[build] adding libposatl_data.in which allows us to fetch data files from Github without using the API

This commit is contained in:
Al
2018-12-13 13:35:01 -05:00
parent 641395e811
commit 9542c8c435
3 changed files with 45 additions and 268 deletions

View File

@@ -1,232 +0,0 @@
#!/bin/sh
set -e
if [ "$#" -lt 3 ]; then
echo "Usage: ./libpostal_data [upload|download] [base|parser|language_classifier|all] data_dir"
exit 1
fi
LIBPOSTAL_VERSION_STRING="v1"
LIBPOSTAL_RELEASE_VERSION_STRING="v1.0.0"
LIBPOSTAL_REPO_NAME="openvenues/libpostal"
LIBPOSTAL_S3_BUCKET_NAME="libpostal"
LIBPOSTAL_S3_KEY="s3://$LIBPOSTAL_S3_BUCKET_NAME"
GITHUB_API_URL="https://api.github.com"
LIBPOSTAL_RELEASE_API_URL="$GITHUB_API_URL/repos/$LIBPOSTAL_REPO_NAME/releases"
LIBPOSTAL_DATA_FILE="libpostal_data.tar.gz"
LIBPOSTAL_PARSER_FILE="parser.tar.gz"
LIBPOSTAL_LANG_CLASS_FILE="language_classifier.tar.gz"
LIBPOSTAL_DATA_S3_PREFIX="$LIBPOSTAL_LATEST_MODEL_VERSION_STRING/parser.tar.gz"
LIBPOSTAL_PARSER_S3_PREFIX="$LIBPOSTAL_LATEST_DATA_VERSION_STRING/libpostal_data.tar.gz"
LIBPOSTAL_LANG_CLASS_S3_PREFIX="$LIBPOSTAL_LATEST_MODEL_VERSION_STRING/language_classifier.tar.gz"
COMMAND=$1
FILE=$2
LIBPOSTAL_DATA_DIR=$3
LIBPOSTAL_DATA_VERSION_FILE=$LIBPOSTAL_DATA_DIR/data_version
LIBPOSTAL_DATA_DIR_VERSION=
mkdir -p $LIBPOSTAL_DATA_DIR
LIBPOSTAL_DATA_UPDATED_PATH=$LIBPOSTAL_DATA_DIR/last_updated
LIBPOSTAL_PARSER_UPDATED_PATH=$LIBPOSTAL_DATA_DIR/last_updated_parser
LIBPOSTAL_LANG_CLASS_UPDATED_PATH=$LIBPOSTAL_DATA_DIR/last_updated_language_classifier
BASIC_MODULE_DIRS="address_expansions numex transliteration"
PARSER_MODULE_DIR=address_parser
LANGUAGE_CLASSIFIER_MODULE_DIR=language_classifier
export LC_ALL=C
EPOCH_DATE="Jan 1 00:00:00 1970"
MB=$((1024*1024))
CHUNK_SIZE=$((64*$MB))
LARGE_FILE_SIZE=$((CHUNK_SIZE*2))
NUM_WORKERS=12
kill_background_processes() {
jobs -p | xargs kill;
exit
}
trap kill_background_processes INT
PART_MSG='echo "Downloading part $1: filename=$5, offset=$2, max=$3"'
PART_CURL='curl -L $4 --silent -H"Range:bytes=$2-$3" --retry 3 --retry-delay 2 -o $5'
DOWNLOAD_PART="$PART_MSG;$PART_CURL"
download_multipart() {
url=$1
filename=$2
size=$3
num_chunks=$((size/CHUNK_SIZE))
echo "Downloading multipart: $url, size=$size, num_chunks=$num_chunks"
offset=0
i=0
while [ $i -lt $num_chunks ]; do
i=$((i+1))
part_filename="$filename.$i"
if [ $i -lt $num_chunks ]; then
max=$((offset+CHUNK_SIZE-1));
else
max=$size;
fi;
printf "%s\0%s\0%s\0%s\0%s\0" "$i" "$offset" "$max" "$url" "$part_filename"
offset=$((offset+CHUNK_SIZE))
done | xargs -0 -n 5 -P $NUM_WORKERS sh -c "$DOWNLOAD_PART" --
> $local_path
i=0
while [ $i -lt $num_chunks ]; do
i=$((i+1))
part_filename="$filename.$i"
cat $part_filename >> $local_path
rm $part_filename
done;
}
download_file() {
updated_path=$1
data_dir=$2
metadata_url=$3
url=$4
size=$5
filename=$6
name=$7
shift 7
subdirs=$@
local_path=$data_dir/$filename
if [ ! -e $updated_path ]; then
echo "$EPOCH_DATE" > $updated_path;
fi;
echo "Checking for new libpostal $name..."
if [ $(curl -LsI $metadata_url -z "$(cat $updated_path)" --remote-time -w %{http_code} -o /dev/null | grep "^200$") ]; then
echo "New libpostal $name available"
if [ $size -ge $LARGE_FILE_SIZE ]; then
download_multipart $url $local_path $size
else
curl -L $url --retry 3 --retry-delay 2 -o $local_path
fi
if date -ud "@$(date -ur . +%s)" >/dev/null 2>&1; then
echo $(date -ud "$(date -ud "@$(date -ur $local_path +%s)") + 1 second") > $updated_path;
elif stat -f %Sm . >/dev/null 2>&1; then
echo $(date -ur $(stat -f %m $local_path) -v+1S) > $updated_path;
fi;
for subdir in $subdirs; do
rm -rf $data_dir/$subdir;
done
tar -xvzf $local_path --no-same-owner -C $data_dir;
rm $local_path;
else
echo "libpostal $name up to date"
fi
}
if [ $COMMAND = "download" ]; then
if [ -e $LIBPOSTAL_DATA_VERSION_FILE ]; then
LIBPOSTAL_DATA_DIR_VERSION=$(cat $LIBPOSTAL_DATA_VERSION_FILE)
fi
if [ "$LIBPOSTAL_DATA_DIR_VERSION" != "$LIBPOSTAL_VERSION_STRING" ]; then
echo "Old version of datadir detected, removing..."
for subdir in $BASIC_MODULE_DIRS $PARSER_MODULE_DIR $LANGUAGE_CLASSIFIER_MODULE_DIR; do
rm -rf $LIBPOSTAL_DATA_DIR/$subdir;
done
# Legacy, blow it away too to be nice
if [ -e $LIBPOSTAL_DATA_DIR/geodb ]; then
rm -rf $LIBPOSTAL_DATA_DIR/geodb;
fi
rm -f $LIBPOSTAL_DATA_UPDATED_PATH
rm -f $LIBPOSTAL_LANG_CLASS_UPDATED_PATH
rm -f $LIBPOSTAL_PARSER_UPDATED_PATH
fi
mkdir -p $LIBPOSTAL_DATA_DIR
release_id=$(curl -s $LIBPOSTAL_RELEASE_API_URL/tags/$LIBPOSTAL_RELEASE_VERSION_STRING | grep "\"id\"" | head -n1 | grep -o '[0-9][0-9]*')
release_assets="$(curl -s $LIBPOSTAL_RELEASE_API_URL/$release_id/assets)"
asset_names_tempfile="$LIBPOSTAL_DATA_DIR/asset_names.tmp"
echo "$release_assets" | grep -o '"name": *"[^"]*"' | grep -o '"[^"]*"$' | tr -d '"' > $asset_names_tempfile
asset_metadata_tempfile="$LIBPOSTAL_DATA_DIR/asset_metadata.tmp"
echo "$release_assets" | grep -o '"url": *"[^"]*/releases/assets/[0-9]*"' | grep -o '"[^"]*"$' | tr -d '"' > $asset_metadata_tempfile
asset_urls_tempfile="$LIBPOSTAL_DATA_DIR/asset_urls.tmp"
echo "$release_assets" | grep -o '"browser_download_url": *"[^"]*"' | grep -o '"[^"]*"$' | tr -d '"' > $asset_urls_tempfile
asset_sizes_tempfile="$LIBPOSTAL_DATA_DIR/asset_sizes.tmp"
echo "$release_assets" | grep -o '"size": *[0-9]*' | grep -o '[0-9]*$' > $asset_sizes_tempfile
assets_tempfile="$LIBPOSTAL_DATA_DIR/assets.tmp"
paste -d' ' $asset_names_tempfile $asset_metadata_tempfile $asset_urls_tempfile $asset_sizes_tempfile > $assets_tempfile
rm $asset_names_tempfile $asset_metadata_tempfile $asset_urls_tempfile $asset_sizes_tempfile
while read -r line; do
asset=$(echo "$line" | cut -f1 -d' ')
asset_metadata_url=$(echo "$line" | cut -f2 -d' ')
asset_url=$(echo "$line" | cut -f3 -d' ')
asset_size=$(echo "$line" | cut -f4 -d' ')
if [ $asset = $LIBPOSTAL_DATA_FILE ] && ([ $FILE = "base" ] || [ $FILE = "all" ]); then
download_file $LIBPOSTAL_DATA_UPDATED_PATH $LIBPOSTAL_DATA_DIR $asset_metadata_url $asset_url $asset_size $LIBPOSTAL_DATA_FILE "data file" $BASIC_MODULE_DIRS
fi
if [ $asset = $LIBPOSTAL_PARSER_FILE ] && ([ $FILE = "parser" ] || [ $FILE = "all" ]); then
download_file $LIBPOSTAL_PARSER_UPDATED_PATH $LIBPOSTAL_DATA_DIR $asset_metadata_url $asset_url $asset_size $LIBPOSTAL_PARSER_FILE "parser data file" $PARSER_MODULE_DIR
fi
if [ $asset = $LIBPOSTAL_LANG_CLASS_FILE ] && ([ $FILE = "language_classifier" ] || [ $FILE = "all" ]); then
download_file $LIBPOSTAL_LANG_CLASS_UPDATED_PATH $LIBPOSTAL_DATA_DIR $asset_metadata_url $asset_url $asset_size $LIBPOSTAL_LANG_CLASS_FILE "language classifier data file" $LANGUAGE_CLASSIFIER_MODULE_DIR
fi
if [ "$LIBPOSTAL_DATA_DIR_VERSION" != "$LIBPOSTAL_VERSION_STRING" ]; then
echo $LIBPOSTAL_VERSION_STRING > $LIBPOSTAL_DATA_VERSION_FILE;
fi
done < $assets_tempfile;
rm $assets_tempfile
elif [ $COMMAND = "upload" ]; then
echo "upload not implemented yet"
#if [ $FILE = "base" ] || [ $FILE = "all" ]; then
# tar -C $LIBPOSTAL_DATA_DIR -cvzf $LIBPOSTAL_DATA_DIR/$LIBPOSTAL_DATA_FILE $BASIC_MODULE_DIRS
# aws s3 cp --acl=public-read $LIBPOSTAL_DATA_DIR/$LIBPOSTAL_DATA_FILE $LIBPOSTAL_S3_KEY/$LIBPOSTAL_DATA_S3_PREFIX/
#fi
#if [ $FILE = "parser" ] || [ $FILE = "all" ]; then
# latest_parser=$(curl --silent $LIBPOSTAL_S3_BUCKET_URL/$LIBPOSTAL_PARSER_S3_PREFIX/latest)
# tar -C $LIBPOSTAL_DATA_DIR -cvzf $LIBPOSTAL_DATA_DIR/$LIBPOSTAL_PARSER_FILE $PARSER_MODULE_DIR
# parser_s3_dir="$LIBPOSTAL_S3_KEY/$LIBPOSTAL_PARSER_S3_PREFIX/$latest_parser/"
# aws s3 cp --acl=public-read $LIBPOSTAL_DATA_DIR/$LIBPOSTAL_PARSER_FILE $parser_s3_dir
#fi
#if [ $FILE = "language_classifier" ] || [ $FILE = "all" ]; then
# latest_lang_class=$(curl --silent $LIBPOSTAL_S3_BUCKET_URL/$LIBPOSTAL_LANG_CLASS_S3_PREFIX/latest)
# tar -C $LIBPOSTAL_DATA_DIR -cvzf $LIBPOSTAL_DATA_DIR/$LIBPOSTAL_LANG_CLASS_FILE $LANGUAGE_CLASSIFIER_MODULE_DIR
# lang_class_s3_dir="$LIBPOSTAL_S3_KEY/$LIBPOSTAL_LANG_CLASS_S3_PREFIX/$latest_lang_class/"
# aws s3 cp --acl=public-read $LIBPOSTAL_DATA_DIR/$LIBPOSTAL_LANG_CLASS_FILE $lang_class_s3_dir
#fi
else
echo "Invalid command: $COMMAND"
exit 1
fi

View File

@@ -178,27 +178,6 @@ if [ $COMMAND = "download" ]; then
download_release $LIBPOSTAL_LANG_CLASS_MODEL_VERSION_PATH $LIBPOSTAL_LANG_CLASS_MODEL_LATEST_VERSION $LIBPOSTAL_DATA_DIR $LIBPOSTAL_LANG_CLASS_MODEL_CHUNKS $LIBPOSTAL_LANG_CLASS_FILE "language classifier data file" $LANGUAGE_CLASSIFIER_MODULE_DIR
fi
elif [ $COMMAND = "upload" ]; then
echo "upload not implemented yet"
#if [ $FILE = "base" ] || [ $FILE = "all" ]; then
# tar -C $LIBPOSTAL_DATA_DIR -cvzf $LIBPOSTAL_DATA_DIR/$LIBPOSTAL_DATA_FILE $BASIC_MODULE_DIRS
# aws s3 cp --acl=public-read $LIBPOSTAL_DATA_DIR/$LIBPOSTAL_DATA_FILE $LIBPOSTAL_S3_KEY/$LIBPOSTAL_DATA_S3_PREFIX/
#fi
#if [ $FILE = "parser" ] || [ $FILE = "all" ]; then
# latest_parser=$(curl --silent $LIBPOSTAL_S3_BUCKET_URL/$LIBPOSTAL_PARSER_S3_PREFIX/latest)
# tar -C $LIBPOSTAL_DATA_DIR -cvzf $LIBPOSTAL_DATA_DIR/$LIBPOSTAL_PARSER_FILE $PARSER_MODULE_DIR
# parser_s3_dir="$LIBPOSTAL_S3_KEY/$LIBPOSTAL_PARSER_S3_PREFIX/$latest_parser/"
# aws s3 cp --acl=public-read $LIBPOSTAL_DATA_DIR/$LIBPOSTAL_PARSER_FILE $parser_s3_dir
#fi
#if [ $FILE = "language_classifier" ] || [ $FILE = "all" ]; then
# latest_lang_class=$(curl --silent $LIBPOSTAL_S3_BUCKET_URL/$LIBPOSTAL_LANG_CLASS_S3_PREFIX/latest)
# tar -C $LIBPOSTAL_DATA_DIR -cvzf $LIBPOSTAL_DATA_DIR/$LIBPOSTAL_LANG_CLASS_FILE $LANGUAGE_CLASSIFIER_MODULE_DIR
# lang_class_s3_dir="$LIBPOSTAL_S3_KEY/$LIBPOSTAL_LANG_CLASS_S3_PREFIX/$latest_lang_class/"
# aws s3 cp --acl=public-read $LIBPOSTAL_DATA_DIR/$LIBPOSTAL_LANG_CLASS_FILE $lang_class_s3_dir
#fi
else
echo "Invalid command: $COMMAND"
exit 1