troggle-unchained/parsers/survex.py
2020-06-25 03:17:56 +01:00

583 lines
28 KiB
Python

import sys
import os
import re
import time
from datetime import datetime, timedelta
from subprocess import call, Popen, PIPE
from django.utils.timezone import get_current_timezone
from django.utils.timezone import make_aware
import troggle.settings as settings
import troggle.core.models as models
import troggle.core.models_caves as models_caves
import troggle.core.models_survex as models_survex
from troggle.parsers.people import GetPersonExpeditionNameLookup
from troggle.core.views_caves import MapLocations
survexblockroot = None
ROOTBLOCK = "rootblock"
class SurvexLeg():
"""No longer a models.Model subclass, so no longer a database table
"""
tape = 0.0
compass = 0.0
clino = 0.0
class LoadSurvex():
"""A 'survex block' is a *begin...*end set of cave data.
A survex file can contain many begin-end blocks, which can be nested, and which can *include
other survex files.
A 'scansfolder' is what we today call a "survey scans folder" or a "wallet".
"""
# This interprets the survex "*data normal" command which sets out the order of the fields in the data, e.g.
# *DATA normal from to length gradient bearing ignore ignore ignore ignore
stardatadefault = {"type":"normal", "t":"leg", "from":0, "to":1, "tape":2, "compass":3, "clino":4}
stardataparamconvert = {"length":"tape", "bearing":"compass", "gradient":"clino"}
rx_linelen = re.compile(r"[\d\-+.]+$")
rx_team = re.compile(r"(?i)(Insts|Notes|Tape|Dog|Useless|Pics|Helper|Disto|Consultant)\s+(.*)$")
rx_person = re.compile(r"(?i) and | / |, | & | \+ |^both$|^none$")
rx_qm = re.compile(r'^\s*QM(\d)\s+?([a-dA-DxX])\s+([\w\-]+)\.(\d+)\s+(([\w\-]+)\.(\d+)|\-)\s+(.+)$')
# remember there is also QM_PATTERN used in views_other and set in settings.py
rx_comment = re.compile(r"([^;]*?)\s*(?:;\s*(.*))?\n?$")
rx_ref = re.compile(r'^\s*ref[\s.:]*(\d+)\s*#\s*(X)?\s*(\d+)')
rx_star = re.compile(r'\s*\*[\s,]*(\w+)\s*(.*?)\s*(?:;.*)?$')
rx_starref = re.compile(r'(?i)^\s*\*ref[\s.:]*((?:19[6789]\d)|(?:20[0123]\d))\s*#?\s*(X)?\s*(.*?\d+.*?)$')
rx_argsref = re.compile(r'(?i)^[\s.:]*((?:19[6789]\d)|(?:20[0123]\d))\s*#?\s*(X)?\s*(.*?\d+.*?)$')
survexlegsalllength = 0.0
survexlegsnumber = 0
depthbegin = 0
lineno = 0
insp = ""
callcount = 0
stardata ={}
def __init__(self):
pass
def LoadSurvexIgnore(self, survexblock, line, cmd):
if cmd == "title":
pass # unused in troggle today - but will become text list on SurvexBlock
elif cmd == "require":
pass # should we check survex version available for processing?
elif cmd in ["equate", "fix", "alias", "calibrate", "cs","entrance", "export", "case",
"declination", "infer","instrument", "sd", "units"]:
pass # we ignore all these, which is fine.
else:
if cmd in ["include", "data", "flags", "title", "set", "ref"]:
message = "! Unparsed [*{}]: '{}' {}".format(cmd, line, survexblock.survexfile.path)
print((self.insp+message))
models.DataIssue.objects.create(parser='survex', message=message)
else:
message = "! Bad svx command: [*{}] {} ({}) {}".format(cmd, line, survexblock, survexblock.survexfile.path)
print((self.insp+message))
models.DataIssue.objects.create(parser='survex', message=message)
def LoadSurvexTeam(self, survexblock, line):
teammembers = [ ]
mteammember = self.rx_team.match(line)
if mteammember:
for tm in self.rx_person.split(mteammember.group(2)):
if tm:
personexpedition = survexblock.expedition and GetPersonExpeditionNameLookup(survexblock.expedition).get(tm.lower())
if (personexpedition, tm) not in teammembers:
teammembers.append((personexpedition, tm))
personrole = models_survex.SurvexPersonRole(survexblock=survexblock, nrole=mteammember.group(1).lower(), personexpedition=personexpedition, personname=tm)
personrole.expeditionday = survexblock.expeditionday
if personexpedition:
personrole.person=personexpedition.person
personrole.save()
def LoadSurvexDate(self, survexblock, line):
# we should make this a date range for everything
if len(line) == 10:
survexblock.date = make_aware(datetime.strptime(re.sub(r"\.", "-", line), '%Y-%m-%d'), get_current_timezone())
expeditions = models.Expedition.objects.filter(year=line[:4])
if expeditions:
assert len(expeditions) == 1
survexblock.expedition = expeditions[0]
survexblock.expeditionday = survexblock.expedition.get_expedition_day(survexblock.date)
survexblock.save()
def LoadSurvexLineLeg(self, survexblock, svxline, sline, comment):
"""This reads compass, clino and tape data but only keeps the tape lengths,
the rest is discarded after error-checking.
"""
stardata = self.stardata
survexleg = SurvexLeg()
ls = sline.lower().split()
# this next fails for two surface survey svx files which use / for decimal point
# e.g. '29/09' in the tape measurement, or use decimals but in brackets, e.g. (06.05)
if stardata["type"] == "normal": # should use current flags setting for this
# print(" !! lineno '{}'\n !! svxline '{}'\n !! sline '{}'\n !! ls '{}'\n !! stardata {}".format(self.lineno, svxline, sline, ls,stardata))
tape = ls[stardata["tape"]]
tape = tape.replace("(","")
tape = tape.replace(")","")
tape = tape.replace("/",".")
try:
survexleg.tape = float(tape)
self.survexlegsnumber += 1
except ValueError:
print(("! Tape misread in", survexblock.survexfile.path))
print((" Stardata:", stardata))
print((" Line:", ls))
message = ' ! Value Error: Tape misread in line %s in %s' % (ls, survexblock.survexfile.path)
models.DataIssue.objects.create(parser='survex', message=message)
survexleg.tape = 0
try:
survexblock.totalleglength += survexleg.tape
self.survexlegsalllength += survexleg.tape
except ValueError:
message = ' ! Value Error: Tape length not added %s in %s' % (ls, survexblock.survexfile.path)
models.DataIssue.objects.create(parser='survex', message=message)
try:
lclino = ls[stardata["clino"]]
except:
print(("! Clino misread in", survexblock.survexfile.path))
print((" Stardata:", stardata))
print((" Line:", ls))
message = ' ! Value Error: Clino misread in line %s in %s' % (ls, survexblock.survexfile.path)
models.DataIssue.objects.create(parser='survex', message=message)
lclino = error
try:
lcompass = ls[stardata["compass"]]
except:
print(("! Compass misread in", survexblock.survexfile.path))
print((" Stardata:", stardata))
print((" Line:", ls))
message = ' ! Value Error: Compass misread in line %s in %s' % (ls, survexblock.survexfile.path)
models.DataIssue.objects.create(parser='survex', message=message)
lcompass = error
if lclino == "up":
survexleg.compass = 0.0
survexleg.clino = 90.0
elif lclino == "down":
survexleg.compass = 0.0
survexleg.clino = -90.0
elif lclino == "-" or lclino == "level":
try:
survexleg.compass = float(lcompass)
except ValueError:
print(("! Compass misread in", survexblock.survexfile.path))
print((" Stardata:", stardata))
print((" Line:", ls))
message = ' ! Value Error: line %s in %s' % (ls, survexblock.survexfile.path)
models.DataIssue.objects.create(parser='survex', message=message)
survexleg.compass = 1000
survexleg.clino = -90.0
else:
assert self.rx_linelen.match(lcompass), ls
assert self.rx_linelen.match(lclino) and lclino != "-", ls
survexleg.compass = float(lcompass)
survexleg.clino = float(lclino)
# delete the object so that django autosaving doesn't save it.
survexleg = None
def LoadSurvexRef(self, survexblock, args):
# *REF but also ; Ref years from 1960 to 2039
if len(args)< 4:
message = " ! Empty or BAD *REF command '{}' in '{}'".format(args, survexblock.survexfile.path)
print((self.insp+message))
models.DataIssue.objects.create(parser='survex', message=message)
return
argsgps = self.rx_argsref.match(args)
if argsgps:
yr, letterx, wallet = argsgps.groups()
else:
message = " ! BAD *REF command '{}' in '{}'".format(args, survexblock.survexfile.path)
print((self.insp+message))
models.DataIssue.objects.create(parser='survex', message=message)
return
if not letterx:
letterx = ""
else:
letterx = "X"
if len(wallet)<2:
wallet = "0" + wallet
assert (int(yr)>1960 and int(yr)<2039), "Wallet year out of bounds: %s" % yr
refscan = "%s#%s%s" % (yr, letterx, wallet)
try:
if int(wallet)>100:
message = " ! Wallet *REF {} - too big in '{}'".format(refscan, survexblock.survexfile.path)
print((self.insp+message))
models.DataIssue.objects.create(parser='survex', message=message)
except:
message = " ! Wallet *REF {} - not numeric in '{}'".format(refscan, survexblock.survexfile.path)
print((self.insp+message))
models.DataIssue.objects.create(parser='survex', message=message)
manyscansfolders = models_survex.ScansFolder.objects.filter(walletname=refscan)
if manyscansfolders:
survexblock.scansfolder = manyscansfolders[0]
survexblock.save()
if len(manyscansfolders) > 1:
message = " ! Wallet *REF {} - {} scan folders from DB search in {}".format(refscan, len(manyscansfolders), survexblock.survexfile.path)
print((self.insp+message))
models.DataIssue.objects.create(parser='survex', message=message)
else:
message = " ! Wallet *REF '{}' - NOT found in DB search '{}'".format(refscan, survexblock.survexfile.path)
print((self.insp+message))
models.DataIssue.objects.create(parser='survex', message=message)
def LoadSurvexQM(self, survexblock, qmline):
insp = self.insp
qm_no = qmline.group(1)
qm_grade = qmline.group(2)
if qmline.group(3): # usual closest survey station
qm_nearest = qmline.group(3)
if qmline.group(4):
qm_nearest = qm_nearest +"."+ qmline.group(4)
if qmline.group(6) and qmline.group(6) != '-':
qm_resolve_station = qmline.group(6)
if qmline.group(7):
qm_resolve_station = qm_resolve_station +"."+ qmline.group(7)
else:
qm_resolve_station = ""
qm_notes = qmline.group(8)
# Spec of QM in SVX files:
# ;Serial number grade(A/B/C/D/X) nearest-station resolution-station description
# ;QM1 a hobnob_hallway_2.42 hobnob-hallway_3.42 junction of keyhole passage
# ;QM1 a hobnob_hallway_2.42 - junction of keyhole passage
# NB none of the SurveyStations are in the DB now, so if we want to link to aSurvexStation
# we would have to create one. But that is not obligatory and no QMs loaded from CSVs have one
try:
qm = models_caves.QM.objects.create(number=qm_no,
# nearest_station=a_survex_station_object, # can be null
nearest_station_description=qm_resolve_station,
nearest_station_name=qm_nearest,
grade=qm_grade.upper(),
location_description=qm_notes)
qm.save
# message = " ! QM{} '{}' CREATED in DB in '{}'".format(qm_no, qm_nearest,survexblock.survexfile.path)
# print(insp+message)
# models.DataIssue.objects.create(parser='survex', message=message)
except:
message = " ! QM{} FAIL to create {} in'{}'".format(qm_no, qm_nearest,survexblock.survexfile.path)
print(insp+message)
models.DataIssue.objects.create(parser='survex', message=message)
def LoadSurvexComment(self, survexblock, comment):
# ignore all comments except ;ref and ;QM
refline = self.rx_ref.match(comment)
if refline:
comment = comment.replace("ref","").strip()
self.LoadSurvexRef(survexblock, comment)
qmline = self.rx_qm.match(comment)
if qmline:
self.LoadSurvexQM(survexblock, qmline)
def LoadSurvexFlags(self, line, cmd):
# Here we could set on/off 'splay', 'not splay', 'surface', 'not surface', or 'duplicate'
# but this data is only used for sense-checking not to actually calculate anything important
pass
def LoadSurvexSetup(self,survexblock, survexfile):
self.depthbegin = 0
self.stardata = self.stardatadefault
blocklegs = self.survexlegsnumber
print(self.insp+" - MEM:{:.3f} Reading. parent:{} <> {} ".format(models.get_process_memory(),survexblock.survexfile.path, survexfile.path))
self.lineno = 0
sys.stderr.flush();
self.callcount +=1
if self.callcount >=10:
self.callcount=0
print(".", file=sys.stderr,end='')
# Try to find the cave in the DB if not use the string as before
path_match = re.search(r"caves-(\d\d\d\d)/(\d+|\d\d\d\d-?\w+-\d+)/", survexblock.survexfile.path)
if path_match:
pos_cave = '%s-%s' % (path_match.group(1), path_match.group(2))
cave = models_caves.getCaveByReference(pos_cave)
if cave:
survexfile.cave = cave
def RecursiveLoad(self, survexblock, survexfile, fin):
"""Follows the *include links in all the survex files from the root file 1623.svx
and reads in the survex blocks, other data and the wallet references (scansfolder) as it
goes. This part of the data import process is where the maximum memory is used and where it
crashes on memory-constrained machines. Begin-end blocks may also be nested.
"""
self.LoadSurvexSetup(survexblock, survexfile)
insp =self.insp
previousnlegs = 0
svxlines = fin.read().splitlines()
# cannot close file now as may be recursively called with the same fin if nested *begin-end
for svxline in svxlines:
self.lineno += 1
sline, comment = self.rx_comment.match(svxline.strip()).groups()
if comment:
self.LoadSurvexComment(survexblock, comment)
if not sline:
continue # skip blank lines
# detect the star command
mstar = self.rx_star.match(sline)
if mstar: # yes we are reading a *cmd
cmd, args = mstar.groups()
cmd = cmd.lower()
if re.match("include$(?i)", cmd):
includepath = os.path.normpath(os.path.join(os.path.split(survexfile.path)[0], re.sub(r"\.svx$", "", args)))
print((insp+' - INCLUDE-go path found, including - ' + includepath))
# Try to find the cave in the DB. if not, use the string as before
path_match = re.search(r"caves-(\d\d\d\d)/(\d+|\d\d\d\d-?\w+-\d+)/", includepath)
if path_match:
pos_cave = '%s-%s' % (path_match.group(1), path_match.group(2))
print((insp+' - Match in DB (i) for cave {}.'.format(pos_cave)))
cave = models_caves.getCaveByReference(pos_cave)
if cave:
survexfile.cave = cave
else:
print((insp+' - NO Match in DB (i) for a cave for {}'.format(includepath)))
includesurvexfile = models_survex.SurvexFile(path=includepath)
includesurvexfile.save()
includesurvexfile.SetDirectory()
if includesurvexfile.exists():
survexblock.save()
self.insp += "> "
#--------------------------------------------------------
fininclude = includesurvexfile.OpenFile()
self.RecursiveLoad(survexblock, includesurvexfile, fininclude)
fininclude.close()
#--------------------------------------------------------
self.insp = self.insp[2:]
insp = self.insp
print((insp+' - INCLUDE-return from include - ' + includepath))
else:
print((insp+' ! ERROR *include file not found for %s' % includesurvexfile))
elif re.match("begin$(?i)", cmd):
# On a *begin statement we start a new survexblock.
# There should not be any *include inside a begin-end block, so this is a simple
# load not a recursive fileload. But there may be many blocks nested to any depth in one file.
if args:
newsvxpath = os.path.join(os.path.split(survexfile.path)[0], re.sub(r"\.svx$", "", args))
# Try to find the cave in the DB if not use the string as before
path_match = re.search(r"caves-(\d\d\d\d)/(\d+|\d\d\d\d-?\w+-\d+)/", newsvxpath)
if path_match:
pos_cave = '%s-%s' % (path_match.group(1), path_match.group(2))
# print(insp+pos_cave)
cave = models_caves.getCaveByReference(pos_cave)
if cave:
survexfile.cave = cave
else:
print((insp+' - No match (b) for %s' % newsvxpath))
previousnlegs = self.survexlegsnumber
name = args.lower()
print(insp+' - Begin found for:{}, creating new SurvexBlock '.format(name))
# the recursive call re-reads the entire file. This is wasteful. We should pass in only
# the un-parsed part of the file.
survexblockdown = models_survex.SurvexBlock(name=name, parent=survexblock,
survexpath=survexblock.survexpath+"."+name,
cave=survexfile.cave, survexfile=survexfile,
legsall=0, legssplay=0, legssurfc=0, totalleglength=0.0)
survexblockdown.save()
survexblock.save()
survexblock = survexblockdown
print(insp+" - BLOCK-enter nested *begin/*end block: '{}'".format(name))
self.insp += "> "
#--------------------------------------------------------
self.RecursiveLoad(survexblockdown, survexfile, fin)
#--------------------------------------------------------
# do not close the file as there may be more blocks in this one
# and it is re-read afresh with every nested begin-end block.
self.insp = self.insp[2:]
insp = self.insp
else:
self.depthbegin += 1
elif re.match("end$(?i)", cmd):
if self.depthbegin:
print(insp+" - BLOCK-return from nested *begin/*end block: '{}'".format(args))
self.depthbegin -= 1
else:
legsinblock = self.survexlegsnumber - previousnlegs
print(insp+" - LEGS: {} (previous: {}, now:{})".format(legsinblock,previousnlegs,self.survexlegsnumber))
survexblock.legsall = legsinblock
survexblock.save()
return
elif cmd == "ref":
self.LoadSurvexRef(survexblock, args)
elif cmd == "flags":
self.LoadSurvexFlags(args, cmd)
elif cmd == "data":
ls = args.lower().split()
stardata = { "type":ls[0] }
for i in range(0, len(ls)):
stardata[self.stardataparamconvert.get(ls[i], ls[i])] = i - 1
self.stardata = stardata
if ls[0] in ["normal", "cartesian", "nosurvey"]:
assert (("from" in stardata and "to" in stardata) or "station" in stardata), args
elif ls[0] == "default":
stardata = self.stardatadefault
else:
assert ls[0] == "passage", args
elif cmd == "set" and re.match("names(?i)", args):
pass
elif re.match("date$(?i)", cmd):
self.LoadSurvexDate(survexblock, args)
elif re.match("team$(?i)", cmd):
self.LoadSurvexTeam(survexblock, args)
else:
self.LoadSurvexIgnore(survexblock, args, cmd)
else: # not a *cmd so we are reading data OR rx_comment failed
if "from" in self.stardata: # only interested in survey legs
self.LoadSurvexLineLeg(survexblock, svxline, sline, comment)
else:
pass # ignore all other sorts of data
def FindAndLoadAllSurvex(survexblockroot, survexfileroot):
"""Follows the *include links recursively to find files
"""
print(' - redirecting stdout to svxblks.log...')
stdout_orig = sys.stdout
# Redirect sys.stdout to the file
sys.stdout = open('svxblks.log', 'w')
svxl = LoadSurvex()
finroot = survexfileroot.OpenFile()
svxl.RecursiveLoad(survexblockroot, survexfileroot, finroot)
finroot.close()
survexlegsnumber = svxl.survexlegsnumber
survexlegsalllength = svxl.survexlegsalllength
# Close the logging file, Restore sys.stdout to our old saved file handle
sys.stdout.close()
print("+", file=sys.stderr)
sys.stderr.flush();
sys.stdout = stdout_orig
return (survexlegsnumber, survexlegsalllength)
def LoadAllSurvexBlocks():
print(' - Flushing All Survex Blocks...')
models_survex.SurvexBlock.objects.all().delete()
models_survex.SurvexFile.objects.all().delete()
models_survex.SurvexDirectory.objects.all().delete()
models_survex.SurvexPersonRole.objects.all().delete()
models_survex.SurvexStation.objects.all().delete()
print(" - survex Data Issues flushed")
models.DataIssue.objects.filter(parser='survex').delete()
survexfileroot = models_survex.SurvexFile(path=settings.SURVEX_TOPNAME, cave=None)
survexfileroot.save()
survexfileroot.SetDirectory()
survexblockroot = models_survex.SurvexBlock(name=ROOTBLOCK, survexpath="", cave=None, survexfile=survexfileroot,
legsall=0, legssplay=0, legssurfc=0, totalleglength=0.0)
# this is the first so id=1
survexblockroot.save()
print(' - Loading All Survex Blocks...')
memstart = models.get_process_memory()
survexlegsnumber, survexlegsalllength = FindAndLoadAllSurvex(survexblockroot, survexfileroot)
memend = models.get_process_memory()
print(" - MEMORY start:{:.3f} MB end:{:.3f} MB increase={:.3f} MB".format(memstart,memend, memend-memstart))
survexblockroot.totalleglength = survexlegsalllength
survexblockroot.legsall = survexlegsnumber
survexblockroot.save()
print(" - total number of survex legs: {}m".format(survexlegsnumber))
print(" - total leg lengths loaded: {}m".format(survexlegsalllength))
print(' - Loaded All Survex Blocks.')
poslineregex = re.compile(r"^\(\s*([+-]?\d*\.\d*),\s*([+-]?\d*\.\d*),\s*([+-]?\d*\.\d*)\s*\)\s*([^\s]+)$")
def LoadPos():
"""Run cavern to produce a complete .3d file, then run 3dtopos to produce a table of
all survey point positions. Then lookup each position by name to see if we have it in the database
and if we do, then save the x/y/z coordinates.
If we don't have it in the database, print an error message and discard it.
This is ONLY ever used for entrance and fixedpts locations for the prospecting map:
about 600 points out of 32,000.
"""
topdata = settings.SURVEX_DATA + settings.SURVEX_TOPNAME
print((' - Generating a list of Pos from %s.svx and then loading...' % (topdata)))
found = 0
skip = {}
print("\n") # extra line because cavern overwrites the text buffer somehow
# cavern defaults to using same cwd as supplied input file
call([settings.CAVERN, "--output=%s.3d" % (topdata), "%s.svx" % (topdata)])
call([settings.THREEDTOPOS, '%s.3d' % (topdata)], cwd = settings.SURVEX_DATA)
mappoints = {}
for pt in MapLocations().points():
svxid, number, point_type, label = pt
mappoints[svxid]=True
posfile = open("%s.pos" % (topdata))
posfile.readline() #Drop header
try:
survexblockroot = models_survex.SurvexBlock.objects.get(name=ROOTBLOCK)
except:
try:
survexblockroot = models_survex.SurvexBlock.objects.get(id=1)
except:
message = ' ! FAILED to find root SurvexBlock'
print(message)
models.DataIssue.objects.create(parser='survex', message=message)
raise
for line in posfile.readlines():
r = poslineregex.match(line)
if r:
x, y, z, id = r.groups()
for sid in mappoints:
if id.endswith(sid):
blockpath = "." + id[:-len(sid)].strip(".")
try:
sbqs = models_survex.SurvexBlock.objects.filter(survexpath=blockpath)
if len(sbqs)==1:
sb = sbqs[0]
if len(sbqs)>1:
message = ' ! MULTIPLE SurvexBlocks matching Entrance point {} {}'.format(blockpath, sid)
print(message)
models.DataIssue.objects.create(parser='survex', message=message)
sb = sbqs[0]
elif len(sbqs)<=0:
message = ' ! ZERO SurvexBlocks matching Entrance point {} {}'.format(blockpath, sid)
print(message)
models.DataIssue.objects.create(parser='survex', message=message)
sb = survexblockroot
except:
message = ' ! FAIL in getting SurvexBlock matching Entrance point {} {}'.format(blockpath, sid)
print(message)
models.DataIssue.objects.create(parser='survex', message=message)
try:
ss = models_survex.SurvexStation(name=id, block=sb)
ss.x = float(x)
ss.y = float(y)
ss.z = float(z)
ss.save()
found += 1
except:
message = ' ! FAIL to create SurvexStation Entrance point {} {}'.format(blockpath, sid)
print(message)
models.DataIssue.objects.create(parser='survex', message=message)
raise
print(" - {} SurvexStation entrances found.".format(found))