def set_eindeutige_stop_id(self):
"""
Setze eindeutige stop_id
"""
cur = self.conn.cursor()
sql = """
-- setze für noch nicht gefundenen H_IDs Haltestellennummern aus verschiedenen Haltestellentabellen
UPDATE trips f
SET stop_id = h."H_ID"
FROM
(
SELECT
h."H_Name"
FROM stops h
GROUP BY h."H_Name"
HAVING count(*) = 1 ) h1,
stops h
WHERE h."H_Name" = h1."H_Name"
AND f.stop_id IS NULL
AND f."H_Name" = h."H_Name";
"""
cur.execute(sql)
logger.info('eindeutige stop_ids aus Haltestellen: %s' %
cur.statusmessage)
def save_haltestellen_id(self):
"""
save H_ID into stop_id before deleting a trip
"""
cur = self.conn.cursor()
sql = """
-- sichere vor dem Löschen H_ID, die bei doppelten Fahrten gefunden wurden in Spalte stop_id
UPDATE trips f
SET stop_id = f2."H_ID"
FROM
(
SELECT a.abfahrt_id_final,
f.fahrt_index,
max(f."H_ID") AS "H_ID"
FROM trips f,
departures a
WHERE f.abfahrt_id = a.abfahrt_id
GROUP BY a.abfahrt_id_final,
f.fahrt_index
) f2
WHERE f.abfahrt_id = f2.abfahrt_id_final
AND f.fahrt_index = f2.fahrt_index
AND f2."H_ID" IS NOT NULL;
"""
cur.execute(sql)
logger.info('save haltestellen_ids: %s' % cur.statusmessage)
def set_stop_id(self):
sql = """
UPDATE trips SET stop_id = "H_ID" WHERE "H_ID" IS NOT NULL;
UPDATE trips SET stop_id = NULL WHERE stop_id >= 10000000;
DELETE FROM stops WHERE "H_ID" >= 10000000;
"""
cur = self.conn.cursor()
cur.execute(sql)
logger.info('Set Stop ID: %s' % cur.statusmessage)
def delete_invalid_fahrten(self):
sql = """
DELETE FROM trips WHERE abfahrt_id IN (
SELECT f.abfahrt_id FROM trips f LEFT JOIN departures a
ON f.abfahrt_id = a.abfahrt_id
WHERE a.abfahrt_id IS NULL) ;
"""
cur = self.conn.cursor()
cur.execute(sql)
logger.info('Delete invalid trips: %s' % cur.statusmessage)
def test_for_negative_travel_times(self):
cur = self.conn.cursor()
sql = """
SELECT count(*) FROM (
SELECT *,
lag("H_Abfahrt") OVER(PARTITION BY abfahrt_id ORDER BY fahrt_index) AS ab0
FROM trips f )f
where "H_Ankunft" < ab0;
"""
cur.execute(sql)
result = cur.fetchone()
sql = """
SELECT count(*) FROM trips f
where "H_Abfahrt" < "H_Ankunft";
"""
cur.execute(sql)
result2 = cur.fetchone()
if result[0] or result2[0]:
logger.warning('''There exist still negative travel times with
departures before arrivals!''')
if result[0]:
sql = """
SELECT abfahrt_id FROM (
SELECT *,
lag("H_Abfahrt") OVER(PARTITION BY abfahrt_id ORDER BY fahrt_index) AS ab0
FROM trips f )f
WHERE "H_Ankunft" < ab0;
"""
cur.execute(sql)
abfahrten = np.array(cur.fetchall())
logger.warning('%s' % abfahrten)
if result2[0]:
sql = """
SELECT abfahrt_id FROM trips f
WHERE "H_Abfahrt" < "H_Ankunft";
"""
cur.execute(sql)
abfahrten = np.array(cur.fetchall())
logger.warning('%s' % abfahrten)
# raise ValueError('''There exist still negative travel times with
# departures before arrivals!''')
else:
logger.info('no negative travel times')
def copy_layer(self, schema, layer, dest_schema):
"""
copy layer
Parameters
----------
layer : str
"""
cmd = '{OGR2OGR} -overwrite -geomfield geom -nln {layer} {srid_option} -lco FEATURE_DATASET="{dest_schema}" -f "FileGDB" {path} PG:"host={host} port={port} user={user} dbname={db}" "{schema}.{layer}"'
if self.gdbname is None:
gdbname = '{db}.gdb'.format(db=self.destination_db)
else:
gdbname = self.gdbname
if not gdbname.endswith('.gdb'):
gdbname += '.gdb'
# get srid
if self.target_srid is None:
srid = self.get_target_srid_from_dest_db()
srid_option = '-a_srs EPSG:{srid}'.format(srid=srid)
else:
srid_option = '-t_srs EPSG:{srid}'.format(
srid=self.target_srid)
folder = os.path.join(self.folder,
'projekte',
self.destination_db,
'fgdb', )
self.make_folder(folder)
path = os.path.join(folder, gdbname)
full_cmd = cmd.format(OGR2OGR=self.OGR2OGRPATH,
layer=layer,
srid_option=srid_option,
path=path,
host=self.login.host,
port=self.login.port,
user=self.login.user,
db=self.login.db,
schema=schema,
dest_schema=dest_schema,
)
logger.info(full_cmd)
ret = subprocess.call(full_cmd, shell=self.SHELL)
if ret:
raise IOError('Layer {layer} could not be copied to FGDB'.format(layer=layer))
def reset_ein_aus(self):
sql1 = """
UPDATE trips
SET "H_Abfahrt" = NULL
WHERE ein = FALSE;
"""
sql2 = """
UPDATE trips
SET "H_Ankunft" = NULL
WHERE aus = FALSE;
"""
cur = self.conn.cursor()
cur.execute(sql1)
logger.info('no Entry: %s' % cur.statusmessage)
cur.execute(sql2)
logger.info('no Exit: %s' % cur.statusmessage)
def set_departure_to_day_before(self):
"""
"""
sql = """
--setze Abfahrtstag für Ankunftszeiten zwischen 12:00 und 23:59 auf Folgetag
UPDATE trips f
SET
"H_Abfahrt" = "H_Abfahrt" - interval '1 day'
FROM
--SELECT * FROM trips f,
(SELECT abf.abfahrt_id, abf.d0, abf.h0, ank.d1, ank.h1
FROM
(SELECT abfahrt_id, h1, d1
FROM (
SELECT
abfahrt_id,
extract(hour from "H_Ankunft") AS h1 ,
extract(day from "H_Ankunft") AS d1, "H_Ankunft",
fahrt_index, max(fahrt_index) over (PARTITION BY abfahrt_id) AS maxindex
FROM trips
) f1
WHERE fahrt_index = maxindex) ank,
(SELECT
abfahrt_id,
extract(hour from "H_Abfahrt") AS h0,
extract(day from "H_Abfahrt") AS d0
FROM trips WHERE fahrt_index = 1
) AS abf,
departures a
WHERE abf.abfahrt_id = ank.abfahrt_id
AND a.abfahrt_id = ank.abfahrt_id
-- wenn der Tag gleich ist
AND abf.d0 = ank.d1
AND abf.d0 = ank.d1
-- und die Ankunft an Endhaltestelle nach der Abfahrt an der Starthaltestelle erfolgt
AND ank.h1 < abf.h0
ORDER BY abf.abfahrt_id) b
WHERE f.abfahrt_id = b.abfahrt_id
AND extract(hour from "H_Abfahrt") >= 12;
"""
cur = self.conn.cursor()
cur.execute(sql)
logger.info('set_departure_to_day_before: %s' % cur.statusmessage)
def shift_trips(self):
sql1 = """
-- korrigiere Fahrten über Tagesgrenzen hinaus
--verschiebe Fahrten nach vorne, so dass nur Fahrten heute und morgen auftreten
UPDATE trips f
SET "H_Abfahrt" = "H_Abfahrt" + '1 day'::INTERVAL,
"H_Ankunft" = "H_Ankunft" + '1 day'::INTERVAL
FROM (
SELECT DISTINCT abfahrt_id
FROM trips
WHERE date_part('day', "H_Abfahrt") < {today}) v
WHERE f.abfahrt_id = v.abfahrt_id
;
""".format(today=self.today.day)
sql2 = """
--verschiebe Fahrten nach hinten, so dass nur Fahrten heute und morgen auftreten
UPDATE trips f
SET "H_Abfahrt" = "H_Abfahrt" - '1 day'::INTERVAL,
"H_Ankunft" = "H_Ankunft" - '1 day'::INTERVAL
FROM (
SELECT DISTINCT abfahrt_id
FROM trips
WHERE date_part('day', "H_Ankunft") > {tomorrow}) v
WHERE f.abfahrt_id = v.abfahrt_id
;
""".format(tomorrow=self.today.day + 1)
cur = self.conn.cursor()
n_updated = 1
while n_updated:
cur.execute(sql1)
msg = cur.statusmessage
n_updated = int(msg.split(' ')[1])
logger.info('Shift Trips forward: %s' % cur.statusmessage)
n_updated = 1
while n_updated:
cur.execute(sql2)
msg = cur.statusmessage
n_updated = int(msg.split(' ')[1])
logger.info('Shift Trips backward: %s' % cur.statusmessage)
def copy2pbf(self):
"""
copy the according schema to a pbf with osmosis
"""
fn = f'{self.login.db}_{self.network}'
folder = os.path.join(self.folder,
'projekte',
self.login.db,
self.subfolder,
)
self.make_folder(folder)
file_path = os.path.join(folder, fn)
if self.as_xml:
to_xml = f' --tee --write-xml file={file_path}.osm.bz2 '
else:
to_xml = ''
cmd = ('{OSMOSIS} -v '
'--read-pgsql '
'postgresSchema={schema} '
'authFile="{authfile}" '
'host={host}:{port} user={user} database={db} '
'--dataset-dump {to_xml}'
'--write-pbf omitmetadata=true file={fn}.osm.pbf')
full_cmd = cmd.format(OSMOSIS=self.OSMOSISPATH,
schema=self.schema,
authfile=self.AUTHFILE,
host=self.login.host,
port=self.login.port,
user=self.login.user,
db=self.login.db,
fn=file_path,
to_xml=to_xml,
)
logger.info(full_cmd)
ret = subprocess.call(full_cmd, shell=self.SHELL)
if ret:
layer = 'pbf'
raise IOError(f'Layer {layer} could not becopied to pbf-file')
def delete_multiple_abfahrten(self):
"""
delete abfahrten marked with keep=False
"""
cur = self.conn.cursor()
sql = """
-- lösche doppelte Fahrten
DELETE FROM departures WHERE keep = FALSE
OR keep IS NULL;
"""
cur.execute(sql)
logger.info('abfahrten deleted: %s' % cur.statusmessage)
sql = """
-- remove copied dummy-haltestellen-ids
UPDATE trips f
SET stop_id = NULL WHERE stop_id > 10000000;
"""
cur.execute(sql)
logger.info('dummy-haltestellen_ids removed: %s' % cur.statusmessage)
def show_multiple_trips(self):
cur = self.conn.cursor()
sql = '''
SELECT
f.*,
a.keep
FROM trips f,
(SELECT
abfahrt_id,
abfahrt_id_final
FROM departures
WHERE keep = False) a1,
departures a
WHERE f.abfahrt_id = a.abfahrt_id
AND a1.abfahrt_id_final = a.abfahrt_id_final
ORDER BY f."Fahrt_Name", f.abfahrt_id, f.fahrt_index;
'''
cur.execute(sql)
rows = cur.fetchall()
msg = '{0.Fahrt_Name}\t{0.abfahrt_id}\t{0.fahrt_index}\t{0.H_Ankunft} - {0.H_Abfahrt}\t{0.H_ID}\t{0.H_Name}'
for row in rows:
m = msg.format(row)
logger.info(m)
def export_gtfs(self):
"""
exports the data to the path
"""
path = os.path.join(self.base_path, self.destination_db,
self.subfolder)
with Connection(self.login1) as conn:
self.conn = conn
self.set_search_path()
cur = self.conn.cursor()
sql = '''SET CLIENT_ENCODING TO '{encoding}';'''
encoding = 'UTF8'
cur.execute(sql.format(encoding=encoding))
tables = [
'stops', 'agency', 'stop_times', 'routes', 'trips', 'shapes',
'calendar', 'calendar_dates', 'transfers'
]
folder = path.replace('~', os.environ['HOME'])
self.make_folder(folder)
zipfilename = os.path.join(folder,
'{}.zip'.format(self.destination_db))
with zipfile.ZipFile(zipfilename, 'w') as z:
for table in tables:
tn = 'gtfs_{tn}'.format(tn=table)
tablename = '{tn}.txt'.format(tn=table)
fn = os.path.join(folder, tablename)
logger.info('write {}'.format(fn))
with open(fn, 'w') as f:
sql = self.conn.copy_sql.format(tn=tn, fn=fn)
logger.info(sql)
cur.copy_expert(sql, f)
z.write(fn, tablename)
os.remove(fn)
sql = '''RESET CLIENT_ENCODING;'''
cur.execute(sql)
def set_ein_aus(self):
sql1 = """
UPDATE trips
SET ein = FALSE
WHERE "H_Abfahrt" IS NULL
"""
sql2 = """
UPDATE trips
SET aus = FALSE
WHERE "H_Ankunft" IS NULL
"""
cur = self.conn.cursor()
cur.execute(sql1)
logger.info('no Entry: %s' % cur.statusmessage)
cur.execute(sql2)
logger.info('no Exit: %s' % cur.statusmessage)
sql3 = """
UPDATE trips f
SET "H_Abfahrt" = f."H_Ankunft"
FROM
(SELECT abfahrt_id, fahrt_index FROM
(select abfahrt_id,
"H_Ankunft", "H_Abfahrt", fahrt_index,
max(fahrt_index) over (PARTITION BY abfahrt_id) AS maxindex
FROM trips) f1
WHERE fahrt_index < maxindex
AND "H_Ankunft" IS NOT NULL
AND "H_Abfahrt" IS NULL) ank
WHERE ank.abfahrt_id = f.abfahrt_id
AND ank.fahrt_index = f.fahrt_index;
"""
cur.execute(sql3)
logger.info('Missing Departure Time added : %s' % cur.statusmessage)
sql4 = """
UPDATE trips f
SET "H_Ankunft" = "H_Abfahrt"
WHERE fahrt_index > 1
AND "H_Ankunft" IS NULL
AND "H_Abfahrt" IS NOT NULL;
"""
cur.execute(sql4)
logger.info('Missing Arrival Time added : %s' % cur.statusmessage)
def update_stop_id_from_database(self):
"""
Update stop_ids aus Deutchlandweiten Daten
"""
cur = self.conn.cursor()
sql = """
-- setze für noch nicht gefundenen H_IDs Haltestellennummern aus verschiedenen Haltestellentabellen
UPDATE trips f
SET stop_id = h."H_ID"
FROM
stops h
WHERE f.stop_id IS NULL
AND f."H_Name" = h."H_Name";
"""
sql2 = """
SELECT dblink_connect_u('conn', 'dbname=europe');
-- füge fehlende Haltestelle aus der Deutschland-Tabelle hinzu
INSERT INTO stops
("H_ID", "H_Name", geom, kreis)
SELECT "H_ID", "H_Name", st_transform(geom, {srid}) AS geom, kreis
FROM dblink('conn',
'SELECT h."H_ID", h."H_Name", h.geom, h.kreis
FROM timetables.haltestellen AS h') AS hd(
"H_ID" integer,
"H_Name" text,
geom geometry,
kreis text)
WHERE hd."H_ID" NOT IN (SELECT DISTINCT h."H_ID" FROM stops h)
--AND hd."H_ID" IN (SELECT DISTINCT f.stop_id FROM trips AS f)
AND hd."H_Name" IN (SELECT DISTINCT f."H_Name" FROM trips AS f);
"""
query = sql
logger.info(query)
cur.execute(query)
logger.info('{msg}'.format(msg=cur.statusmessage))
query = sql2.format(srid=self.target_srid)
logger.info(query)
cur.execute(query)
logger.info('{msg}'.format(msg=cur.statusmessage))
def count(self):
sql = 'SELECT count(*) FROM stops;'
cur = self.conn.cursor()
cur.execute(sql)
rows = cur.fetchall()
logger.info('Haltestellen: %s, ' % rows[0][0])
sql = 'SELECT count(*) FROM departures;'
cur = self.conn.cursor()
cur.execute(sql)
rows = cur.fetchall()
logger.info('Abfahrten: %s ' % rows[0][0])
sql = 'SELECT count(*) FROM trips;'
cur = self.conn.cursor()
cur.execute(sql)
rows = cur.fetchall()
logger.info('Fahrten: %s' % rows[0][0])
def update_stop_id_from_similar_routes(self):
"""
"""
cur = self.conn.cursor()
sql = """
--
UPDATE trips f
SET stop_id = c.stop1
FROM
(SELECT f.abfahrt_id, f.fahrt_index, b.stop1
FROM
(SELECT a."Fahrt_Name",a.stopname1, a.stopname0, a.stopname2, a.stop1
FROM
(
-- In jeder Zeile steht stop und stopname der Vorhaltestelle, aktuellen Haltestelle und Folgehaltestelle
SELECT
f."Fahrt_Name",
f.abfahrt_id,
f.fahrt_index,
f.stop_id AS stop1,
lag(f.stop_id) OVER(PARTITION BY f.abfahrt_id ORDER BY f.fahrt_index) AS stop0,
lead(f.stop_id) OVER(PARTITION BY f.abfahrt_id ORDER BY f.fahrt_index) AS stop2,
f."H_Name" AS stopname1,
lag(f."H_Name") OVER(PARTITION BY f.abfahrt_id ORDER BY f.fahrt_index) AS stopname0,
lead(f."H_Name") OVER(PARTITION BY f.abfahrt_id ORDER BY f.fahrt_index) AS stopname2
FROM trips f) AS a
-- Suche für jede Fahrt_Name alle eindeutigen Haltestellenfolgen (definiert durch den Haltestellennamen
GROUP BY a."Fahrt_Name",a.stopname1, a.stopname0, a.stopname2, a.stop1) b,
-- Fahrtentabelle mit den Haltestellennamen der Vorgänger und Nachfolgehaltestellen
(SELECT
f.abfahrt_id,
f.fahrt_index,
f."Fahrt_Name",
f.stop_id,
f."H_Name" AS stopname1,
lag(f."H_Name") OVER(PARTITION BY f.abfahrt_id ORDER BY f.fahrt_index) AS stopname0,
lead(f."H_Name") OVER(PARTITION BY f.abfahrt_id ORDER BY f.fahrt_index) AS stopname2
FROM
trips AS f ) AS f
-- Aktualisier unbekannte stop_id in der Fahrtentabelle aus den bekannten Haltestellenfolgen hinzu
WHERE f.stop_id IS NULL
AND f."Fahrt_Name" = b."Fahrt_Name"
AND f.stopname0 = b.stopname0
AND f.stopname2 = b.stopname2
AND f.stopname1 = b.stopname1
AND b.stop1 IS NOT NULL )c
WHERE f.abfahrt_id = c.abfahrt_id
AND f.fahrt_index = c.fahrt_index
;
"""
cur.execute(sql)
logger.info('UPDATE stop_id FROM similar routes: %s' %
cur.statusmessage)
sql = """
-- setze die Haltestellennummer bei Bussen auf die kleinste bekannte "H_ID" für den Haltestellennamen
UPDATE trips f
SET stop_id = a.stop_id
FROM (
SELECT f.abfahrt_id, f.fahrt_index,
min(h."H_ID") as stop_id
FROM trips f, stops h
WHERE f.stop_id IS NULL AND f."H_Name" = h."H_Name"
AND f."Fahrt_Name" LIKE 'Bus%'
GROUP BY f.abfahrt_id, f.fahrt_index )a
WHERE f.abfahrt_id = a.abfahrt_id
AND f.fahrt_index = a.fahrt_index;
"""
cur.execute(sql)
logger.info('UPDATE stop_id for busses: %s' % cur.statusmessage)
sql = """
-- setze die Haltestellennummer bei Nicht-Bussen auf die größte bekannte "H_ID" für den Haltestellennamen
UPDATE trips f
SET stop_id = a.stop_id
FROM (
SELECT f.abfahrt_id, f.fahrt_index,
max(h."H_ID") as stop_id
FROM trips f, stops h
WHERE f.stop_id IS NULL AND f."H_Name" = h."H_Name"
AND f."Fahrt_Name" NOT LIKE 'Bus%'
GROUP BY f.abfahrt_id, f.fahrt_index )a
WHERE f.abfahrt_id = a.abfahrt_id
AND f.fahrt_index = a.fahrt_index;
"""
cur.execute(sql)
logger.info('UPDATE stop_id for busses: %s' % cur.statusmessage)
sql = """
-- Setze die Koordinaten in haltestellen auf 0, wenn unbekannte Halteselle
-- Wenn keine "H_ID" übergeben wird, wird automatisch eine "H_ID" mit einem
-- serial Feld > 50000000 erzeugt
INSERT INTO stops ("H_Name", geom)
SELECT DISTINCT f."H_Name" AS "H_Name",
st_transform(ST_SETSRID(ST_MAKEPOINT(0,0), 4326), {srid}) AS geom
FROM trips f
WHERE f.stop_id IS NULL;
-- setze für Dummy-Haltestellen die stop_id
UPDATE trips f
SET stop_id = h."H_ID"
FROM
stops h
WHERE f.stop_id IS NULL
AND f."H_Name" = h."H_Name";
"""
self.run_query(sql.format(srid=self.target_srid))
sql = """
UPDATE trips SET stop_id_txt = stop_id::text;
"""
self.run_query(sql)
def show_search_path(self):
cur = self.conn.cursor()
sql = 'show search_path ;'
cur.execute(sql)
rows = cur.fetchall()
logger.info(rows)
def count_multiple_abfahrten(self):
"""
count the number of occurences of abfahrten
and save into abfahrt_id_final
"""
cur = self.conn.cursor()
sql = """
-- zähle die doppelten Fahrten durch
UPDATE departures a
SET abfahrt_id_final = b.abfahrt_id_final,
keep =(b.rn = 1) ::bool
FROM (
SELECT a.abfahrt_id,
row_number() OVER(PARTITION BY a."Fahrt_Name", a.sh_name, a.anz_haltestellen, a.sh_ab,
a.eh_name, a.eh_an, a.zwh_name, a.zwh_an) AS rn,
first_value(a.abfahrt_id) OVER(PARTITION BY a."Fahrt_Name", a.sh_name,
a.anz_haltestellen, a.sh_ab, a.eh_name, a.eh_an, a.zwh_name, a.zwh_an) AS
abfahrt_id_final
FROM (
SELECT a.abfahrt_id,
a."Fahrt_Name",
a."Fahrt_Ziel",
a."H_ID",
f."H_ID" AS sh_id,
f."H_Name" AS sh_name,
f."H_Abfahrt" AS sh_ab,
f1.fahrt_index AS anz_haltestellen,
f1."H_ID" AS eh_id,
f1."H_Name" AS eh_name,
f1."H_Ankunft" AS eh_an,
f2."H_ID" AS zwh_id,
f2."H_Name" AS zwh_name,
f2."H_Ankunft" AS zwh_an
FROM departures a,
(
SELECT f.abfahrt_id,
f."H_ID",
f."H_Ankunft",
f."H_Name",
f.fahrt_index,
row_number() OVER(PARTITION BY f.abfahrt_id
ORDER BY f.fahrt_index DESC) rn
FROM trips f
) f1, -- letzte Haltestelle der Fahrt row_number() mit ORDER BY fahrt_index descending, also rückwärte
trips f, -- erste Haltestelle der Fahrt
trips f2 -- zweite Haltestelle der Fahrt
WHERE a.abfahrt_id = f.abfahrt_id
AND a.abfahrt_id = f1.abfahrt_id
AND a.abfahrt_id = f2.abfahrt_id
AND f.fahrt_index = 1 -- erste Haltestelle
AND f1.rn = 1 --letzte Haltestelle
AND f2.fahrt_index = 2 --auch 2. Haltestelle gleich, damit Ringfahrten (U3 Barmbek-Barmbek) nicht aus versehen gelöscht werden
ORDER BY a."Fahrt_Name",
sh_ab,
sh_name
) a
) b
WHERE b.abfahrt_id = a.abfahrt_id;
"""
cur.execute(sql)
logger.info('Multiple_abfahrten: %s' % cur.statusmessage)
sql = """
-- zähle, wie viele doppelte vorhanden
SELECT keep,
count(*)
FROM departures
GROUP BY keep;
"""
cur.execute(sql)
rows = cur.fetchall()
msg = 'keep abfahrten: {0.keep}: {0.count}'
for row in rows:
pass
