D=-dsco NAME="街路名群 Street name groups" \
-dsco DESCRIPTION="See $(subst $(HOME)/,https://www.,$(PWD))"
## Author: Dan Jacobson https://www.jidanni.org/
## Copyright: https://www.gnu.org/licenses/gpl.html
## Created: 2025-04-27T12:55:10+0000
## Last-Updated: 2025-05-10T22:48:48+0000
##     Update #: 178
d=~/Downloads/d/data
#A={{geocodeArea:Beitun District, Taichung}}
A=area(id:3602959946)#what is official way to get these? Wikidata?
# https://community.openstreetmap.org/t/how-to-get-all-street-names-of-city/88436/5
# https://wiki.openstreetmap.org/wiki/Overpass_API/Overpass_QL#CSV_output_mode
W=https://overpass-api.de/api/interpreter
vv=[out:csv(::lon,::lat)]\
[timeout:20];$A->.searchArea; way[\
"highway"]["name"~"^$*"](area.searchArea);out geom;
want = \
安順 崇德 昌平 軍福 敦富 文昌 景賢 南興 祥順 遼陽 松竹 后庄 \
豐樂 太和 敦和 太順 山西 經貿 河北 軍榮 洲際 和祥 三光 松柏 \
豐美 中清 太安 順平 榮德 德和 敦化 太祥 豐德 同榮
ww  = $(addsuffix .wkt,$(want))
$d/%.csv:
	perl -MURI::Escape -wle 'use utf8; print "$W?data=", uri_escape_utf8(q($(vv)))'|\
	xargs wget -O $@
# Concave hull approach:
# %.kml:$d/%.csv
# 	ogr2ogr -q $@ $< -dialect SQLite -of LIBKML \
# 	-oo X_POSSIBLE_NAMES=@lon -oo Y_POSSIBLE_NAMES=@lat -sql \
# 	'SELECT ST_ConcaveHull(ST_Collect(GEOMETRY)), "$*" AS name FROM "$*"'
#street_name_groups.kmz:$(ww); ogrmerge -f LIBKML -overwrite_ds -o $@ $^ $D
# Single point approach:
%.wkt:$d/%.csv
	ogr2ogr -of CSV -q /vsistdout/ $< -dialect SQLite \
	-lco GEOMETRY=AS_WKT -lco STRING_QUOTING=IF_NEEDED \
	-oo X_POSSIBLE_NAMES=@lon -oo Y_POSSIBLE_NAMES=@lat -sql \
	'SELECT ST_Centroid(Extent(GEOMETRY)) AS WKT,'\
	'CONCAT("$*(",COUNT(*),")") AS name FROM "$*"' > $@ #so it won't make a directory.
mm.csv:$(ww); sort -o $@ -ur $^ #to leave just one header line at top
street_name_groups.kml:mm.csv; ogr2ogr -of LIBKML $@ $< -sql 'SELECT name FROM $(basename $<)'

## Here's how we get the list of what are the heavy use prefixes for this area:
HEAVY=3 #minimum number of streets for a group to be interesting to us
B=beitun
b=$d/$B.csv
numbers = 一|二|三|四|五|六|七|八|九|十
directions = 東|西|南|北|中
rank:$b #interesting name stems, by frequency
	LC_ALL=zh_TW.UTF-8 perl -Mutf8 -C -nwle \
	'next unless /..(:?$(numbers)|$(directions))(:?街|路)$$/; /../; $$h{$$&}++;'\
	'END{for(sort{$$h{$$b}<=>$$h{$$a}}keys %h){print "$$_,$$h{$$_}"}}' $< > $@
	test -s $@
r:rank; perl -F, -anwle 'exit unless $$F[1] >= $(HEAVY); print $$F[0]' $<|xargs -n 12
# And then I pasted the list into this Makefile.

R=太 順 祥 和
v:rank; for i in $R; do	printf $$i:; perl -nwe 's/,.*//; print;' $<; done
# Not considering 文昌四巷, 碧柳一巷, etc. non-useful exceptions.
# Note e.g., 崇德 roads might extend between Districts, so perhaps use a bounding box instead.
.PRECIOUS: $d/%.csv
.SECONDARY: