cloverleaf-larry/lib/nc-table.sh

#!/usr/bin/env bash
# nc-table.sh — read and modify Cloverleaf lookup tables (.tbl files).
# Every modification goes through the journal (snapshot + diff + atomic write).
#
# Subcommands:
#   list [--site SITE]                                 list .tbl files
#   show <name> [--site SITE]                          dump the file
#   pairs <name> [--site SITE] [--format csv|tsv]      input→output pairs only
#   lookup <name> <input>                              find output for input
#   reverse-lookup <name> <output>                     find input(s) for output
#   add <name> <input> <output> [--site SITE]          add or update a row (journaled)
#   delete <name> <input> [--site SITE]                delete a row by input (journaled)
#   create <name> --from-csv FILE [opts]               create a new table from CSV
#   replace <name> --from-csv FILE [opts]              replace contents (journaled)
#
# Find tables at: $HCISITEDIR/tables/<name>.tbl  (or $HCIROOT/Tables/ shared)
set -o pipefail

NC_SELF="$0"
LIB_DIR="$(cd "$(dirname "$NC_SELF")" && pwd)"
JOURNAL="$LIB_DIR/journal.sh"
C2T="$LIB_DIR/csv-to-table.sh"
T2C="$LIB_DIR/table-to-csv.sh"

die() { printf 'nc-table: %s\n' "$*" >&2; exit 1; }

[ -f "$JOURNAL" ] && . "$JOURNAL"

locate_table() {
  local name="$1" site="${2:-${HCISITE:-}}"
  # Strip .tbl if user gave it
  name="${name%.tbl}"
  for d in \
    "${HCIROOT:-}/$site/tables" \
    "${HCIROOT:-}/$site/Tables" \
    "${HCIROOT:-}/Tables"; do
    [ -f "$d/${name}.tbl" ] && { printf '%s\n' "$d/${name}.tbl"; return 0; }
  done
  return 1
}

cmd_list() {
  local site="${HCISITE:-}"
  while [ $# -gt 0 ]; do case "$1" in --site) shift; site="$1" ;; esac; shift; done
  for d in "${HCIROOT:-}/$site/tables" "${HCIROOT:-}/$site/Tables" "${HCIROOT:-}/Tables"; do
    [ -d "$d" ] || continue
    find "$d" -maxdepth 1 -name '*.tbl' -type f 2>/dev/null | sort
  done | sort -u
}

cmd_show() {
  local name="$1"; shift
  local site="${HCISITE:-}"
  while [ $# -gt 0 ]; do case "$1" in --site) shift; site="$1" ;; esac; shift; done
  local f; f=$(locate_table "$name" "$site") || die "no such table: $name"
  cat "$f"
}

cmd_pairs() {
  local name="$1"; shift
  local site="${HCISITE:-}"
  local fmt="csv"
  while [ $# -gt 0 ]; do case "$1" in
    --site)   shift; site="$1" ;;
    --format) shift; fmt="$1" ;;
  esac; shift; done
  local f; f=$(locate_table "$name" "$site") || die "no such table: $name"
  case "$fmt" in
    csv) "$T2C" --with-header "$f" ;;
    tsv) "$T2C" --with-header --delim $'\t' "$f" ;;
    *)   die "bad --format" ;;
  esac
}

cmd_lookup() {
  local name="$1" input="$2"
  local site="${HCISITE:-}"
  local f; f=$(locate_table "$name" "$site") || die "no such table: $name"
  "$T2C" "$f" | awk -F',' -v target="$input" '
    BEGIN { found=0 }
    {
      # CSV unquote (basic)
      v=$1; gsub(/^"|"$/, "", v); gsub(/""/, "\"", v)
      if (v == target) { o=$2; gsub(/^"|"$/, "", o); gsub(/""/, "\"", o); print o; found=1; exit }
    }
    END { if (!found) exit 1 }
  '
}

cmd_reverse_lookup() {
  local name="$1" output="$2"
  local site="${HCISITE:-}"
  local f; f=$(locate_table "$name" "$site") || die "no such table: $name"
  "$T2C" "$f" | awk -F',' -v target="$output" '
    {
      v=$2; gsub(/^"|"$/, "", v); gsub(/""/, "\"", v)
      if (v == target) { i=$1; gsub(/^"|"$/, "", i); gsub(/""/, "\"", i); print i }
    }
  '
}

# Modification helpers (journal-backed)
modify_via_csv() {
  local name="$1" csv_path="$2" site="$3" mode="$4"   # mode = add|delete|replace|create
  local action="$mode-table"
  local target

  # Determine target path
  if [ "$mode" = "create" ]; then
    [ -n "$site" ] || die "create: --site required"
    [ -d "${HCIROOT}/$site/tables" ] || mkdir -p "${HCIROOT}/$site/tables"
    target="${HCIROOT}/$site/tables/${name%.tbl}.tbl"
    [ ! -e "$target" ] || die "table already exists: $target (use replace if you want to overwrite)"
  else
    target=$(locate_table "$name" "$site") || die "no such table: $name"
  fi

  local new; new=$(mktemp)
  "$C2T" --has-header --out "$new" < "$csv_path" 2>/dev/null \
    || "$C2T" --has-header "$csv_path" > "$new"

  if declare -f journal_write >/dev/null 2>&1; then
    journal_write "$target" "$new" || { echo "ERROR: journal_write failed — table NOT modified" >&2; return 4; }
  else
    # No journal — direct write with simple backup
    # v0.7.5: strip non-digits from date — Cygwin date.exe could yield
    # "1779999999\r" which would produce a path with embedded CR.
    local _ts; _ts=$(date +%s | tr -cd '0-9')
    [ -f "$target" ] && cp -p "$target" "${target}.larry-bak.${_ts:-0}"
    mv "$new" "$target"
    echo "(no journal available; backup at ${target}.larry-bak.<ts>)"
  fi
  rm -f "$new"
}

cmd_add() {
  local name="$1" input="$2" output="$3"; shift 3
  local site="${HCISITE:-}"
  while [ $# -gt 0 ]; do case "$1" in --site) shift; site="$1" ;; esac; shift; done
  [ -n "$name" ] && [ -n "$input" ] && [ -n "$output" ] || die "usage: add NAME INPUT OUTPUT"
  local f; f=$(locate_table "$name" "$site") || die "no such table: $name"

  # Read current pairs, replace existing input row or append, then rewrite
  local tmp_csv; tmp_csv=$(mktemp)
  {
    "$T2C" --with-header "$f"
    # If the new pair wasn't already in the existing data, it'll be added below
  } > "$tmp_csv"

  local awk_script='
    BEGIN { added=0 }
    NR==1 { print; next }
    {
      n=split($0, c, ",")
      v=c[1]; gsub(/^"|"$/, "", v); gsub(/""/, "\"", v)
      if (v == NEW_IN) {
        printf "%s,%s\n", NEW_IN, NEW_OUT
        added=1
      } else { print }
    }
    END { if (!added) printf "%s,%s\n", NEW_IN, NEW_OUT }
  '
  local newcsv; newcsv=$(mktemp)
  awk -v NEW_IN="$input" -v NEW_OUT="$output" "$awk_script" "$tmp_csv" > "$newcsv"
  modify_via_csv "$name" "$newcsv" "$site" "add"
  rm -f "$tmp_csv" "$newcsv"
}

cmd_delete() {
  local name="$1" input="$2"; shift 2
  local site="${HCISITE:-}"
  while [ $# -gt 0 ]; do case "$1" in --site) shift; site="$1" ;; esac; shift; done
  [ -n "$name" ] && [ -n "$input" ] || die "usage: delete NAME INPUT"
  local f; f=$(locate_table "$name" "$site") || die "no such table: $name"
  local tmp_csv; tmp_csv=$(mktemp)
  "$T2C" --with-header "$f" \
    | awk -F',' -v IN="$input" '
        NR==1 { print; next }
        {
          v=$1; gsub(/^"|"$/, "", v); gsub(/""/, "\"", v)
          if (v == IN) next
          print
        }
      ' > "$tmp_csv"
  modify_via_csv "$name" "$tmp_csv" "$site" "delete"
  rm -f "$tmp_csv"
}

cmd_create() {
  local name="$1"; shift
  local from_csv=""
  local site="${HCISITE:-}"
  while [ $# -gt 0 ]; do case "$1" in
    --from-csv) shift; from_csv="$1" ;;
    --site)     shift; site="$1" ;;
  esac; shift; done
  [ -f "$from_csv" ] || die "create: --from-csv FILE required"
  modify_via_csv "$name" "$from_csv" "$site" "create"
}

cmd_replace() {
  local name="$1"; shift
  local from_csv=""
  local site="${HCISITE:-}"
  while [ $# -gt 0 ]; do case "$1" in
    --from-csv) shift; from_csv="$1" ;;
    --site)     shift; site="$1" ;;
  esac; shift; done
  [ -f "$from_csv" ] || die "replace: --from-csv FILE required"
  modify_via_csv "$name" "$from_csv" "$site" "replace"
}

SUB="${1:-list}"
case "$SUB" in
  list)            shift; cmd_list "$@" ;;
  show)            shift; [ $# -ge 1 ] || die "usage: show NAME"; cmd_show "$@" ;;
  pairs)           shift; [ $# -ge 1 ] || die "usage: pairs NAME"; cmd_pairs "$@" ;;
  lookup)          shift; [ $# -ge 2 ] || die "usage: lookup NAME INPUT"; cmd_lookup "$@" ;;
  reverse-lookup)  shift; [ $# -ge 2 ] || die "usage: reverse-lookup NAME OUTPUT"; cmd_reverse_lookup "$@" ;;
  add)             shift; cmd_add "$@" ;;
  delete)          shift; cmd_delete "$@" ;;
  create)          shift; cmd_create "$@" ;;
  replace)         shift; cmd_replace "$@" ;;
  help|-h|--help)  sed -n '2,20p' "$NC_SELF" ;;
  *)               die "unknown subcommand: $SUB" ;;
esac