Update S3 crawler Dockerfile to skip library dependencies and add bui…

…ld stamps

crawler/s3/Dockerfile

@@ -1,6 +1,8 @@
+# Repo name: arynai/sycamore-crawler-s3
+
 # In the root directory:
 # docker build -t sycamore_crawler_s3 -f crawler/s3/Dockerfile
-# docker run -it -e AWS_ACCESS_KEY_ID -e AWS_SECRET_ACCESS_KEY -e AWS_SESSION_TOKEN sycamore_crawler_s3
+# docker run -it -v crawl_data:/app/.data/.s3 -e AWS_ACCESS_KEY_ID -e AWS_SECRET_ACCESS_KEY -e AWS_SESSION_TOKEN sycamore_crawler_s3
 FROM python:3
 
 WORKDIR /app
@@ -13,11 +15,24 @@ ENV POETRY_NO_INTERACTION=1 \
     POETRY_CACHE_DIR=/tmp/poetry_cache
 
 COPY pyproject.toml poetry.lock crawler/README.md ./
-RUN poetry install --only main,crawler_s3 --no-root && rm -rf $POETRY_CACHE_DIR
+RUN poetry install --only crawler_s3 --no-root && rm -rf $POETRY_CACHE_DIR
 
 COPY crawler/s3 .
 # Hack beccause pyproject.toml expects a sycamore directory
 RUN mkdir sycamore && touch sycamore/__init__.py
 RUN poetry install --only-root && rm -rf $POETRY_CACHE_DIR
 
+ARG GIT_BRANCH="main"
+ARG GIT_COMMIT="unknown"
+ARG GIT_DIFF="unknown"
+
+ENV GIT_BRANCH=${GIT_BRANCH}
+ENV GIT_COMMIT=${GIT_COMMIT}
+ENV GIT_DIFF=${GIT_DIFF}
+
+LABEL org.opencontainers.image.authors="[email protected]"
+LABEL git_branch=${GIT_BRANCH}
+LABEL git_commit=${GIT_COMMIT}
+LABEL git_diff=${GIT_DIFF}
+
 ENTRYPOINT [ "poetry", "run", "python", "./crawler/s3_crawler.py"]

crawler/s3/crawler/s3_crawler.py

@@ -85,6 +85,10 @@ def _get_s3_client(
 
 
 if __name__ == "__main__":
+    print("Version-Info, Sycamore Crawler S3 Branch:", os.environ.get("GIT_BRANCH", "unset"))
+    print("Version-Info, Sycamore Crawler S3 Commit:", os.environ.get("GIT_COMMIT", "unset"))
+    print("Version-Info, Sycamore Crawler S3 Diff:", os.environ.get("GIT_DIFF", "unset"))
+
     if len(sys.argv) > 3 or (len(sys.argv) > 1 and sys.argv[1] == "-h"):
         print("Usage : poetry run python s3_crawler.py bucket_name prefix_value")
     else: