colouring-montreal/etl/extract_mastermap.sh

45 lines
750 B
Bash
Raw Normal View History

2018-09-21 06:10:39 -04:00
#!/usr/bin/env bash
#
# Extract MasterMap
#
: ${1?"Usage: $0 ./path/to/mastermap/dir"}
data_dir=$1
2018-09-21 06:10:39 -04:00
#
# Extract buildings from *.gz to CSV
2018-05-30 15:18:37 -04:00
#
# Features where::
# descriptiveGroup = '(1:Building)'
#
# Use `fid` as source ID, aka TOID.
2018-09-21 06:10:39 -04:00
#
2018-05-30 15:18:37 -04:00
2022-03-10 04:59:22 -05:00
#
# Unzip to GML
#
find $data_dir -type f -name '*.zip' -printf "%f\n" | \
parallel \
unzip -u $data_dir/{} -d $data_dir
find $data_dir -type f -name '*.gz' -printf "%f\n" | \
parallel \
gunzip $data_dir/{} -k -S gml
2018-05-30 15:18:37 -04:00
2018-10-02 16:12:46 -04:00
rename 's/$/.gml/' $data_dir/*[^gzvt]
2018-05-30 15:18:37 -04:00
find $data_dir -type f -name '*.gml' -printf "%f\n" | \
2018-05-30 15:18:37 -04:00
parallel \
ogr2ogr \
2018-09-21 06:10:39 -04:00
-select fid,descriptiveGroup \
-f CSV $data_dir/{}.csv \
$data_dir/{} \
TopographicArea \
-lco GEOMETRY=AS_WKT
rm $data_dir/*.gfs
2018-10-03 15:10:16 -04:00
rm $data_dir/*.gml