From 69bb49f49c15d7ff5dcb253bec166959422f86df Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Motiejus=20Jak=C5=A1tys?= <motiejus@uber.com>
Date: Sat, 13 Mar 2021 16:44:31 +0200
Subject: [PATCH] aggregate and filter rivers

---
 IV/Makefile                |  8 ++------
 IV/aggregate-rivers.sql    | 37 +++++++++++++++++++++++++++++++++++++
 IV/filter-rivers-query.awk | 12 ------------
 3 files changed, 39 insertions(+), 18 deletions(-)
 create mode 100644 IV/aggregate-rivers.sql
 delete mode 100755 IV/filter-rivers-query.awk

diff --git a/IV/Makefile b/IV/Makefile
index 8ced223..96e268a 100644
--- a/IV/Makefile
+++ b/IV/Makefile
@@ -1,11 +1,7 @@
-# HACK HACK
-EMPTY :=
-SPACE := $(EMPTY) $(EMPTY)
-
-RIVERS ?= Visinčia Šalčia Žeimena Lakaja Nemunas
+WHERE ?= name='Visinčia' OR name='Šalčia' OR name='Nemunas'
 
 .faux_filter-rivers: .faux_import-osm
-	./filter-rivers-query.awk $(RIVERS) | ./db -f -
+	./db -v where="$(WHERE)" -f aggregate-rivers.sql
 	touch $@
 
 .faux_import-osm: lithuania-latest.osm.pbf .faux.db
diff --git a/IV/aggregate-rivers.sql b/IV/aggregate-rivers.sql
new file mode 100644
index 0000000..786c4c5
--- /dev/null
+++ b/IV/aggregate-rivers.sql
@@ -0,0 +1,37 @@
+/* Aggregates rivers by name and proximity. */
+drop function if exists aggregate_rivers;
+create function aggregate_rivers() returns table(osm_id bigint, name text, way geometry) as $$
+declare
+  c record;
+  cc record;
+  changed boolean;
+begin
+  while (select count(1) from agg_tmp_objects) > 0 loop
+    select * from agg_tmp_objects limit 1 into c;
+    delete from agg_tmp_objects a where a.osm_id = c.osm_id;
+    changed = true;
+    while changed loop
+      changed = false;
+      for cc in (select * from agg_tmp_objects a where a.name = c.name and st_dwithin(a.way, c.way, 500)) loop
+        c.way = st_linemerge(st_union(c.way, cc.way));
+        delete from agg_tmp_objects a where a.osm_id = cc.osm_id;
+        changed = true;
+      end loop;
+    end loop; -- while changed
+    return query select c.osm_id, c.name, c.way;
+  end loop; -- count(1) from agg_tmp_objects > 0
+  return;
+end
+$$ language plpgsql;
+
+create temporary table agg_tmp_objects (osm_id bigint, name text, way geometry);
+create index agg_tmp_objects_id on agg_tmp_objects(osm_id);
+create index agg_tmp_objects_gix on agg_tmp_objects using gist(way) include(name);
+
+insert into agg_tmp_objects
+  select p.osm_id, p.name, p.way from planet_osm_line p
+    where waterway in ('river', 'stream', 'canal') and :where;
+
+drop table if exists agg_rivers;
+create table agg_rivers as (select * from aggregate_rivers());
+drop table agg_tmp_objects;
diff --git a/IV/filter-rivers-query.awk b/IV/filter-rivers-query.awk
deleted file mode 100755
index 4991147..0000000
--- a/IV/filter-rivers-query.awk
+++ /dev/null
@@ -1,12 +0,0 @@
-#!/usr/bin/awk -f
-
-BEGIN {
-    print "DROP TABLE IF EXISTS rivers;"
-    printf "CREATE TABLE rivers AS SELECT name,way FROM planet_osm_line WHERE "
-    for (i = 1; i < ARGC; i++) {
-        printf "name='%s'", ARGV[i]
-        if (i != ARGC - 1)
-            printf " OR ";
-    }
-    print ";";
-}