Extremely basic e2e test to prove the file is valid

Author: the80srobot

Source/santad/Logs/EndpointSecurity/ParquetLogger/e2e_test/Dockerfile

@@ -0,0 +1,3 @@
+FROM quay.io/jupyter/datascience-notebook:2023-11-17
+
+COPY check_parquet_file.py /home/jovyan/work/check_parquet_file.py

Source/santad/Logs/EndpointSecurity/ParquetLogger/e2e_test/README.md

@@ -0,0 +1,3 @@
+This is a really basic end-to-end test to confirm that the parquet file
+generated with `write_test_file.cc` is valid and can be read by Arrow (via
+pyarrow via pandas via ipython).

Source/santad/Logs/EndpointSecurity/ParquetLogger/e2e_test/check_parquet_file.py

@@ -0,0 +1,11 @@
+#!/usr/bin/env python
+
+import pandas as pd
+import sys
+
+df = pd.read_parquet(sys.argv[1])
+if len(df) != 4:
+    print("Expected 4 rows, got %d" % len(df))
+    sys.exit(1)
+
+sys.exit(0)

Source/santad/Logs/EndpointSecurity/ParquetLogger/e2e_test/e2e_test.sh

@@ -0,0 +1,24 @@
+#!/usr/bin/env bash
+
+set -e
+
+SCRIPT_DIR=$(cd -- "$( dirname -- "${BASH_SOURCE[0]}")" &> /dev/null && pwd )
+ROOT="$(bazel info workspace)"
+TMPDIR="$(mktemp -d)"
+
+# Go to the test directory and build the Docker image with Pandas.
+pushd "${SCRIPT_DIR}" > /dev/null
+docker build . -t parquet_logger_e2e_test
+
+# Generate the test file
+pushd .. > /dev/null
+bazel run :write_test_file -- "${TMPDIR}/test_file.parquet"
+
+# Run the test in a Docker container
+docker run \
+  --rm \
+  -v "${TMPDIR}:/tmp" \
+  parquet_logger_e2e_test \
+  /usr/bin/env python /home/jovyan/work/check_parquet_file.py /tmp/test_file.parquet
+
+popd > /dev/null