import random
import string
import re
-from nose.tools import * # for assert functions
import psycopg2.extras
+from check_functions import Almost
+
class PlaceColumn:
def __init__(self, context, force_name):
self.add_hstore('extratags', key[6:], value)
elif key.startswith('addr+'):
self.add_hstore('address', key[5:], value)
+ elif key in ('name', 'address', 'extratags'):
+ self.columns[key] = eval('{' + value + '}')
else:
- assert_in(key, ('class', 'type'))
+ assert key in ('class', 'type')
self.columns[key] = None if value == '' else value
def set_key_name(self, value):
self.add_hstore('name', 'name', value)
def set_key_osm(self, value):
- assert_in(value[0], 'NRW')
- ok_(value[1:].isdigit())
+ assert value[0] in 'NRW'
+ assert value[1:].isdigit()
self.columns['osm_type'] = value[0]
self.columns['osm_id'] = int(value[1:])
def set_key_geometry(self, value):
self.geometry = self.context.osm.parse_geometry(value, self.context.scene)
- assert_is_not_none(self.geometry)
+ assert self.geometry is not None
def add_hstore(self, column, key, value):
if column in self.columns:
self.columns[column] = { key : value }
def db_insert(self, cursor):
- assert_in('osm_type', self.columns)
+ assert 'osm_type' in self.columns
if self.force_name and 'name' not in self.columns:
self.add_hstore('name', 'name', ''.join(random.choice(string.printable)
for _ in range(int(random.random()*30))))
self.geometry = "ST_SetSRID(ST_Point(%f, %f), 4326)" % pt
else:
- assert_is_not_none(self.geometry, "Geometry missing")
+ assert self.geometry is not None, "Geometry missing"
query = 'INSERT INTO place (%s, geometry) values(%s, %s)' % (
','.join(self.columns.keys()),
','.join(['%s' for x in range(len(self.columns))]),
if self.pid is None:
return "<null>"
+ if self.pid == 0:
+ return "place ID 0"
+
cur = self.conn.cursor()
cur.execute("""SELECT osm_type, osm_id, class
FROM placex WHERE place_id = %s""",
(self.pid, ))
- eq_(1, cur.rowcount, "No entry found for place id %s" % self.pid)
+ assert cur.rowcount == 1, "No entry found for place id %s" % self.pid
return "%s%s:%s" % cur.fetchone()
def compare_place_id(expected, result, column, context):
if expected == '0':
- eq_(0, result,
+ assert result == 0, \
LazyFmt("Bad place id in column %s. Expected: 0, got: %s.",
- column, PlaceObjName(result, context.db)))
+ column, PlaceObjName(result, context.db))
elif expected == '-':
- assert_is_none(result,
+ assert result is None, \
LazyFmt("bad place id in column %s: %s.",
- column, PlaceObjName(result, context.db)))
+ column, PlaceObjName(result, context.db))
else:
- eq_(NominatimID(expected).get_place_id(context.db.cursor()), result,
+ assert NominatimID(expected).get_place_id(context.db.cursor()) == result, \
LazyFmt("Bad place id in column %s. Expected: %s, got: %s.",
- column, expected, PlaceObjName(result, context.db)))
+ column, expected, PlaceObjName(result, context.db))
+
+def check_database_integrity(context):
+ """ Check some generic constraints on the tables.
+ """
+ # place_addressline should not have duplicate (place_id, address_place_id)
+ cur = context.db.cursor()
+ cur.execute("""SELECT count(*) FROM
+ (SELECT place_id, address_place_id, count(*) as c
+ FROM place_addressline GROUP BY place_id, address_place_id) x
+ WHERE c > 1""")
+ assert cur.fetchone()[0] == 0, "Duplicates found in place_addressline"
+
class NominatimID:
""" Splits a unique identifier for places into its components.
if oid is not None:
m = self.id_regex.fullmatch(oid)
- assert_is_not_none(m, "ID '%s' not of form <osmtype><osmid>[:<class>]" % oid)
+ assert m is not None, "ID '%s' not of form <osmtype><osmid>[:<class>]" % oid
self.typ = m.group('tp')
self.oid = m.group('id')
def get_place_id(self, cur):
where, params = self.table_select()
cur.execute("SELECT place_id FROM placex WHERE %s" % where, params)
- eq_(1, cur.rowcount,
- "Expected exactly 1 entry in placex for %s found %s"
- % (str(self), cur.rowcount))
+ assert cur.rowcount == 1, \
+ "Expected exactly 1 entry in placex for %s found %s" % (str(self), cur.rowcount)
return cur.fetchone()[0]
return
if column.startswith('centroid'):
- fac = float(column[9:]) if column.startswith('centroid*') else 1.0
- x, y = value.split(' ')
- assert_almost_equal(float(x) * fac, row['cx'], "Bad x coordinate")
- assert_almost_equal(float(y) * fac, row['cy'], "Bad y coordinate")
+ if value == 'in geometry':
+ query = """SELECT ST_Within(ST_SetSRID(ST_Point({}, {}), 4326),
+ ST_SetSRID('{}'::geometry, 4326))""".format(
+ row['cx'], row['cy'], row['geomtxt'])
+ cur = context.db.cursor()
+ cur.execute(query)
+ assert cur.fetchone()[0], "(Row %s failed: %s)" % (column, query)
+ else:
+ fac = float(column[9:]) if column.startswith('centroid*') else 1.0
+ x, y = value.split(' ')
+ assert Almost(float(x) * fac) == row['cx'], "Bad x coordinate"
+ assert Almost(float(y) * fac) == row['cy'], "Bad y coordinate"
elif column == 'geometry':
geom = context.osm.parse_geometry(value, context.scene)
cur = context.db.cursor()
query = "SELECT ST_Equals(ST_SnapToGrid(%s, 0.00001, 0.00001), ST_SnapToGrid(ST_SetSRID('%s'::geometry, 4326), 0.00001, 0.00001))" % (
geom, row['geomtxt'],)
cur.execute(query)
- eq_(cur.fetchone()[0], True, "(Row %s failed: %s)" % (column, query))
+ assert cur.fetchone()[0], "(Row %s failed: %s)" % (column, query)
elif value == '-':
- assert_is_none(row[column], "Row %s" % column)
+ assert row[column] is None, "Row %s" % column
else:
- eq_(value, str(row[column]),
- "Row '%s': expected: %s, got: %s"
- % (column, value, str(row[column])))
+ assert value == str(row[column]), \
+ "Row '%s': expected: %s, got: %s" % (column, value, str(row[column]))
################################ STEPS ##################################
WHERE class='place' and type='houses' and osm_type='W'
and ST_GeometryType(geometry) = 'ST_LineString'""")
context.db.commit()
- context.nominatim.run_setup_script('index', 'index-noanalyse')
+ context.nominatim.run_setup_script('calculate-postcodes', 'index', 'index-noanalyse')
+ check_database_integrity(context)
@when("updating places")
def update_place_table(context):
if cur.rowcount == 0:
break
+ check_database_integrity(context)
+
+@when("updating postcodes")
+def update_postcodes(context):
+ context.nominatim.run_update_script('calculate-postcodes')
+
@when("marking for delete (?P<oids>.*)")
def delete_places(context, oids):
context.nominatim.run_setup_script(
ST_X(centroid) as cx, ST_Y(centroid) as cy
FROM placex where %s""" % where,
params)
- assert_less(0, cur.rowcount, "No rows found for " + row['object'])
+ assert cur.rowcount > 0, "No rows found for " + row['object']
for res in cur:
if exact:
expected_content.add((res['osm_type'], res['osm_id'], res['class']))
for h in row.headings:
- if h.startswith('name'):
+ if h in ('extratags', 'address'):
+ if row[h] == '-':
+ assert res[h] is None
+ else:
+ vdict = eval('{' + row[h] + '}')
+ assert vdict == res[h]
+ elif h.startswith('name'):
name = h[5:] if h.startswith('name+') else 'name'
- assert_in(name, res['name'])
- eq_(res['name'][name], row[h])
+ assert name in res['name']
+ assert res['name'][name] == row[h]
elif h.startswith('extratags+'):
- eq_(res['extratags'][h[10:]], row[h])
+ assert res['extratags'][h[10:]] == row[h]
elif h.startswith('addr+'):
if row[h] == '-':
if res['address'] is not None:
- assert_not_in(h[5:], res['address'])
+ assert h[5:] not in res['address']
else:
- assert_in(h[5:], res['address'], "column " + h)
- assert_equals(res['address'][h[5:]], row[h],
- "column %s" % h)
+ assert h[5:] in res['address'], "column " + h
+ assert res['address'][h[5:]] == row[h], "column %s" % h
elif h in ('linked_place_id', 'parent_place_id'):
compare_place_id(row[h], res[h], h, context)
else:
if exact:
cur.execute('SELECT osm_type, osm_id, class from placex')
- eq_(expected_content, set([(r[0], r[1], r[2]) for r in cur]))
+ assert expected_content == set([(r[0], r[1], r[2]) for r in cur])
context.db.commit()
ST_GeometryType(geometry) as geometrytype
FROM place where %s""" % where,
params)
- assert_less(0, cur.rowcount, "No rows found for " + row['object'])
+ assert cur.rowcount > 0, "No rows found for " + row['object']
for res in cur:
if exact:
msg = "%s: %s" % (row['object'], h)
if h in ('name', 'extratags', 'address'):
if row[h] == '-':
- assert_is_none(res[h], msg)
+ assert res[h] is None, msg
else:
vdict = eval('{' + row[h] + '}')
- assert_equals(vdict, res[h], msg)
+ assert vdict == res[h], msg
elif h.startswith('name+'):
- assert_equals(res['name'][h[5:]], row[h], msg)
+ assert res['name'][h[5:]] == row[h], msg
elif h.startswith('extratags+'):
- assert_equals(res['extratags'][h[10:]], row[h], msg)
+ assert res['extratags'][h[10:]] == row[h], msg
elif h.startswith('addr+'):
if row[h] == '-':
if res['address'] is not None:
- assert_not_in(h[5:], res['address'])
+ assert h[5:] not in res['address']
else:
- assert_equals(res['address'][h[5:]], row[h], msg)
+ assert res['address'][h[5:]] == row[h], msg
elif h in ('linked_place_id', 'parent_place_id'):
compare_place_id(row[h], res[h], h, context)
else:
if exact:
cur.execute('SELECT osm_type, osm_id, class from place')
- eq_(expected_content, set([(r[0], r[1], r[2]) for r in cur]))
+ assert expected_content, set([(r[0], r[1], r[2]) for r in cur])
context.db.commit()
-@then("search_name contains")
-def check_search_name_contents(context):
+@then("search_name contains(?P<exclude> not)?")
+def check_search_name_contents(context, exclude):
cur = context.db.cursor(cursor_factory=psycopg2.extras.DictCursor)
for row in context.table:
pid = NominatimID(row['object']).get_place_id(cur)
cur.execute("""SELECT *, ST_X(centroid) as cx, ST_Y(centroid) as cy
FROM search_name WHERE place_id = %s""", (pid, ))
- assert_less(0, cur.rowcount, "No rows found for " + row['object'])
+ assert cur.rowcount > 0, "No rows found for " + row['object']
for res in cur:
for h in row.headings:
if h in ('name_vector', 'nameaddress_vector'):
- terms = [x.strip().replace('#', ' ') for x in row[h].split(',')]
+ terms = [x.strip() for x in row[h].split(',') if not x.strip().startswith('#')]
+ words = [x.strip()[1:] for x in row[h].split(',') if x.strip().startswith('#')]
subcur = context.db.cursor()
- subcur.execute("""SELECT word_id, word_token
- FROM word, (SELECT unnest(%s) as term) t
- WHERE word_token = make_standard_name(t.term)""",
- (terms,))
- ok_(subcur.rowcount >= len(terms),
- "No word entry found for " + row[h])
+ subcur.execute(""" SELECT word_id, word_token
+ FROM word, (SELECT unnest(%s::TEXT[]) as term) t
+ WHERE word_token = make_standard_name(t.term)
+ and class is null and country_code is null
+ and operator is null
+ UNION
+ SELECT word_id, word_token
+ FROM word, (SELECT unnest(%s::TEXT[]) as term) t
+ WHERE word_token = ' ' || make_standard_name(t.term)
+ and class is null and country_code is null
+ and operator is null
+ """,
+ (terms, words))
+ if not exclude:
+ assert subcur.rowcount >= len(terms) + len(words), \
+ "No word entry found for " + row[h] + ". Entries found: " + str(subcur.rowcount)
for wid in subcur:
- assert_in(wid[0], res[h],
- "Missing term for %s/%s: %s" % (pid, h, wid[1]))
+ if exclude:
+ assert wid[0] not in res[h], "Found term for %s/%s: %s" % (pid, h, wid[1])
+ else:
+ assert wid[0] in res[h], "Missing term for %s/%s: %s" % (pid, h, wid[1])
else:
assert_db_column(res, h, row[h], context)
context.db.commit()
+@then("location_postcode contains exactly")
+def check_location_postcode(context):
+ cur = context.db.cursor(cursor_factory=psycopg2.extras.DictCursor)
+
+ cur.execute("SELECT *, ST_AsText(geometry) as geomtxt FROM location_postcode")
+ assert cur.rowcount == len(list(context.table)), \
+ "Postcode table has %d rows, expected %d rows." % (cur.rowcount, len(list(context.table)))
+
+ table = list(cur)
+ for row in context.table:
+ for i in range(len(table)):
+ if table[i]['country_code'] != row['country'] \
+ or table[i]['postcode'] != row['postcode']:
+ continue
+ for h in row.headings:
+ if h not in ('country', 'postcode'):
+ assert_db_column(table[i], h, row[h], context)
+
+@then("word contains(?P<exclude> not)?")
+def check_word_table(context, exclude):
+ cur = context.db.cursor(cursor_factory=psycopg2.extras.DictCursor)
+
+ for row in context.table:
+ wheres = []
+ values = []
+ for h in row.headings:
+ wheres.append("%s = %%s" % h)
+ values.append(row[h])
+ cur.execute("SELECT * from word WHERE %s" % ' AND '.join(wheres), values)
+ if exclude:
+ assert cur.rowcount == 0, "Row still in word table: %s" % '/'.join(values)
+ else:
+ assert cur.rowcount > 0, "Row not in word table: %s" % '/'.join(values)
+
@then("place_addressline contains")
def check_place_addressline(context):
cur = context.db.cursor(cursor_factory=psycopg2.extras.DictCursor)
cur.execute(""" SELECT * FROM place_addressline
WHERE place_id = %s AND address_place_id = %s""",
(pid, apid))
- assert_less(0, cur.rowcount,
- "No rows found for place %s and address %s"
- % (row['object'], row['address']))
+ assert cur.rowcount > 0, \
+ "No rows found for place %s and address %s" % (row['object'], row['address'])
for res in cur:
for h in row.headings:
context.db.commit()
+@then("place_addressline doesn't contain")
+def check_place_addressline_exclude(context):
+ cur = context.db.cursor(cursor_factory=psycopg2.extras.DictCursor)
+
+ for row in context.table:
+ pid = NominatimID(row['object']).get_place_id(cur)
+ apid = NominatimID(row['address']).get_place_id(cur)
+ cur.execute(""" SELECT * FROM place_addressline
+ WHERE place_id = %s AND address_place_id = %s""",
+ (pid, apid))
+ assert cur.rowcount == 0, \
+ "Row found for place %s and address %s" % (row['object'], row['address'])
+
+ context.db.commit()
+
@then("(?P<oid>\w+) expands to(?P<neg> no)? interpolation")
def check_location_property_osmline(context, oid, neg):
cur = context.db.cursor(cursor_factory=psycopg2.extras.DictCursor)
nid = NominatimID(oid)
- eq_('W', nid.typ, "interpolation must be a way")
+ assert 'W' == nid.typ, "interpolation must be a way"
cur.execute("""SELECT *, ST_AsText(linegeo) as geomtxt
FROM location_property_osmline
(nid.oid, ))
if neg:
- eq_(0, cur.rowcount)
+ assert cur.rowcount == 0
return
todo = list(range(len(list(context.table))))
else:
assert_db_column(res, h, row[h], context)
- eq_(todo, [])
+ assert not todo
@then("(?P<table>placex|place) has no entry for (?P<oid>.*)")
nid = NominatimID(oid)
where, params = nid.table_select()
cur.execute("SELECT * FROM %s where %s" % (table, where), params)
- eq_(0, cur.rowcount)
+ assert cur.rowcount == 0
context.db.commit()
@then("search_name has no entry for (?P<oid>.*)")
cur = context.db.cursor(cursor_factory=psycopg2.extras.DictCursor)
pid = NominatimID(oid).get_place_id(cur)
cur.execute("SELECT * FROM search_name WHERE place_id = %s", (pid, ))
- eq_(0, cur.rowcount)
+ assert cur.rowcount == 0
context.db.commit()
projects / nominatim.git / blobdiff