run_all.sh

#!/bin/bash
set -e

###############################################################################
#  run_all.sh  --  Full rebuild and verification of all validation results
#
#  This script re-runs every validation and analysis script that underpins the
#  paper's claims, captures pass/fail status for each, then generates SHA-256
#  checksums of every results file so that outputs are reproducible and
#  auditable.
#
#  Usage:   ./run_all.sh          (from the Paper directory, or anywhere)
###############################################################################

# ---------------------------------------------------------------------------
# 0. Resolve the Paper directory from this script's own location
# ---------------------------------------------------------------------------
DIR="$(cd "$(dirname "$0")" && pwd)"
cd "$DIR"

# ---------------------------------------------------------------------------
# 1. Header
# ---------------------------------------------------------------------------
echo "==========================================================================="
echo "  Voynich Manuscript Decoding -- Full Validation Rebuild"
echo "==========================================================================="
echo ""
echo "  Paper directory : $DIR"
echo "  Date            : $(date '+%Y-%m-%d %H:%M:%S %Z')"
echo "  Python          : $(python3 --version 2>&1)"
echo ""
echo "  This script runs every validation and analysis script in scripts/,"
echo "  reports PASS/FAIL for each, and writes SHA-256 checksums of all"
echo "  result files to results/CHECKSUMS.sha256."
echo ""
echo "==========================================================================="
echo ""

# ---------------------------------------------------------------------------
# 2. Define the ordered list of scripts to run
# ---------------------------------------------------------------------------
SCRIPTS=(
    "scripts/validate_coverage.py"
    "scripts/validate_vowel_final.py"
    "scripts/validate_phonotactics.py"
    "scripts/validate_domain_clustering.py"
    "scripts/validate_external_pharma.py"
    "scripts/keyword_section_clustering.py"
    "scripts/sov_syntax_test.py"
    "scripts/collocation_test.py"
    "scripts/dual_null_model_test.py"
    "scripts/stability_robustness_test.py"
    "scripts/multiple_testing_correction.py"
    "scripts/holdout_validation.py"
)

# Scripts that are known to be long-running
declare -A SLOW_SCRIPTS
SLOW_SCRIPTS["scripts/dual_null_model_test.py"]=1
SLOW_SCRIPTS["scripts/stability_robustness_test.py"]=1

# ---------------------------------------------------------------------------
# 3. Run each script, track results
# ---------------------------------------------------------------------------
TOTAL=0
PASSED=0
FAILED=0
SKIPPED=0

FAILED_NAMES=()

for script in "${SCRIPTS[@]}"; do
    # Check existence first
    if [[ ! -f "$DIR/$script" ]]; then
        echo "  WARNING: $script not found -- skipping"
        echo ""
        SKIPPED=$((SKIPPED + 1))
        continue
    fi

    TOTAL=$((TOTAL + 1))

    echo "-------------------------------------------------------------------"
    echo "  Running: $script"

    # Print slow-script advisory
    if [[ -n "${SLOW_SCRIPTS[$script]+_}" ]]; then
        echo "  NOTE: This script may take several minutes (long-running test)."
    fi

    echo "-------------------------------------------------------------------"

    # Run the script; capture exit code without letting set -e kill us
    set +e
    python3 "$DIR/$script"
    EXIT_CODE=$?
    set -e

    if [[ $EXIT_CODE -eq 0 ]]; then
        echo ""
        echo "  >>> PASS: $script (exit code 0)"
        PASSED=$((PASSED + 1))
    else
        echo ""
        echo "  >>> FAIL: $script (exit code $EXIT_CODE)"
        FAILED=$((FAILED + 1))
        FAILED_NAMES+=("$script")
    fi
    echo ""
done

# ---------------------------------------------------------------------------
# 4. Generate SHA-256 checksums for every file in results/
# ---------------------------------------------------------------------------
echo "==========================================================================="
echo "  Generating SHA-256 checksums for results/"
echo "==========================================================================="
echo ""

mkdir -p "$DIR/results"

# Remove old checksums file so it is not included in its own checksum
rm -f "$DIR/results/CHECKSUMS.sha256"

# Only checksum if there are files present
RESULT_FILES=("$DIR"/results/*.txt)
if [[ -e "${RESULT_FILES[0]}" ]]; then
    sha256sum "$DIR"/results/*.txt > "$DIR/results/CHECKSUMS.sha256"
    echo "  Checksums written to: $DIR/results/CHECKSUMS.sha256"
else
    echo "  WARNING: No .txt files found in results/ -- no checksums generated."
fi

echo ""

# ---------------------------------------------------------------------------
# 5. Summary
# ---------------------------------------------------------------------------
TIMESTAMP="$(date '+%Y-%m-%d %H:%M:%S %Z')"

echo "==========================================================================="
echo "  SUMMARY"
echo "==========================================================================="
echo ""
echo "  Total scripts run : $TOTAL"
echo "  Passed            : $PASSED"
echo "  Failed            : $FAILED"
echo "  Skipped (missing) : $SKIPPED"
echo ""

if [[ $FAILED -gt 0 ]]; then
    echo "  Failed scripts:"
    for name in "${FAILED_NAMES[@]}"; do
        echo "    - $name"
    done
    echo ""
fi

echo "  Checksums file    : $DIR/results/CHECKSUMS.sha256"
echo "  Completed at      : $TIMESTAMP"
echo ""
echo "==========================================================================="

# Exit with non-zero status if any script failed
if [[ $FAILED -gt 0 ]]; then
    exit 1
fi

exit 0