#!/bin/bash

# Bash script to download geoscrub_input table dump (created by AMK) from the

# vegbien database, and load it into the geoscrub database (i.e., the

# postgis database prepped with geonames.org data, GADM2 data, and

# associated mapping tables).

#

# Won't be necessary if we end up injecting all of the geoscrubbing and

# geovalidation functionality directly into vegbien itself. And if we

# end up implementing this stuff as a standalone service instead, we'd

# need to rethink (and generalize) how the input data is handled. But

# for now, this should at least serve as a placeholder that could be

# tweaked manually to load any arbitrary geoscrub input data table.

#

# Jim Regetz

# NCEAS

# Created Nov 2012

# 

# Paul Sarando

# iPlant Collaborative

# Updated Oct 2013

# Note, to force data to download from DATA_URL, ensure the DATAFILE is deleted

# before running this script.

DB_NAME="geoscrub"

DB_USER="bien"

DB_HOST="localhost"

SCRIPT_DIR=$(dirname $0)

DATA_URL="http://fs.vegpath.org/exports/geoscrub_input.no_header.cols=country,stateProvince,county,decimalLatitude,decimalLongitude.csv"

DATADIR="${SCRIPT_DIR}/input"

DATAFILE="${DATADIR}/geoscrub-corpus.csv"

if [[ ! -d "$DATADIR" ]]; then

    echo "making directory ${DATADIR}"

    mkdir -p "$DATADIR"

    if [[ $? != 0 ]]; then

        echo "Could not create directory ${DATADIR}"

        exit 1

    fi

fi

if [[ ! -r "$DATAFILE" ]]; then

    # download distinct records from vegbien

    wget -O "$DATAFILE" "$DATA_URL"

    if [[ $? != 0 ]]; then

        echo "Could not download input to ${DATAFILE}"

        exit 1

    fi

fi

echo "Loading vegbien data from ${DATAFILE}"

# clear previous data

psql -e -U "$DB_USER" -h "$DB_HOST" -d "$DB_NAME" --set ON_ERROR_STOP=1 < "${SCRIPT_DIR}/truncate.vegbien_geoscrub.sql"

if [[ $? != 0 ]]; then

    echo "Could not clear data from vegbien_geoscrub tables."

    exit 1

fi

# load

psql -U "$DB_USER" -h "$DB_HOST" -d "$DB_NAME" -c "\COPY vegbien_geoscrub FROM '${DATAFILE}' WITH CSV"