Dockerfile.catalog-sync

# Dockerfile.catalog-sync
#
# Multi-stage build that pulls operator catalog index images from
# registry.redhat.io via FROM instructions (uses node-level CRI-O
# credentials in CI, bypassing the need for explicit auth files).
#
# Extracts /configs/ from each image, processes them with
# catalog_metadata.py to generate operators.json, dependencies.json,
# catalog-info.json, and assembles catalog-index.json.
#
# This Dockerfile is used by ci-operator in the catalog-sync variant.
# The output image provides catalog-data/ which is injected into the
# main mirror-gui image build via inputs.paths.
#
# For local development, use the regular Dockerfile instead.

# --- Stage 1: Pull all 18 catalog index images ---
FROM registry.redhat.io/redhat/redhat-operator-index:v4.16 AS redhat-4-16
FROM registry.redhat.io/redhat/redhat-operator-index:v4.17 AS redhat-4-17
FROM registry.redhat.io/redhat/redhat-operator-index:v4.18 AS redhat-4-18
FROM registry.redhat.io/redhat/redhat-operator-index:v4.19 AS redhat-4-19
FROM registry.redhat.io/redhat/redhat-operator-index:v4.20 AS redhat-4-20
FROM registry.redhat.io/redhat/redhat-operator-index:v4.21 AS redhat-4-21
FROM registry.redhat.io/redhat/certified-operator-index:v4.16 AS certified-4-16
FROM registry.redhat.io/redhat/certified-operator-index:v4.17 AS certified-4-17
FROM registry.redhat.io/redhat/certified-operator-index:v4.18 AS certified-4-18
FROM registry.redhat.io/redhat/certified-operator-index:v4.19 AS certified-4-19
FROM registry.redhat.io/redhat/certified-operator-index:v4.20 AS certified-4-20
FROM registry.redhat.io/redhat/certified-operator-index:v4.21 AS certified-4-21
FROM registry.redhat.io/redhat/community-operator-index:v4.16 AS community-4-16
FROM registry.redhat.io/redhat/community-operator-index:v4.17 AS community-4-17
FROM registry.redhat.io/redhat/community-operator-index:v4.18 AS community-4-18
FROM registry.redhat.io/redhat/community-operator-index:v4.19 AS community-4-19
FROM registry.redhat.io/redhat/community-operator-index:v4.20 AS community-4-20
FROM registry.redhat.io/redhat/community-operator-index:v4.21 AS community-4-21

# --- Stage 2: Extract and process catalog data ---
FROM node:22-slim AS processor
WORKDIR /app

RUN apt-get update && \
    apt-get install -y --no-install-recommends jq python3 python3-yaml && \
    rm -rf /var/lib/apt/lists/*

COPY scripts/catalog_metadata.py ./scripts/catalog_metadata.py

COPY --from=redhat-4-16 /configs catalog-data-synced/redhat-operator-index/v4.16/configs/
COPY --from=redhat-4-17 /configs catalog-data-synced/redhat-operator-index/v4.17/configs/
COPY --from=redhat-4-18 /configs catalog-data-synced/redhat-operator-index/v4.18/configs/
COPY --from=redhat-4-19 /configs catalog-data-synced/redhat-operator-index/v4.19/configs/
COPY --from=redhat-4-20 /configs catalog-data-synced/redhat-operator-index/v4.20/configs/
COPY --from=redhat-4-21 /configs catalog-data-synced/redhat-operator-index/v4.21/configs/
COPY --from=certified-4-16 /configs catalog-data-synced/certified-operator-index/v4.16/configs/
COPY --from=certified-4-17 /configs catalog-data-synced/certified-operator-index/v4.17/configs/
COPY --from=certified-4-18 /configs catalog-data-synced/certified-operator-index/v4.18/configs/
COPY --from=certified-4-19 /configs catalog-data-synced/certified-operator-index/v4.19/configs/
COPY --from=certified-4-20 /configs catalog-data-synced/certified-operator-index/v4.20/configs/
COPY --from=certified-4-21 /configs catalog-data-synced/certified-operator-index/v4.21/configs/
COPY --from=community-4-16 /configs catalog-data-synced/community-operator-index/v4.16/configs/
COPY --from=community-4-17 /configs catalog-data-synced/community-operator-index/v4.17/configs/
COPY --from=community-4-18 /configs catalog-data-synced/community-operator-index/v4.18/configs/
COPY --from=community-4-19 /configs catalog-data-synced/community-operator-index/v4.19/configs/
COPY --from=community-4-20 /configs catalog-data-synced/community-operator-index/v4.20/configs/
COPY --from=community-4-21 /configs catalog-data-synced/community-operator-index/v4.21/configs/

SHELL ["/bin/bash", "-c"]
RUN set -euo pipefail && \
    CATALOG_DATA_DIR=./catalog-data-synced && \
    OCP_VERSIONS="4.16 4.17 4.18 4.19 4.20 4.21" && \
    CATALOG_TYPES="redhat-operator-index certified-operator-index community-operator-index" && \
    TOTAL=0 && SUCCESS=0 && FAILED=0 && \
    for ocp_version in $OCP_VERSIONS; do \
      for catalog_type in $CATALOG_TYPES; do \
        TOTAL=$((TOTAL + 1)); \
        catalog_dir="${CATALOG_DATA_DIR}/${catalog_type}/v${ocp_version}"; \
        if [ ! -d "${catalog_dir}/configs" ]; then \
          echo "WARN: No configs for ${catalog_type} v${ocp_version}, skipping"; \
          FAILED=$((FAILED + 1)); \
          continue; \
        fi; \
        operators_file="${catalog_dir}/operators.json"; \
        dependencies_file="${catalog_dir}/dependencies.json"; \
        if python3 ./scripts/catalog_metadata.py generate \
            --catalog-dir "$catalog_dir" \
            --catalog-type "$catalog_type" \
            --ocp-version "v${ocp_version}" \
            --operators-file "$operators_file" \
            --dependencies-file "$dependencies_file"; then \
          operator_count=$(jq '. | length' "$operators_file" 2>/dev/null || echo "0"); \
          synced_at=$(date -u +"%Y-%m-%dT%H:%M:%SZ"); \
          printf '{"catalog_type":"%s","ocp_version":"v%s","catalog_url":"registry.redhat.io/redhat/%s:v%s","operator_count":%s,"digest":"from-build","synced_at":"%s"}\n' \
            "$catalog_type" "$ocp_version" "$catalog_type" "$ocp_version" "$operator_count" "$synced_at" \
            > "${catalog_dir}/catalog-info.json"; \
          rm -rf "${catalog_dir}/configs"; \
          SUCCESS=$((SUCCESS + 1)); \
        else \
          echo "ERROR: Failed to process ${catalog_type} v${ocp_version}"; \
          rm -rf "${catalog_dir}/configs" 2>/dev/null; \
          FAILED=$((FAILED + 1)); \
        fi; \
      done; \
    done && \
    if [ "$FAILED" -gt 0 ]; then \
      echo "ERROR: ${FAILED}/${TOTAL} catalogs failed to process"; \
      exit 1; \
    fi && \
    OCP_JSON=$(echo $OCP_VERSIONS | tr ' ' '\n' | sed 's/.*/"&"/' | paste -sd,) && \
    CAT_JSON=$(echo $CATALOG_TYPES | tr ' ' '\n' | sed 's/.*/"&"/' | paste -sd,) && \
    printf '{"ocp_versions":[%s],"catalog_types":[%s],"catalogs":[]}\n' "$OCP_JSON" "$CAT_JSON" \
      > "${CATALOG_DATA_DIR}/catalog-index.json" && \
    for ocp_version in $OCP_VERSIONS; do \
      for catalog_type in $CATALOG_TYPES; do \
        info="${CATALOG_DATA_DIR}/${catalog_type}/v${ocp_version}/catalog-info.json"; \
        if [ -f "$info" ]; then \
          jq --argjson entry "$(cat "$info")" '.catalogs += [$entry]' \
            "${CATALOG_DATA_DIR}/catalog-index.json" > "${CATALOG_DATA_DIR}/catalog-index.json.tmp" && \
            mv "${CATALOG_DATA_DIR}/catalog-index.json.tmp" "${CATALOG_DATA_DIR}/catalog-index.json"; \
        fi; \
      done; \
    done && \
    find "${CATALOG_DATA_DIR}" -type d -name configs -exec rm -rf {} + 2>/dev/null; \
    echo "Completed: ${SUCCESS}/${TOTAL} catalogs successful, ${FAILED} failed" && \
    echo "--- catalog-data contents ---" && \
    find "${CATALOG_DATA_DIR}" -type f && \
    cat "${CATALOG_DATA_DIR}/catalog-index.json"