troggle-unchained/parsers/survex.py

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.parsers.logbooks import GetCaveLookup
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 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 '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'(?i)^\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_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_ref     = re.compile(r'(?i)^\s*ref[\s.:]*(\d+)\s*#\s*(X)?\s*(\d+)')
    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+.*?)$')

    survexlegsalllength = 0.0
    survexlegsnumber = 0
    depthbegin = 0
    depthinclude = 0
    stackbegin =[]
    stackinclude = []
    stacksvxfiles = []
    svxfileslist = []
    svxdirs = {}
    survexdict = {} # each key is a directory, and its value is a list of files
    lineno = 0
    insp = ""
    callcount = 0
    stardata ={}
    ignoreprefix = ["surface", "kataster", "fixedpts", "gpx"]
    ignorenoncave = ["caves-1623", "caves-1623/2007-neu"]
    includedfilename =""
    currentsurvexblock = None
    currentsurvexfile = None
    currentcave = None

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

    def LoadSurvexIgnore(self, survexblock, line, cmd):
        if 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.save()
                        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.
        """
        # Check first to see if we are in a splay and abort if so.
        # TO DO splay abort
        
        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: lcompass:'{}' line {} in '{}'".format(lcompass, 
                            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 LoadSurvexDataCmd(self,survexblock,args):
        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

    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 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.
        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:
            print('    ! No regex 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()] = models_survex.SurvexDirectory(path=headpath, primarysurvexfile=self.currentsurvexfile)
            self.svxdirs[headpath.lower()].save()
        return self.svxdirs[headpath.lower()]

    def ReportNonCaveIncludes(self, headpath, includelabel):
        """Ignore surface, kataser and gps *include survex files
        """
        if headpath in self.ignorenoncave:
            return
        for i in self.ignoreprefix:
            if headpath.startswith(i):
                return
        message = " ! {} is not a cave. (while creating '{}' sfile & sdirectory)".format(headpath, includelabel)
        print("\n"+message)
        print("\n"+message,file=sys.stderr)
        models.DataIssue.objects.create(parser='survex', message=message)
        
    def LoadSurvexFile(self, includelabel):
        """Creates SurvexFile in the database, and SurvexDirectory if needed
        with links to 'cave'
        Creates a new current survexblock with valid .survexfile and valid .survexdirectory
        The survexblock passed-in is not necessarily the parent. FIX THIS.
        """
        depth = " " * self.depthbegin
        print("{:2}{}   - NEW survexfile:'{}'".format(self.depthbegin, depth, includelabel))
        headpath, tail = os.path.split(includelabel)

        newfile = models_survex.SurvexFile(path=includelabel)
        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] = [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)
            models.DataIssue.objects.create(parser='survex', message=message)

        if cave:
            newdirectory.cave = cave
            newfile.cave   = cave
            #print("\n"+str(newdirectory.cave),file=sys.stderr)
        else:
            self.ReportNonCaveIncludes(headpath, includelabel)

        if not newfile.survexdirectory:
            message = " ! SurvexDirectory NOT SET in new SurvexFile {} ".format(includelabel)
            print(message)
            print(message,file=sys.stderr)
            models.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

    def ProcessIncludeLine(self, included):
        svxid = included.groups()[0]
        #depth = " " * self.depthbegin
        #print("{:2}{}   - Include survexfile:'{}'".format(self.depthbegin, depth,  svxid))
        self.LoadSurvexFile(svxid)
        self.stacksvxfiles.append(self.currentsurvexfile)

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

    def LoadSurvexComment(self, survexblock, comment):
        # ignore all comments except ;ref and ;QM and ;*include (for collated survex file)
        refline = self.rx_ref.match(comment)
        if refline:
            comment = re.sub('(?i)\s*ref[.;]?',"",comment.strip())
            self.LoadSurvexRef(survexblock, comment)

        qmline = self.rx_qm.match(comment)
        if qmline:
            self.LoadSurvexQM(survexblock, qmline)
            
        included = self.rx_comminc.match(comment)
        # ;*include means we have been included; not 'proceed to include' which *include means
        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.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 ==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 = models_caves.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 recursively.
        """
        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 !?
        
        blockcount = 0
        def tickle():
            nonlocal blockcount
            blockcount +=1
            if blockcount % 10 ==0 :
                print(".", file=sys.stderr,end='')
            if blockcount % 500 ==0 :
                print("\n", file=sys.stderr,end='')
            sys.stderr.flush();

        for svxline in svxlines:
            sline, comment = self.rx_comment.match(svxline.strip()).groups()
            if comment:
                self.LoadSurvexComment(survexblock, comment) # this catches the ;*include and ;*edulcni lines too
            if not sline:
                continue # skip blank lines

            # detect a star command
            mstar = self.rx_star.match(sline)
            if mstar: # yes we are reading a *cmd
                cmd, args = mstar.groups()
                cmd = cmd.lower()

                # ------------------------BEGIN
                if re.match("begin$(?i)", cmd):
                    self.depthbegin += 1
                    depth = " " * self.depthbegin
                    blockid = args.lower()
                    self.stackbegin.append(blockid)

                    previousnlegs = self.survexlegsnumber
                    print("{:2}{}   - Begin for :'{}'".format(self.depthbegin,depth, blockid))
                    pathlist = ""
                    for id in self.stackbegin:
                        if len(id) > 0:
                            pathlist += "." + id
                    newsurvexblock = models_survex.SurvexBlock(name=blockid, parent=survexblock, 
                            survexpath=pathlist, 
                            title = survexblock.title, # copy parent inititally
                            cave=self.currentcave, survexfile=self.currentsurvexfile, 
                            legsall=0, legssplay=0, legssurfc=0, totalleglength=0.0)
                    newsurvexblock.save()
                    survexblock = newsurvexblock
                    # survexblock.survexfile.save() 
                    survexblock.save() # django insists on this , but we want to save at the end !
                    tickle()

                # ---------------------------END
                elif re.match("end$(?i)", cmd):
                    depth = " " * self.depthbegin

                    print("{:2}{}   - End   from:'{}'".format(self.depthbegin,depth,args))
                    legsinblock = self.survexlegsnumber - previousnlegs
                    print("{:2}{}   - LEGS: {} (previous: {}, now:{})".format(self.depthbegin,
                        depth,legsinblock,previousnlegs,self.survexlegsnumber))
                    survexblock.legsall = legsinblock
                    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
                    self.currentsurvexblock = survexblock.parent
                    survexblock = survexblock.parent
                    blockid = self.stackbegin.pop()
                    self.depthbegin -= 1

                # -----------------------------
                elif re.match("(?i)title$", cmd):
                    survexblock.title = args # block has own title, overwrite that from parent
                elif re.match("(?i)ref$", cmd):
                    self.LoadSurvexRef(survexblock, args)
                elif re.match("(?i)flags$", cmd):
                    self.LoadSurvexFlags(args, cmd)
                elif re.match("(?i)data$", cmd):
                    self.LoadSurvexDataCmd(survexblock, args)
                elif re.match("(?i)date$", cmd):
                    self.LoadSurvexDate(survexblock, args)
                elif re.match("(?i)team$", cmd):
                    self.LoadSurvexTeam(survexblock, args)
                elif re.match("(?i)set$", cmd) and re.match("(?i)names", args):
                    pass
                elif re.match("(?i)include$", cmd):
                    message = " ! -ERROR *include command not expected here {}. Re-run a full Survex import.".format(path)
                    print(message)
                    print(message,file=sys.stderr)
                    models.DataIssue.objects.create(parser='survex', message=message)
                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 RecursiveScan(self, survexblock, survexfile, 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
        """
        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 survexfile in self.svxfileslist:
            message = " * Warning. Survex file already seen: {}".format(survexfile.path)
            print(message)
            print(message,file=flinear)
            print(message,file=sys.stderr)
            models.DataIssue.objects.create(parser='survex', message=message)
            if self.svxfileslist.count(survexfile) > 20:
                message = " ! ERROR. Survex file already seen 20x. Probably an infinite loop so fix your *include statements that include this. Aborting. {}".format(survexfile.path)
                print(message)
                print(message,file=flinear)
                print(message,file=sys.stderr)
                models.DataIssue.objects.create(parser='survex', message=message)
                return
        self.svxfileslist.append(survexfile)
        
        svxlines = fin.read().splitlines()
        for svxline in svxlines:
            self.lineno += 1
            includestmt =self.rx_include.match(svxline)
            if not includestmt:
                fcollate.write("{}\n".format(svxline))

            sline, comment = self.rx_comment.match(svxline.strip()).groups()
            mstar = self.rx_star.match(sline)
            if mstar: # yes we are reading a *cmd
                cmd, args = mstar.groups()
                cmd = cmd.lower()
                if re.match("(?i)include$", cmd):
                    includepath = os.path.normpath(os.path.join(os.path.split(survexfile.path)[0], re.sub(r"\.svx$", "", args)))
                    path_match = re.search(r"caves-(\d\d\d\d)/(\d+|\d\d\d\d-?\w+-\d+)/", includepath)

                    includesurvexfile = models_survex.SurvexFile(path=includepath)
                    includesurvexfile.save()

                    if includesurvexfile.exists():
                        # do not create SurvexFile in DB here by doing includesurvexfile.save(). Do it when reading data.
                        #--------------------------------------------------------
                        self.depthinclude += 1
                        fininclude = includesurvexfile.OpenFile()
                        fcollate.write(";*include {}\n".format(includesurvexfile.path))
                        flinear.write("{:2} {} *include {}\n".format(self.depthinclude, indent, includesurvexfile.path))
                        push = includesurvexfile.path.lower()
                        self.stackinclude.append(push)
                        #-----------------
                        self.RecursiveScan(survexblock, includesurvexfile, fininclude, flinear, fcollate)
                        #-----------------
                        pop = self.stackinclude.pop()
                        if pop != push:
                            message = "!!!!!!!    ERROR pop != push {} != {} {}".format(pop, push, self.stackinclude)
                            print(message)
                            print(message,file=flinear)
                            print(message,file=sys.stderr)
                            models.DataIssue.objects.create(parser='survex', message=message)
                        includesurvexfile.path += "-TEMP"
                        includesurvexfile = None
                        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(includesurvexfile, includepath)
                        print(message)
                        print(message,file=sys.stderr)
                        models.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 BEGIN/END pop != push {} != {}\n{}".format(popargs, args, self. stackbegin)
                        print(message)
                        print(message,file=flinear)
                        print(message,file=sys.stderr)
                        models.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 FindAndLoadSurvex(survexblockroot):
    """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')

    print(' - SCANNING All Survex Blocks...',file=sys.stderr)
    survexfileroot = survexblockroot.survexfile

    collatefilename = "_" + survexfileroot.path + ".svx"

    svx_scan = LoadingSurvex()
    svx_scan.callcount = 0
    svx_scan.depthinclude = 0
    indent=""
    fcollate = open(collatefilename, 'w')

    mem0 = models.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))

    finroot = survexfileroot.OpenFile()
    fcollate.write(";*include {}\n".format(survexfileroot.path))
    flinear.write("{:2} {} *include {}\n".format(svx_scan.depthinclude, indent, survexfileroot.path))
    #----------------------------------------------------------------
    svx_scan.RecursiveScan(survexblockroot, survexfileroot, finroot, flinear, fcollate)
    #----------------------------------------------------------------
    flinear.write("{:2} {} *edulcni {}\n".format(svx_scan.depthinclude, indent, survexfileroot.path))
    fcollate.write(";*edulcni {}\n".format(survexfileroot.path))
    mem1 = models.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()
    svx_scan = None
    print("\n -  {:,} survex files in linear include list \n".format(len(svxfileslist)),file=sys.stderr)

    mem1 = models.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 (LinearRecursive)',file=sys.stderr)
    svx_load = LoadingSurvex()
    
    svx_load.svxdirs[""] = survexfileroot.survexdirectory
    with open(collatefilename, "r") as fcollate:
            svxlines = fcollate.read().splitlines()
    #----------------------------------------------------------------
    svx_load.LinearLoad(survexblockroot,survexfileroot.path, svxlines)
    #----------------------------------------------------------------

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

    survexlegsnumber = svx_load.survexlegsnumber
    survexlegsalllength = svx_load.survexlegsalllength
    mem1 = models.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))
    svx_load = None

    # 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 MakeSurvexFileRoot():
    survexfileroot = models_survex.SurvexFile(path=settings.SURVEX_TOPNAME, cave=None)
    survexfileroot.save()
    survexdirectoryroot = models_survex.SurvexDirectory(path=settings.EXPOWEB, cave=None, primarysurvexfile=survexfileroot)
    survexdirectoryroot.save()
    survexfileroot.survexdirectory = survexdirectoryroot
    survexfileroot.save() # mutually dependent objects need a double-save like this

    return survexfileroot

def LoadSurvexBlocks():

    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 = MakeSurvexFileRoot()
    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 Survex Blocks...')
    memstart = models.get_process_memory()
    survexlegsnumber, survexlegsalllength = FindAndLoadSurvex(survexblockroot)
    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: {}".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():
    """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.
    """
    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(".")
                    # 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 = models_survex.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)
                                models.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)
                                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=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)
                        models.DataIssue.objects.create(parser='survex', message=message)
                        raise
    print(" - {} SurvexStation entrances found.".format(found))