#!/bin/bash
#
# File: check-index-errors
# Version: 0.95a
#
# Check whether the index.html file in a volume has some common errors
# (c) 2020-2025 by Manfred Jeusfeld. This script is made available under the
# Creative Commons Attribution-ShareAlike CC-BY-SA 4.0 license.
#
# The BASH script is part of the scripts used for CEUR-WS.org. No warrantee whatsoever. No support. 
#
# Note that this script is updated on a regular basis, in particular to cover changes with
# the layout for index files ceur-ws.org/Vol-XXX/index.html.
#
# Call this script in the directory that contains the index.html file that you want to check.
#
# Manfred 2020-03-20 (2026-03-15)
#

if [ ! -f index.html ]; then
    echo "No file index.html in this directory. Call this script in the directory that contains your submissions files."
    exit 1
fi

if [[ -f index.html.orig && "$1" != "pre" ]]; then
    echo "Using check-index-errors-after for rules holding for edited index.html files ... "
    check-index-errors-after "$@"
    exit 0
fi


echo "check-index-errors V0.95a (2026-03-15) CC-BY-SA 4.0"
echo ""

# to distinguish regular index files for "semantified" index files
indexformat=regular
if grep -q "foaf:" index.html;
then indexformat=semantic
fi

if grep -q "Workshop on Publishing Papers" index.html;
then echo " ";
     echo " ==========> ERROR (A1) in index.html!!!!";
     echo "String 'Workshop on Publishing Papers' found in index.html";
fi

if grep -q "OPub" index.html;
then echo " ";
     echo " ==========> ERROR (A2) in index.html!!!!";
     echo "String 'OPub' found in index.html";
fi

if grep -q -E "Coeditor|Carlos Nombre|Anne Foé" index.html;
then echo " ";
     echo " ==========> ERROR (A3) in index.html!!!!";
     echo "String 'Coeditor' or 'Carlos Nombre' or 'Anne Foé' left over from template Vol-XXX found in index.html";
fi

if grep -q "<li>" index.html;
then echo " ";
     echo " ==========> ERROR (A4) in index.html!!!!";
     echo "<li> element without ID found in index.html";
fi

if grep -q "workshop-website.org/loc" index.html;
then echo " ";
     echo " ==========> ERROR (A5) in index.html!!!!";
     echo "wrong URL workshop-website.org/loc found in index.html";
fi


if grep -q "Copying permitted for private and academic purposes" index.html;
then echo " ";
     echo " ==========> ERROR (A6) in index.html!!!!";
     echo "Old non-CC-BY copyright statement";
fi

grep -q  "../ceur-ws.css" index.html
if [[ $? != 0 ]];
then echo " ";
     echo " ==========> ERROR (A7) in index.html!!!!";
     echo "wrong stylesheet; must be ../ceur-ws.css";
fi

if grep -q "JJJJ" index.html;
then echo " ";
     echo " ==========> ERROR (A8) in index.html!!!!";
     echo "Coypyright year JJJJ must be set in the header";
fi


file index.html | grep -q "UTF-16"
if [[ $? == 0 ]];
then echo " ";
     echo " ==========> ERROR (A9) in index.html!!!!";
     echo "index.html should not be encoded in UTF-16; use iconv to convert to UTF-8";
fi

grep -q  "urn:nbn:de:0074" index.html
if [[ $? != 0 ]];
then echo " ";
     echo " ==========> ERROR (A10) in index.html!!!!";
     echo "Prefix urn:nbn:de:0074 missing in URN";
fi

grep -q  "CEURTOC" index.html
if [[ $? != 0 ]];
then echo " ";
     echo " ==========> ERROR (A11) in index.html!!!!";
     echo "DIV section CEURTOC missing; compare to Vol-XXX/index.html";
fi


if grep -q "<script" index.html;
then echo " ";
     echo " ==========> ERROR (A12) in index.html!!!!";
     echo "index.html contains a script; scripts are forbidden";
fi

grep -q  "CEURLOCTIME" index.html
if [[ $? != 0 ]];
then echo " ";
     echo " ==========> ERROR (A13) in index.html!!!!";
     echo "CEURLOCTIME element is missing; compare to Vol-XXX/index.html";
fi

if [ $indexformat == regular ];
then
  if grep -o -P '(?<=CEURAUTHOR">).*(?=</span)' index.html | grep -v ', Jr.' | grep ',' ;
  then echo " ";
       echo " ==========> ERROR (A14) in index.html!!!!";
       echo "Some CEURAUTHOR lines contain commas. Use one line per CEURAUTHOR. A CEURAUTHOR is a single person. Use multiple CEURAUTHOR fields when a paper has multiple authors.";
       grep -o -P '(?<=CEURAUTHOR">).*(?=</span)' index.html | grep ','
  fi

  if grep -o -P '(?<=class="CEUR).*(?=</span)' index.html | grep '<' | grep -v '</span' ;
  then echo " ";
       echo " ==========> ERROR (A15) in index.html!!!!";
       echo "Some CEUR span field contains '<' e.g. for a HTML tag. CEUR span fields may not contain HTML tags!";
       grep -o -P '(?<=class="CEUR).*(?=</span)' index.html | grep '<' | grep -v '</span'
  fi
fi

if grep -q "http://ceur-ws.org/ceur-ws.css" index.html;
then echo " ";
     echo " ==========> ERROR (A16) in index.html!!!!";
     echo "index.html contains absolute path http//ceur-ws.org/ceur-ws.css; must be ../ceur-ws.css";
fi

for KEYWORD in CEURVERSION CEURVOLTITLE CEURFULLTITLE CEURURN CEURPUBYEAR CEURVOLACRONYM CEURTOC CEURLOCTIME CEURPUBDATE CEURVOLNR
do
  NRWORDS=`grep -wc $KEYWORD index.html`
  if [ "$NRWORDS" -ne "1" ]
  then 
    echo " ==========> ERROR (A17) in index.html!!!!";
    echo "The label $KEYWORD must have exactly one value in index.html. It your case, the value is $NRWORDS."
    grep $KEYWORD index.html
  fi
done

for KEYWORD in CEURCOLOCATED
do
  NRWORDS=`grep -wc $KEYWORD index.html`
  if [ "$NRWORDS" -gt "1" ]
  then 
    echo " ==========> ERROR (A17a) in index.html!!!!";
    echo "The label $KEYWORD must have exactly one value in index.html. It your case, the value is $NRWORDS."
    grep $KEYWORD index.html
  fi
done


for KEYWORD in CEURVOLEDITOR
do
  NRWORDS=`grep -wc $KEYWORD index.html`
  if [ "$NRWORDS" == "0" ]
  then 
    echo " ==========> ERROR (A18) in index.html!!!!";
    echo "The label $KEYWORD must have at least one value in index.html. It your case, the value is $NRWORDS."
    echo " ";
  fi
done


grep -q  "%2FVol-XXX%2F" index.html
if [[ $? != 0 ]];
then echo " ";
     echo " ==========> ERROR (A19) in index.html!!!!";
     echo "The validator link at the end of index.html must initially include the parameter 'Vol-XXX'. Otherwise, our scripts cannot insert the right volume number here!";
     echo " ";
fi


if grep CEURAUTHOR index.html | grep -q " </span>";
then echo " ";
     echo " ==========> ERROR (A20) in index.html!!!!";
     echo "index.html contains CEURAUTHOR SPAN elements that end with a blank character. This creates unwanted variants for labels such as author names.";
       grep -n -P ' </span>' index.html
fi

if grep CEURAUTHOR index.html | grep -q " and";
then echo " ";
     echo " ==========> ERROR (A21) in index.html!!!!";
     echo "index.html contains CEURAUTHOR elements using 'and' as separator instead ','.";
     grep -n -P 'CEURAUTHOR.*</span>, and' index.html;
     echo " ";
fi

# extract certain CEUR fields from the index.html file
# ceurpubyr=`grep -oE 'class="CEURPUBYEAR"[^<>]*>[^<>]+' index.html |  cut -d'>' -f2`
# Could be used for some plausibility checks, e.g. CEURPUBYEAR should normally occur in CEURLOCTIME


if grep -q "Mary Y. Writter" index.html;
then echo " ";
     echo " ==========> ERROR (A22) in index.html!!!!";
     echo "index.html contains a synthetic author name from Vol-XXX, here Mary Y. Writter";
fi

DUPPDFLINKS=`grep "<a href=" index.html | grep pdf | sed "s/<a href/\\n<a href/g" | sed 's/\"/\"><\/a>\n/2' | grep href | sort | uniq --count | grep -v "1 <a"`
if [ "$DUPPDFLINKS" != "" ]
then echo " ";
     echo " ==========> ERROR (A23) in index.html!!!!";
     echo " index.html contains multiple links to the same PDF file!";
     echo $DUPPDFS
fi

DUPTITLES=`grep "CEURTITLE" index.html | sed "s/<a href/\\n<a href/g" | sed 's/\"/\"><\/a>\n/2' | grep href | sort | uniq --count | grep -Ev "^[[:space:]]*1 "`
if [ "$DUPTITLES" != "" ]
then echo " ";
     echo " ==========> ERROR (A24) in index.html!!!!";
     echo " index.html contains multiple CEURTITLEs with the same label!";
     echo $DUPTITLES
fi


if [ $indexformat == regular ];
then
  if grep -o -P '(?<=li id=).*(?=pdf)' index.html | grep '/' ;
  then echo " ";
       echo " ==========> ERROR (A25) in index.html!!!!";
       echo "Some papers appear to be in subdirectories or use absolute URLs. They shall be in the same directory as index.html and use relative URLs";
       grep -o -P '(?<=li id=).*(?=pdf)' index.html | grep '/'
  fi
fi

grep "CEURVOLACRONYM" index.html |  grep -q  " 2"
if [[ $? != 0 ]];
then echo " ";
     echo " ==========> ERROR (A26) in index.html!!!!";
     echo "CEURVOLACRONYM must contain the year in which the event took place like in 'ABCD 2023'";
fi

for KEYWORD in CEURTITLE CEURAUTHOR CEURVOLTITLE
do
  if grep $KEYWORD index.html | grep -v 'Jr.' | grep -q "\.</span>";
  then echo " ";
       echo " ==========> ERROR (A27) in index.html!!!!";
       echo "$KEYWORD ends with a '.'. It should be just a label not ending with a dot.";
       grep -n -P $KEYWORD index.html |  grep -v 'Jr.' | grep  '\.</span>'
  fi
done


grep "submitted by" index.html |  grep -q  ","
if [[ $? != 0 ]];
then echo " ";
     echo " ==========> ERROR (A28) in index.html!!!!";
     echo "The name of the person after 'submitted by' in end the end of index.html must be followed by a comma ','.";
fi

grep -q  "<body>" index.html
if [[ $? != 0 ]];
then echo " ";
     echo " ==========> ERROR (A29) in index.html!!!!";
     echo "The index.html file must contain the tag 'body' without any parameters.";
fi


ALLPDFSLINKED="yes"
for f in *.pdf
do
  # All PDF files must be linked in index.html
  grep -q ""$f"" index.html
  if [[ $? != 0 ]];
  then
    ALLPDFSLINKED="unlinkedpdffound";
    echo "PDF file $f is not linked in index.html";
  fi
done

if [[ $ALLPDFSLINKED == "unlinkedpdffound" ]] ; then
  echo " ==========> ERROR (A30) in index.html!!!!";
  echo "Make sure all paper PDFs are linked in index.html";
  echo " ";
fi


# All links to PDF files in index.html must point to a file in the same directory
pdf_links=$(grep -o 'href="[^"]*\.pdf"' index.html | sed 's/href="//' | sed 's/"$//')
ALLPDFSEXISTS="yes"
for f in $pdf_links
do
  if [[ $f != http* ]]  && [[ ! -f "$f" ]]; then
    echo "File $f referenced in index.html does not exist in the same directory"
    ALLPDFSEXISTS="no"
  fi
done

if [[ $ALLPDFSEXISTS == "no" ]] ; then
  echo " ==========> ERROR (A31) in index.html!!!!";
  echo "Make sure all references to paper PDFs in index.html also exist as file in the same directory"
  echo " "
fi


# A PDF file should be referenced only once in index.html
DUPCOUNTRESULT=`grep -o '"[[:alnum:]_-]\+\.pdf"' index.html | sed 's/<[^>]*>//g' | tr -s '[:space:]' '\n' | grep -v '^$' | sort | uniq -c | awk '$1 > 1'`
if [[ "$DUPCOUNTRESULT" != "" ]];
then echo " ";
     echo " ==========> ERROR (A32) in index.html!!!!";
     echo "The index.html file has multiple references to the same PDF file.";
fi

if grep -q "<script" index.html;
then echo " ";
     echo " ==========> ERROR (A33) in index.html!!!!";
     echo "Script keyword found in index.html; we do not allow JavaScript on our pages";
fi

if grep -q "http://" index.html;
then echo " ";
     echo " ==========> ERROR (A34) in index.html!!!!";
     echo "Change occurrences of 'http://' to 'https://' if the URL is also working with https://";
fi

if grep -q "0000-0000-0000-0000" index.html;
then echo " ";
     echo " ==========> ERROR (A35) in index.html!!!!";
     echo "The index.html file has a dummy ORCID 0000-0000-0000-0000. Either correct or remove the tags itemcope itemtype and itemid";
fi

grep -q  "0074-XXX-C" index.html
if [[ $? != 0 ]];
then echo " ";
     echo " ==========> ERROR (A36) in index.html!!!!";
     echo "Do not change the URN template from Vol-XXX in the submission index.html file; we shall assign the URN automatically";
     echo " ";
fi

grep -q  "\"CEURVOLNR\">Vol-XXX" index.html
if [[ $? != 0 ]];
then echo " ";
     echo " ==========> ERROR (A37) in index.html!!!!";
     echo "Do not change the volume number Vol-XXX in the submission index.html file; we shall assign it automatically";
     echo " ";
fi

# Check if the file contains the opening tag but NOT the closing tag on the same line
# This is necessary for a proper execution of the tests on ORCIDs
grep -qP '(?=.*<span class="(CEURAUTHOR|CEURVOLEDITOR)")^(?!.*</span>)' index.html
if [[ $? -eq 0 ]]; then
     echo " "
     grep -nP '(?=.*<span class="(CEURAUTHOR|CEURVOLEDITOR)")^(?!.*</span>)' index.html | head -n 1
     echo " ==========> ERROR (A38) in index.html!!!!"
     echo "The <span> expression for CEURAUTHOR and CEURVOLEDITOR must be on a single line!"
fi




# 2025-11-16: Check whether ORCIDs in index.html have matching authors names
ORCIDSFOUND="true"
grep -q  "itemid=\"https://orcid.org/" index.html
if [[ $? != 0 ]];
then ORCIDSFOUND="false"
fi


# Function to normalize names: remove accents, lowercase, collapse spaces
normalize() {
    echo "$1" | iconv -f utf-8 -t ascii//TRANSLIT 2>/dev/null | tr '[:upper:]' '[:lower:]' | tr -s ' ' | sed 's/^[[:space:]]*//;s/[[:space:]]*$//'
}

echo " "
if [[ "$ORCIDSFOUND" == "true" ]]; then
  echo " "
  echo "(*) Checking whether author names match their profile name on ORCID.org"
  ORCIDMISMATCH="no"


  # Loop over all CEUR authors/editors
  while read -r span; do
    # Extract the author/editor name
    name=$(echo "$span" | sed -E 's/.*>([^<]+)<\/span>/\1/' | sed 's/&[a-zA-Z0-9#]\+;/ /g')

    # Extract the ORCID if present
    orcid=$(echo "$span" | grep -oP 'itemid="https://orcid\.org/\K[0-9\-X]+')
    [[ -z "$orcid" ]] && continue

    # Query ORCID Public API
    json=$(curl -s -H "Accept: application/json" "https://pub.orcid.org/v3.0/$orcid/person")

    # Extract main given and family names
    api_given=$(echo "$json" | grep -oP '"given-names":\s*\{"value":\s*"\K[^"]+')
    api_family=$(echo "$json" | grep -oP '"family-name":\s*\{"value":\s*"\K[^"]+')
    api_full="$api_given $api_family"

    # Extract other-names / aliases
    aliases=$(echo "$json" | grep -oP '"other-name":\s*\{"content":\s*"\K[^"]+' | tr '\n' ' ')

    # Build array of all names to check
    name_list=("$api_full")
    for alias in $aliases; do
        name_list+=("$alias")
    done

    # Normalize CEUR name
    ceur_norm=$(normalize "$name")
    match_found=0

    # Check against all ORCID names
    for n in "${name_list[@]}"; do
        n_norm=$(normalize "$n")
        n_words=($n_norm)
        if [[ ${#n_words[@]} -ge 2 ]]; then
            given=${n_words[0]}
            family=${n_words[-1]}
            if [[ "$ceur_norm" =~ $given ]] && [[ "$ceur_norm" =~ $family ]]; then
                match_found=1
                break
            fi
        fi
    done

    # Only output mismatches
    if [[ $match_found -eq 0 ]]; then
        ORCIDMISMATCH="yes"
        echo "In index.html: $name → $orcid → On ORCID.org: $api_full (aliases: $aliases)"
    fi

  done < <(grep -oP '<span class="CEUR(?:AUTHOR|VOLEDITOR)".*?>.*?</span>' index.html)

  if [[ "$ORCIDMISMATCH" == "yes" ]] ; then
    echo " "
    echo " ===> Make sure that the CEURAUTHOR names with ORCIDs match the profile name or its aliases on ORCID.org."
    echo "      It may be that the CEURAUTHOR name in index.html has a spelling error, or the ORCID is false, or"
    echo "      the name in the ORCID profile has a spelling error."
    echo "      The profile should include suitable aliases ('also known as') of the author name. If a mismatch"
    echo "      cannot be corrected, then remove the ORCID tags from the CEURAUTHOR element in index.html."
    echo " "
  else
    echo "ok"
    echo " "
  fi

fi


echo " "
echo "(*) Checking mismatches of LI id tags with PDF filenames"
# Variable to store all error messages encountered during the check
IDTAGMISMATCHES=""

# Use "index.html" directly. 
# We capture the entire awk output (which contains potential errors) into IDTAGMISMATCHES.
IDTAGMISMATCHES=$(grep -E '<li.*id=".*".*href=".*\.pdf"' "index.html" | awk '
{
    # 1. Extract the ID value: id="VALUE"
    if (match($0, /id="([^"]*)"/, id_match)) {
        LI_ID = id_match[1];
    } else {
        next; 
    }

    # 2. Extract the Filename value: href="FILENAME.pdf"
    if (match($0, /href="([^"]*\.pdf)"/, href_match)) {
        PDF_FULL_NAME = href_match[1];
    } else {
        next; 
    }

    # 3. Strip the ".pdf" extension from the full filename
    PDF_BASE_NAME = substr(PDF_FULL_NAME, 1, length(PDF_FULL_NAME) - 4);
    
    # 4. Compare the extracted ID and the base filename
    if (LI_ID != PDF_BASE_NAME) {
        
        # Output a single, descriptive line for the error.
        printf "Error in index.html: Mismatch of filename \"%s.pdf\" with its LI id tag \"%s\".\n", \
               PDF_BASE_NAME, LI_ID;
        
    }
}' | grep . || true) # The grep . || true ensures IDTAGMISMATCHES remains empty if no output exists


if [[ -z "$IDTAGMISMATCHES" ]]; then
    echo "ok"
else
    echo -e "$IDTAGMISMATCHES"
    echo " ===> Make sure that LI id tag for papers matches the PDF filename."
    echo "      For example if the filename is paper12.pdf then the LI id tag must be paper12."
fi

echo " "

# Do the official W3C HTML validation check
echo "(*) Running W3C HTML5 validation..."
if [ -f  $HOME/bin/vnu.jar ]; then
  java -jar $HOME/bin/vnu.jar --errors-only index.html
  if [ $? -eq 0 ]; then
      echo "ok"
  fi
else
  echo "Download vnu.jar from https://github.com/validator/validator/releases/download/latest/vnu.jar into $HOME/bin for a full check"
fi



# 2026-02-28: Check for consecutive page numbers (A39)
echo " "
echo "(*) Check for consecutive page numbers..."

perl -0777 -Mutf8 -ne '
  binmode(STDOUT, ":utf8");
  
  # REGEX FIX: Removed the requirement for whitespace before the tag
  # It now finds <span class="CEURPAGES"> even if it is glued directly to </a>
  while (/<span class="CEURPAGES">\s*(\d+)\s*(?:[^0-9<]+\s*(\d+))?\s*<\/span>/gs) {
      my $start = $1;
      my $end = defined($2) ? $2 : $1;

      if (defined $last_end) {
          if ($start != $last_end + 1) {
              # print "\n ==========> ERROR (A39) in index.html!!!!\n";
              print "Inconsistent CEURPAGES in index.html: Previous paper ended at $last_end, but current starts at $start.\n";
              $error = 1;
          }
      }
      $last_end = $end;
  }
  exit 1 if $error;
' index.html

if [ $? -eq 0 ]; then echo "ok"; fi


# 2026-02-28: Verify physical PDF matches the CEURPAGES metadata (A40)
# 2026-03-04: Fixed Robust PDF Match (A40)
echo " "
echo "(*) Checking if PDF page counts match CEURPAGES..."

perl -0777 -Mutf8 -ne '
  binmode(STDOUT, ":utf8");
  
  # NEW REGEX: (?:(?!href=).)*? 
  # This ensures we dont jump over the next paper link to find a page tag.
  while (/href="([^"]+\.pdf)"(?:(?!href=).)*?<span class="CEURPAGES">\s*(\d+)\s*(?:[^0-9<]+\s*(\d+))?\s*<\/span>/gs) {
      my ($pdf, $start, $end_val) = ($1, $2, $3);
      my $end = defined($end_val) ? $end_val : $start;
      my $expected = $end - $start + 1;
      
      if (-f $pdf) {
          my $info = `pdfinfo "$pdf" 2>/dev/null`;
          if ($info =~ /^Pages:\s+(\d+)/m) {
              my $actual = $1;
              if ($actual != $expected) {
                  print "$pdf: CEURPAGES in index.html claims $expected pages ($start-$end), but PDF has $actual.\n";
                  $error = 1;
              }
          }
      }
  }
  exit 1 if $error;
' index.html

if [ $? -eq 0 ]; then echo "ok"; fi

echo " "


# 2026-03-07: Diversity rule: Check if a CEURVOLEDITOR is CEURAUTHOR of more than one paper (A41)
echo "(*) Checking diversity rule for editor/author overlap (A41) ..."

EDITORS=$(grep 'class="CEURVOLEDITOR"' index.html | sed -e 's/.*class="CEURVOLEDITOR"[^>]*>//' -e 's/<\/span>.*//' | sed 's/^[ \t]*//;s/[ \t]*$//' | sort -u)
AUTHORS=$(sed -n '/<li id="preface">/,/<\/li>/!p' index.html | grep 'class="CEURAUTHOR"' | sed -e 's/.*class="CEURAUTHOR">//' -e 's/<\/span>.*//' | sed 's/^[ \t]*//;s/[ \t]*$//')

EDITOR_OVERLAP_ERROR="no"

if [ -n "$EDITORS" ]; then
    while IFS= read -r editor; do
        if [ -n "$editor" ]; then
            # Count occurrences of the editor name in the research paper author list 
            COUNT=$(echo "$AUTHORS" | grep -Fxc "$editor")
            
            if [ "$COUNT" -gt 1 ]; then
                EDITOR_OVERLAP_ERROR="yes"
                echo " Editor '$editor' is an author of $COUNT research papers."
            fi
        fi
    done <<< "$EDITORS"
fi

if [[ "$EDITOR_OVERLAP_ERROR" == "no" ]]; then
    echo "ok"
else
    echo " ==========> ERROR (A41) in index.html!!!!"
fi

echo " "


echo "(*) Checking minimum number of papers ..."
NRWORDS=`grep -wc CEURTITLE index.html`
if [[ "$NRWORDS" -ge "6" ]]; then 
    echo "ok"
else
#    echo " ==========> ERROR (A42) in index.html!!!!";
    echo "The index.html file lists $NRWORDS papers with a CEURTITLE. Minimum is 6."
fi

echo " "