#!/bin/sh

# Public domain notice for all NCBI EDirect scripts is located at:
# https://www.ncbi.nlm.nih.gov/books/NBK179288/#chapter6.Public_Domain_Notice

doall=false
dofresh=false
flag="none"
asn=false
pme=false
turbo=false

while [ $# -gt 0 ]
do
  case "$1" in
    -strict )
      flag="strict"
      shift
      ;;
    -mixed )
      flag="mixed"
      shift
      ;;
    -path )
      shift
      ;;
    -all )
      doall=true
      shift
      ;;
    -fresh )
      dofresh=true
      shift
      ;;
    -asn )
      asn=true
      shift
      ;;
    -xml )
      asn=false
      shift
      ;;
    -pme )
      pme=true
      shift
      ;;
    -turbo )
      turbo=true
      shift
      ;;
    -* )
      exec >&2
      echo "$0: Unrecognized option $1"
      exit 1
      ;;
    * )
      break
      ;;
  esac
done

# get local master and working volumes from database
ev=$( rchive -local pubmed )
if [ -n "$ev" ]
then
  # only care about master volume and its Archive subfolder
  target="${ev%:*}"
fi

if [ -z "$target" ]
then
  echo "ERROR: Must supply path to local data by setting EDIRECT_LOCAL_MASTER environment variable" >&2
  exit 1
fi

osname=`uname -s | sed -e 's/_NT-.*$/_NT/; s/^MINGW[0-9]*/CYGWIN/'`
if [ "$osname" = "CYGWIN_NT" -a -x /bin/cygpath ]
then
  target=`cygpath -w "$target"`
fi

# remove trailing slash
target=${target%/}

if [ ! -d "$target" ]
then
  echo "ERROR: Local archive ${target} is not mounted" >&2
  exit 1
fi

target=$( echo "${target}/Archive")

if [ ! -d "$target" ]
then
  echo "ERROR: Archive directory ${target} is not present" >&2
  exit 1
fi

HEAD=$(cat <<EOF
<?xml version="1.0" encoding="UTF-8"?>
<!DOCTYPE PubmedArticleSet PUBLIC "-//NLM//DTD PubMedArticle, 1st January 2019//EN" "https://dtd.nlm.nih.gov/ncbi/pubmed/out/pubmed_190101.dtd">
<PubmedArticleSet>
EOF
)

TAIL=$(cat <<EOF
</PubmedArticleSet>
EOF
)

if [ "$doall" = true ]
then
  mx=$(
    for dr in "$target"/*
    do
      if [ -d "$dr" ]
      then
        bs=$(basename "$dr")
        echo "$bs"
      fi
    done |
    sort -n | tail -n 1
  )
  maximum=$(((mx + 1) * 1000000))
  fr=0
  chunk_size=250000
  to=$((chunk_size - 1))
  loop_max=$((maximum / chunk_size))
  echo "$HEAD"
  seq 1 $loop_max | while read n
  do
   seq -f "%0.f" $fr $to | rchive -gzip -db pubmed -fetch "$target"
   fr=$((fr + chunk_size))
   to=$((to + chunk_size))
   perl -e "select(undef, undef, undef, 0.1)"
  done
  echo "$TAIL"
elif [ "$dofresh" = true ]
then
  mx=$(
    for dr in "$target"/*
    do
      if [ -d "$dr" ]
      then
        bs=$(basename "$dr")
        echo "$bs"
      fi
    done |
    sort -n | tail -n 1
  )
  maximum=$(((mx + 1) * 1000000))
  fr=0
  chunk_size=250000
  to=$((chunk_size - 1))
  loop_max=$((maximum / chunk_size))
  seq 1 $loop_max | while read n
  do
    base=$(printf pubmed%03d $n)
    if [ -f "$base.xml.gz" ]
    then
      fr=$((fr + chunk_size))
      to=$((to + chunk_size))
      continue
    fi
    if [ -f "$base.xml" ]
    then
      continue
    fi
    echo "$base XML"
    seconds_start=$(date "+%s")
    seq -f "%0.f" $fr $to | rchive -gzip -db pubmed -fetch "$target" -head "<PubmedArticleSet>" -tail "</PubmedArticleSet>" > "$base.xml"
    fr=$((fr + chunk_size))
    to=$((to + chunk_size))
    seconds_end=$(date "+%s")
    seconds=$((seconds_end - seconds_start))
    echo "$seconds seconds"
    fsize=$(wc -c < "$base.xml")
    if [ "$fsize" -le 300 ]
    then
      rm "$base.xml.gz"
      exit 0
    fi
    perl -e "select(undef, undef, undef, 0.1)"
  done
elif [ "$asn" = true ]
then
  rchive -gzip -asn -flag "$flag" -db pubmed -fetch "$target"
elif [ "$pme" = true ]
then
  rchive -gzip -flag "$flag" -db pubmed -fetch "$target" |
  pma2pme -std
elif [ "$turbo" = true ]
then
  echo "$HEAD"
  rchive -gzip -db pubmed -flag "$flag" -turbo -db pubmed -fetch "$target"
  echo "$TAIL"
else
  echo "$HEAD"
  rchive -gzip -db pubmed -flag "$flag" -db pubmed -fetch "$target"
  echo "$TAIL"
fi
