troggle/parsers/survex.py

from __future__ import absolute_import, division, print_function

import os
import re
import sys
import time
from datetime import datetime, timedelta
from subprocess import PIPE, Popen, call

from django.utils.timezone import get_current_timezone, make_aware

import troggle.settings as settings
import troggle.core.models as models
from troggle.parsers.people import GetPersonExpeditionNameLookup

"""A 'survex block' is a *begin...*end set of cave data.
A 'survexscansfolder' is what we today call a "survey scans folder" or a "wallet".
"""

line_leg_regex = re.compile(r"[\d\-+.]+$")

def LoadSurvexLineLeg(survexblock, stardata, sline, comment, cave):
    # The try catches here need replacing as they are relatively expensive
    ls = sline.lower().split()
    ssfrom = survexblock.MakeSurvexStation(ls[stardata["from"]])
    ssto = survexblock.MakeSurvexStation(ls[stardata["to"]])

    survexleg = models.SurvexLeg(block=survexblock, stationfrom=ssfrom, stationto=ssto)
    # 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":
        try:
            survexleg.tape = float(ls[stardata["tape"]])
        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 = 1000
        try:
            lclino = ls[stardata["clino"]]
        except:
            print("! Clino misread in", survexblock.survexfile.path)
            print("  Stardata:", stardata)
            print("  Line:", ls)
            message = ' ! Value Error: Clino misread in line %s in %s' % (ls, survexblock.survexfile.path)
            models.DataIssue.objects.create(parser='survex', message=message)
            lclino = error
        try:
            lcompass = ls[stardata["compass"]]
        except:
            print("! Compass misread in", survexblock.survexfile.path)
            print("  Stardata:", stardata)
            print("  Line:", ls)
            message = ' ! Value Error: Compass misread in line %s in %s' % (ls, survexblock.survexfile.path)
            models.DataIssue.objects.create(parser='survex', message=message)
            lcompass = error
        if lclino == "up":
            survexleg.compass = 0.0
            survexleg.clino = 90.0
        elif lclino == "down":
            survexleg.compass = 0.0
            survexleg.clino = -90.0
        elif lclino == "-" or lclino == "level":
            try:
                survexleg.compass = float(lcompass)
            except ValueError:
                print("! Compass misread in", survexblock.survexfile.path)
                print("  Stardata:", stardata)
                print("  Line:", ls)
                message = ' ! Value Error: line %s in %s' % (ls, survexblock.survexfile.path)
                models.DataIssue.objects.create(parser='survex', message=message)
                survexleg.compass = 1000
            survexleg.clino = -90.0
        else:
            assert line_leg_regex.match(lcompass), ls
            assert line_leg_regex.match(lclino) and lclino != "-", ls
            survexleg.compass = float(lcompass)
            survexleg.clino = float(lclino)

        if cave:
            survexleg.cave = cave

        # only save proper legs
        survexleg.save()

    itape = stardata.get("tape")
    if itape:
        try:
            survexblock.totalleglength += float(ls[itape])
        except ValueError:
            print("! Length not added")
        survexblock.save()


def LoadSurvexEquate(survexblock, sline):
    #print sline #
    stations = sline.split()
    assert len(stations) > 1
    for station in stations:
        survexblock.MakeSurvexStation(station)


def LoadSurvexLinePassage(survexblock, stardata, sline, comment):
    pass

stardatadefault = {"type":"normal", "t":"leg", "from":0, "to":1, "tape":2, "compass":3, "clino":4}
stardataparamconvert = {"length":"tape", "bearing":"compass", "gradient":"clino"}

regex_comment = re.compile(r"([^;]*?)\s*(?:;\s*(.*))?\n?$")
regex_ref     = re.compile(r'.*?ref.*?(\d+)\s*#\s*(X)?\s*(\d+)')
regex_star    = re.compile(r'\s*\*[\s,]*(\w+)\s*(.*?)\s*(?:;.*)?$')
# years from 1960 to 2039
regex_starref = re.compile(r'^\s*\*ref[\s.:]*((?:19[6789]\d)|(?:20[0123]\d))\s*#?\s*(X)?\s*(.*?\d+.*?)$(?i)')
# regex_starref = re.compile("""?x   # VERBOSE mode - can't get this to work
# ^\s*\*ref       # look for *ref at start of line
# [\s.:]*         # some spaces, stops or colons
# ((?:19[6789]\d)|(?:20[0123]\d)) # a date from 1960 to 2039 - captured as one field
# \s*#            # spaces then hash separator 
# ?\s*(X)         # optional X - captured
# ?\s*(.*?\d+.*?) # maybe a space, then at least one digit in the string - captured
# $(?i)""", re.X) # the end  (do the whole thing case insensitively)

 
regex_team    = re.compile(r"(Insts|Notes|Tape|Dog|Useless|Pics|Helper|Disto|Consultant)\s+(.*)$(?i)")
regex_team_member        = re.compile(r" and | / |, | & | \+ |^both$|^none$(?i)")
regex_qm      = re.compile(r'^\s*QM(\d)\s+?([a-dA-DxX])\s+([\w\-]+)\.(\d+)\s+(([\w\-]+)\.(\d+)|\-)\s+(.+)$')

insp = ""
callcount = 0
def RecursiveLoad(survexblock, survexfile, fin, textlines):
    """Follows the *include links in all the survex files from the root file 1623.svx
    and reads in the survex blocks, other data and the wallet references (survexscansfolder) as it
    goes. This part of the data import process is where the maximum memory is used and where it
    crashes on memory-constrained machines.
    """
    iblankbegins = 0
    text = [ ]
    stardata = stardatadefault
    teammembers = [ ]
    global insp
    global callcount

    # uncomment to print out all files during parsing 
    print(insp+"  - Reading file: " + survexblock.survexfile.path + " <> " + survexfile.path)
    stamp = datetime.now()
    lineno = 0
    
    sys.stderr.flush();
    callcount +=1
    if callcount >=10:
        callcount=0
        print(".", file=sys.stderr,end='')

    # Try to find the cave in the DB if not use the string as before
    path_match = re.search(r"caves-(\d\d\d\d)/(\d+|\d\d\d\d-?\w+-\d+)/", survexblock.survexfile.path)
    if path_match:
        pos_cave = '%s-%s' % (path_match.group(1), path_match.group(2))
        # print(insp+'Match')
        # print(insp+os_cave)
        cave = models.getCaveByReference(pos_cave)
        if cave:
            survexfile.cave = cave
    svxlines = ''
    svxlines = fin.read().splitlines()
    # print(insp+'Cave - preloop ' + str(survexfile.cave))
    # print(insp+survexblock)
    for svxline in svxlines:

        # print(insp+survexblock)

        # print(insp+svxline)
        # if not svxline:
        #     print(insp+'   - Not survex')
        #     return
        # textlines.append(svxline)

        lineno += 1

        # print(insp+'   - Line: %d' % lineno)

        # break the line at the comment
        sline, comment = regex_comment.match(svxline.strip()).groups()
        # detect ref line pointing to the scans directory
        mref = comment and regex_ref.match(comment)
        if mref:
            yr, letterx, wallet = mref.groups()
            if not letterx:
                letterx = ""
            else:
                letterx = "X"
            if len(wallet)<2:
                wallet = "0" + wallet
            refscan = "%s#%s%s" % (yr, letterx, wallet )
            #print(insp+' - Wallet ;ref - %s - looking for survexscansfolder' % refscan)
            survexscansfolders = models.SurvexScansFolder.objects.filter(walletname=refscan)
            if survexscansfolders:
                survexblock.survexscansfolder = survexscansfolders[0]
                #survexblock.refscandir = "%s/%s%%23%s" % (mref.group(1), mref.group(1), mref.group(2))
                survexblock.save()
                # print(insp+' - Wallet ; ref - %s - found in survexscansfolders' % refscan)
            else:
                message = ' ! Wallet ; ref - %s - NOT found in survexscansfolders %s-%s-%s' % (refscan,yr,letterx,wallet)
                print(insp+message)
                models.DataIssue.objects.create(parser='survex', message=message)

        # This whole section should be moved if we can have *QM become a proper survex command
        # Spec of QM in SVX files, currently commented out need to add to survex
        # needs to match regex_qm
        # ;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
        qmline = comment and regex_qm.match(comment)
        if qmline:
            # print(insp+qmline.groups())
            #(u'1', u'B', u'miraclemaze', u'1.17', u'-', None, u'\tcontinuation of rift')
            qm_no = qmline.group(1)
            qm_grade = qmline.group(2)
            qm_from_section = qmline.group(3)
            qm_from_station = qmline.group(4)
            qm_resolve_section = qmline.group(6)
            qm_resolve_station = qmline.group(7)
            qm_notes = qmline.group(8)

            # print(insp+'Cave - %s' % survexfile.cave)
            # print(insp+'QM no %d' % int(qm_no))
            # print(insp+'QM grade %s' % qm_grade)
            # print(insp+'QM section %s' % qm_from_section)
            # print(insp+'QM station %s' % qm_from_station)
            # print(insp+'QM res section %s' % qm_resolve_section)
            # print(insp+'QM res station %s' % qm_resolve_station)
            # print(insp+'QM notes %s' % qm_notes)

            # If the QM isn't resolved (has a resolving station) then load it
            if not qm_resolve_section or qm_resolve_section is not '-' or qm_resolve_section is not 'None':
                from_section = models.SurvexBlock.objects.filter(name=qm_from_section)
                # If we can find a section (survex note chunck, named)
                if len(from_section) > 0:
                    # print(insp+from_section[0])
                    from_station = models.SurvexStation.objects.filter(block=from_section[0], name=qm_from_station)
                    # If we can find a from station then we have the nearest station and can import it
                    if len(from_station) > 0:
                        # print(insp+from_station[0])
                        qm = models.QM.objects.create(number=qm_no,
                                                      nearest_station=from_station[0],
                                                      grade=qm_grade.upper(),
                                                      location_description=qm_notes)
            else:
                # print(insp+' - QM found but resolved')
                pass

        #print(insp+'Cave -sline ' + str(cave))
        if not sline:
            continue

        # detect the star ref command 
        mstar = regex_starref.match(sline)
        if mstar:
            yr,letterx,wallet = mstar.groups()
            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
            assert (int(wallet)<100), "Wallet number more than 100: %s" % wallet
            refscan = "%s#%s%s" % (yr, letterx, wallet)
            survexscansfolders = models.SurvexScansFolder.objects.filter(walletname=refscan)
            if survexscansfolders:
                survexblock.survexscansfolder = survexscansfolders[0]
                survexblock.save()
                # print(insp+' - Wallet *REF - %s - found in survexscansfolders' % refscan)
            else:
                message = ' ! Wallet *REF - %s - NOT found in survexscansfolders %s-%s-%s' % (refscan,yr,letterx,wallet)
                print(insp+message)
                models.DataIssue.objects.create(parser='survex', message=message)
            continue

        # detect the star command
        mstar = regex_star.match(sline)
        if not mstar:
            if "from" in stardata:
                # print(insp+'Cave ' + str(survexfile.cave))
                # print(insp+survexblock)
                LoadSurvexLineLeg(survexblock, stardata, sline, comment, survexfile.cave)
                # print(insp+'   - From: ')
                # print(insp+stardata)
                pass
            elif stardata["type"] == "passage":
                LoadSurvexLinePassage(survexblock, stardata, sline, comment)
                # print(insp+'   - Passage: ')
            #Missing "station" in stardata.
            continue

        # detect the star command
        cmd, line = mstar.groups()
        cmd = cmd.lower()
        if re.match("include$(?i)", cmd):
            includepath = os.path.join(os.path.split(survexfile.path)[0], re.sub(r"\.svx$", "", line))
            print(insp+'   - Include path found including - ' + includepath)
            # Try to find the cave in the DB if not use the string as before
            path_match = re.search(r"caves-(\d\d\d\d)/(\d+|\d\d\d\d-?\w+-\d+)/", includepath)
            if path_match:
                pos_cave = '%s-%s' % (path_match.group(1), path_match.group(2))
                # print(insp+pos_cave)
                cave = models.getCaveByReference(pos_cave)
                if cave:
                    survexfile.cave = cave
            else:
                print(insp+'    - No match in DB (i) for %s, so loading..' % includepath)
            includesurvexfile = models.SurvexFile(path=includepath)
            includesurvexfile.save()
            includesurvexfile.SetDirectory()
            if includesurvexfile.exists():
                survexblock.save()
                fininclude = includesurvexfile.OpenFile()
                insp += "> "
                RecursiveLoad(survexblock, includesurvexfile, fininclude, textlines)
                insp = insp[2:]

        elif re.match("begin$(?i)", cmd):
            if line:
                newsvxpath = os.path.join(os.path.split(survexfile.path)[0], re.sub(r"\.svx$", "", line))
                # Try to find the cave in the DB if not use the string as before
                path_match = re.search(r"caves-(\d\d\d\d)/(\d+|\d\d\d\d-?\w+-\d+)/", newsvxpath)
                if path_match:
                    pos_cave = '%s-%s' % (path_match.group(1), path_match.group(2))
                    # print(insp+pos_cave)
                    cave = models.getCaveByReference(pos_cave)
                    if cave:
                        survexfile.cave = cave
                else:
                    print(insp+'    - No match (b) for %s' % newsvxpath)

                name = line.lower()
                print(insp+'   - Begin found for: ' + name)
                # print(insp+'Block cave: ' + str(survexfile.cave))
                survexblockdown = models.SurvexBlock(name=name, begin_char=fin.tell(), parent=survexblock, survexpath=survexblock.survexpath+"."+name, cave=survexfile.cave, survexfile=survexfile, totalleglength=0.0)
                survexblockdown.save()
                survexblock.save()
                survexblock = survexblockdown
                # print(insp+survexblockdown)
                textlinesdown = [ ]
                insp += "> "
                RecursiveLoad(survexblockdown, survexfile, fin, textlinesdown)
                insp = insp[2:]
            else:
                iblankbegins += 1

        elif re.match("end$(?i)", cmd):
            if iblankbegins:
                iblankbegins -= 1
            else:
                survexblock.text = "".join(textlines)
                survexblock.save()
                # print(insp+'   - End found: ')
                endstamp = datetime.now()
                timetaken = endstamp - stamp
                # print(insp+'   - Time to process: ' + str(timetaken))
                return

        elif re.match("date$(?i)", cmd):
            if len(line) == 10:
                #print(insp+'   - Date found: ' + line)
                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()

        elif re.match("team$(?i)", cmd):
            pass
            # print(insp+'   - Team found: ')
            mteammember = regex_team.match(line)
            if mteammember:
                for tm in regex_team_member.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.SurvexPersonRole(survexblock=survexblock, nrole=mteammember.group(1).lower(), personexpedition=personexpedition, personname=tm)
                            personrole.expeditionday = survexblock.expeditionday
                            if personexpedition:
                                personrole.person=personexpedition.person
                            personrole.save()

        elif cmd == "title":
            #print(insp+'   - Title found: ')
            survextitle = models.SurvexTitle(survexblock=survexblock, title=line.strip('"'), cave=survexfile.cave)
            survextitle.save()
            pass

        elif cmd == "require":
            # should we check survex version available for processing?
            pass

        elif cmd == "data":
            #print(insp+'   - Data found: ')
            ls = line.lower().split()
            stardata = { "type":ls[0] }
            #print(insp+'    - Star data: ', stardata)
            #print(insp+ls)
            for i in range(0, len(ls)):
                stardata[stardataparamconvert.get(ls[i], ls[i])] = i - 1
            if ls[0] in ["normal", "cartesian", "nosurvey"]:
                assert (("from" in stardata and "to" in stardata) or "station" in stardata), line
            elif ls[0] == "default":
                stardata = stardatadefault
            else:
                assert ls[0] == "passage", line

        elif cmd == "equate":
            #print(insp+'   - Equate found: ')
            LoadSurvexEquate(survexblock, line)

        elif cmd == "fix":
            #print(insp+'   - Fix found: ')
            survexblock.MakeSurvexStation(line.split()[0])

        else:
            #print(insp+'   - Stuff')
            if cmd not in ["sd", "include", "units", "entrance", "data", "flags", "title", "export", "instrument",
                           "calibrate", "set", "infer", "alias", "cs", "declination", "case"]:
                message = "! Bad svx command in line:%s %s %s %s" % (cmd, line, survexblock, survexblock.survexfile.path)
                print(insp+message)
                models.DataIssue.objects.create(parser='survex', message=message)

        endstamp = datetime.now()
        timetaken = endstamp - stamp
        # print(insp+'   - Time to process: ' + str(timetaken))

def LoadAllSurvexBlocks():

    print(' - Flushing All Survex Blocks...')

    models.SurvexBlock.objects.all().delete()
    models.SurvexFile.objects.all().delete()
    models.SurvexDirectory.objects.all().delete()
    models.SurvexEquate.objects.all().delete()
    models.SurvexLeg.objects.all().delete()
    models.SurvexTitle.objects.all().delete()
    models.SurvexPersonRole.objects.all().delete()
    models.SurvexStation.objects.all().delete()

    print(" - Data flushed")
    # Clear the data issues as we are reloading
    models.DataIssue.objects.filter(parser='survex').delete()
    print(' - Loading All Survex Blocks...')
    
    print('  - redirecting stdout to loadsurvexblks.log...')
    stdout_orig = sys.stdout
    # Redirect sys.stdout to the file
    sys.stdout = open('loadsurvexblks.log', 'w')

    survexfile = models.SurvexFile(path=settings.SURVEX_TOPNAME, cave=None)
    survexfile.save()
    survexfile.SetDirectory()

    #Load all
    survexblockroot = models.SurvexBlock(name="root", survexpath="", begin_char=0, cave=None, survexfile=survexfile, totalleglength=0.0)
    survexblockroot.save()
    fin = survexfile.OpenFile()
    textlines = [ ]
    # The real work starts here
    RecursiveLoad(survexblockroot, survexfile, fin, textlines)
    fin.close()
    survexblockroot.text = "".join(textlines)
    survexblockroot.save()
    
    # Close the file
    sys.stdout.close()
    print("+", file=sys.stderr)
    sys.stderr.flush();
    
    # Restore sys.stdout to our old saved file handler
    sys.stdout = stdout_orig
    print(' - Loaded All Survex Blocks.')


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


def LoadPos():
    """Run cavern to produce a complete .3d file, then run 3dtopos to produce a table of 
    all survey point positions. Then lookup each position by name to see if we have it in the database 
    and if we do, then save the x/y/z coordinates.
    If we don't have it in the database, print an error message and discard it.
    """
    topdata = settings.SURVEX_DATA + settings.SURVEX_TOPNAME
    print(' - Generating a list of Pos from %s.svx and then loading...' % (topdata))
    
    # Be careful with the cache file. 
    # If LoadPos has been run before, 
    # but without cave import being run before,
    # then *everything* may be in the fresh  'not found' cache file. 
    
    cachefile = settings.SURVEX_DATA + "posnotfound.cache"
    notfoundbefore = {}
    if os.path.isfile(cachefile):
        # this is not a good test. 1623.svx may never change but *included files may have done.
        # When the *include is unrolled, we will be able to get a proper timestamp to use
        # and can increase the timeout from 3 days to 30 days.
        updtsvx = os.path.getmtime(topdata + ".svx")
        updtcache = os.path.getmtime(cachefile)
        age = updtcache - updtsvx
        print('   svx: %s    cache: %s    not-found cache is fresher by: %s' % (updtsvx, updtcache, str(timedelta(seconds=age) )))
        
        now = time.time()
        if now - updtcache > 3*24*60*60:
            print( "   cache is more than 3 days old. Deleting.")
            os.remove(cachefile)
        elif age < 0 :
            print("   cache is stale. Deleting.")
            os.remove(cachefile)
        else:
            print("   cache is fresh. Reading...")
            try:
                with open(cachefile, "r") as f:
                    for line in f:
                        l = line.rstrip()
                        if l in notfoundbefore:
                            notfoundbefore[l] +=1 # should not be duplicates
                            print(" DUPLICATE ", line, notfoundbefore[l])
                        else:
                            notfoundbefore[l] =1
            except:
                print("   FAILURE READ opening cache file %s" % (cachefile))
                raise
            
    
    notfoundnow =[]
    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)
    print("  - This next bit takes a while. Matching ~32,000 survey positions. Be patient...")

    posfile = open("%s.pos" % (topdata))
    posfile.readline() #Drop header
    for line in posfile.readlines():
        r = poslineregex.match(line)
        if r:
            x, y, z, name = r.groups() # easting, northing, altitude
            if name in notfoundbefore:
                skip[name] = 1
            else:
                try:
                    ss = models.SurvexStation.objects.lookup(name)
                    ss.x = float(x)
                    ss.y = float(y)
                    ss.z = float(z) 
                    ss.save()
                    found += 1
                except:
                    notfoundnow.append(name)
    print(" - %s stations not found in lookup of SurvexStation.objects. %s found. %s skipped." % (len(notfoundnow),found, len(skip)))

    if found > 10: # i.e. a previous cave import has been done
        try:
            with open(cachefile, "w") as f:
                c = len(notfoundnow)+len(skip)
                for i in notfoundnow:
                    f.write("%s\n" % i)
                for j in skip:
                    f.write("%s\n" % j) # NB skip not notfoundbefore
                print('   Not-found cache file written: %s entries' % c)
        except:
            print("   FAILURE WRITE opening cache file %s" % (cachefile))
            raise