textblock

textblock

#!/bin/sh
# textblock - extract indentation-delimited (or pattern-delimited) blocks from text
set -e

usage() {
    cat <<'EOF'
Usage: textblock [OPTIONS] START [file ...]

Split text into blocks. A block begins on any line matching START and, by
default, continues on subsequent lines whose indentation is greater than
the start line. With -e, blocks instead continue until an end-pattern is
matched.

Required:
  START            Regex to identify the start of a block

Options:
  -n               Number each output block
  -q               Output each block on a single shell-quoted line
  -g               Do not start a new block while already inside one
  -m PATTERN       Only include blocks whose full text matches PATTERN
  -s PATTERN       Exclude blocks if any line matches PATTERN
  -e PATTERN       Keep block open until a line matches PATTERN
  -h, --help       Show this help
EOF
}

# --- parse options ---
opt_n=0
opt_q=0
opt_g=0
opt_m=""
opt_s=""
opt_e=""

while [ $# -gt 0 ]; do
    case "$1" in
        -h|--help) usage; exit 0 ;;
        -n) opt_n=1; shift ;;
        -q) opt_q=1; shift ;;
        -g) opt_g=1; shift ;;
        -m) opt_m="$2"; shift 2 ;;
        -s) opt_s="$2"; shift 2 ;;
        -e) opt_e="$2"; shift 2 ;;
        --) shift; break ;;
        -*) echo "textblock: unknown option: $1" >&2; usage >&2; exit 1 ;;
        *)  break ;;
    esac
done

if [ $# -lt 1 ]; then
    echo "textblock: missing START pattern" >&2
    usage >&2
    exit 1
fi

start_pat="$1"
shift
# remaining args are files (or empty => stdin)

exec awk \
    -v start_pat="$start_pat" \
    -v opt_n="$opt_n" \
    -v opt_q="$opt_q" \
    -v opt_g="$opt_g" \
    -v opt_m="$opt_m" \
    -v opt_s="$opt_s" \
    -v opt_e="$opt_e" \
'
function indent_of(line,    i, c) {
    i = 0
    while (i < length(line)) {
        c = substr(line, i + 1, 1)
        if (c == " " || c == "\t")
            i++
        else
            break
    }
    return i
}

# Shell-quote a string using $'"'"'...'"'"' syntax
function shell_quote(s,    out, i, c, n, SQ) {
    SQ = "\047"
    if (s == "") return "$" SQ SQ
    out = ""
    n = length(s)
    for (i = 1; i <= n; i++) {
        c = substr(s, i, 1)
        if (c == "\\")       out = out "\\\\"
        else if (c == SQ)    out = out "\\" SQ
        else if (c == "\n")  out = out "\\n"
        else if (c == "\t")  out = out "\\t"
        else if (c == "\r")  out = out "\\r"
        else                 out = out c
    }
    return "$" SQ out SQ
}

function flush_block(    i, text) {
    if (blk_len == 0) return

    # -s: exclude if any line matches
    if (opt_s != "") {
        for (i = 0; i < blk_len; i++) {
            if (blk[i] ~ opt_s) {
                reset_block()
                return
            }
        }
    }

    # -m: include only if the full block text matches
    if (opt_m != "") {
        text = ""
        for (i = 0; i < blk_len; i++) {
            if (i > 0) text = text "\n"
            text = text blk[i]
        }
        if (text !~ opt_m) {
            reset_block()
            return
        }
    }

    # separate successive blocks with a blank line (unless -q)
    if (block_count > 0 && opt_q == 0) printf "\n"
    block_count++

    if (opt_q) {
        text = ""
        for (i = 0; i < blk_len; i++) {
            if (i > 0) text = text "\n"
            text = text blk[i]
        }
        print shell_quote(text)
    } else if (opt_n) {
        printf "%6d\t%s\n", block_count, blk[0]
        for (i = 1; i < blk_len; i++) {
            printf "\t%s\n", blk[i]
        }
    } else {
        for (i = 0; i < blk_len; i++) {
            print blk[i]
        }
    }

    reset_block()
}

function reset_block() {
    blk_len = 0
    in_block = 0
    base_indent = -1
}

BEGIN {
    in_block = 0
    blk_len = 0
    base_indent = -1
    block_count = 0
}

{
    line = $0
    ind = indent_of(line)

    if (in_block) {
        # -e mode: block ends when end-pattern matches
        if (opt_e != "") {
            if (line ~ opt_e) {
                flush_block()
                # End-pattern line is NOT part of the closed block,
                # but it might start a new block.
                if (line ~ start_pat) {
                    in_block = 1
                    base_indent = ind
                    blk[0] = line
                    blk_len = 1
                }
                next
            }
            if (opt_g == 0 && line ~ start_pat) {
                flush_block()
                in_block = 1
                base_indent = ind
                blk[0] = line
                blk_len = 1
                next
            }
            blk[blk_len++] = line
            next
        }

        # Default (indent) mode
        if (ind > base_indent) {
            if (opt_g == 0 && line ~ start_pat) {
                flush_block()
                in_block = 1
                base_indent = ind
                blk[0] = line
                blk_len = 1
                next
            }
            blk[blk_len++] = line
            next
        }

        # Indentation <= base_indent: block ends
        flush_block()
        # Fall through to check for new block start
    }

    if (line ~ start_pat) {
        in_block = 1
        base_indent = ind
        blk[0] = line
        blk_len = 1
    }
}

END {
    flush_block()
}
' "$@"