#!/bin/sh
      test $# -gt 0||exec echo "
      10 requests per second
      500 requests per day without acct 
      tokens last 2 weeks

      usage: ${0##*/} =querystring -- search
      usage: ${0##*/} resume < 1.json >> 1.json -- continuation search
      usage: ${0##*/} 12345678 -- info
      usage: ${0##*/} 12345678 pdf -- download pdf
      usage: ${0##*/} -h -- search/api help
      example: ${0##*/} =created_at%3a%5bNOW-1DAY+TO+*%5d -- all docs created in last 24h
      "

      x=https://api.www.documentcloud.org/api/documents
      y=""
      case $1 in =*)
      ftp -4o\|yy059 "$x/search/?q$1&expand=user%2Corganization&version=2.0&hl=true&per_page=100"|tmux loadb -b dcl /dev/stdin
      tmux saveb -b dcl /dev/stdout
      while true;do
      y=$(tmux saveb -b dcl /dev/stdout|sed -n '/\"next\":\"/{s///;s/\"//p;}')
      test ${#y} -ge 140||exit
      ftp -4o\|yy059 "$y"|tmux loadb -b dcl /dev/stdin
      tmux saveb -b dcl /dev/stdout
      done 
      ;;resume)
      y=$(grep \"next\":|tail -1|sed -n '/\"next\":\"/{s///;s/\"//p;}');
      ftp -4o\|yy059 "$x/search/?q$1&expand=user%2Corganization&version=2.0&hl=true&per_page=100"|tmux loadb -b dcl /dev/stdin
      tmux saveb -b dcl /dev/stdout
      while true;do
      y=$(tmux saveb -b dcl /dev/stdout|sed -n '/\"next\":\"/{s///;s/\"//p;}')
      test ${#y} -ge 140||exit
      ftp -4o\|yy059 "$y" |tmux loadb -b dcl /dev/stdin 
      tmux saveb -b dcl /dev/stdout
      done
      ;;-h)
      exec echo '
      https://api.www.documentcloud.org/pages/help/search/
      https://api.www.documentcloud.org/pages/help/api/
      https://solr.apache.org/guide/solr/latest/query-guide/common-query-parameters.html
      '
      esac 

      ftp -4o\|yy059 $x/$1/?expand=user%2Corganization%2Cnotes%2Csections%2Cnotes.organization%2Cnotes.user \
      |if test $# -gt 1;
       then y=$(sed -n -e "/\"canonical_url\":/{ s///;s/\"//g;s/www/s3/; s/-/\//;s/$/.$2/p;}");
       ftp -4o $1.$2 $y;
       else sed;exit;
       fi 
         
         

could use curl, python, etc. to send token instead of haproxy

http-request add-header cookie "csrftoken=...; sessionid=..." if { hdr(host) api.www.documentcloud.org }

could use curl, etc. instead of ftp (tnftp)

yy059 is a filter that prints JSON left justified according to personal preference

could use jq, etc. instead of yy059

could use a temp file instead of tmux buffers

verdict: documentcloud.org is a welcome alternative to proprietary algo search engines and opinionated news, not to mention ads and tracking

interesting fact: the site actually offers a free "scraping service" if sign up (not for me; dont like python)