#!/usr/bin/env bash

# annex-ec: Use erasure codes for more efficient storage use in git-annex
# Copyright (C) 2026  Scott Worley

# This program is free software: you can redistribute it and/or modify
# it under the terms of the GNU Affero General Public License as
# published by the Free Software Foundation, either version 3 of the
# License, or (at your option) any later version.

set -euo pipefail

die() {
  echo "$*" >&2
  exit 1
}


parse_volume_list() {
  this_volume_name=$(join -j1 <(git config get annex.uuid) <(git cat-file -p git-annex:uuid.log|sort)  | cut -d' ' -f2)
  tmp=()
  mapfile -d , tmp <<< "$1"
  volumes=()
  for x in "${tmp[@]}";do
    x=${x%,}
    x=${x%$'\n'}
    if [[ "$x" ]];then
      if [[ "$x" == "$this_volume_name" ]];then
        volumes+=( "here" )
      else
        volumes+=( "$x" )
      fi
    fi
  done
}

make_name() {
  FILENAME_MAX=255
  EXAMPLE_SUFFIX='.vol0000+9999.par2'
  EXAMPLE_ANNEX_INTERNAL_PREFIX='ingest-'
  EXAMPLE_ANNEX_INTERNAL_SUFFIX='-1-1173fd7'
  num_separating_dashes=$((N - 1))
  overhead=$(( ${#EXAMPLE_SUFFIX} + ${#EXAMPLE_ANNEX_INTERNAL_PREFIX} + ${#EXAMPLE_ANNEX_INTERNAL_SUFFIX} + num_separating_dashes ))
  available=$((FILENAME_MAX - overhead))
  len=$((available / N))
  name=$(find "$@" -printf '%l\n' | sed -r 's/.*SHA256E-s[0-9]+--//;s/\..*//' | cut -c-$len | tr \\n -)
  name=${name%-}
}

volumes=()
redundancy=1
max_block_size=$((128*1024*1024))
block_size_is_a_multiple_of=4  # par2 requires that this be at least 4
blocks_per_file=10
while getopts b:m:r:v:x: opt;do
 case $opt in
  b)    blocks_per_file=$OPTARG;;
  m)    block_size_is_a_multiple_of=$OPTARG;;
  r)    redundancy=$OPTARG;;
  v)    parse_volume_list "$OPTARG";;
  x)    max_block_size=$OPTARG;;
  *)    echo 'usage: annex-ec [-v remote1,remote2,...] [-r N] [-b N] [-x N] [-m N] file file...' >&2; exit 1;;
 esac
done
shift $((OPTIND - 1))

if (( ${#volumes[@]} == 0 ));then
  parse_volume_list "here,$(git remote | tr \\n ,)"
fi

N=$((${#volumes[@]} - redundancy))

(( $# == N )) || die "Expected $N files in this group ($N + $redundancy = ${#volumes[@]}), but got $#"


git annex get -- "$@"

max_size=$(find -L "$@" -printf '%s\n' | sort -nr | head -n1)
while true;do
  block_size=$(( ((max_size/(block_size_is_a_multiple_of*blocks_per_file))+1) * block_size_is_a_multiple_of))
  if (( block_size < max_block_size ));then break;fi
  blocks_per_file=$((blocks_per_file + 1))
done

make_name "$@"

if [[ ! -d ec ]];then
  mkdir ec
  # TODO: Make this robust against being interrupted here
  echo '* annex.numcopies=1' >> ec/.gitattributes
  git add ec/.gitattributes
fi

par2 c -u -n"$redundancy" -c"$((blocks_per_file * redundancy))" -s"$block_size" "$name.par2" "$@"
rm "$name.par2"
mv "$name.vol"* ec/

i=0
for f;do
  target_volume="${volumes[i]}"
  for volume in "${volumes[@]}";do
    if [[ "$volume" != here ]];then
      if [[ "$volume" == "$target_volume" ]]; then
        git annex copy --to "$volume" "$f"
      fi
    fi
  done
  i=$((i+1))
done

for f in ec/"$name.vol"*;do
  target_volume="${volumes[i]}"
  git annex add "$f"
  if [[ "$target_volume" != here ]];then
    git annex move --to "$target_volume" "$f"
  fi
  i=$((i+1))
done

for f;do
  echo "${f// /[[:space:]]} annex.numcopies=1" >> .gitattributes
done

for volume in here "${volumes[@]}";do
  i=0
  to_drop=()
  for f;do
    target_volume="${volumes[i]}"
    if [[ "$volume" != "$target_volume" ]]; then
      to_drop+=( "$f" )
    fi
    i=$((i+1))
  done
  if [[ "$volume" == here ]];then
    git annex drop "${to_drop[@]}"
  else
    git annex drop --from "$volume" "${to_drop[@]}"
  fi
done
git add .gitattributes

(
  flock 1
  echo "$name"
  for f;do
    echo " $f"
  done
) >> ec/.meta
git add ec/.meta