From 1a66f49d15cc885d3e668c8087c4fc77db68d127 Mon Sep 17 00:00:00 2001
From: Christopher Tabone <christabone@gmail.com>
Date: Tue, 17 Feb 2026 00:34:08 +0000
Subject: [PATCH] Add scRNAseq metadata copy-forward script

Shell script that copies scRNAseq metadata files from the previous
release to the current one, renaming them appropriately. These files
don't change between releases (until a new scRNAseq curator is hired).

Reads RELEASE, PREV_RELEASE, and BUILD_DIR from GoCD environment.
Intended to be wired up as a GoCD stage before other bulk report scripts.
---
 src/copy_scrna_metadata.sh | 65 ++++++++++++++++++++++++++++++++++++++
 1 file changed, 65 insertions(+)
 create mode 100755 src/copy_scrna_metadata.sh

diff --git a/src/copy_scrna_metadata.sh b/src/copy_scrna_metadata.sh
new file mode 100755
index 0000000..0f926d1
--- /dev/null
+++ b/src/copy_scrna_metadata.sh
@@ -0,0 +1,65 @@
+#!/bin/bash
+#
+# copy_scrna_metadata.sh
+#
+# Copies scRNAseq metadata files forward from the previous release,
+# renaming them for the current release. These files don't change
+# between releases (until a new scRNAseq curator is hired at Cambridge).
+#
+# Expects GoCD environment variables:
+#   RELEASE          - current release (e.g., 2025_06)
+#   PREV_RELEASE     - previous release (e.g., 2025_05)
+#   BUILD_DIR        - build directory (e.g., /data/build-public-release)
+#
+# Usage:
+#   copy_scrna_metadata.sh
+#
+
+set -euo pipefail
+
+# Validate required environment variables
+for var in RELEASE PREV_RELEASE BUILD_DIR; do
+    if [ -z "${!var:-}" ]; then
+        echo "ERROR: Required environment variable $var is not set." >&2
+        exit 1
+    fi
+done
+
+PREV_DIR="${BUILD_DIR}/fb_${PREV_RELEASE}_reporting/bulk_reports"
+CURR_DIR="${BUILD_DIR}/fb_${RELEASE}_reporting/bulk_reports"
+
+FILES=(
+    "scRNAseq_metadata_fb_YYYY_MM_reporting.json"
+    "scRNAseq_metadata_schema_fb_YYYY_MM_reporting.yaml"
+)
+
+echo "Copying scRNAseq metadata from ${PREV_RELEASE} to ${RELEASE}..."
+echo "  Source: ${PREV_DIR}"
+echo "  Dest:   ${CURR_DIR}"
+
+# Verify source directory exists
+if [ ! -d "${PREV_DIR}" ]; then
+    echo "ERROR: Previous release directory not found: ${PREV_DIR}" >&2
+    exit 1
+fi
+
+# Verify destination directory exists
+if [ ! -d "${CURR_DIR}" ]; then
+    echo "ERROR: Current release directory not found: ${CURR_DIR}" >&2
+    exit 1
+fi
+
+for template in "${FILES[@]}"; do
+    src_file="${PREV_DIR}/${template//YYYY_MM/${PREV_RELEASE}}"
+    dst_file="${CURR_DIR}/${template//YYYY_MM/${RELEASE}}"
+
+    if [ ! -f "${src_file}" ]; then
+        echo "ERROR: Source file not found: ${src_file}" >&2
+        exit 1
+    fi
+
+    cp "${src_file}" "${dst_file}"
+    echo "  OK: $(basename "${dst_file}") ($(stat -c%s "${dst_file}") bytes)"
+done
+
+echo "Done. scRNAseq metadata files copied successfully."