shellfu-bash-uripecker/src/uripecker.sh.skel

#!/bin/bash

shellfu import pretty


uripecker() {
    #
    # Scan stdin for what looks like URI, ID or keyword
    #
    # Usage:
    #     <SOURCE_TEXT uripecker MAP
    #
    # Search through SOURCE_TEXT and output in following order:
    #
    #  1. apparent URIs,
    #  2. hash IDs ("bug#1234"),
    #  3. equal sign expressions ("bug = 1234"),
    #  4. "tags" ("bug1234")
    #  5. keyword expressions ("BUG 1234" or "g hello world"),
    #
    # all (except the first one, obviously) converted to URI using mappings
    # from MAP file.
    #
    # Note that keyword expressions (e.g. "bug 123") work only if they start
    # the line;  rest of the line is taken as query argument and URL-quoted,
    # so that "g what is bmo" would work as expected (given 'g' is defined
    # as Google search).
    #
    # Apply this filter to args or clipboard, and either use head -1 or
    # if you are brave, open all URIs.
    #
    # The MAP file can contain any number of query mappings in format:
    #
    #     NAME = URI_PATTERN
    #
    # where NAME is any string without spaces, equal sign or dot, and
    # URI_PATTERN is a string with precisely one instance of '%s'.
    #
    # For example, given this MAP file:
    #
    #     issue = http://gitlab.example.org/issue?id=%s
    #     faq = http://faq.example.org/faq/%s
    #
    # following text
    #
    #     issue = 1
    #     faq#225
    #     issue42
    #     http://other.example.com/
    #     faq 14
    #     or faq 15
    #
    # would return
    #
    #     http://other.example.com/
    #     http://faq.example.org/faq/225
    #     http://gitlab.example.org/issue?id=1
    #     http://gitlab.example.org/issue?id=42
    #     http://faq.example.org/faq/14
    #
    # Note that the URI_PATTERN can be any kind of URI, such as ftp:// URI,
    # (or any string, actually) but the '%s' is converted using HTTP URI rules.
    #
    local MapFile=$1; shift
    test -n "$MapFile" || {
        warn "usage: uripecker MAP"
        return 2
    }
    local MapBody
    local tmp   # temporary pipe storage
    tmp=$(mktemp -d -t uripecker.XXXXXXXX)
    local maybe_uris="$tmp/maybe_uris"
    local maybe_ids="$tmp/maybe_ids"
    local maybe_exps="$tmp/maybe_exps"
    local maybe_tags="$tmp/maybe_tags"
    local maybe_kws="$tmp/maybe_kws"
    local uris="$tmp/uris"
    local uris_from_ids="$tmp/uris_from_ids"
    local uris_from_exps="$tmp/uris_from_exps"
    local uris_from_tags="$tmp/uris_from_tags"
    local uris_from_kws="$tmp/uris_from_kws"
    MapBody=$(<"$MapFile") || {
        warn "error reading map file: $MapFile"
        return 3
    }
    ##
    # heat up and fill pipes
    #
    mkfifo "$maybe_uris" "$maybe_ids" "$maybe_exps" "$maybe_tags" "$maybe_kws" \
           "$uris" "$uris_from_ids" "$uris_from_exps" "$uris_from_tags" "$uris_from_kws"
    __uripecker__minimize | tee "$maybe_uris" "$maybe_ids" "$maybe_exps" "$maybe_tags" "$maybe_kws" \
        >/dev/null &
    ##
    # process each pipe *async* by different filter
    #
    < "$maybe_uris" __uripecker__flt_uris                          > "$uris" &
    < "$maybe_ids"  __uripecker__flt_ids  | __uripecker__deref > "$uris_from_ids" &
    < "$maybe_exps" __uripecker__flt_exps | __uripecker__deref > "$uris_from_exps" &
    < "$maybe_tags" __uripecker__flt_tags | __uripecker__deref > "$uris_from_tags" &
    < "$maybe_kws"  __uripecker__flt_kws  | __uripecker__deref > "$uris_from_kws" &
    ##
    # print result *sync* in correct order
    #
    {
        cat "$uris"
        cat "$uris_from_ids"
        cat "$uris_from_exps"
        cat "$uris_from_tags"
        cat "$uris_from_kws"
    } | grep .  # throw away empties; add missing LF
    rm -rf "$tmp"
}

__uripecker__map_lookup() {
    #
    # Look up query $1 in $Map
    #
    local key=${1,,}
    sed '
            s/^ *//
            s/ *$//
            s/ *= */=/
        ' <<<"$MapBody" \
      | grep "^$key=" \
      | cut -d= -f2-
}

__uripecker__deref() {
    #
    # Turn query (like "g hello" for google) to URI
    #
    local kw        # keyword part, eg. "g"             or "bug"
    local query     # query part,   eg. "hello+dolly"   or "1234"
    local fmt       # for queries: format string (from ini) to pass to printf
    while read -r kw query;
    do
        debug -v kw query
        fmt=$(__uripecker__map_lookup "$kw") || return 1
        debug -v fmt
        #shellcheck disable=SC2059
        printf "$fmt\n" "$(__uripecker__urlquote "$query")"
    done
}

__uripecker__flt_exps() {
    #
    # Hack expressions like bug = 123 out of the text
    #
    sed -e 's/(\d)\</\n/;' \
      | perl -CS -ne '
            next unless m/\b([a-zA-Z]\w*\s*=\s*[[:alnum:]][[:alnum:]_#-]*)\b/;
            print "$1\n";
        ' \
      | sed -re 's/\s*=\s*/ /'
}

__uripecker__flt_ids() {
    #
    # Hack doer-like id's (ID#123) out of the text
    #
    tr ' ' '\n' \
      | perl -CS -ne '
            next unless m/\b([a-zA-Z]\w*#[[:alnum:]][[:alnum:]_#-]*)\b/;
            print "$1\n";
        ' \
      | tr '#' ' '
}

__uripecker__flt_kws() {
    #
    # Hack out lines that look like kw expressions (word space text
    #
    # Eg. 'wiki hello world'
    #
    grep -Ee '^\s*[a-zA-Z][[:alpha:]_]*\s+[^=]'
}

__uripecker__flt_tags() {
    #
    # Hack "tags" like bug123 out of the text
    #
    tr -c '[:alnum:]' '\n' \
      | grep -E '^[a-zA-Z]+[0-9]+$' \
      | sed -r 's/([a-zA-Z]+)([0-9])/\1 \2/'
}

__uripecker__flt_uris() {
    #
    # Hack URIs out of the text.
    #
    # Uses URL parser regex Found as SO answer[1] and adapted to include
    # comments from the original Imme's gist[2].
    #
    #   [1]: https://stackoverflow.com/a/30408189/835945
    #   [2]: https://gist.github.com/imme-emosol/731338/810d83626a6d79f40a251f250ed4625cac0e731f
    #
    local uri_re    # Imme Emosol's URI pattern
    local py_code   # code to do the matching
    uri_re=$(
        echo -n '\b'
                    ## protocol identifier
                    #
        echo -n     '(?:(?:https?|ftp)://)'
                    ## user:pass authentication
                    #
        echo -n     '(?:\S+(?::\S*)?@)?'
        echo -n     '(?:'
                        ## IP address dotted notation octets
                        #
        echo -n         '(?:[1-9]\d?|1\d\d|2[01]\d|22[0-3])'
        echo -n         '(?:\.(?:1?\d{1,2}|2[0-4]\d|25[0-5])){2}'
        echo -n         '(?:\.(?:[1-9]\d?|1\d\d|2[0-4]\d|25[0-4]))'
        echo -n      '|'
                        ## host name
                        #
        echo -n         '(?:(?:[a-z\u00a1-\uffff0-9]+-?)*[a-z\u00a1-\uffff0-9]+)'
                        ## domain name
                        #
        echo -n         '(?:\.(?:[a-z\u00a1-\uffff0-9]+-?)*[a-z\u00a1-\uffff0-9]+)*'
                        ## TLD identifier
                        #
        echo -n         '(?:\.(?:[a-z\u00a1-\uffff]{2,}))'
        echo -n      '|'
                        ## "localhost"
                        #
        echo -n         '(?:localhost)'
        echo -n     ')'
                    ## port number
                    #
        echo -n     '(?::\d{2,5})?'
                    ## resource path
                    #
        echo -n     '(?:/[^\s]*)?'
        echo -n '\b'
    )
    py_code=$(
        echo 'import os'
        echo 'import re'
        echo 'import sys'
        echo
        echo 'for uri in re.findall(os.environ["URI_RE"], sys.stdin.read()):'
        echo '    print(uri)'
    )
    URI_RE=$uri_re $__URIPECKER_PYBIN -c "$py_code"
}

__uripecker__minimize() {
    #
    # Strip, squash spaces and replace tabs
    #
    sed '
        s/\t/ /g
        s/  */ /g
        s/^ //
        s/ $//
    '
}

__uripecker__urlquote() {
    #
    # URL-quote query $1 and print result
    #
    local query=$1
    debug -v query
    #shellcheck disable=SC2028
    LC_ALL=en_US.UTF-8 $__URIPECKER_PYBIN -c "$(
        echo "import $__URIPECKER_PYMOD"
        echo 'import sys'
        echo "print($__URIPECKER_PYMOD.quote_plus(sys.argv[1]))"
    )" "$query"
    #FIXME: There should be proper way w/o touching LC_ALL
}

__shellfu_uripecker__init() {
    #
    # See what python version is here
    #
    if python3 --version 2>/dev/null \
        | grep -qF 'Python 3.'; then
        __URIPECKER_PYBIN=python3
        __URIPECKER_PYMOD=urllib.parse
    else
        __URIPECKER_PYBIN=python
        __URIPECKER_PYMOD=urllib
    fi
}

#shellfu module-version=__MKIT_PROJ_VERSION__