source: trunk/SRC/Documentation/xmldoc/pro2href.sh

#! /bin/sh
#
# DESCRIPTION
# ===========
#
# replace <element>ginette</element> by a sequence like with element given
# in argument <a href="./ginette.html">ginette</a>
# in all html files in dirhtml given in argument
# We have to deal with the path of ginette.html in refhtml also given in
# argument.
#
# see call in savesaxo.sh
#
# EVOLUTIONS
# ==========
# $Id$
#
# fplod 2007-08-21T09:06:58Z aedon.locean-ipsl.upmc.fr (Darwin)
# add -e (element) parameter to define element syntax like
# <pro></pro> or <proidl>...</proidl>
# add -i parameter : this directory is the one where we are looking for
# html files to be modified. so replace output by dirhtml
# add -r parameter :  this directory is the one where we are looking for
# html files to be linked
# remove lowercase translation (because of idl help files which are in
# uppercase)
# fplod 2007-08-20T11:25:39Z aedon.locean-ipsl.upmc.fr (Darwin)
# correction from
# 4.11. How do I match only the first occurrence of a pattern?
# in sed faq http://www.student.northpark.edu/pemente/sed/sedfaq4.html#s4.11
#++ still not working perfectly because path of links on the different
# occurrences of <pro>...</pro> may not be the same
# ex: see restoreboxparam.pro : domdef and saveboxparam
# fplod 2007-06-26T13:32:06Z aedon.locean-ipsl.upmc.fr (Darwin)
# improvement for multiple occurrences of <pro>...</pro> on one line
# see http://www.gentoo.org/doc/en/articles/l-sed2.xml
# sed by example, Part 2
# especially for directory-overview files
# fplod 2007-03-20T14:02:14Z aedon.locean-ipsl.upmc.fr (Darwin)
# creation
#
set -o posix
set -u
#
command=$(basename ${0} .sh)
log_date=$(date -u +"%Y%m%dT%H%M%SZ")
log=/tmp/${command}.${log_date}
#
usage=" Usage : ${command} -i dirhtml -r refhtml -e element"
#
while [ ${#} -gt 0 ]
do
    case ${1} in
        -i)
            dirhtml=${2}
            shift
        ;;
        -r)
            refhtml=${2}
            shift
        ;;
        -e)
            element=${2}
            shift
        ;;
        -h)
            echo "${usage}"
            exit 0
        ;;
        *)
            # other choice
            echo "${usage}"
            exit 1
        ;;
    esac
    # next flag
    shift
done
#
# check for dirhtml
if [ ! -d ${dirhtml} ]
then
    echo "eee : ${dirhtml} not found"
    exit 1
fi
# ++ check for readable/writable
#
# check for refhtml
if [ ! -d ${refhtml} ]
then
    echo "eee : ${refhtml} not found"
    exit 1
fi
# ++ check for readable
#
# check for element
if [ "${element}" == "" ]
then
    echo "eee : ${element} empty"
    exit 1
fi
#
# first find all files containing <element>...</element>
list_html_element=$(find ${dirhtml} -name "*.html" -exec grep -l "<${element}>.*</${element}>" {} \;)
if [ "${list_html_element}" == "" ]
then
    echo "iii : no <${element}>...</${element}> in html files"
else
    for file_html in ${list_html_element}
    do
        #echo "file_html ${file_html}"
        #read a
        fpath=$(dirname ${file_html} | sed -e "s+\(${dirhtml}/\)\(.*\)+\2+")
        #echo "fpath ${fpath}"
        #read a
        list_link=$(tr -s " " "\n" < ${file_html} | grep "<${element}>.*</${element}>" | sed -e "s/^.*<${element}>//" -e "s/<\/${element}>.*$//")
        # echo "liste link" ${list_link}
        # read a
        for link in ${list_link}
        do
            # replace <element>something</element> by
            #<a href="something.html">something</a>
            # modulehtml is the html file name to be used
            modulehtml=${link}.html
            # lpath is the path on module relatively to the location of
            # the html file containing the <element>...</element>
            lpath=$(find ${refhtml} -name "${modulehtml}")
            if [ "${lpath}" = "" ]
            then
                echo "eee : path of ${modulehtml} not found under ${refhtml}"
                echo "eee : ${link} is used in ${file_html}"
            else
                lpath=$(dirname ${lpath} | sed -e "s+\(${refhtml}/\)\(.*\)+\2+")
                #echo "path du fichier html ${fpath}"
                #echo "lpath ${lpath}"
                #read a
                if [ "${lpath}" = "${fpath}" ]
                then
                    path="./"
                fi
                if [ "${lpath:0:1}" = "/" ]
                then
                    # absolute path
                    path=${lpath}
                else
                    nblev=$(echo ${fpath} |  sed -e "s@/\$@@" | awk -F "/" '{print NF}')
                    relpath=""
                    ilev=1
                    while [ ${ilev} -le ${nblev} ]
                    do
                        relpath="${relpath}../"
                        ilev=$(( ${ilev} + 1 ))
                    done
                    path=${relpath}/${lpath}
                fi
                #echo "path ${path}"
                #read a
                cat <<EOF > /tmp/pro2href${$}.sed
1{x;s@^@first@;x;}
1,/<${element}>${link}<\/${element}>/{x;/first/s///;x;s@<${element}>${link}<\/${element}>@<a href="${path}/${link}.html">${link}<\/a>@;}
EOF
                sed -f /tmp/pro2href${$}.sed \
                ${file_html} > ${file_html}_modify
                #diff  ${file_html}  ${file_html}_modify
                #read a
                mv ${file_html}_modify ${file_html}
                rm /tmp/pro2href${$}.sed
            fi
        done
    done
fi
exit 0