tommy/libpostal
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

[geoplanet] fixing geoplanet aliases insert warning

Browse Source
This commit is contained in:
Al
2016-12-08 02:31:10 -05:00
parent 763c86dcd4
commit 28d9ef12c0
1 changed files with 5 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files

6
scripts/geodata/geoplanet/create_geoplanet_db.sh
View File

@@ -22,10 +22,11 @@ GEOPLANET_ZIP_FILE="geoplanet_data_7.10.0.zip"
GEOPLANET_URL="https://archive.org/download/$GEOPLANET_ZIP_FILE/$GEOPLANET_ZIP_FILE"
GEOPLANET_ORIGINAL_PLACES_FILE="geoplanet_places_7.10.0.tsv"
GEOPLANET_ADMINS_FILE="geoplanet_admins_7.10.0.tsv"
GEOPLANET_ALIASES_FILE="geoplanet_aliases_7.10.0.tsv"
GEOPLANET_ORIGINAL_ALIASES_FILE="geoplanet_aliases_7.10.0.tsv"
GEOPLANET_PLACES_FILE="geoplanet_places.tsv"
GEOPLANET_POSTAL_CODES_FILE="geoplanet_postal_codes.tsv"
GEOPLANET_ALIASES_FILE="geoplanet_aliases.tsv"
GEOPLANET_GEONAMES_CONCORDANCE_FILE="geonames-geoplanet-matches.csv"
GEOPLANET_GEONAMES_CONCORDANCE_URL="https://github.com/blackmad/geoplanet-concordance/raw/master/current/$GEOPLANET_GEONAMES_CONCORDANCE_FILE"
@@ -57,6 +58,9 @@ awk -F'\t' 'BEGIN{OFS="\t";} {if ($5 == "Zip") print $0;}' $GEOPLANET_ORIGINAL_P
echo "Creating GeoPlanet places file"
awk -F'\t' 'BEGIN{OFS="\t";} {if ($5 == "Continent" || $5 == "Country" || $5 == "Nationality" || $5 == "State" || $5 == "County" || $5 == "Town" || $5 == "LocalAdmin" || $5 == "Island" || $5 == "Suburb") print $0;}' $GEOPLANET_ORIGINAL_PLACES_FILE > $GEOPLANET_PLACES_FILE
echo "Creating GeoPlanet aliases file"
tail +2 $GEOPLANET_ORIGINAL_ALIASES_FILE > $GEOPLANET_ALIASES_FILE
echo "Fetching GeoNames concordance"
download_file $GEOPLANET_GEONAMES_CONCORDANCE_URL $GEOPLANET_GEONAMES_CONCORDANCE_FILE