chaloBEST/chaloBEST/imports/import_atlas.py
2012-02-17 13:34:28 +05:30

332 lines
13 KiB
Python

from settings import PROJECT_ROOT
from os.path import join
import json
import csv
import pdb #debugger
from mumbai.models import *
from fuzzywuzzy import process as fuzzprocess
import datetime
'''
Convert Atlas.csv file (obtained from BEST) into first stage Atlas.json
(step 1)
'''
def csvToJSON():
atlasCSV = csv.reader(open(join(PROJECT_ROOT, "../db_csv_files/Atlas.csv"), "r"), delimiter="\t")
atlasDict = {}
previousRoute = None
for a in atlasCSV:
routeNo = a[1].strip()
# print a
if routeNo != '':
if atlasDict.has_key(routeNo):
atlasDist[routeNo].append(a)
else:
atlasDict[routeNo] = [a]
previousRoute = routeNo
else:
atlasDict[previousRoute].append(a)
# print atlasDict
jsonFile = open(join(PROJECT_ROOT, "../db_csv_files/Atlas.json"), "w")
jsonFile.write(json.dumps(atlasDict, indent=2))
jsonFile.close()
'''
function to copy over values of AM N PM + Schedule from previous row, reading from Atlas.json, writing to atlasCopied.json
(fill in blank rows where 'copy from previous' is assumed, and create new json file - step 2)
NOTE: this omits routes for which we dont have routeAlias to code mapping for in routeMapping.json
'''
def processJSON():
routeErrors = {'routes': [], 'others': []}
routeMapping = json.loads(open(join(PROJECT_ROOT, "../db_csv_files/routeMapping.json")).read())
routes = json.loads(open(join(PROJECT_ROOT, "../db_csv_files/Atlas.json")).read())
previousRow = []
outDict = {}
for key in routes.keys():
print key
if key not in routeMapping: #make note of routeNames we dont have routeAlias for yet.
routeErrors['routes'].append(key)
else: #else, go ahead ..
routeAlias = routeMapping[key]
thisRoute = routes[key]
#handle copying over empty values from previous rows
outDict[key] = []
for row in thisRoute:
# pdb.set_trace()
if len(row) < 7:
routeErrors['others'].append({key: row})
break
for i in range(2,4):
if row[i].strip() == '':
row[i] = previousRow[i]
try:
if row[-5].strip() == '':
row[-5] = previousRow[-5] #What is -5 ?
except:
pdb.set_trace()
previousRow = row
outDict[key].append(row)
atlasRouteErrors = open("atlasRouteErrors.json", "w")
atlasRouteErrors.write(json.dumps(routeErrors, indent=2))
atlasRouteErrors.close()
atlasCopied = open(join(PROJECT_ROOT, "../db_csv_files/atlasCopied.json"), "w")
atlasCopied.write(json.dumps(outDict, indent=2))
atlasCopied.close()
'''
function to group atlasCopied.json to uniqueRoutes (uniqueRoutes.json)
(step 3)
'''
def groupUnique():
routes = json.loads(open(join(PROJECT_ROOT, "../db_csv_files/atlasCopied.json")).read())
errors = {}
outDict = {}
for key in routes.keys():
outDict[key] = []
for row in routes[key]:
print key
d = {
'from': row[7],
'to': row[10],
'span': row[13],
'is_full': False,
# 'schedule': row[28],
# 'rows': {
# row[-5]: row
# }
}
matchedRow = isNotUnique(d, outDict[key])
schedule = row[-5]
if matchedRow is not None:
outDict[key][matchedRow]['rows'][schedule] = row
else:
if isLargestSpan(d, routes[key]):
d['is_full'] = True
outDict[key].append(d)
if not outDict[key][-1].has_key('rows'):
outDict[key][-1]['rows'] = {}
outDict[key][-1]['rows'][schedule] = row
outFile = open(join(PROJECT_ROOT, "../db_csv_files/uniqueRoutes.json"), "w")
outFile.write(json.dumps(outDict, indent=2))
outFile.close()
'''
Go through uniqueRoutes.json and actually import atlas data into the db
(step 4)
'''
def importUniqueRoutes():
data = json.load(open(join(PROJECT_ROOT, "../db_csv_files/uniqueRoutes.json")))
routeMapping = json.load(open(join(PROJECT_ROOT, "../db_csv_files/routeMapping.json")))
routeDoesNotExistErrors = [] #route codes for which there are entries in routeMapping.json and in Atlas, but which do not exist in RouteMaster
stopMapping = {} #FIXME
stopErrors = [] #This should ideally never happen, and any errors here are bad and would indicate problems with the fuzzy matching logic, most likely.
for route in data.keys():
if routeMapping.has_key(route):
routeCode = routeMapping[route]
try:
routeObj = Route.objects.get(code=routeCode)
except:
routeDoesNotExistErrors.append({'routeCode': routeCode, 'routeAlias': route})
continue
else:
try:
routeObj = Route.objects.get(alias=route)
except:
routeDoesNotExistErrors.append({'routeAlias': route})
continue
for unique_route in data[route]:
thisRoute = unique_route #FIXME
try:
distance = float(thisRoute['span'])
except:
distance = 0
obj = UniqueRoute(route=routeObj, is_full=thisRoute['is_full'], distance=distance, from_stop_txt=thisRoute['from'], to_stop_txt=thisRoute['to'])
if obj.is_full: #If the route is the primary route, we can get stop codes easily from RouteDetails first / last stop
from_to = getFromToStopsForRoute(routeObj)
if not from_to:
routeDoesNotExistErrors.append({'from_to_not_found': route})
continue
obj.from_stop = from_to[0]
if not stopMapping.has_key(obj.from_stop_txt):
stopMapping[obj.from_stop_txt] = from_to[0].name
obj.to_stop = from_to[1]
if not stopMapping.has_key(obj.to_stop_txt):
stopMapping[obj.to_stop_txt] = from_to[1].name
else: #Else we do fuzzy string matching against all possible values for stopname got from RouteDetails
stopnames = []
stopcodes = []
if RouteDetail.objects.filter(route=routeObj).count() == 0:
routeDoesNotExistErrors.append({'routeDetailDoesNotExist': routeObj.code})
continue
for r in RouteDetail.objects.filter(route=routeObj):
stopnames.append(r.stop.name)
stopcodes.append(r.stop.code)
try:
from_fuzz = fuzzprocess.extractOne(thisRoute['from'], stopnames)
to_fuzz = fuzzprocess.extractOne(thisRoute['to'], stopnames)
except:
stopErrors.append(thisRoute)
continue
#pdb.set_trace()
try:
obj.from_stop = Stop.objects.filter(name=from_fuzz[0]).filter(code__in=stopcodes)[0]
obj.to_stop = Stop.objects.filter(name=to_fuzz[0]).filter(code__in=stopcodes)[0]
except:
stopErrors.append([thisRoute['from'], thisRoute['to']])
continue
obj.save()
#pdb.set_trace()
# print thisRoute['rows'].keys()
for schedule in thisRoute['rows'].keys(): #loop through each schedule per UniqueRoute and save it
row = thisRoute['rows'][schedule]
try:
depot = Depot.objects.get(code=row[6])
except:
depot = None #FIXME!! Catch depot errors based on findings
#pdb.set_trace()
routeScheduleObj = RouteSchedule(unique_route=obj, schedule_type=schedule, busesAM=noneInt(row[2]), busesN=noneInt(row[3]), busesPM=noneInt(row[4]), bus_type=row[5], depot_txt=row[6], depot=depot, first_from=formatTime(row[8]), last_from=formatTime(row[9]), first_to=formatTime(row[11]), last_to=formatTime(row[12]), runtime1=noneInt(row[14]), runtime2=noneInt(row[15]), runtime3=noneInt(row[16]), runtime4=noneInt(row[17]), headway1=noneInt(row[18]), headway2=noneInt(row[19]), headway3=noneInt(row[20]), headway4=noneInt(row[21]), headway5=noneInt(row[22]))
routeScheduleObj.save()
#done saving things - write out error files:
errors = open(join(PROJECT_ROOT, "../errors/routeMasterMissingRoutes.json"), "w")
errors.write(json.dumps(routeDoesNotExistErrors, indent=2))
errors.close()
stopMappingFile = open(join(PROJECT_ROOT, "../db_csv_files/stopMapping.json"), "w")
stopMappingFile.write(json.dumps(stopMapping, indent=2))
stopMappingFile.close()
stopErrorsFile = open(join(PROJECT_ROOT, "../errors/atlasStopErrors.json"), "w")
stopErrorsFile.write(json.dumps(stopErrors, indent=2))
stopErrorsFile.close()
'''
>>>formatTime("06.40")
>>>time(6,40)
'''
def formatTime(s):
if s.strip() == '':
return datetime.time(0,0)
split = s.split(".")
try:
hour = int(split[0])
except:
hour = 0
try:
minutes = int(split[1])
except:
minutes = 0 #FIXME: ugly!!
#pdb.set_trace()
try:
return datetime.time(hour, minutes)
except:
return datetime.time(0,0)
'''
Silly function to deal wth invalid strings in the data that need to go in as Integers into the db
passed a string, it will either return int(string) or None if that fails for any reason
FIXME: find a more elegant way to do this
'''
def noneInt(val):
try:
return int(val)
except:
return None
'''
Passed a route code, it gets stop codes for the first and last stop
'''
def getFromToStopsForRoute(route):
routeDetails = RouteDetail.objects.filter(route=route).order_by('serial')
if routeDetails.count() == 0:
return None
fromStop = routeDetails[0].stop
toStop = routeDetails[routeDetails.count() -1].stop
return (fromStop, toStop,)
'''
checks whether the row in a set of rows for a route has the largest 'span' value, useful to tell if a row belongs to a primary route
params:
data - dict with a span attribute
arr - array of rows to check if data['span'] is greater than. span is at row[13]
'''
def isLargestSpan(data, arr):
span = data['span']
for a in arr:
try:
arrSpan = float(a[13])
except:
arrSpan = 0
try:
dataSpan = float(data['span'])
except:
dataSpan = 0
if arrSpan > dataSpan:
return False
return True
'''
returns index of row if not unique, else None
'''
def isNotUnique(data, arr):
i = 0
for a in arr:
if a['from'] == data['from'] and a['to'] == data['to']:
return i
i += 1
return None
'''
Create routeMapping.json file to map route aliases to route codes
TODO: add mappings from hard coded routes
'''
def getRouteCodes():
atlasRawCSV = csv.reader(open(join(PROJECT_ROOT, "../db_csv_files/AtlasRaw.csv"), "r"), delimiter="\t")
atlasDict = json.loads(open(join(PROJECT_ROOT, "../db_csv_files/Atlas.json")).read())
mapping = {}
missingFile = open("/tmp/routesMissing.txt", "w")
for row in atlasRawCSV:
routeAlias = row[0]
routeCode = row[1]
print routeCode
if routeCode in atlasDict:
mapping[routeCode] = routeAlias
else:
missingFile.write(routeCode + "\n")
missingFile.close()
mappingFile = open(join(PROJECT_ROOT, "../db_csv_files/routeMapping.json"), "w")
mappingFile.write(json.dumps(mapping, indent=2))
mappingFile.close()
'''
Import RouteMaster into db
'''
def importRouteMaster():
CsvFile = csv.reader(open(join(PROJECT_ROOT, "../db_csv_files/Route.csv"), "r"), delimiter='\t')
test = CsvFile.next()
stop_errors = []
print test
for row in CsvFile:
if len(row) < 1:
continue
from_to = getFromToStopsForRoute(row[0])
if from_to is None:
stop_errors.append(row[0])
continue
print row[0]
obj = Route(code=row[0], alias=row[1], from_stop_txt=row[2], to_stop_txt=row[3], from_stop=from_to[0], to_stop=from_to[1], distance=float(row[4]), stages=int(row[5]))
obj.save()
errors = open(join(PROJECT_ROOT, "../errors/RouteErrors.json"), "w")
errors.write(json.dumps(stop_errors, indent=2))
errors.close()
def csvClean1():
atlasCSV = csv.reader(open(join(PROJECT_ROOT, "../db_csv_files/Atlas.csv"), "r"), delimiter="\t")