view backend/contrib/shpimporter/dgm.py @ 7602:c50dbbe17950

issue1596: Store table (cell) data twice: Once as (formatted) string as coming from server, once transformed into float (or string). The benefit is that now we can sort table data numerically, while keeping the formatted and i18ned display of values.
author Felix Wolfsteller <felix.wolfsteller@intevation.de>
date Wed, 27 Nov 2013 14:55:25 +0100
parents bb5bd48b80f6
children
line wrap: on
line source
# -*- coding: latin-1 -*-

import codecs
import utils
import datetime

def latin(string):
    return unicode(string, "latin1")

import logging
logger = logging.getLogger("DGM")


# <dbfield> : (<csvfield>, conversion function)
DGM_MAP = {
    "projection"      : "Projektion",
    "elevation_state" : latin("H�henstatus"),
    "format"          : "Format",
    "border_break"    : ("Bruchkanten",
        lambda x: True if x.lower() == "Ja" else False),
    "resolution"      : (latin("Aufl�sung"), lambda x: x),
#   "description"     : 
    "srid"            : "SRID",
    "path"            : ("Pfad_Bestand", lambda x: x),
    }

SQL_INSERT_DGT = "INSERT INTO dem (river_id, name," \
        " time_interval_id, range_id, " + ", ".join(DGM_MAP.keys()) + \
        ") VALUES (%s, %s, %s, %s, %s, %s, %s, %s, %s, %s, %s)"
SQL_INSERT_DGT_ORA = "INSERT INTO dem (river_id, name," \
        " time_interval_id, range_id, " + ", ".join(DGM_MAP.keys()) + \
        ") VALUES (:s, :s, :s, :s, :s, :s, :s, :s, :s, :s, :s)"
SQL_SELECT_TIME_ID = """
SELECT id FROM time_intervals WHERE start_time = %s AND stop_time = %s
"""
SQL_INSERT_TIME_ID = """
INSERT INTO time_intervals (id, start_time, stop_time) VALUES (%s, %s, %s)
"""
SQL_SELECT_TIME_ID_ORA = """
SELECT id FROM time_intervals WHERE start_time = :s AND stop_time = :s
"""
SQL_INSERT_TIME_ID_ORA = """
INSERT INTO time_intervals (id, start_time, stop_time) VALUES (:s, :s, :s)
"""
SQL_SELECT_RANGE_ID = """
SELECT id FROM ranges WHERE river_id = %s AND a = %s AND b = %s
"""
SQL_INSERT_RANGE_ID = """
INSERT INTO ranges (id, river_id, a, b) VALUES (%s, %s, %s, %s)
"""
SQL_SELECT_RANGE_ID_ORA = """
SELECT id FROM ranges WHERE river_id = :s AND a = :s AND b = :s
"""
SQL_INSERT_RANGE_ID_ORA = """
INSERT INTO ranges (id, river_id, a, b) VALUES (:s, :s, :s, :s)
"""
SQL_NEXT_ID     = "select nextval('%s_ID_SEQ')"
SQL_NEXT_ID_ORA = "select %s_ID_SEQ.nextval FROM dual"

def next_id(cur, relation, oracle):
    if oracle:
        cur.execute(SQL_NEXT_ID_ORA % relation.upper())
    else:
        cur.execute(SQL_NEXT_ID % relation.upper())
    idx = cur.fetchone()[0]
    return idx

def get_range_id(cur, river_id, a, b, oracle):
    if oracle:
        cur.execute(SQL_SELECT_RANGE_ID_ORA, (river_id, a, b))
    else:
        cur.execute(SQL_SELECT_RANGE_ID, (river_id, a, b))
    row = cur.fetchone()
    if row: return row[0]
    idx = next_id(cur, "ranges", oracle)
    if oracle:
        cur.execute(SQL_INSERT_RANGE_ID_ORA, (idx, river_id, a, b))
    else:
        cur.execute(SQL_INSERT_RANGE_ID, (idx, river_id, a, b))
    cur.connection.commit()
    return idx

def get_time_interval_id(cur, a, b, oracle):
    if not a or not b:
        return None
    if oracle:
        cur.execute(SQL_SELECT_TIME_ID_ORA, (a, b))
    else:
        cur.execute(SQL_SELECT_TIME_ID, (a, b))
    row = cur.fetchone()
    if row: return row[0]
    idx = next_id(cur, "time_intervals", oracle)
    if oracle:
        cur.execute(SQL_INSERT_TIME_ID_ORA, (idx, a, b))
    else:
        cur.execute(SQL_INSERT_TIME_ID, (idx, a, b))
    cur.connection.commit()
    return idx

def insertRiverDgm(dbconn, dgmfile, river_name, dry_run, oracle):
    with codecs.open(dgmfile, "r", "latin1") as csvfile:
        firstline = csvfile.readline()
        names = firstline.split(";")
        namedict = {}
        field_nr = 0
        for name in names:
            namedict[name] = field_nr
            field_nr += 1

        river_id = utils.getRiverId(dbconn, river_name, oracle)
        for line in csvfile:
            fields = line.split(";")
            if not fields: continue
            if fields[namedict[latin("Gew�sser")]] != \
                    unicode(utils.getUTF8(river_name),'UTF-8'):
                continue
            else:
                values=[]
                for key, val in DGM_MAP.items():
                    if isinstance(val, tuple):
                        values.append(val[1](fields[namedict[val[0]]]))
                    else:
                        values.append(unicode.encode(
                            fields[namedict[val]], "UTF-8"))
                km_von = min(float(fields[namedict["km_von"]]), 
                    float(fields[namedict["km_bis"]]))
                km_bis = max(float(fields[namedict["km_von"]]), 
                    float(fields[namedict["km_bis"]]))
                year_from = None
                year_to = None
                try:
                    year_from = datetime.datetime(
                        int(fields[namedict["Jahr_von"]]), 1, 1)
                    year_to = datetime.datetime(
                        int(fields[namedict["Jahr_bis"]]),1 ,1)
                except ValueError:
                    logger.warn("Invalid numbers (or none) found in year_from and year_to")

                name = "%s KM %s - %s" % (unicode(river_name, "latin1"), km_von, km_bis)
                cur = dbconn.cursor()
                range_id = get_range_id(cur, river_id, float(km_von),
                    float(km_bis), oracle)
                time_interval_id = get_time_interval_id(cur, year_from,
                    year_to, oracle)

                if oracle:
                    stmt = SQL_INSERT_DGT_ORA
                else:
                    stmt = SQL_INSERT_DGT

                cur.execute(stmt, [river_id, name, time_interval_id,
                    range_id] + values)

        if not dry_run:
            dbconn.commit()

http://dive4elements.wald.intevation.org