scripts/run.sh

#!/usr/bin/env bash
# One-click runner: reproduce all restricted MVEB subsets from raw sources.
#
# Each subset directory under scripts/ ships its own process.sh which
#   1. downloads raw data   -> ${MVEB_ROOT}/downloads/<subset>
#   2. extracts sources     -> ${MVEB_ROOT}/source/<subset>
#   3. packs parquet        -> ${MVEB_TRAIN_DIR}/<Subset> , ${MVEB_TEST_DIR}/<Subset>
#
# Environment variables (all optional):
#   MVEB_ROOT       scratch root for downloads/ source/ logs/
#                   (default: two levels above scripts/)
#   MVEB_TRAIN_DIR  MVEB-train repo dir receiving train subsets
#                   (default: parent of scripts/, i.e. this repo)
#   MVEB_TEST_DIR   MVEB-test repo dir receiving test subsets
#                   (default: ${MVEB_ROOT}/MVEB-test)
#
# Usage:
#   bash run.sh                    # process all subsets
#   bash run.sh DukeMTMC Inshop    # process selected subsets only
#   bash run.sh --list             # list available subsets and exit
#
# Notes:
#   - Kaggle subsets need a configured API token (~/.kaggle/kaggle.json).
#   - A failing subset does not abort the others; a summary is printed at
#     the end and the exit code is non-zero if any subset failed.
#   - Per-subset logs: <root>/logs/<Subset>.log

set -uo pipefail

SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
# Scratch root (downloads/source/logs); forwarded to every process.sh.
export MVEB_ROOT="${MVEB_ROOT:-$(cd "${SCRIPT_DIR}/../.." && pwd)}"
# Split output dirs; read by the packing python via env.
export MVEB_TRAIN_DIR="${MVEB_TRAIN_DIR:-$(cd "${SCRIPT_DIR}/.." && pwd)}"
export MVEB_TEST_DIR="${MVEB_TEST_DIR:-${MVEB_ROOT}/MVEB-test}"
ROOT_DIR="${MVEB_ROOT}"
LOG_DIR="${ROOT_DIR}/logs"

# ---------- discover subsets (dirs containing process.sh) ----------
ALL_SUBSETS=()
for dir in "${SCRIPT_DIR}"/*/; do
    name="$(basename "${dir}")"
    [[ -f "${dir}/process.sh" ]] && ALL_SUBSETS+=("${name}")
done

if [[ ${#ALL_SUBSETS[@]} -eq 0 ]]; then
    echo "No subset with process.sh found under ${SCRIPT_DIR}" >&2
    exit 1
fi

if [[ "${1:-}" == "--list" ]]; then
    printf '%s\n' "${ALL_SUBSETS[@]}"
    exit 0
fi

# ---------- resolve target subsets ----------
if [[ $# -gt 0 ]]; then
    TARGETS=()
    for name in "$@"; do
        if [[ -f "${SCRIPT_DIR}/${name}/process.sh" ]]; then
            TARGETS+=("${name}")
        else
            echo "Unknown subset: ${name} (see: bash run.sh --list)" >&2
            exit 1
        fi
    done
else
    TARGETS=("${ALL_SUBSETS[@]}")
fi

# ---------- pre-flight: collect required commands from process.sh ----------
echo "==> MVEB_ROOT:      ${MVEB_ROOT}"
echo "==> MVEB_TRAIN_DIR: ${MVEB_TRAIN_DIR}"
echo "==> MVEB_TEST_DIR:  ${MVEB_TEST_DIR}"
echo "==> Pre-flight check"
MISSING=()
for name in "${TARGETS[@]}"; do
    while read -r cmd; do
        command -v "${cmd}" >/dev/null 2>&1 || MISSING+=("${name}: ${cmd}")
    done < <(grep -oP '^need_cmd \K\w+' "${SCRIPT_DIR}/${name}/process.sh" | sort -u)
done
if [[ ${#MISSING[@]} -gt 0 ]]; then
    echo "Missing commands (install them or the related subset will fail):" >&2
    printf '    %s\n' "${MISSING[@]}" >&2
fi

# ---------- run ----------
mkdir -p "${LOG_DIR}"
PASSED=()
FAILED=()
total=${#TARGETS[@]}
idx=0

for name in "${TARGETS[@]}"; do
    idx=$((idx + 1))
    log_file="${LOG_DIR}/${name}.log"
    echo ""
    echo "==> [${idx}/${total}] ${name} (log: ${log_file})"
    start_ts=$(date +%s)
    if bash "${SCRIPT_DIR}/${name}/process.sh" >"${log_file}" 2>&1; then
        elapsed=$(( $(date +%s) - start_ts ))
        echo "    OK (${elapsed}s)"
        PASSED+=("${name}")
    else
        elapsed=$(( $(date +%s) - start_ts ))
        echo "    FAILED after ${elapsed}s, last lines of log:"
        tail -n 5 "${log_file}" | sed 's/^/    | /'
        FAILED+=("${name}")
    fi
done

# ---------- summary ----------
echo ""
echo "=================================================="
echo "Summary: ${#PASSED[@]} passed, ${#FAILED[@]} failed (of ${total})"
[[ ${#PASSED[@]} -gt 0 ]] && echo "  passed: ${PASSED[*]}"
if [[ ${#FAILED[@]} -gt 0 ]]; then
    echo "  failed: ${FAILED[*]}"
    echo "  check logs under ${LOG_DIR}/ and re-run, e.g.:"
    echo "      bash run.sh ${FAILED[*]}"
    exit 1
fi
echo "All subsets processed. Outputs: ${MVEB_TRAIN_DIR}/<Subset> , ${MVEB_TEST_DIR}/<Subset>"