import sys import os import re import time from datetime import datetime, timedelta from subprocess import call, Popen, PIPE from django.utils.timezone import get_current_timezone from django.utils.timezone import make_aware import troggle.settings as settings import troggle.core.models as models import troggle.core.models_caves as models_caves import troggle.core.models_survex as models_survex from troggle.parsers.people import GetPersonExpeditionNameLookup from troggle.core.views_caves import MapLocations """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". """ rx_braskets= re.compile(r"[()]") rx_line_length = re.compile(r"[\d\-+.]+$") survexlegsalllength = 0.0 survexlegsnumber = 0 def LoadSurvexLineLeg(survexblock, stardata, sline, comment, cave): global survexlegsalllength global survexlegsnumber # 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_survex.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": tape = rx_braskets.sub("",ls[stardata["tape"]]) tape = tape.replace("/",".") try: survexleg.tape = float(tape) 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: 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 rx_line_length.match(lcompass), ls assert rx_line_length.match(lclino) and lclino != "-", ls survexleg.compass = float(lcompass) survexleg.clino = float(lclino) if cave: survexleg.cave = cave # only save proper legs # No need to save as we are measuring lengths only on parsing now. # delete the object so that django autosaving doesn't save it. survexleg = None #survexleg.save() itape = stardata.get("tape") if itape: try: survexblock.totalleglength += float(ls[itape]) survexlegsalllength += float(ls[itape]) except ValueError: print("! Length not added") # No need to save as we are measuring lengths only on parsing now. #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): # do not import *data passage.. data which is LRUD not tape/compass/clino pass # 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_comment = re.compile(r"([^;]*?)\s*(?:;\s*(.*))?\n?$") rx_ref = re.compile(r'.*?ref.*?(\d+)\s*#\s*(X)?\s*(\d+)') rx_star = re.compile(r'\s*\*[\s,]*(\w+)\s*(.*?)\s*(?:;.*)?$') # years from 1960 to 2039 rx_starref = re.compile(r'^\s*\*ref[\s.:]*((?:19[6789]\d)|(?:20[0123]\d))\s*#?\s*(X)?\s*(.*?\d+.*?)$(?i)') # rx_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 # $(?i)""", re.X) # the end (do the whole thing case insensitively) rx_team = re.compile(r"(Insts|Notes|Tape|Dog|Useless|Pics|Helper|Disto|Consultant)\s+(.*)$(?i)") rx_team_member = re.compile(r" and | / |, | & | \+ |^both$|^none$(?i)") rx_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 global survexlegsnumber 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)) cave = models_caves.getCaveByReference(pos_cave) if cave: survexfile.cave = cave svxlines = '' svxlines = fin.read().splitlines() for svxline in svxlines: lineno += 1 # break the line at the comment sline, comment = rx_comment.match(svxline.strip()).groups() # detect ref line pointing to the scans directory mref = comment and rx_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 ) survexscansfolders = models_survex.SurvexScansFolder.objects.filter(walletname=refscan) if survexscansfolders: survexblock.survexscansfolder = survexscansfolders[0] survexblock.save() else: message = ' ! Wallet ; ref {} - NOT found in survexscansfolders {}'.format(refscan, survexblock.survexfile.path) 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 rx_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 rx_qm.match(comment) if qmline: 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 != '-' or qm_resolve_section != 'None': from_section = models_survex.SurvexBlock.objects.filter(name=qm_from_section) # If we can find a section (survex note chunck, named) if len(from_section) > 0: from_station = models_survex.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: qm = models_caves.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 if not sline: continue # detect the star ref command mstar = rx_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_survex.SurvexScansFolder.objects.filter(walletname=refscan) if survexscansfolders: survexblock.survexscansfolder = survexscansfolders[0] survexblock.save() else: message = ' ! Wallet *REF {} - NOT found in survexscansfolders {}'.format(refscan, survexblock.survexfile.path) print((insp+message)) models.DataIssue.objects.create(parser='survex', message=message) continue # detect the star command mstar = rx_star.match(sline) if not mstar: if "from" in stardata: LoadSurvexLineLeg(survexblock, stardata, sline, comment, survexfile.cave) pass elif stardata["type"] == "passage": LoadSurvexLinePassage(survexblock, stardata, sline, comment) #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.normpath(os.path.join(os.path.split(survexfile.path)[0], re.sub(r"\.svx$", "", line))) print((insp+' - Include path found including - ' + includepath)) # Try to find the cave in the DB if not use the string as before path_match = re.search(r"caves-(\d\d\d\d)/(\d+|\d\d\d\d-?\w+-\d+)/", includepath) if path_match: pos_cave = '%s-%s' % (path_match.group(1), path_match.group(2)) # print(insp+pos_cave) cave = models_caves.getCaveByReference(pos_cave) if cave: survexfile.cave = cave else: print((insp+' - No match in DB (i) for %s, so loading..' % includepath)) includesurvexfile = models_survex.SurvexFile(path=includepath) includesurvexfile.save() includesurvexfile.SetDirectory() if includesurvexfile.exists(): survexblock.save() fininclude = includesurvexfile.OpenFile() insp += "> " RecursiveLoad(survexblock, includesurvexfile, fininclude, textlines) insp = insp[2:] elif re.match("begin$(?i)", cmd): if line: newsvxpath = os.path.join(os.path.split(survexfile.path)[0], re.sub(r"\.svx$", "", line)) # Try to find the cave in the DB if not use the string as before path_match = re.search(r"caves-(\d\d\d\d)/(\d+|\d\d\d\d-?\w+-\d+)/", newsvxpath) if path_match: pos_cave = '%s-%s' % (path_match.group(1), path_match.group(2)) # print(insp+pos_cave) cave = models_caves.getCaveByReference(pos_cave) if cave: survexfile.cave = cave else: print((insp+' - No match (b) for %s' % newsvxpath)) previousnlegs = survexlegsnumber name = line.lower() print((insp+' - Begin found for: ' + name)) survexblockdown = models_survex.SurvexBlock(name=name, begin_char=fin.tell(), parent=survexblock, survexpath=survexblock.survexpath+"."+name, cave=survexfile.cave, survexfile=survexfile, totalleglength=0.0) survexblockdown.save() survexblock.save() survexblock = survexblockdown textlinesdown = [ ] insp += "> " RecursiveLoad(survexblockdown, survexfile, fin, textlinesdown) insp = insp[2:] else: iblankbegins += 1 elif re.match("end$(?i)", cmd): if iblankbegins: iblankbegins -= 1 else: # .text not used, using it for number of legs per block legsinblock = survexlegsnumber - previousnlegs print(insp+"LEGS: {} (previous: {}, now:{})".format(legsinblock,previousnlegs,survexlegsnumber)) survexblock.text = str(legsinblock) survexblock.save() endstamp = datetime.now() timetaken = endstamp - stamp return elif re.match("date$(?i)", cmd): if len(line) == 10: survexblock.date = make_aware(datetime.strptime(re.sub(r"\.", "-", line), '%Y-%m-%d'), get_current_timezone()) expeditions = models.Expedition.objects.filter(year=line[:4]) if expeditions: assert len(expeditions) == 1 survexblock.expedition = expeditions[0] survexblock.expeditionday = survexblock.expedition.get_expedition_day(survexblock.date) survexblock.save() elif re.match("team$(?i)", cmd): pass # print(insp+' - Team found: ') mteammember = rx_team.match(line) if mteammember: for tm in rx_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_survex.SurvexPersonRole(survexblock=survexblock, nrole=mteammember.group(1).lower(), personexpedition=personexpedition, personname=tm) personrole.expeditionday = survexblock.expeditionday if personexpedition: personrole.person=personexpedition.person personrole.save() elif cmd == "title": survextitle = models_survex.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": ls = line.lower().split() stardata = { "type":ls[0] } 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": LoadSurvexEquate(survexblock, line) elif cmd == "set" and re.match("names(?i)", line): pass elif cmd == "flags": # 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 elif cmd == "fix": survexblock.MakeSurvexStation(line.split()[0]) elif cmd in ["alias", "calibrate", "cs","entrance", "export", "case", "declination", "infer","instrument", "sd", "units"]: # we ignore all these, which is fine. pass else: if cmd not in ["include", "data", "flags", "title", "set", "ref"]: message = "! Bad svx command: [*{}] {} ({}) {}".format(cmd, line, survexblock, survexblock.survexfile.path) print((insp+message)) models.DataIssue.objects.create(parser='survex', message=message) else: message = "! Unparsed [*{}]: '{}' {}".format(cmd, line, 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(): global survexlegsalllength global survexlegsnumber 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.SurvexEquate.objects.all().delete() models_survex.SurvexLeg.objects.all().delete() models_survex.SurvexTitle.objects.all().delete() models_survex.SurvexPersonRole.objects.all().delete() models_survex.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_survex.SurvexFile(path=settings.SURVEX_TOPNAME, cave=None) survexfile.save() survexfile.SetDirectory() #Load all # this is the first so id=1 survexblockroot = models_survex.SurvexBlock(name="rootblock", 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.totalleglength = survexlegsalllength survexblockroot.text = str(survexlegsnumber) #survexblockroot.text = "".join(textlines) these are all blank 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(" - total number of survex legs: {}m".format(survexlegsnumber)) print(" - total leg lengths loaded: {}m".format(survexlegsalllength)) print(' - Loaded All Survex Blocks.') poslineregex = re.compile(r"^\(\s*([+-]?\d*\.\d*),\s*([+-]?\d*\.\d*),\s*([+-]?\d*\.\d*)\s*\)\s*([^\s]+)$") def LoadPos(): """Run cavern to produce a complete .3d file, then run 3dtopos to produce a table of all survey point positions. Then lookup each position by name to see if we have it in the database and if we do, then save the x/y/z coordinates. If we don't have it in the database, print an error message and discard it. This is ONLY ever used for entrance and fixedpts locations for the prospecting map: about 600 points out of 32,000. """ topdata = settings.SURVEX_DATA + settings.SURVEX_TOPNAME print((' - Generating a list of Pos from %s.svx and then loading...' % (topdata))) # TO DO - remove the cache file apparatus. Not needed. Only laser points and entrances loaded now. # 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...") mappoints = {} for pt in MapLocations().points(): svxid, number, point_type, label = pt mappoints[svxid]=True posfile = open("%s.pos" % (topdata)) posfile.readline() #Drop header for line in posfile.readlines(): r = poslineregex.match(line) if r: x, y, z, id = r.groups() if id in notfoundbefore: skip[id] = 1 else: for sid in mappoints: if id.endswith(sid): try: ss = models_survex.SurvexStation.objects.lookup(id) ss.x = float(x) ss.y = float(y) ss.z = float(z) ss.save() found += 1 except: notfoundnow.append(id) print(" - %s failed lookups 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