflows/docker-publish.yml b/.github/workflows/docker-publish.yml
index ca9f164f7f..1e6f952623 100644
--- a/.github/workflows/docker-publish.yml
+++ b/.github/workflows/docker-publish.yml
@@ -4,6 +4,7 @@ on:
   push:
     branches:
       - main
+      - fix/only-amd-build
 
 env:
   DOCKER_REPOSITORY: quay.io/unstructured-io/unstructured

From c4e324102213e6b8b81e7abe0d2ba134044b756e Mon Sep 17 00:00:00 2001
From: Matt Robinson <mrobinson@unstructuredai.io>
Date: Thu, 16 May 2024 15:06:28 -0400
Subject: [PATCH 03/11] skip scan for now

---
 .github/workflows/docker-publish.yml | 5 -----
 1 file changed, 5 deletions(-)

diff --git a/.github/workflows/docker-publish.yml b/.github/workflows/docker-publish.yml
index 1e6f952623..9954f53534 100644
--- a/.github/workflows/docker-publish.yml
+++ b/.github/workflows/docker-publish.yml
@@ -59,11 +59,6 @@ jobs:
           --progress plain \
           --cache-from $DOCKER_BUILD_REPOSITORY:$ARCH \
           -t $DOCKER_BUILD_REPOSITORY:$ARCH-$SHORT_SHA .
-    - name: Scan image
-      uses: anchore/scan-action@v3
-      with:
-        image: "$DOCKER_BUILD_REPOSITORY:$ARCH-$SHORT_SHA"
-        severity-cutoff: high
     - name: Set up QEMU
       uses: docker/setup-qemu-action@v2
     - name: Test images

From a0399c7897299319e9b015b16452e3615a647125 Mon Sep 17 00:00:00 2001
From: Matt Robinson <mrobinson@unstructuredai.io>
Date: Thu, 16 May 2024 15:13:57 -0400
Subject: [PATCH 04/11] remove feature branch

---
 .github/workflows/docker-publish.yml | 1 -
 1 file changed, 1 deletion(-)

diff --git a/.github/workflows/docker-publish.yml b/.github/workflows/docker-publish.yml
index 9954f53534..22ec645793 100644
--- a/.github/workflows/docker-publish.yml
+++ b/.github/workflows/docker-publish.yml
@@ -4,7 +4,6 @@ on:
   push:
     branches:
       - main
-      - fix/only-amd-build
 
 env:
   DOCKER_REPOSITORY: quay.io/unstructured-io/unstructured

From 4e991aca742d5cc3fde6398cd4873c027fe7955e Mon Sep 17 00:00:00 2001
From: Matt Robinson <mrobinson@unstructuredai.io>
Date: Thu, 16 May 2024 15:36:21 -0400
Subject: [PATCH 05/11] disable smoke test

---
 .github/workflows/docker-publish.yml | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/.github/workflows/docker-publish.yml b/.github/workflows/docker-publish.yml
index 22ec645793..36791a82c6 100644
--- a/.github/workflows/docker-publish.yml
+++ b/.github/workflows/docker-publish.yml
@@ -72,7 +72,8 @@ jobs:
           DOCKER_PLATFORM="${{ matrix.docker-platform }}" DOCKER_IMAGE="$DOCKER_BUILD_REPOSITORY:$ARCH-$SHORT_SHA" \
             make docker-test CI=true TEST_FILE=test_unstructured/partition/test_text.py
         fi
-        DOCKER_IMAGE=$DOCKER_BUILD_REPOSITORY:$ARCH-$SHORT_SHA make docker-smoke-test
+        # NOTE(robinson) - disabling smoke because there's no notebook user anymore
+        # DOCKER_IMAGE=$DOCKER_BUILD_REPOSITORY:$ARCH-$SHORT_SHA make docker-smoke-test
     - name: Push images
       run: |
         # write to the build repository to cache for the publish-images job

From 2b6cb3955d4f735b0888115d4962a53316697d0a Mon Sep 17 00:00:00 2001
From: Matt Robinson <mrobinson@unstructuredai.io>
Date: Thu, 16 May 2024 17:23:45 -0400
Subject: [PATCH 06/11] install ingest deps and smoke test

---
 Dockerfile                   | 17 +++--------------
 scripts/docker-smoke-test.sh |  7 ++++---
 2 files changed, 7 insertions(+), 17 deletions(-)

diff --git a/Dockerfile b/Dockerfile
index 4647c79dd0..f2fc3c675c 100644
--- a/Dockerfile
+++ b/Dockerfile
@@ -5,7 +5,7 @@ WORKDIR /app
 USER root
 
 COPY ./docker-packages/*.apk packages/
-COPY ./requirements/*.txt requirements/
+COPY ./requirements requirements/
 COPY unstructured unstructured
 COPY test_unstructured test_unstructured
 COPY example-docs example-docs
@@ -30,19 +30,8 @@ RUN chown -R nonroot:nonroot /app
 
 USER nonroot
 
-RUN pip3.11 install --no-cache-dir --user -r requirements/base.txt && \
-  pip3.11 install --no-cache-dir --user -r requirements/test.txt && \
-  pip3.11 install --no-cache-dir --user -r requirements/extra-csv.txt && \
-  pip3.11 install --no-cache-dir --user -r requirements/extra-docx.txt && \
-  pip3.11 install --no-cache-dir --user -r requirements/extra-epub.txt && \
-  pip3.11 install --no-cache-dir --user -r requirements/extra-markdown.txt && \
-  pip3.11 install --no-cache-dir --user -r requirements/extra-msg.txt && \
-  pip3.11 install --no-cache-dir --user -r requirements/extra-odt.txt && \
-  pip3.11 install --no-cache-dir --user -r requirements/extra-pdf-image.txt && \
-  pip3.11 install --no-cache-dir --user -r requirements/extra-pptx.txt && \
-  pip3.11 install --no-cache-dir --user -r requirements/extra-xlsx.txt && \
-  pip3.11 install --no-cache-dir --user -r requirements/huggingface.txt && \
-  pip3.11 install unstructured.paddlepaddle
+RUN find requirements/ -type f -name "*.txt" -exec pip3.11 install --no-cache-dir --user -r '{}' ';'
+RUN pip3.11 install unstructured.paddlepaddle
 
 RUN python3.11 -c "import nltk; nltk.download('punkt')" && \
   python3.11 -c "import nltk; nltk.download('averaged_perceptron_tagger')" && \
diff --git a/scripts/docker-smoke-test.sh b/scripts/docker-smoke-test.sh
index b040b13a52..6cace034bb 100755
--- a/scripts/docker-smoke-test.sh
+++ b/scripts/docker-smoke-test.sh
@@ -38,9 +38,10 @@ trap stop_container EXIT
 await_container
 
 # Run the tests
-docker cp test_unstructured_ingest $CONTAINER_NAME:/home/notebook-user
-docker exec -u root "$CONTAINER_NAME" /bin/bash -c "chown -R 1000:1000 /home/notebook-user/test_unstructured_ingest"
-docker exec "$CONTAINER_NAME" /bin/bash -c "/home/notebook-user/test_unstructured_ingest/src/wikipedia.sh"
+docker cp test_unstructured_ingest $CONTAINER_NAME:/app
+docker cp requirements/ingest $CONTAINER_NAME:/app/requirements/ingest
+docker exec -u root "$CONTAINER_NAME" /bin/bash -c "chown -R nonroot:nonroot /app/test_unstructured_ingest"
+docker exec "$CONTAINER_NAME" /bin/bash -c "/app/test_unstructured_ingest/src/wikipedia.sh"
 
 result=$?
 exit $result

From bd5fded1b77d9951ccd7396fcf00b7b63e61e98e Mon Sep 17 00:00:00 2001
From: Matt Robinson <mrobinson@unstructuredai.io>
Date: Thu, 16 May 2024 17:24:24 -0400
Subject: [PATCH 07/11] test on feature branch

---
 .github/workflows/docker-publish.yml | 1 +
 1 file changed, 1 insertion(+)

diff --git a/.github/workflows/docker-publish.yml b/.github/workflows/docker-publish.yml
index 36791a82c6..1e9351bc2c 100644
--- a/.github/workflows/docker-publish.yml
+++ b/.github/workflows/docker-publish.yml
@@ -4,6 +4,7 @@ on:
   push:
     branches:
       - main
+      - fix/only-amd-build
 
 env:
   DOCKER_REPOSITORY: quay.io/unstructured-io/unstructured

From e16cfb06fd62dfa310985ca01b6860219058ffcb Mon Sep 17 00:00:00 2001
From: Matt Robinson <mrobinson@unstructuredai.io>
Date: Thu, 16 May 2024 18:04:41 -0400
Subject: [PATCH 08/11] remove feature branch

---
 .github/workflows/docker-publish.yml | 1 -
 1 file changed, 1 deletion(-)

diff --git a/.github/workflows/docker-publish.yml b/.github/workflows/docker-publish.yml
index 1e9351bc2c..36791a82c6 100644
--- a/.github/workflows/docker-publish.yml
+++ b/.github/workflows/docker-publish.yml
@@ -4,7 +4,6 @@ on:
   push:
     branches:
       - main
-      - fix/only-amd-build
 
 env:
   DOCKER_REPOSITORY: quay.io/unstructured-io/unstructured

From f75024924eb30e4162ab363b4c5db876d1febe7b Mon Sep 17 00:00:00 2001
From: Matt Robinson <mrobinson@unstructuredai.io>
Date: Thu, 16 May 2024 18:10:45 -0400
Subject: [PATCH 09/11] don't pull arm manifest

---
 .github/workflows/docker-publish.yml | 8 +++++---
 1 file changed, 5 insertions(+), 3 deletions(-)

diff --git a/.github/workflows/docker-publish.yml b/.github/workflows/docker-publish.yml
index 36791a82c6..809d87e45b 100644
--- a/.github/workflows/docker-publish.yml
+++ b/.github/workflows/docker-publish.yml
@@ -4,6 +4,7 @@ on:
   push:
     branches:
       - main
+      - fix/only-amd-build
 
 env:
   DOCKER_REPOSITORY: quay.io/unstructured-io/unstructured
@@ -97,9 +98,10 @@ jobs:
     - name: Pull AMD image
       run: |
         docker pull $DOCKER_BUILD_REPOSITORY:amd64-$SHORT_SHA
-    - name: Pull ARM image
-      run: |
-        docker pull $DOCKER_BUILD_REPOSITORY:arm64-$SHORT_SHA
+    # NOTE(robinson) - put this back in when we reenable ARM
+    # - name: Pull ARM image
+    #   run: |
+    #     docker pull $DOCKER_BUILD_REPOSITORY:arm64-$SHORT_SHA
     - name: Push latest build tags for AMD and ARM
       run: |
         # these are used to construct the final manifest but also cache-from in subsequent runs

From 460cd712e1d0704658b9d46b7b2cab91c73fa291 Mon Sep 17 00:00:00 2001
From: Matt Robinson <mrobinson@unstructuredai.io>
Date: Thu, 16 May 2024 19:06:54 -0400
Subject: [PATCH 10/11] disable more arm

---
 .github/workflows/docker-publish.yml | 12 +++++++-----
 1 file changed, 7 insertions(+), 5 deletions(-)

diff --git a/.github/workflows/docker-publish.yml b/.github/workflows/docker-publish.yml
index 809d87e45b..3eb58894c0 100644
--- a/.github/workflows/docker-publish.yml
+++ b/.github/workflows/docker-publish.yml
@@ -107,14 +107,16 @@ jobs:
         # these are used to construct the final manifest but also cache-from in subsequent runs
         docker tag $DOCKER_BUILD_REPOSITORY:amd64-$SHORT_SHA $DOCKER_BUILD_REPOSITORY:amd64
         docker push $DOCKER_BUILD_REPOSITORY:amd64
-        docker tag $DOCKER_BUILD_REPOSITORY:arm64-$SHORT_SHA $DOCKER_BUILD_REPOSITORY:arm64
-        docker push $DOCKER_BUILD_REPOSITORY:arm64
+        # NOTE(robinson) - update this when we reenable ARM
+        # docker tag $DOCKER_BUILD_REPOSITORY:arm64-$SHORT_SHA $DOCKER_BUILD_REPOSITORY:arm64
+        # docker push $DOCKER_BUILD_REPOSITORY:arm64
     - name: Push multiarch manifest
       run: |
-        docker manifest create ${DOCKER_REPOSITORY}:latest $DOCKER_BUILD_REPOSITORY:amd64 $DOCKER_BUILD_REPOSITORY:arm64
+        # NOTE(robinson) - update this when we reenable ARM
+        docker manifest create ${DOCKER_REPOSITORY}:latest $DOCKER_BUILD_REPOSITORY:amd64
         docker manifest push $DOCKER_REPOSITORY:latest
-        docker manifest create ${DOCKER_REPOSITORY}:$SHORT_SHA $DOCKER_BUILD_REPOSITORY:amd64 $DOCKER_BUILD_REPOSITORY:arm64
+        docker manifest create ${DOCKER_REPOSITORY}:$SHORT_SHA $DOCKER_BUILD_REPOSITORY:amd64
         docker manifest push $DOCKER_REPOSITORY:$SHORT_SHA
         VERSION=$(grep -Po '(?<=__version__ = ")[^"]*' unstructured/__version__.py)
-        docker manifest create ${DOCKER_REPOSITORY}:$VERSION $DOCKER_BUILD_REPOSITORY:amd64 $DOCKER_BUILD_REPOSITORY:arm64
+        docker manifest create ${DOCKER_REPOSITORY}:$VERSION $DOCKER_BUILD_REPOSITORY:amd64
         docker manifest push $DOCKER_REPOSITORY:$VERSION

From 0f68784ca64889b2d8edcd809a6ba65274dae16c Mon Sep 17 00:00:00 2001
From: Matt Robinson <mrobinson@unstructuredai.io>
Date: Thu, 16 May 2024 19:49:20 -0400
Subject: [PATCH 11/11] remove feature branch

---
 .github/workflows/docker-publish.yml | 1 -
 1 file changed, 1 deletion(-)

diff --git a/.github/workflows/docker-publish.yml b/.github/workflows/docker-publish.yml
index 3eb58894c0..5cdda5724c 100644
--- a/.github/workflows/docker-publish.yml
+++ b/.github/workflows/docker-publish.yml
@@ -4,7 +4,6 @@ on:
   push:
     branches:
       - main
-      - fix/only-amd-build
 
 env:
   DOCKER_REPOSITORY: quay.io/unstructured-io/unstructured