sh 从archive.org下载Wayback Machine

Posted 2021-05-07

#!/bin/bash

url=http://redefininggod.com
webarchive=https://web.archive.org
wget="wget -e robots=off -nv"
tab="$(printf '\t')"
additional_url=url.list

# Construct listing.txt from url.list
# The list of archived pages, including some wildcard url
# each line contains some fields separated by tabs:
# - the last capture date (opaque format, if different, the last year index file
#   will be redownloaded)
# - the first capture year (hint for which is the oldest index to query)
# - the last capture year (hint for which is the latest index to query)
# - the url, starting with "/web/YYYYMMDDHHMMSS/" or "/web/*" (only) 

: >listing.txt

# Add url.list to listing.txt
while read url; do
  
  if [[ -z "$url" ]]; then continue; fi
  
  if [[ $url != ${url#http*/web.archive.org} ]]; then
    url="${url#http*/web.archive.org}"
  elif [[ $url != ${url%/\*} ]]; then
    mkdir -p "$(dirname "./web/*/$url")"
    $wget "$webarchive/web/*/$url" -O "./web/*/$url.html"
    # <listing.html fgrep 'href="/web/' | cut -d'"' -f2 >listing.txt
    <"./web/*/$url.html" sed -r -e '
      /<table id="resultsUrl">/,/<\/table>/ {
        /a href/ {
          s/.*href="(.*)".*/\1/;
          h
        };
        /dateFrom/ {
          s/.*([0-9]{4})<\/td>.*/\1/;
          x;
          H
        };
        /dateTo/ {
          s/.*>(.*)([0-9]{4})<\/td>.*/\1\2\n\2/;
          x;
          H
        };
        /<\/tr>/ {
          x;
          s/(.*)\n(.*)\n(.*)\n(.*)/\1\t\3\t\2\t\4/;
          p
        }  
      };
      d' >"./web/*/$url.txt"
    cat "./web/*/$url.txt" >>listing.txt
    continue
  else
    url="/web/*/$url"
  fi
  
  printf "%s\t%s\t%s\t%s\n" "$(date)" 1996 2014 "$url" >>listing.txt
  
done <"$additional_url"

# Construct listing2.txt
# Remove the wildcard url and fetch all the versions from index
# Lines only contains the URL starting with "/web/YYYYMMDDHHMMSS/" (only)
# It may contains duplicates
: >listing2.txt

while read line; do
  
  if [[ -z "$line" ]]; then continue; fi

  #printf "%s\n" "$line"
  
  oldifs="$IFS"
  IFS="$tab" elems=($line)
  IFS="$oldifs"
  lastcap="${elems[0]}"
  firstyear="${elems[1]}"
  lastyear="${elems[2]}"
  mainurl="${elems[3]}"

  #echo "Main URL: $firstyear->$lastyear $mainurl"

  if [[ $mainurl =~ '/web/*/' ]]; then
    listing="./$mainurl.txt"
    mkdir -p "$(dirname "$listing")"
    : >"$listing"
    oldlastcap="$(cat "./$mainurl.lastcap.txt" 2>/dev/null)"
    oldlastyear="$(cat "./$mainurl.lastyear.txt" 2>/dev/null)"
    : ${oldlastyear:=$lastyear}
    for y in $(seq $firstyear $lastyear); do
      u="/web/${y}0101000000*/${mainurl#/web/*/}"
      mkdir -p "$(dirname "./$u.html")"
      if ! [[ -s "./$u.html" ]] || ([[ $y -ge $oldlastyear ]] && [[ $lastcap != $oldlastcap ]]) ; then
        $wget "$webarchive$u" -O "./$u.html"
      fi
      #<"./$u.html" egrep 'href="/web/[0-9]+\*' | sed -r 's/.*href="([^"]*)".*/\1/' >"$d/$f.txt"
      <"./$u.html" egrep 'href="/web/[0-9]*/' | sed -r 's/.*href="([^"]*)".*/\1/' >>"$listing"
    done
    printf %s "$lastcap" >"./$mainurl.lastcap.txt"
    printf %s "$lastyear" >"./$mainurl.lastyear.txt"
    <"$listing" | sort | uniq >>listing2.txt
  else
    echo "$mainurl" >>listing2.txt
  fi

done <listing.txt

# Construct listing3.txt
# sort, uniq, use unmodified page appending id_ to the timestamp
# URL must start with "/web/YYYYMMDDHHMMSSid_/" only.
# This is the list of files that needs to be downloaded (if not already present)

<listing2.txt sort | uniq | sed -r 's:^/web/([0-9]*)/:/web/\1id_/:' >listing3.txt

# Download listing3

while read url; do

  if [[ $url != ${url%/} ]]; then
    f="./$url/.index"
  else
    f="./$url"
  fi
    
  mkdir -p "$(dirname "$f")"
  if ! [[ -s "$f" ]]; then
    $wget "$webarchive$url" -O "./$f"
  fi

done <listing3.txt