feat(docker): respect pip mirrors with uv (datahub-project#9963)

alexs-101 · Mar 1, 2024 · 5a516f8 · 5a516f8
1 parent 4aa1dc3
commit 5a516f8
Show file tree

Hide file tree

Showing 3 changed files with 29 additions and 24 deletions.
diff --git a/docker/datahub-ingestion-base/Dockerfile b/docker/datahub-ingestion-base/Dockerfile
@@ -5,7 +5,7 @@ ARG BASE_IMAGE=base
 ARG ALPINE_REPO_URL=http://dl-cdn.alpinelinux.org/alpine
 ARG GITHUB_REPO_URL=https://github.com
 ARG DEBIAN_REPO_URL=https://deb.debian.org/debian
-ARG PIP_MIRROR_URL=null
+ARG PIP_MIRROR_URL=https://pypi.python.org/simple
 
 FROM golang:1-alpine3.18 AS dockerize-binary
 
@@ -26,15 +26,18 @@ RUN go install github.com/jwilder/dockerize@$DOCKERIZE_VERSION
 
 FROM python:3.10 as base
 
-ARG DEBIAN_REPO_URL
-ARG PIP_MIRROR_URL
 ARG GITHUB_REPO_URL
 
 ENV DEBIAN_FRONTEND noninteractive
 
-# Optionally set corporate mirror for apk and pip
+# Optionally set corporate mirror for deb
+ARG DEBIAN_REPO_URL
 RUN if [ "${DEBIAN_REPO_URL}" != "http://deb.debian.org/debian" ] ; then sed -i "s#http.*://deb.debian.org/debian#${DEBIAN_REPO_URL}#g" /etc/apt/sources.list.d/debian.sources ; fi
-RUN if [ "${PIP_MIRROR_URL}" != "null" ] ; then pip config set global.index-url ${PIP_MIRROR_URL} ; fi
+
+# Optionally set corporate mirror for pip
+ARG PIP_MIRROR_URL
+RUN if [ "${PIP_MIRROR_URL}" != "https://pypi.python.org/simple" ] ; then pip config set global.index-url ${PIP_MIRROR_URL} ; fi
+ENV UV_INDEX_URL=${PIP_MIRROR_URL}
 
 RUN apt-get update && apt-get install -y -qq \
     python3-ldap \
@@ -67,8 +70,7 @@ USER datahub
 ENV VIRTUAL_ENV=/datahub-ingestion/.venv
 ENV PATH="${VIRTUAL_ENV}/bin:$PATH"
 RUN python3 -m venv $VIRTUAL_ENV && \
-    uv pip install --no-cache -r requirements.txt && \
-    pip uninstall -y acryl-datahub
+    uv pip install --no-cache -r requirements.txt
 
 ENTRYPOINT [ "/entrypoint.sh" ]
 

diff --git a/docker/datahub-ingestion/Dockerfile b/docker/datahub-ingestion/Dockerfile
@@ -2,12 +2,22 @@
 ARG APP_ENV=full
 ARG BASE_IMAGE=acryldata/datahub-ingestion-base
 ARG DOCKER_VERSION=head
-ARG PIP_MIRROR_URL=null
 ARG DEBIAN_REPO_URL=https://deb.debian.org/debian
+ARG PIP_MIRROR_URL=https://pypi.python.org/simple
 
 FROM $BASE_IMAGE:$DOCKER_VERSION as base
+
+# Optionally set corporate mirror for deb
+USER 0
+ARG DEBIAN_REPO_URL
+RUN if [ "${DEBIAN_REPO_URL}" != "http://deb.debian.org/debian" ] ; then sed -i "s#http.*://deb.debian.org/debian#${DEBIAN_REPO_URL}#g" /etc/apt/sources.list.d/debian.sources ; fi
 USER datahub
 
+# Optionally set corporate mirror for pip
+ARG PIP_MIRROR_URL
+RUN if [ "${PIP_MIRROR_URL}" != "https://pypi.python.org/simple" ] ; then pip config set global.index-url ${PIP_MIRROR_URL} ; fi
+ENV UV_INDEX_URL=${PIP_MIRROR_URL}
+
 COPY --chown=datahub ./metadata-ingestion /datahub-ingestion
 COPY --chown=datahub ./metadata-ingestion-modules/airflow-plugin /datahub-ingestion/airflow-plugin
 
@@ -19,23 +29,17 @@ RUN sed -i.bak "s/__version__ = \"1\!0.0.0.dev0\"/__version__ = \"$(echo $RELEAS
     cat airflow-plugin/src/datahub_airflow_plugin/__init__.py | grep __version__
 
 FROM base as slim-install
-ARG PIP_MIRROR_URL
 
-RUN if [ "${PIP_MIRROR_URL}" != "null" ] ; then pip config set global.index-url ${PIP_MIRROR_URL} ; fi
 RUN uv pip install --no-cache "acryl-datahub[base,datahub-rest,datahub-kafka,snowflake,bigquery,redshift,mysql,postgres,hive,clickhouse,glue,dbt,looker,lookml,tableau,powerbi,superset,datahub-business-glossary] @ ."
 
 FROM base as full-install-build
-ARG PIP_MIRROR_URL
-ARG DEBIAN_REPO_URL
 
 USER 0
-RUN if [ "${DEBIAN_REPO_URL}" != "http://deb.debian.org/debian" ] ; then sed -i "s#http.*://deb.debian.org/debian#${DEBIAN_REPO_URL}#g" /etc/apt/sources.list.d/debian.sources ; fi
 RUN apt-get update && apt-get install -y -qq maven
 
 USER datahub
 COPY ./docker/datahub-ingestion/pyspark_jars.sh .
 
-RUN if [ "${PIP_MIRROR_URL}" != "null" ] ; then pip config set global.index-url ${PIP_MIRROR_URL} ; fi
 RUN uv pip install --no-cache "acryl-datahub[base,all] @ ." "acryl-datahub-airflow-plugin[plugin-v2] @ ./airflow-plugin" && \
     datahub --version
 RUN ./pyspark_jars.sh

diff --git a/docker/datahub-ingestion/Dockerfile-slim-only b/docker/datahub-ingestion/Dockerfile-slim-only
@@ -1,26 +1,25 @@
 # Defining environment
 ARG BASE_IMAGE=acryldata/datahub-ingestion-base
 ARG DOCKER_VERSION=head-slim
-ARG PIP_MIRROR_URL=null
+ARG PIP_MIRROR_URL=https://pypi.python.org/simple
 
 FROM $BASE_IMAGE:$DOCKER_VERSION as base
-USER 0
+USER datahub
+
+# Optionally set corporate mirror for apk and pip
+ARG PIP_MIRROR_URL
+RUN if [ "${PIP_MIRROR_URL}" != "https://pypi.python.org/simple" ] ; then pip config set global.index-url ${PIP_MIRROR_URL} ; fi
+ENV UV_INDEX_URL=${PIP_MIRROR_URL}
 
-COPY ./metadata-ingestion /datahub-ingestion
+COPY --chown=datahub ./metadata-ingestion /datahub-ingestion
 
 ARG RELEASE_VERSION
 WORKDIR /datahub-ingestion
 RUN sed -i.bak "s/__version__ = \"1\!0.0.0.dev0\"/__version__ = \"$(echo $RELEASE_VERSION|sed s/-/+/)\"/" src/datahub/__init__.py && \
-    cat src/datahub/__init__.py && \
-    chown -R datahub /datahub-ingestion
-
-USER datahub
+    cat src/datahub/__init__.py
 
 FROM base as slim-install
 
-ARG PIP_MIRROR_URL
-
-RUN if [ "${PIP_MIRROR_URL}" != "null" ] ; then pip config set global.index-url ${PIP_MIRROR_URL} ; fi
 RUN uv pip install --no-cache "acryl-datahub[base,datahub-rest,datahub-kafka,snowflake,bigquery,redshift,mysql,postgres,hive,clickhouse,glue,dbt,looker,lookml,tableau,powerbi,superset,datahub-business-glossary] @ ." && \
     datahub --version