lncdwiki_weekly_summary

#!/usr/bin/env bash

#
# fetch and format recent wiki changes for a slack message
# as promotion for wiki
#
# set TIMEDIFF and/or PASS environ to change behave
#   TIMEDIFF="1 week ago - 1 day" # pased onto 'date -d' for decoding
#   PASS="$(pass wiki)" #looks like 'user:pass'
#
# when PASS is set, uses curl to create cookie
#
# used in cron, piped to slack-msg
#  00 12 * * 5  PASS=$(pass wiki) lncdwiki_weekly_summary | slack-msg random

# 20231013WF - added code header; password; links in top of message


TIMEDIFF=${TIMEDIFF:-1 week ago - 1 day}
EDITS_STARTING_ON="$(date -d "$TIMEDIFF" +%F)"

export EDITS_STARTING_ON # export for $ENV perl line in get_wiki_updates

PASS=${PASS:-} # default to no pass
COOKIEARGS=()  # and no cookie jar args for curl

# originally thought creds in url like https://${PASS}@lncd.pitt.edu...
# would work. but need cookie login
# NB. cookie file hanging out in /tmp/ is maybe not great
auth_cookie(){
   declare -g COOKIEARGS
   jar=${1:-/tmp/cjar}
   if [ -n "$PASS" -a -z "${COOKIEARGS[*]}" ]; then
      u=${PASS/:*/}
      p=${PASS/*:/}
      curl -Lsd "u=$u&p=$p" --cookie-jar "$jar" https://lncd.pitt.edu/wiki/doku.php >/dev/null
      COOKIEARGS=(--cookie "$jar" --cookie-jar "$jar")
   fi
}
wiki_curl(){
   declare -g COOKIEARGS
   auth_cookie /tmp/cjar # populates COOKIEARGS
   curl "${COOKIEARGS[@]}" -Ls  "$@"
}

# yq's jq for xml, "xq," is xq-python in debian
xq=xq
command -v $xq >/dev/null || xq="xq-python"

get_wiki_updates(){
   #echo "cookies: ${COOKIEARGS[*]}" >&2
   saveto=${1:-/tmp/wiki-results.tsv}
   wiki_curl "https://lncd.pitt.edu/wiki/feed.php?num=500" | tee /tmp/wiki-results.html |
     $xq  --arg DATE "$EDITS_STARTING_ON" -r \
     '."rdf:RDF".item[]|
      select(."dc:date" > $DATE )|
      select((if .title == null then "no title" else .title end) |test("Page moved from")|not)|
      [(."dc:creator"|sub(" .*";"")),
       .link         |sub("http.*?=|&.*";"";"g")]|
      @tsv' | tee "$saveto" |
     grep -v %3A # dont show images
}
wiki_revisions(){
  local id="$1"
  wiki_curl "https://lncd.pitt.edu/wiki/doku.php?id=${id}&do=revisions" |
   grep 'div.*no.*li>'|sed 's:/li>:/li>\n:g'|
   perl -lne \
   '$date=$1=~s:/:-:gr if m:class="date">([0-9/]+) [^<]+</span:;
    $user=$1 if m:<bdi>(.*)</bdi:;
    $url=$1 if m/doku.php\?id=(.*?)&rev/;
    print "$user\t$url" if m:/li: and $user and $url and $date gt $ENV{EDITS_STARTING_ON}'
}
wiki_user_path_edits(){
  get_wiki_updates /tmp/wiki-results.tsv |
  while read -r user path; do
     echo -e "$user\t$path"
     wiki_revisions "$path" || :
  done |
  tee /tmp/wiki-results_revs.tsv |
  sort -u
}
wiki_main(){
# message top with very visible links to wiki
echo "<https://lncd.pitt.edu/wiki/doku.php|Wiki> Edits <https://lncd.pitt.edu/wiki/doku.php?id=start&do=recent|this week> (since $EDITS_STARTING_ON) below."
#echo "Other entry points: <https://lncd.pitt.edu/wiki/doku.php?id=admin:website:wiki:missing|missing>, <https://lncd.pitt.edu/wiki/doku.php?id=notebooks|notebook>, <https://lncd.pitt.edu/wiki/doku.php?do=randompage|random article>"

# getting a large amount(500)  of b/c
  wiki_user_path_edits |     # dont care about repeat edits
  perl -slane \
    'push @{$a{$F[0]}}, $F[1];
     END{print " :star: ", $#{$a{$_}}+1,
               " $_ ",
               join(",", map {$_=" <https://lncd.pitt.edu/wiki/doku.php?id=$_|`:$_`>"} @{$a{$_}}) for(keys %a)}'|
  tee /tmp/wiki-results-counts.tsv |
  grep -v " Anonymous " | # where did anonymous come from?
  sort -t' ' -nrk3,3
}

# iffmain from lncdtools
eval "$(iffmain wiki_main)"