troggle-unchained/parsers/survex.py

import sys
import os
import re
import time
import copy
import subprocess

from pathlib import Path
from datetime import datetime, timedelta, date

from django.utils.timezone import get_current_timezone
from django.utils.timezone import make_aware

import troggle.settings as settings
from troggle.core.models.caves import Cave, Entrance, QM, LogbookEntry
from troggle.core.utils import get_process_memory, chaosmonkey
from troggle.parsers.people import GetPersonExpeditionNameLookup
from troggle.parsers.logbooks import GetCaveLookup
from troggle.core.models.troggle import DataIssue, Expedition
from troggle.core.models.survex import SurvexPersonRole, Wallet, SurvexDirectory, SurvexFile, SurvexBlock, SurvexStation

'''Imports the tree of survex files following form a defined root .svx file
It does also NOT scan the Loser repo for all the svx files - though it should !
'''

todo = '''Also walk the entire tree in the :loser: repo looking for unconnected survex files
- add them to the system so that they can be reported-on
- produce a parser report and create a troggle report page (some are OK, e.g. futility series replaced by ARGE survey in 115)

-       If you look at e.g. http://expo.survex.com/survexfile/161#T_caves-1623/161/lhr/alllhr
        you will see than have the team members are recognised by this parser, but not recognised by the
        wider troggle system (the name is not a hyperlink) - apparently randomly. 
        GetPersonExpeditionNameLookup() needs to be fixed.
        
-#BUG, if *date comes after *team, the person's date is not set at all. It needs re-setting at the endof the block.

- LoadSurvexFile() Creates a new current survexfile and valid .survexdirectory
        The survexblock passed-in is not necessarily the parent. FIX THIS.
        
- rx_qm recognises only simple survey point ids. EXTEND to cover more naming formats and test fully for 2023
'''
survexblockroot = None
ROOTBLOCK = "rootblock"
METRESINFEET = 3.28084

debugprint = False # Turns on debug printout for just one *include file
debugprinttrigger = "!"
# debugprinttrigger = "caves-1623/40/old/EisSVH"

class MapLocations(object):
    p = [
        ("laser.0_7", "BNase", "Reference", "Bräuning Nase laser point"),
        ("226-96", "BZkn", "Reference", "Bräuning Zinken trig point"),
        ("vd1","VD1","Reference", "VD1 survey point"),
        ("laser.kt114_96","HSK","Reference", "Hinterer Schwarzmooskogel trig point"), 
        ("2000","Nipple","Reference", "Nipple (Weiße Warze)"),
        ("3000","VSK","Reference", "Vorderer Schwarzmooskogel summit"),
        ("topcamp", "OTC", "Reference", "Old Top Camp"),
        ("laser.0", "LSR0", "Reference", "Laser Point 0"),
        ("laser.0_1", "LSR1", "Reference", "Laser Point 0/1"),
        ("laser.0_3", "LSR3", "Reference", "Laser Point 0/3"),
        ("laser.0_5", "LSR5", "Reference", "Laser Point 0/5"),
        ("225-96", "BAlm", "Reference", "Bräuning Alm trig point")
    ]
    def points(self):
        for ent in Entrance.objects.all():
            if ent.best_station():
                try:
                    k = ent.caveandentrance_set.all()[0].cave
                except:
                    message = " ! Failed to get Cave linked to Entrance:{} from:{} best:{}".format(ent.name, ent.filename, ent.best_station())
                    DataIssue.objects.create(parser='entrances', message=message)
                    print(message)
                    raise
                try:
                    areaName = k.getArea().short_name
                except:
                    message = " ! Failed to get Area on cave '{}' linked to Entrance:{} from:{} best:{}".format(cave, ent.name, ent.filename, ent.best_station())
                    DataIssue.objects.create(parser='entrances', message=message)
                    print(message)
                    raise
                self.p.append((ent.best_station(), "%s-%s" % (areaName, str(ent)[5:]), ent.needs_surface_work(), str(ent)))
        message = f" -  {len(self.p)} entrances linked to caves."
        print(message)
        return self.p

    def __str__(self):
        return "{} map locations".format(len(self.p))

class SurvexLeg():
    """No longer a models.Model subclass, so no longer a database table
    """
    tape        = 0.0
    compass     = 0.0
    clino       = 0.0

class LoadingSurvex():
    """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 'scanswallet' is what we today call a "survey scans folder" or a "wallet".
    """
    rx_begin   = re.compile(r'(?i)begin')
    rx_end     = re.compile(r'(?i)end$')
    rx_title   = re.compile(r'(?i)title$')
    rx_ref     = re.compile(r'(?i)ref$')
    rx_data    = re.compile(r'(?i)data$')
    rx_flags   = re.compile(r'(?i)flags$')
    rx_alias   = re.compile(r'(?i)alias$')
    rx_entrance   = re.compile(r'(?i)entrance$')
    rx_date    = re.compile(r'(?i)date$')
    rx_units   = re.compile(r'(?i)units$')
    rx_team    = re.compile(r'(?i)team$')
    rx_set     = re.compile(r'(?i)set$')

    rx_names   = re.compile(r'(?i)names')
    rx_flagsnot= re.compile(r"not\s")
    rx_linelen = re.compile(r"[\d\-+.]+$")
    instruments = "(waiting_patiently|slacker|Useless|nagging|unknown|Inst|instrument|rig|rigger|rigging|helper|something| compass|comp|clino|Notes|sketch|book|Tape|Dog|Pics|photo|drawing|Helper|GPS|Disto|Distox|Distox2|topodroid|point|Consultant|nail|polish|nail_polish_bitch|nail_polish_monkey|varnish|nail_polish|nail_varnish|bitch|monkey|PowerDrill|drill)"
    rx_teammem = re.compile(r"(?i)"+instruments+"?(?:es|s)?\s+(.*)"+instruments+"?(?:es|s)?$")
    rx_person  = re.compile(r"(?i) and | / |, | & | \+ |^both$|^none$")
    rx_qm      = re.compile(r'(?i)^\s*QM(\d)\s+?([a-dA-DxX])\s+([\w\-]+)\.(\d+)\s+(([\w\-]+)\.(\d+)|\-)\s+(.+)$')
    # does not recognise non numeric suffix survey point ids
    rx_qm0     = re.compile(r'(?i)^\s*QM(\d)\s+(.+)$')
#   remember there is also QM_PATTERN used in views.other and set in settings.py
    rx_tapelng = re.compile(r'(?i).*(tape|length).*$')

    rx_cave    = re.compile(r'(?i)caves-(\d\d\d\d)/([-\d\w]+|\d\d\d\d-?\w+-\d+)')
    rx_comment = re.compile(r'([^;]*?)\s*(?:;\s*(.*))?\n?$')
    rx_comminc = re.compile(r'(?i)^\*include[\s]*([-\w/]*).*$') # inserted by linear collate ;*include
    rx_commcni = re.compile(r'(?i)^\*edulcni[\s]*([-\w/]*).*$') # inserted by linear collate ;*edulcni
    rx_include = re.compile(r'(?i)^\s*(\*include[\s].*)$')
    rx_commref = re.compile(r'(?i)^\s*ref(?:erence)?[\s.:]*(\d+)\s*#\s*(X)?\s*(\d+)')
    rx_wallet  = re.compile(r'(?i)^\s*wallet[\s.:]*(\d+)\s*#\s*(X)?\s*(\d+)')
    rx_implicit= re.compile(r'(?i)^[\s.:]*(\d+)\s*#\s*(X)?\s*(\d+)')
    rx_ref_text= re.compile(r'(?i)^\s*\"[^"]*\"\s*$')
    rx_star    = re.compile(r'(?i)\s*\*[\s,]*(\w+)\s*(.*?)\s*(?:;.*)?$')
    rx_starref = re.compile(r'(?i)^\s*\*ref[\s.:]*((?:19[6789]\d)|(?:20[0123]\d))\s*#?\s*(X)?\s*(.*?\d+.*?)$')
    rx_argsref = re.compile(r'(?i)^[\s.:]*((?:19[6789]\d)|(?:20[0123]\d))\s*#?\s*(X)?\s*(.*?\d+.*?)$')
    rx_badmerge= re.compile(r'(?i).*(\>\>\>\>\>)|(\=\=\=\=\=)|(\<\<\<\<\<).*$')
    rx_ref2 = re.compile(r'(?i)\s*ref[.;]?')
    rx_ref3 = re.compile(r'(?i)\s*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
    datastardefault = {"type":"normal", "from":0, "to":1, "tape":2, "compass":3, "clino":4}
    flagsdefault = {"duplicate":False, "surface":False, "splay":False, "skiplegs":False, "splayalias":False}

    datastar ={}
    flagsstar = {}
    units = "metres"
    unitsfactor = None
    slength = 0.0
    legsnumber = 0
    depthbegin = 0
    depthinclude = 0
    unitsstack = []
    legsnumberstack = []
    slengthstack = []
    personexpedstack = []
    stackbegin =[]
    flagsstack =[]
    datastack =[]
    includestack = []
    stacksvxfiles = []
    svxfileslist = []
    svxdirs = {} 
    uniquename = {}
    expos = {}
    survexdict = {} # each key is a directory, and its value is a list of files
    lineno = 0
    insp = ""
    callcount = 0
    caverncount = 0
    ignoreprefix = ["surface", "kataster", "fixedpts", "gpx"]
    ignorenoncave = ["caves-1623", "caves-1623/2007-neu"]
    includedfilename =""
    currentsurvexblock = None
    currentsurvexfile = None
    currentcave = None
    caverndate = None
    currentpersonexped = []

    def __init__(self):
        self.caveslist = GetCaveLookup()
        pass

    def LoadSurvexFallThrough(self, survexblock, line, cmd):
        if cmd == "require":
            pass # should we check survex version available for processing?
        elif cmd in ["equate", "fix", "calibrate", "cs", "export", "case", 
                "declination", "infer","instrument", "sd"]:
            pass # we ignore all these, which is fine.
        else:
            if cmd in ["include", "data", "flags", "title", "entrance","set", "units", "alias", "ref"]:
                message = "! Unparsed [*{}]: '{}' {} - not an error (probably)".format(cmd, line, survexblock.survexfile.path)
                print((self.insp+message))
                DataIssue.objects.create(parser='survex', message=message)
            else:
                message = "! Bad svx command: [*{}] {} ({}) {}".format(cmd, line, survexblock, survexblock.survexfile.path)
                print((self.insp+message))
                DataIssue.objects.create(parser='survex', message=message)

    def LoadSurvexTeam(self, survexblock, line):
        """Interpeting the *team fields means interpreting older style survex as well as current survex standard,
        *team Insts Anthony Day   - this is how most of our files specify the team member
        *team "Anthony Day" notes pictures tape    - this is how the survex documentation says it should be done
        We have a huge variety of abbreviations and mispellings. The most laconic being 
        *team gb, bl
        If you look at e.g. http://expo.survex.com/survexfile/161#T_caves-1623/161/lhr/alllhr
        you will see than have the team members are recognised by this parser, but not recognised by the
        wider troggle system (the name is not a hyperlink) - apparently randomly. 
        GetPersonExpeditionNameLookup() needs to be fixed.
        """
        teammembers = [ ]
        mteammember = self.rx_teammem.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 = SurvexPersonRole(survexblock=survexblock, personexpedition=personexpedition, personname=tm)
                        personrole.save()
                        personrole.expeditionday = survexblock.expeditionday #BUG, if *date comes after *team, this is NOT SET.
                        if personexpedition:
                            personrole.person=personexpedition.person
                            self.currentpersonexped.append(personexpedition)
                        personrole.save()

    def LoadSurvexEntrance(self, survexblock, line):
        # Not using this yet
        pass
        
    def LoadSurvexAlias(self, survexblock, line):
        # *alias station - ..
        splayalias = re.match("(?i)station\s*\-\s*\.\.\s*$",line)
        if splayalias:
            self.flagsstar["splayalias"] = True
        else:
            message = "! Bad *ALIAS: '{}' ({}) {}".format(line, survexblock, survexblock.survexfile.path)
            print((self.insp+message))
            DataIssue.objects.create(parser='survex', message=message)

    def LoadSurvexUnits(self, survexblock, line):
        # all for 4 survex files with measurements in feet. bugger.
        tapeunits = self.rx_tapelng.match(line) # tape|length
        if not tapeunits:
            return
        convert = re.match("(?i)(\w*)\s*([\.\d]+)\s*(\w*)",line)
        if convert:
            factor = convert.groups()[1]
            self.unitsfactor = float(factor)
            if debugprint:
                message = "! *UNITS NUMERICAL conversion [{}x] '{}' ({}) {}".format(factor, line, survexblock, survexblock.survexfile.path)
                print((self.insp+message))
                DataIssue.objects.create(parser='survexunits', message=message)

        feet = re.match("(?i).*feet$",line)
        metres = re.match("(?i).*(METRIC|METRES|METERS)$",line)
        if feet:
            self.units = "feet"
        elif metres:
            self.units = "metres"
        else:
            message = "! *UNITS in YARDS!? - not converted '{}' ({}) {}".format(line, survexblock, survexblock.survexfile.path)
            print((self.insp+message))
            DataIssue.objects.create(parser='survexunits', message=message)

    def LoadSurvexDate(self, survexblock, line):
        # we should make this a date RANGE for everything
        def findexpedition(year):
            return Expedition.objects.filter(year=year)
            
        def setdate(year):
            # cacheing to save DB query on every block and to prepare for django-less troggle in future
            if year in self.expos:
                expo = self.expos[year]
            else:
                expeditions = findexpedition(year)
                if len(expeditions) != 1 :
                    message = f"! More than one expedition in year {year} '{line}' ({survexblock}) {survexblock.survexfile.path}"
                    print((self.insp+message))
                    DataIssue.objects.create(parser='survexunits', message=message)
                   
                expo= expeditions[0]
                self.expos[year]= expo

            survexblock.expedition = expo
            survexblock.expeditionday = survexblock.expedition.get_expedition_day(survexblock.date)
            survexblock.save()

        if len(line) > 10: 
            if line[10] == "-":
                line = line[0:10]
        if len(line) == 10: 
            year = line[:4]
            # TO DO set to correct Austrian timezone Europe/Vienna
            # %m and %d need leading zeros. Source svx files require them.
            survexblock.date = datetime.strptime(re.sub(r"\.", "-", line), '%Y-%m-%d')
            setdate(year)
        elif len(line) == 7: 
            year = line[:4]
            survexblock.date = datetime.strptime(re.sub(r"\.", "-", line), '%Y-%m') # sets to first of month
            setdate(year)
        elif len(line) == 4: 
            year = line[:4]
            survexblock.date = datetime.strptime(line, '%Y') # sets to January 1st
            setdate(year)
        else:
            message = "! DATE unrecognised '{}' ({}) {}".format(line, survexblock, survexblock.survexfile.path)
            print((self.insp+message))
            DataIssue.objects.create(parser='survex', message=message)

    def LoadSurvexLeg(self, survexblock, sline, comment):
        """This reads compass, clino and tape data but only keeps the tape lengths,
        the rest is discarded after error-checking.
        Now skipping the error checking - returns as soon as the leg is not one we count.
        """
        invalid_clino = 180.0
        invalid_compass = 720.0
        invalid_tape = 0.0

        if self.flagsstar["skiplegs"]:
            if debugprint:
                print("skip in ", self.flagsstar, survexblock.survexfile.path)
            return

        if debugprint:
            print("! LEG datastar type:{}++{}\n{} ".format(self.datastar["type"].upper(), survexblock.survexfile.path, sline))
        if self.datastar["type"] == "passage":
            return
        if self.datastar["type"] == "cartesian":
            return
        if self.datastar["type"] == "nosurvey":
            return
        if self.datastar["type"] == "diving":
            return
        if self.datastar["type"] == "cylpolar":
            return
        if debugprint:
            print(" !! LEG data lineno:{}\n !! sline:'{}'\n !! datastar['tape']: {}".format(self.lineno, sline, self.datastar["tape"]))
           
        if self.datastar["type"] != "normal": 
            return
 
        datastar = self.datastar # shallow copy: alias but the things inside are the same things
        survexleg = SurvexLeg()
 
        ls = sline.lower().split()

        # skip all splay legs
        if ls[datastar["from"]] == ".." or ls[datastar["from"]] == ".":
            if debugprint:
                print("Splay in ", survexblock.survexfile.path)
            return
        if ls[datastar["to"]] == ".." or ls[datastar["to"]] == ".":
            if debugprint:
                print("Splay in ", survexblock.survexfile.path)
            return
        if self.flagsstar["splayalias"]:
            if ls[datastar["from"]] == "-":
                if debugprint:
                    print("Aliased splay in ", survexblock.survexfile.path)
                return
            if ls[datastar["to"]] == "-":
                if debugprint:
                    print("Aliased splay in ", survexblock.survexfile.path)
                return

        try:
            tape = ls[datastar["tape"]]
        except:
            message = ' ! datastar parsing incorrect in line %s in %s' % (ls, survexblock.survexfile.path)
            print((self.insp+message))
            DataIssue.objects.create(parser='survexleg', message=message)
            survexleg.tape = invalid_tape
            return
        # e.g. '29/09' or '(06.05)' in the tape measurement
        # tape = tape.replace("(","")  # edited original file (only one) instead
        # tape = tape.replace(")","")  # edited original file (only one) instead
        # tape = tape.replace("/",".") # edited original file (only one) instead.
        try:
            if self.unitsfactor:
                tape = float(tape) * self.unitsfactor
                if debugprint:
                    message = " ! Units: Length scaled {}m '{}' in ({})  units:{} factor:{}x".format(tape, ls, survexblock.survexfile.path, self.units, self.unitsfactor)
                    print((self.insp+message))
                    DataIssue.objects.create(parser='survexleg', message=message)
            if self.units =="feet":
                tape = float(tape) / METRESINFEET
                if debugprint:
                    message = " ! Units: converted  to {:.3f}m from {} '{}' in ({})".format(tape, self.units, ls, survexblock.survexfile.path)
                    print((self.insp+message))
                    DataIssue.objects.create(parser='survexleg', message=message)
            survexleg.tape = float(tape)
            self.legsnumber += 1
        except ValueError:
            message = " ! Value Error: Tape misread in line'{}' in {} units:{}".format(ls, survexblock.survexfile.path, self.units)
            print((self.insp+message))
            DataIssue.objects.create(parser='survexleg', message=message)
            survexleg.tape = invalid_tape
        try:
            survexblock.legslength += survexleg.tape
            self.slength   += survexleg.tape
        except ValueError:
            message = " ! Value Error: Tape length not added '{}' in {} units:{}".format(ls, survexblock.survexfile.path, self.units)
            print((self.insp+message))
            DataIssue.objects.create(parser='survexleg', message=message)

        try:
            lcompass = ls[datastar["compass"]]
        except:
            message = ' ! Value Error: Compass not found in line %s in %s' % (ls, survexblock.survexfile.path)
            print((self.insp+message))
            DataIssue.objects.create(parser='survexleg', message=message)
            lcompass = invalid_compass

        try:
            lclino = ls[datastar["clino"]]
        except:
            print(("! Clino misread in", survexblock.survexfile.path))
            print(("  datastar:", datastar))
            print(("  Line:", ls))
            message = ' ! Value Error: Clino misread in line %s in %s' % (ls, survexblock.survexfile.path)
            DataIssue.objects.create(parser='survexleg', message=message)
            lclino = invalid_clino

        if lclino == "up":
            survexleg.clino = 90.0
            lcompass = invalid_compass
        elif lclino == "down":
            survexleg.clino = -90.0
            lcompass = invalid_compass
        elif lclino == "-" or lclino == "level":
            survexleg.clino = -90.0

        try:
            survexleg.compass = float(lcompass)
        except ValueError:
            print(("! Compass misread in", survexblock.survexfile.path))
            print(("  datastar:", datastar))
            print(("  Line:", ls))
            message = " ! Value Error: lcompass:'{}' line {} in '{}'".format(lcompass, 
                    ls, survexblock.survexfile.path)
            DataIssue.objects.create(parser='survexleg', message=message)
            survexleg.compass = invalid_compass

        # delete the object to save memory
        survexleg = None

    def LoadSurvexRef(self, survexblock, args):
        #print(self.insp+ "*REF ---- '"+ args +"'")
        url=f'/survexfile/{survexblock.survexfile.path}'
        # *REF but also ; Ref      years from 1960 to 2039
        refline = self.rx_ref_text.match(args)
        if refline:
            # a textual reference such as  "1996-1999 Not-KH survey book pp 92-95"
            print(self.insp+ "*REF quoted text so ignored:"+ args)
            return
        
        if len(args)< 4:
            message = " ! Empty or BAD *REF statement '{}' in '{}'".format(args, survexblock.survexfile.path)
            print((self.insp+message))
            DataIssue.objects.create(parser='survex', message=message, url=url)
            return

        argsgps = self.rx_argsref.match(args)
        if argsgps:
            yr, letterx, wallet = argsgps.groups()
        else:
            message = " ! Wallet *REF '{}' malformed id in '{}' ".format(args, survexblock.survexfile.path)
            print(self.insp+message)
            DataIssue.objects.create(parser='survex', message=message, url=url)
            return

        if not letterx:
            letterx = ""
        else:
            letterx = "X"
        if len(wallet)<2:
            wallet = "0" + wallet
        if not (int(yr)>1960 and int(yr)<2039):
                message = " ! Wallet year out of bounds {yr} '{refscan}' {survexblock.survexfile.path}"
                print((self.insp+message))
                DataIssue.objects.create(parser='survex', message=message, url=url)
            
        refscan = "%s#%s%s" % (yr, letterx, wallet)
        try:
            if int(wallet)>99:
                message = " ! Wallet *REF {} - too big in '{}'".format(refscan, survexblock.survexfile.path)
                print((self.insp+message))
                DataIssue.objects.create(parser='survex', message=message, url=url)
        except:
            message = " ! Wallet *REF {} - not numeric in '{}'  -- parsing continues".format(refscan, survexblock.survexfile.path)
            print((self.insp+message))
            DataIssue.objects.create(parser='survex', message=message, url=url)
        manywallets = Wallet.objects.filter(walletname=refscan)
        if manywallets:
            survexblock.scanswallet = manywallets[0] # this is a ForeignKey field
            print(manywallets[0])
            survexblock.save()
            if len(manywallets) > 1:
                message = " ! Wallet *REF {} - more than one found {} scan folders  in {}".format(refscan, len(manywallets), survexblock.survexfile.path)
                print((self.insp+message))
                DataIssue.objects.create(parser='survex', message=message, url=url)
        else:
            message = " ! Wallet *REF '{}' - NOT found '{}'".format(refscan, survexblock.survexfile.path)
            print((self.insp+message))
            DataIssue.objects.create(parser='survex', message=message, url=url)

    def LoadSurvexQM(self, survexblock, qmline):
        insp = self.insp

        qm_no = qmline.group(1) # this may not be unique across multiple survex files
        
        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
        
        # Older troggle/CSV assumes a logbook entry 'found_by' for each QM, with a date. 
        # We don't need this anymore so we don't need to create a placeholder logbook entry.
        qmyear = str(survexblock.date)[:4]
        blockname = survexblock.name[:7]
        logslug = f'D{int(qmyear)}_{blockname}_{int(qm_no):03d}'
        if survexblock.survexfile.cave:
            caveslug = survexblock.survexfile.cave.slug()
            place = survexblock.survexfile.cave
        else:
            caveslug = "ugh"
            place = "oops"
       
        
        # message = f' ! - logbook dummy "{logslug}" {str(survexblock.date)[:11]} for cave "{caveslug}" created.'

        # placeholder, hadToCreate = LogbookEntry.objects.get_or_create(date__year=qmyear, 
                        # place=place, 
                        # title="placeholder for survex QM", 
                        # text=message, 
                        # entry_type="DUMMY", 
                        # expedition_id=1, 
                        # defaults={"date": survexblock.date,"cave_slug":caveslug, "slug": logslug})
        # print(insp+message)
        # DataIssue.objects.create(parser='survex', message=message)
  
        try:
            qm = 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,
                                              blockname = blockname, # only set for survex-imported QMs
                                              # found_by = placeholder,
                                              expoyear = str(survexblock.date.year),
                                              cave = survexblock.survexfile.cave)
            qm.save
            # message = " ! QM{} '{}' CREATED in DB in '{}'".format(qm_no, qm_nearest,survexblock.survexfile.path)
            # print(insp+message)
            # 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)
            DataIssue.objects.create(parser='survex', message=message)

    def LoadSurvexDataNormal(self,survexblock,args):
        """Sets the order for data elements in this and following blocks, e.g.
        *data normal from to compass clino tape
        *data normal from to tape compass clino
        We are only collecting length data so we are disinterested in from, to, LRUD etc.
        """
        # datastardefault = { # included here as reference to help understand the code
                        # "type":"normal", 
                        # "t":"leg", 
                        # "from":0, 
                        # "to":1, 
                        # "tape":2, 
                        # "compass":3, 
                        # "clino":4}
        datastar = copy.deepcopy(self.datastardefault)
        if args == "":
            # naked '*data' which is relevant only for passages. Ignore. Continue with previous settings.
            return
        # DEFAULT | NORMAL | CARTESIAN| NOSURVEY |PASSAGE | TOPOFIL | CYLPOLAR | DIVING  
        ls = args.lower().split()   
        if ls[0] == "default":
            self.datastar = copy.deepcopy(self.datastardefault)
        elif ls[0] == "normal" or ls[0] == "topofil":
            if not ("from" in datastar and "to" in datastar):
                message = " ! - Unrecognised *data normal statement '{}' {}|{}".format(args, survexblock.name, survexblock.survexpath)
                print(message)
                print(message,file=sys.stderr)
                DataIssue.objects.create(parser='survex', message=message)
                return
            else:
                datastar = self.datastardefault
                # ls = ["normal", "from", "to", "tape", "compass", "clino" ]
                for i in range(1, len(ls)): # len[0] is "normal"
                    if ls[i] in ["bearing","compass"]:
                        datastar["compass"] = i-1
                    if ls[i] in ["clino","gradient"]:
                        datastar["clino"] = i-1
                    if ls[i] in ["tape","length"]:
                        datastar["tape"] = i-1
                self.datastar = copy.deepcopy(datastar)
                return
        elif ls[0] == "cartesian" or ls[0] == "nosurvey" or ls[0] == "diving" or ls[0] == "cylpolar" or ls[0] == "passage":
            # message = " ! - *data {}  blocks ignored. {}|{}"   '{}' .format(ls[0].upper(), survexblock.name, survexblock.survexpath, args)
            # print(message)
            # print(message,file=sys.stderr)
            # DataIssue.objects.create(parser='survex', message=message)
            self.datastar["type"] = ls[0]
        else:
            message = " ! - Unrecognised *data statement '{}' {}|{}".format(args, survexblock.name, survexblock.survexpath)
            print(message)
            print(message,file=sys.stderr)
            DataIssue.objects.create(parser='survex', message=message)

    def LoadSurvexFlags(self, args):
        # Valid flags are DUPLICATE, SPLAY, and SURFACE, and a flag may be preceded with NOT to turn it off.
        # Default values are NOT any of them
        self.flagsstar = copy.deepcopy(self.flagsdefault)
        flags = []
        
        args = self.rx_flagsnot.sub("not",args)
        argslist = args.split()
        for s in argslist:
            flags.append(s)
        if debugprint:
            print(" ^ flagslist:{}".format(flags),)
        
        if "duplicate" in flags:
            self.flagsstar["duplicate"] = True
        if "surface" in flags:
            self.flagsstar["surface"] = True
        if "splay" in flags:
            self.flagsstar["splay"] = True

        if "notduplicate" in flags:
            self.flagsstar["duplicate"] = False
        if "notsurface" in flags:
            self.flagsstar["surface"] = False
        if "notsplay" in flags:
            self.flagsstar["splay"] = False

        # if self.flagsstar["duplicate"] == True or self.flagsstar["surface"] == True or self.flagsstar["splay"] == True:
        # actually we do want to count duplicates as this is for "effort expended in surveying underground"
        if self.flagsstar["surface"] == True or self.flagsstar["splay"] == True:
            self.flagsstar["skiplegs"] = True
        if debugprint:
            print(" $ flagslist:{}".format(flags),)


    def IdentifyCave(self, cavepath):
        if cavepath.lower() in self.caveslist:
            return self.caveslist[cavepath.lower()]
        # TO DO - some of this is already done in generating self.caveslist so simplify this
        # esp. as it is in a loop.
        # TO DO recognise cave if different name, e.g. gruenstein == 281
        path_match = self.rx_cave.search(cavepath)
        if path_match:
            sluggy = '{}-{}'.format(path_match.group(1), path_match.group(2))
            guesses = [sluggy.lower(), path_match.group(2).lower()]
            for g in guesses:
                if g in self.caveslist:
                    self.caveslist[cavepath] = self.caveslist[g]
                    return self.caveslist[g]
            print('    ! Failed to find cave for {}'.format(cavepath.lower()))
        else:
            # not a cave, but that is fine.
            # print('    ! No regex(standard identifier) cave match for %s' % cavepath.lower())
            return None

    def GetSurvexDirectory(self, headpath):
        """This creates a SurvexDirectory if it has not been seen before, and on creation
        it sets the primarysurvexfile. This is correct as it should be set on the first file
        in the directory, where first is defined by the *include ordering. Which is what we
        are doing.
        """
        if not headpath:
            return self.svxdirs[""]
        if headpath.lower() not in self.svxdirs:
            self.svxdirs[headpath.lower()] = SurvexDirectory(path=headpath, primarysurvexfile=self.currentsurvexfile)
            self.svxdirs[headpath.lower()].save()
            self.survexdict[self.svxdirs[headpath.lower()]] = [] # list of the files in the directory
        return self.svxdirs[headpath.lower()]

    def ReportNonCaveIncludes(self, headpath, includelabel):
        """Ignore surface, kataser and gpx *include survex files
        """
        if headpath in self.ignorenoncave:
            #message = f" - {headpath} is <ignorenoncave> (while creating '{includelabel}' sfile & sdirectory)"
            #print("\n"+message)
            #print("\n"+message,file=sys.stderr)
            return
        for i in self.ignoreprefix:
            if headpath.startswith(i):
                message = f" - {headpath} starts with <ignoreprefix> (while creating '{includelabel}' sfile & sdirectory)"
                #print("\n"+message)
                #print("\n"+message,file=sys.stderr)
                return
        message = f" ! Error: '{headpath}' FAILURE (while creating '{includelabel}' in db - not a cave or in the ignore list of surface surveys. )"
        # getting this triggered for gpx/2018 (cavern error) but not for gpx/2017 (no content).
        print("\n"+message)
        print("\n"+message,file=sys.stderr)
        DataIssue.objects.create(parser='survex', message=message)
        print(f' # datastack in  LoadSurvexFile:{includelabel} type:', end="",file=sys.stderr)
        for dict in self.datastack:
            print(f'{dict["type"].upper()}   ', end="",file=sys.stderr)
        

    def LoadSurvexFile(self, svxid):
        """Creates SurvexFile in the database, and SurvexDirectory if needed
        with links to 'cave'
        Creates a new current survexfile and valid .survexdirectory
        The survexblock passed-in is not necessarily the parent. FIX THIS.
        """
        if debugprint:
            print(" # datastack in  LoadSurvexFile:{} 'type':".format(svxid), end="")
            for dict in self.datastack:
                print("'{}'   ".format(dict["type"].upper()), end="")
            print("")


        depth = " " * self.depthbegin
        # print("{:2}{}   - NEW survexfile:'{}'".format(self.depthbegin, depth, svxid))
        headpath = os.path.dirname(svxid)

        newfile = SurvexFile(path=svxid)
        newfile.save() # until we do this there is no internal id so no foreign key works
        self.currentsurvexfile = newfile 
        newdirectory = self.GetSurvexDirectory(headpath)
        newdirectory.save() 
        newfile.survexdirectory = newdirectory
        self.survexdict[newdirectory].append(newfile)
        cave = self.IdentifyCave(headpath) # cave already exists in db

        if not newdirectory:
            message = " ! 'None' SurvexDirectory returned from GetSurvexDirectory({})".format(headpath)
            print(message)
            print(message,file=sys.stderr)
            DataIssue.objects.create(parser='survex', message=message)

        if cave:
            newdirectory.cave = cave
            newfile.cave   = cave
            # print(f"\n - New directory {newdirectory} for cave {newdirectory.cave}",file=sys.stderr)
        else: # probably a surface survey
            self.ReportNonCaveIncludes(headpath, svxid)
            
        if not newfile.survexdirectory:
            message = " ! SurvexDirectory NOT SET in new SurvexFile {} ".format(svxid)
            print(message)
            print(message,file=sys.stderr)
            DataIssue.objects.create(parser='survex', message=message)
        self.currentsurvexfile.save() # django insists on this although it is already saved !?
        try:
            newdirectory.save()
        except:
            print(newdirectory, file=sys.stderr)
            print(newdirectory.primarysurvexfile, file=sys.stderr)
            raise
            
        if debugprint:
            print(" # datastack end LoadSurvexFile:{} 'type':".format(svxid), end="")
            for dict in self.datastack:
                print("'{}'   ".format(dict["type"].upper()), end="")
            print("")

    def ProcessIncludeLine(self, included):
        global debugprint
        svxid = included.groups()[0]
        if svxid.lower() == debugprinttrigger.lower():
            debugprint = True
        self.LoadSurvexFile(svxid)
        self.stacksvxfiles.append(self.currentsurvexfile)

    def ProcessEdulcniLine(self, edulcni):
        """Saves the current survexfile in the db
        """
        global debugprint
        svxid = edulcni.groups()[0]
        if debugprint:
            depth = " " * self.depthbegin
            print("{:2}{}   - Edulcni  survexfile:'{}'".format(self.depthbegin, depth, svxid))
        if svxid.lower() == debugprinttrigger.lower():
            debugprint = False
        self.currentsurvexfile.save()
        self.currentsurvexfile = self.stacksvxfiles.pop()

    def LoadSurvexComment(self, survexblock, comment):
        # ignore all comments except ;ref, ; wallet and ;QM and ;*include (for collated survex file)
        #    rx_ref2 = re.compile(r'(?i)\s*ref[.;]?')
        #    rx_ref3 = re.compile(r'(?i)\s*wallet[.;]?')
        
        # This should also check that the QM survey point rxists in the block

        refline = self.rx_commref.match(comment)
        if refline:
            #comment = re.sub('(?i)\s*ref[.;]?',"",comment.strip())
            comment = self.rx_ref2.sub("",comment.strip())
            self.LoadSurvexRef(survexblock, comment)
        walletline = self.rx_wallet.match(comment)
        if walletline:
            #comment = re.sub('(?i)\s*wallet[.;]?',"",comment.strip())
            comment = self.rx_ref3.sub("",comment.strip())
            self.LoadSurvexRef(survexblock, comment)
        implicitline = self.rx_implicit.match(comment)
        if implicitline:
            self.LoadSurvexRef(survexblock, comment)

        qml = self.rx_qm0.match(comment)
        if qml:
            qmline = self.rx_qm.match(comment)
            if qmline:
                self.LoadSurvexQM(survexblock, qmline)
            else:
                message = f' ! QM Unrecognised as a valid QM in "{survexblock.survexfile.path}" QM{qml.group(1)} {qml.group(2)}'
                print(message)
                DataIssue.objects.create(parser='survex', message=message, url=f'/survexfile/{survexblock.survexfile.path}.svx')

            
        included = self.rx_comminc.match(comment)
        # ;*include means 'we have been included'; whereas *include means 'proceed to include' 
        if included:
            self.ProcessIncludeLine(included)

        edulcni = self.rx_commcni.match(comment)
        # ;*edulcni means we are returning from an included file
        if edulcni:
            self.ProcessEdulcniLine(edulcni)

    def LoadSurvexSetup(self,survexblock, survexfile):
        self.depthbegin = 0
        self.datastar = self.datastardefault
        blocklegs = self.legsnumber
        print(self.insp+"  - MEM:{:.3f} Reading. parent:{}  <> {} ".format(get_process_memory(),survexblock.survexfile.path, survexfile.path))
        self.lineno = 0
        sys.stderr.flush();
        self.callcount +=1
        if self.callcount % 10 ==0 :
            print(".", file=sys.stderr,end='')
        if self.callcount % 500 ==0 :
            print("\n", 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 = getCaveByReference(pos_cave)
            if cave:
                survexfile.cave = cave

    def LinearLoad(self, survexblock, path, svxlines):
        """Loads a single survex file. Usually used to import all the survex files which have been collated
        into a single file. Loads the begin/end blocks using a stack for labels.
        """
        blkid = None
        pathlist = None
        args = None
        oldflags = None
        blockcount = 0
        self.lineno = 0
        slengthtotal = 0.0
        nlegstotal = 0
        self.relativefilename = path
        cave = self.IdentifyCave(path) # this will produce null for survex files which are geographic collections
        
        self.currentsurvexfile = survexblock.survexfile
        self.currentsurvexfile.save() # django insists on this although it is already saved !?

        self.datastar = copy.deepcopy(self.datastardefault)
        self.flagsstar = copy.deepcopy(self.flagsdefault)

        def tickle():
            nonlocal blockcount
            
            blockcount +=1
            if blockcount % 10 ==0 :
                print(".", file=sys.stderr,end='')
            if blockcount % 200 ==0 :
                print("\n", file=sys.stderr,end='')
                print(" - MEM:{:7.3f} MB in use".format(get_process_memory()),file=sys.stderr)
                print("    ", file=sys.stderr,end='')
            sys.stderr.flush()

        def printbegin():
            nonlocal blkid
            nonlocal pathlist

            depth = " " * self.depthbegin
            self.insp = depth
            if debugprint:
                print("{:2}{}   - Begin for :'{}'".format(self.depthbegin,depth, blkid))
            pathlist = ""
            for id in self.stackbegin:
                if len(id) > 0:
                    pathlist += "." + id

        def printend():
            nonlocal args

            depth = " " * self.depthbegin
            if debugprint:
                print("{:2}{}   - End   from:'{}'".format(self.depthbegin,depth,args))
                print("{:2}{}   - LEGS: {} (n: {}, length:{} units:{})".format(self.depthbegin,
                        depth, self.slength, self.slength, self.legsnumber, self.units))

        def pushblock():
            nonlocal blkid
            if debugprint:
                print(" # datastack at  1 *begin {} 'type':".format(blkid), end="")
                for dict in self.datastack:
                    print("'{}'   ".format(dict["type"].upper()), end="")
                print("")
                print("'{}' self.datastar  ".format(self.datastar["type"].upper()))
            # ------------ * DATA
            self.datastack.append(copy.deepcopy(self.datastar))
            # ------------ * DATA
            if debugprint:
                print(" # datastack at  2 *begin {} 'type':".format(blkid), end="")
                for dict in self.datastack:
                    print("'{}'   ".format(dict["type"].upper()), end="")
                print("")
                print("'{}' self.datastar  ".format(self.datastar["type"].upper()))
            
            # ------------ * FLAGS
            self.flagsstack.append(copy.deepcopy(self.flagsstar))
            # ------------ * FLAGS
            pass

        def popblock():
            nonlocal blkid
            nonlocal oldflags
            if debugprint:
                print(" # datastack  at  *end '{} 'type':".format(blkid), end="")
                for dict in self.datastack:
                    print("'{}'   ".format(dict["type"].upper()), end="")
                print("")
                print("'{}' self.datastar  ".format(self.datastar["type"].upper()))
            # ------------ * DATA
            self.datastar  = copy.deepcopy(self.datastack.pop())
            # ------------ * DATA
            if debugprint:
                print(" # datastack  after *end '{} 'type':".format(blkid), end="")
                for dict in self.datastack:
                    print("'{}'   ".format(dict["type"].upper()), end="")
                print("")
                print("'{}' self.datastar  ".format(self.datastar["type"].upper()))
            
            # ------------ * FLAGS
            self.flagsstar = copy.deepcopy(self.flagsstack.pop()) 
            # ------------ * FLAGS
            if debugprint:
                if oldflags["skiplegs"] != self.flagsstar["skiplegs"]:
                    print(" # POP  'any' flag now:'{}'  was:{} ".format(self.flagsstar["skiplegs"], oldflags["skiplegs"]))

        def starstatement(star):
            nonlocal survexblock
            nonlocal blkid
            nonlocal pathlist
            nonlocal args
            nonlocal oldflags
            nonlocal slengthtotal
            nonlocal nlegstotal

            cmd, args = star.groups()
            cmd = cmd.lower()

            # ------------------------BEGIN
            if self.rx_begin.match(cmd):
                blkid = args.lower()
                # PUSH state ++++++++++++++
                self.stackbegin.append(blkid)
                self.unitsstack.append((self.units, self.unitsfactor))
                self.legsnumberstack.append(self.legsnumber)
                self.slengthstack.append(self.slength)
                self.personexpedstack.append(self.currentpersonexped)
                pushblock()
                # PUSH state ++++++++++++++
                self.legsnumber = 0
                self.slength = 0.0
                self.units = "metres"
                self.currentpersonexped = []
                printbegin()
                newsurvexblock = SurvexBlock(name=blkid, parent=survexblock, 
                        survexpath=pathlist, 
                        cave=self.currentcave, survexfile=self.currentsurvexfile, 
                        legsall=0, legslength=0.0)
                newsurvexblock.save()
                newsurvexblock.title = "("+survexblock.title+")" # copy parent inititally, overwrite if it has its own
                survexblock = newsurvexblock
                survexblock.save() # django insists on this , but we want to save at the end !
                tickle()

            # ---------------------------END
            elif self.rx_end.match(cmd):
                survexblock.legsall = self.legsnumber
                survexblock.legslength = self.slength
                printend()
                slengthtotal += self.slength
                nlegstotal += self.legsnumber

                try:
                    survexblock.parent.save() # django insists on this although it is already saved !?
                except:
                    print(survexblock.parent, file=sys.stderr)
                    raise
                try:
                    survexblock.save() # save to db at end of block
                except:
                    print(survexblock, file=sys.stderr)
                    raise
               # POP  state ++++++++++++++
                popblock()
                self.currentpersonexped = self.personexpedstack.pop()
                self.legsnumber = self.legsnumberstack.pop()
                self.units, self.unitsfactor = self.unitsstack.pop()
                self.slength = self.slengthstack.pop()
                blkid = self.stackbegin.pop()
                self.currentsurvexblock = survexblock.parent
                survexblock = survexblock.parent
                oldflags = self.flagsstar
                self.depthbegin -= 1
                # POP  state ++++++++++++++

            # -----------------------------
            elif self.rx_title.match(cmd):
                quotedtitle = re.match("(?i)^\"(.*)\"$",args)
                if quotedtitle:
                    survexblock.title = quotedtitle.groups()[0]
                else:
                    survexblock.title = args 
            elif self.rx_ref.match(cmd):
                self.LoadSurvexRef(survexblock, args)
            elif self.rx_flags.match(cmd):
                oldflags = self.flagsstar
                self.LoadSurvexFlags(args)
                if debugprint:
                    if oldflags["skiplegs"] != self.flagsstar["skiplegs"]:
                       print(" # CHANGE 'any' flag now:'{}'  was:{} ".format(self.flagsstar["skiplegs"], oldflags["skiplegs"]))

            elif self.rx_data.match(cmd):
                self.LoadSurvexDataNormal(survexblock, args)
            elif self.rx_alias.match(cmd):
                self.LoadSurvexAlias(survexblock, args)
            elif self.rx_entrance.match(cmd):
                self.LoadSurvexEntrance(survexblock, args)
            elif self.rx_date.match(cmd):
                self.LoadSurvexDate(survexblock, args)
            elif self.rx_units.match(cmd):
                self.LoadSurvexUnits(survexblock, args)
            elif self.rx_team.match(cmd):
                self.LoadSurvexTeam(survexblock, args)
            elif self.rx_set.match(cmd) and self.rx_names.match(cmd):
                pass
            elif self.rx_include.match(cmd):
                message = " ! -ERROR *include command not expected here {}. Re-run a full Survex import.".format(path)
                print(message)
                print(message,file=sys.stderr)
                DataIssue.objects.create(parser='survex', message=message)
            else:
                self.LoadSurvexFallThrough(survexblock, args, cmd)


        for svxline in svxlines:
            self.lineno += 1
            sline, comment = self.rx_comment.match(svxline).groups()
            if comment:
                # this catches the ;*include NEWFILE and ;*edulcni ENDOFFILE lines too
                self.LoadSurvexComment(survexblock, comment) 

            if not sline:
                continue # skip blank lines

            # detect a merge failure inserted by version control
            mfail = self.rx_badmerge.match(sline)
            if mfail: 
                message = f"\n ! - ERROR version control merge failure\n   - '{sline}'\n"
                message = message + f"   - line {self.lineno} in {blkid} in {survexblock}\n   - NERD++ needed to fix it"
                print(message)
                print(message,file=sys.stderr)
                DataIssue.objects.create(parser='survex', message=message)
                continue # skip this line

            # detect a star command
            star = self.rx_star.match(sline)
            if star: 
                # yes we are reading a *command
                starstatement(star)
            else: # not a *cmd so we are reading data OR a ";" rx_comment failed. We hope.
                self.LoadSurvexLeg(survexblock, sline, comment)

        self.legsnumber = nlegstotal
        self.slength = slengthtotal  

    def PushdownStackScan(self, survexblock, path, fin, flinear, fcollate):
        """Follows the *include links in all the survex files from the root file 1623.svx
        and reads only the *include and *begin and *end statements. It produces a linearised
        list of the include tree and detects blocks included more than once.
        """
        thissvxline = 0
        indent = " " * self.depthinclude
        sys.stderr.flush();
        self.callcount +=1
        if self.callcount % 10 ==0 :
            print(".", file=sys.stderr,end='')
        if self.callcount % 500 ==0 :
            print("\n    ", file=sys.stderr,end='')

        if path in self.svxfileslist:
            message = " * Warning. Duplicate detected in *include list at callcount:{} depth:{} file:{}".format(self.callcount, self.depthinclude, path)
            print(message)
            print(message,file=flinear)
            print("\n"+message,file=sys.stderr)
            DataIssue.objects.create(parser='survex', message=message)
            if self.svxfileslist.count(path) > 20:
                message = " ! ERROR. Survex file already seen 20x. Probably an infinite loop so fix your *include statements that include this. Aborting. {}".format(path)
                print(message)
                print(message,file=flinear)
                print(message,file=sys.stderr)
                DataIssue.objects.create(parser='survex', message=message)
                return
        self.svxfileslist.append(path)

        try:
            svxlines = fin.read().splitlines()
        except UnicodeDecodeError:
            # some bugger put an umlaut in a non-UTF survex file ?!
            message = f"    ! ERROR *include file '{path}' in '{survexblock}' has UnicodeDecodeError"
            print(message)
            print(message,file=sys.stderr)
            offendingfile = "/survexfile/" + path + ".svx"
            DataIssue.objects.create(parser='survex', message=message,  url=offendingfile)
            return # skip this survex file and all things *included in it
       
        
        for svxline in svxlines:
            self.lineno += 1
            thissvxline += 1
            # detect a merge failure inserted by version control
            mfail = self.rx_badmerge.match(svxline)
            if mfail: 
                message = f"\n!! - ERROR version control merge failure\n   - '{svxline}'\n"
                message = message + f"   - in '{path}' at line {thissvxline}\n"
                message = message + f"   - line {self.lineno}  {survexblock}\n   - Parsing aborted. NERD++ needed to fix it"
                print(message)
                print(message,file=sys.stderr)
                DataIssue.objects.create(parser='survex', message=message)
                return # skip this survex file and all things *included in it

            includestmt =self.rx_include.match(svxline)
            if not includestmt:
                fcollate.write("{}\n".format(svxline.strip()))

            sline, comment = self.rx_comment.match(svxline.strip()).groups()
            star = self.rx_star.match(sline)
            if star: # yes we are reading a *cmd
                cmd, args = star.groups()
                cmd = cmd.lower()
                if re.match("(?i)include$", cmd):
                    includepath = os.path.normpath(os.path.join(os.path.split(path)[0], re.sub(r"\.svx$", "", args)))

                    fullpath = os.path.join(settings.SURVEX_DATA, includepath + ".svx")
                    self.RunSurvexIfNeeded(os.path.join(settings.SURVEX_DATA, includepath))
                    self.checkUniqueness(os.path.join(settings.SURVEX_DATA, includepath))
                    if os.path.isfile(fullpath):
                        #--------------------------------------------------------
                        self.depthinclude += 1
                        fininclude = open(fullpath,'r')
                        fcollate.write(";*include {}\n".format(includepath))
                        flinear.write("{:2} {} *include {}\n".format(self.depthinclude, indent, includepath))
                        push = includepath.lower()
                        self.includestack.append(push)
                        #-----------------
                        self.PushdownStackScan(survexblock, includepath, fininclude, flinear, fcollate)
                        #-----------------
                        pop = self.includestack.pop()
                        if pop != push:
                            message = "!! ERROR mismatch *include pop!=push  {}".format(pop, push, self.includestack)
                            print(message)
                            print(message,file=flinear)
                            print(message,file=sys.stderr)
                            DataIssue.objects.create(parser='survex', message=message)
                        flinear.write("{:2} {} *edulcni {}\n".format(self.depthinclude, indent, pop))
                        fcollate.write(";*edulcni {}\n".format(pop))
                        fininclude.close()
                        self.depthinclude -= 1
                        #--------------------------------------------------------
                    else:
                        message = "    ! ERROR *include file not found for:'{}'".format(includepath)
                        print(message)
                        print(message,file=sys.stderr)
                        DataIssue.objects.create(parser='survex', message=message)
                elif re.match("(?i)begin$", cmd):
                    self.depthbegin += 1
                    depth = " " * self.depthbegin
                    if args:
                        pushargs = args
                    else:
                        pushargs = " "
                    self.stackbegin.append(pushargs.lower())
                    flinear.write("            {:2} {} *begin {}\n".format(self.depthbegin, depth, args))
                    pass
                elif re.match("(?i)end$", cmd):
                    depth = " " * self.depthbegin
                    flinear.write("            {:2} {} *end   {}\n".format(self.depthbegin, depth, args))
                    if not args:
                        args = " "
                    popargs = self.stackbegin.pop()
                    if popargs != args.lower():
                        message = "!! ERROR mismatch in BEGIN/END labels pop!=push '{}'!='{}'\n{}".format(popargs, args, self. stackbegin)
                        print(message)
                        print(message,file=flinear)
                        print(message,file=sys.stderr)
                        DataIssue.objects.create(parser='survex', message=message)

                    self.depthbegin -= 1
                    pass
                elif re.match("(?i)title$", cmd):
                    depth = " " * self.depthbegin
                    flinear.write("                    {:2} {} *title {}\n".format(self.depthbegin, depth, args))
                    pass


    def checkUniqueness(self,fullpath):
        fn = Path(fullpath).name
        if fn not in self.uniquename:
            self.uniquename[fn] = [fullpath]
        else:
            self.uniquename[fn].append(fullpath)
            # This is not an error now that we are moving .3d files to the :loser: directory tree
            # message = f" ! NON-UNIQUE survex filename, '{fn}' - '{self.uniquename[fn]}' #{len(self.uniquename[fn])}"
            # print(message)
            # DataIssue.objects.create(parser='survex', message=message)
            message = f" NOTE:  non-unique survex filename, '{fn}' - '{self.uniquename[fn]}' #{len(self.uniquename[fn])}"
            print(message)

    
    def RunSurvexIfNeeded(self,fullpath):
        now = time.time()
        cav_t = now - 365*24*3600
        log_t = now - 365*24*3600
        svx_t = now - 365*24*3600

        def runcavern():
            '''regenerates the .3d file from the .svx if it is older than the svx file, or older than the software,
            or randomly using chaosmonkey() just to keep things ticking over.
            '''
            print(f" -  Regenerating stale (or chaos-monkeyed) cavern .log and .3d for '{fullpath}'\n     at '{logpath}'\n")
            print(f"days svx old: {(svx_t - log_t)/(24*3600):.1f}  cav:{(cav_t - log_t)/(24*3600):.1f}   log old: { (now - log_t)/(24*3600):.1f}")
            
            outputdir = Path(str(f'{fullpath}.svx')).parent
            sp = subprocess.run([settings.CAVERN, "--log", f'--output={outputdir}', f'{fullpath}.svx'],
                                    capture_output=True, check=False, text=True)
            if sp.returncode != 0:
                message = f' ! Error running {settings.CAVERN}: {fullpath}'
                url = f'/survexfile{fullpath}.svx'.replace(settings.SURVEX_DATA, "")
                DataIssue.objects.create(parser='xEntrances', message=message, url=url)
                print(message)
                print(f'stderr:\n\n' + str(sp.stderr) + '\n\n' + str(sp.stdout)  + '\n\nreturn code: ' + str(sp.returncode))
            self.caverncount += 1
            
            # should also collect all the .err files too and create a DataIssue for each one which 
            # - is nonzero in size AND
            # - has Error greater than 5% anywhere, or some other more serious error
            
            errpath = Path(fullpath + ".err")
            if errpath.is_file():
                if errpath.stat().st_size == 0:
                    errpath.unlink() # delete empty closure error file


        svxpath = Path(fullpath + ".svx")
        logpath = Path(fullpath + ".log")
        outputdir = Path(svxpath).parent

        if not logpath.is_file(): # always run if logfile not there
            runcavern()
            return

        self.caverndate = now - 2*365*24*3600

        if not self.caverndate:
            sp = subprocess.run(["which", "{}".format(settings.CAVERN)], 
                                    capture_output=True, check=False, text=True)
            if sp.returncode != 0:
                message = f' ! Error running "which" on {settings.CAVERN}' 
                DataIssue.objects.create(parser='entrances', message=message)
                print(message)
                print(f'stderr:\n\n' + str(sp.stderr) + '\n\n' + str(sp.stdout)  + '\n\nreturn code: ' + str(sp.returncode))
                self.caverndate = os.path.getmtime(sp.stdout.strip())
            else:
                self.caverndate = now - 2*365*24*3600
        cav_t = self.caverndate
        log_t = os.path.getmtime(logpath)
        svx_t = os.path.getmtime(svxpath)
        now = time.time()

        if svx_t - log_t > 0:          # stale, svx file is newer than log
            runcavern()
            return
        if now - log_t > 60 *24*60*60: # >60 days, re-run anyway
            runcavern()
            return
        if cav_t - log_t > 0:          # new version of cavern
            runcavern()
            return
        if chaosmonkey(350):           # one in every 350 runs
            runcavern()

def FindAndLoadSurvex(survexblockroot):
    """Follows the *include links successively to find files in the whole include tree
    """
    print('  - redirecting stdout to svxblks.log...')
    stdout_orig = sys.stdout
    # Redirect sys.stdout to the file
    sys.stdout = open('svxblks.log', 'w')

    print('  - Scanning Survex Blocks tree from {}.svx ...'.format(settings.SURVEX_TOPNAME),file=sys.stderr)
    survexfileroot = survexblockroot.survexfile # i.e. SURVEX_TOPNAME only
    collatefilename = "_" + survexfileroot.path + ".svx"

    svx_scan = LoadingSurvex()
    svx_scan.callcount = 0
    svx_scan.depthinclude = 0
    fullpathtotop = os.path.join(survexfileroot.survexdirectory.path, survexfileroot.path)
    
    # Rather than do this check for the presence of the .log and .3d files synchronously here,
    # we should instead run this in a separate thread asynchronously.
    print("  - RunSurvexIfNeeded cavern on '{}'".format(fullpathtotop), file=sys.stderr)
    svx_scan.RunSurvexIfNeeded(fullpathtotop)
    svx_scan.checkUniqueness(fullpathtotop)
    
    indent=""
    fcollate = open(collatefilename, 'w')

    mem0 = get_process_memory()
    print(" - MEM:{:7.2f} MB START".format(mem0),file=sys.stderr)
    flinear = open('svxlinear.log', 'w')
    flinear.write("    - MEM:{:7.2f} MB START {}\n".format(mem0,survexfileroot.path))
    print("    ", file=sys.stderr,end='')

    finroot = survexfileroot.OpenFile()
    fcollate.write(";*include {}\n".format(survexfileroot.path))
    flinear.write("{:2} {} *include {}\n".format(svx_scan.depthinclude, indent, survexfileroot.path))

    import cProfile, pstats
    from pstats import SortKey
    pr = cProfile.Profile()
    pr.enable()
    #----------------------------------------------------------------
    svx_scan.PushdownStackScan(survexblockroot, survexfileroot.path, finroot, flinear, fcollate)
    #----------------------------------------------------------------
    pr.disable()
    with open('PushdownStackScan.prof', 'w') as f:
        ps = pstats.Stats(pr, stream=f)
        ps.sort_stats(SortKey.CUMULATIVE)
        ps.print_stats()
        
    flinear.write("{:2} {} *edulcni {}\n".format(svx_scan.depthinclude, indent, survexfileroot.path))
    fcollate.write(";*edulcni {}\n".format(survexfileroot.path))
    mem1 = get_process_memory()
    flinear.write("\n    - MEM:{:.2f} MB STOP {}\n".format(mem1,survexfileroot.path))
    flinear.write("    - MEM:{:.3f} MB USED\n".format(mem1-mem0))
    svxfileslist = svx_scan.svxfileslist
    flinear.write("    - {:,} survex files in linear include list \n".format(len(svxfileslist)))
    flinear.close()
    fcollate.close()
    print("\n -  {:,} runs of survex 'cavern' refreshing .3d files \n".format(svx_scan.caverncount),file=sys.stderr)
    svx_scan = None # Hmm. Does this actually delete all the instance variables if they are lists, dicts etc.? 
    print("\n -  {:,} survex files in linear include list \n".format(len(svxfileslist)),file=sys.stderr)


    mem1 = get_process_memory()
    print(" - MEM:{:7.2f} MB END ".format(mem0),file=sys.stderr)
    print(" - MEM:{:7.3f} MB USED".format(mem1-mem0),file=sys.stderr)
    svxfileslist = [] # free memory
 
    # Before doing this, it would be good to identify the *equate and *entrance we need that are relevant to the
    # entrance locations currently loaded after this by LoadPos(), but could better be done before ?
    # look in MapLocations() for how we find the entrances
   
    print('\n  - Loading All Survex Blocks (LinearLoad)',file=sys.stderr)
    svx_load = LoadingSurvex()

    svx_load.survexdict[survexfileroot.survexdirectory] = []
    svx_load.survexdict[survexfileroot.survexdirectory].append(survexfileroot)
    svx_load.svxdirs[""] = survexfileroot.survexdirectory

    # This next should be rewritten to use a generator so that only one
    # line is held in memory at a time:
    with open(collatefilename, "r") as fcollate:
            svxlines = fcollate.read().splitlines()
    #pr2 = cProfile.Profile()
    #pr2.enable()
    print("    ", file=sys.stderr,end='')
    #----------------------------------------------------------------
    svx_load.LinearLoad(survexblockroot,survexfileroot.path, svxlines)
    #----------------------------------------------------------------
    #pr2.disable()
    # with open('LinearLoad.prof', 'w') as f:
        # ps = pstats.Stats(pr2, stream=f)
        # ps.sort_stats(SortKey.CUMULATIVE)
        # ps.print_stats()

    print("\n - MEM:{:7.2f} MB STOP".format(mem1),file=sys.stderr)
    print(" - MEM:{:7.3f} MB USED".format(mem1-mem0),file=sys.stderr)

    # 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

    legsnumber = svx_load.legsnumber
    mem1 = get_process_memory()

    print("  - Number of SurvexDirectories: {}".format(len(svx_load.survexdict)))
    tf=0
    for d in svx_load.survexdict:
        tf += len(svx_load.survexdict[d])
    print("  - Number of SurvexFiles: {}".format(tf))
    print(f"  - Number of Survex legs: {legsnumber}")
    svx_load = None

    return legsnumber

def MakeSurvexFileRoot():
    """Returns a file_object.path = SURVEX_TOPNAME associated with directory_object.path = SURVEX_DATA
    """
    # find a cave, any cave..
    c = Cave.objects.all()
    
    fileroot = SurvexFile(path=settings.SURVEX_TOPNAME, cave=None)
    fileroot.save()
    directoryroot = SurvexDirectory(path=settings.SURVEX_DATA, cave=c[0], primarysurvexfile=fileroot)
    # MariaDB doesn't like this hack. Complains about non-null cave_id EVEN THOUGH our model file says this is OK:
    # cave = models.ForeignKey('Cave', blank=True, null=True,on_delete=models.SET_NULL)
    directoryroot.save()
    fileroot.survexdirectory = directoryroot # i.e. SURVEX_DATA/SURVEX_TOPNAME
    fileroot.save() # mutually dependent objects need a double-save like this
    return fileroot

def LoadSurvexBlocks():

    print(' - Flushing All Survex Blocks...')
    SurvexBlock.objects.all().delete()
    SurvexFile.objects.all().delete()
    SurvexDirectory.objects.all().delete()
    SurvexPersonRole.objects.all().delete()
    SurvexStation.objects.all().delete()
    print(" - survex Data Issues flushed")
    DataIssue.objects.filter(parser='survex').delete()
    DataIssue.objects.filter(parser='survexleg').delete()
    DataIssue.objects.filter(parser='survexunits').delete()
    DataIssue.objects.filter(parser='entrances').delete()
    DataIssue.objects.filter(parser='xEntrances').delete()

    survexfileroot = MakeSurvexFileRoot()
    # this next makes a block_object assciated with a file_object.path = SURVEX_TOPNAME
    survexblockroot = SurvexBlock(name=ROOTBLOCK, survexpath="", cave=None, survexfile=survexfileroot, 
            legsall=0, legslength=0.0)
    survexblockroot.save()

    print(' - Loading Survex Blocks...')
    memstart = get_process_memory()
    #----------------------------------------------------------------
    FindAndLoadSurvex(survexblockroot)
    #----------------------------------------------------------------
    memend = get_process_memory()
    print(" - MEMORY start:{:.3f} MB end:{:.3f} MB increase={:.3f} MB".format(memstart,memend, memend-memstart))
    
    survexblockroot.save()

    print(' - Loaded All Survex Blocks.')

poslineregex = re.compile(r"^\(\s*([+-]?\d*\.\d*),\s*([+-]?\d*\.\d*),\s*([+-]?\d*\.\d*)\s*\)\s*([^\s]+)$")

def LoadPositions():
    """First load the survex stations for entrances and fixed points (about 600) into the database.
    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. This gives us coordinates of the entrances.
    If we don't have it in the database, print an error message and discard it.
    """
    svx_t = 0
    d3d_t = 0
    def runcavern3d():
        outputdir = Path(str(f'{topdata}.svx')).parent

        # print(" -  Regenerating stale cavern .log and .3d for '{}'\n    days old: {:.1f}    {:.1f}    {:.1f}".
        #    format(topdata, (svx_t - d3d_t)/(24*3600), (cav_t - d3d_t)/(24*3600), (now - d3d_t)/(24*3600)))

        file3d = Path(f'{topdata}.3d')
        try:
            sp = subprocess.run([settings.CAVERN, "--log", f"--output={outputdir}", f"{topdata}.svx"], 
                                    capture_output=True, check=False, text=True) #check=False means exception not raised
            if sp.returncode != 0:
                message = f' ! Error: cavern: creating {file3d} in runcavern3()'
                DataIssue.objects.create(parser='survex', message=message)
                print(message)
                
                # find the errors in the 1623.log file
                sp = subprocess.run(["grep", "error:", f"{topdata}.log"], 
                                        capture_output=True, check=False, text=True) #check=False means exception not raised
                message = f' ! Error: cavern: {sp.stdout}'
                DataIssue.objects.create(parser='survex', message=message)
                print(message)

        except: 
            message = " ! CalledProcessError 'cavern' in runcavern3() at {topdata}."
            DataIssue.objects.create(parser='survex', message=message)
            print(message)
            
            if file3d.is_file():
                message = " ! CalledProcessError. File permissions {file3d.stat().st_mode} on {str(file3d)}"
                DataIssue.objects.create(parser='survex', message=message)
                print(message)
                
        if file3d.is_file(): # might be an old one though
            try:
                # print(" -  Regenerating {} {}.3d  in  {}".format(settings.SURVEXPORT, topdata, settings.SURVEX_DATA))
                sp = subprocess.run([settings.SURVEXPORT, '--pos', f'{file3d}'], cwd = settings.SURVEX_DATA, 
                                        capture_output=True, check=False, text=True)
                if sp.returncode != 0:
                    print(f' ! Error: survexport creating {topdata}.pos in runcavern3().\n\n' + str(sp.stdout)  + '\n\nreturn code: ' + str(sp.returncode))
            except: 
                message = " ! CalledProcessError 'survexport' in runcavern3() at {file3d}." 
                DataIssue.objects.create(parser='entrances', message=message)
                print(message)
        else:
            message = f" ! Failed to find {file3d} so aborting generation of new .pos, using old one if present"
            DataIssue.objects.create(parser='entrances', message=message)
            print(message)

    topdata = os.fspath(Path(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

    completed_process = subprocess.run(["which", "{}".format(settings.CAVERN)], 
                            capture_output=True, check=True, text=True)
    cav_t = os.path.getmtime(completed_process.stdout.strip())

    svxpath = topdata + ".svx"
    d3dpath = topdata + ".3d"
    pospath = topdata + ".pos"

    svx_t = os.path.getmtime(svxpath)

    if os.path.isfile(d3dpath):
        # always fails to find log file if a double directory, e.g. caves-1623/B4/B4/B4.svx Why ?
        d3d_t = os.path.getmtime(d3dpath)

    now = time.time()
    if not os.path.isfile(pospath):
        runcavern3d()
    if not os.path.isfile(d3dpath):
        runcavern3d()
    elif d3d_t - svx_t > 0:          # stale, 3d older than svx file
        runcavern3d()
    elif now - d3d_t> 60 *24*60*60: # >60 days old, re-run anyway
        runcavern3d()
    elif cav_t - d3d_t > 0:          # new version of cavern
        runcavern3d()

    mappoints = {}
    for pt in MapLocations().points():
        svxid, number,  point_type, label = pt
        mappoints[svxid]=True

    if not Path(pospath).is_file():
        message = f" ! Failed to find {pospath} so aborting generation of entrance locations. "
        DataIssue.objects.create(parser='entrances', message=message)
        print(message)
        return

    posfile = open(pospath)
    posfile.readline() #Drop header
    try:
        survexblockroot = SurvexBlock.objects.get(name=ROOTBLOCK)
    except:
        try:
            survexblockroot = SurvexBlock.objects.get(id=1)
        except:
            message = ' ! FAILED to find root SurvexBlock'
            print(message)
            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(".")
                    # But why are we doing this? Why do we need the survexblock id for each of these ?
                    # ..because mostly they don't actually appear in any SVX file. We should match them up
                    # via the cave data, not by this half-arsed syntactic match which almost never works. PMS.
                    if False:
                        try:
                            sbqs = SurvexBlock.objects.filter(survexpath=blockpath)
                            if len(sbqs)==1:
                                sb = sbqs[0]
                            if len(sbqs)>1:
                                message = " ! MULTIPLE SurvexBlocks {:3} matching Entrance point {} {} '{}'".format(len(sbqs), blockpath, sid, id)
                                print(message)
                                DataIssue.objects.create(parser='survex', message=message)
                                sb = sbqs[0]
                            elif len(sbqs)<=0:
                                message = " ! ZERO SurvexBlocks matching Entrance point {} {} '{}'".format(blockpath, sid, id)
                                print(message)
                                DataIssue.objects.create(parser='survex', message=message)
                                sb = survexblockroot
                        except:
                            message = ' ! FAIL in getting SurvexBlock matching Entrance point {} {}'.format(blockpath, sid)
                            print(message)
                            DataIssue.objects.create(parser='survex', message=message)
                    try:
                        ss = SurvexStation(name=id, block=survexblockroot)
                        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)
                        DataIssue.objects.create(parser='survex', message=message)
                        raise
    print(" -  {} SurvexStation entrances found.".format(found))