import copy import os import re import subprocess import sys import time from datetime import datetime, timezone from pathlib import Path import troggle.settings as settings from troggle.core.models.caves import Cave, Entrance, GetCaveLookup from troggle.core.models.logbooks import QM from troggle.core.models.survex import SurvexBlock, SurvexDirectory, SurvexFile, SurvexPersonRole, SurvexStation from troggle.core.models.wallets import Wallet from troggle.core.models.troggle import DataIssue, Expedition from troggle.core.utils import chaosmonkey, get_process_memory #from troggle.parsers.logbooks import GetCaveLookup from troggle.parsers.caves import create_new_cave from troggle.parsers.people import GetPersonExpeditionNameLookup, known_foreigner """Imports the tree of survex files following from a defined root .svx file It also scans the Loser repo for all the svx files, which it loads individually afterwards. """ todo = """ - Lots to do to cut down on unnecessary .save() calls to avoid hitting the db so much. SHould speed it up noticably. - Obscure bug in the *team inheritance and rootblock initialization needs tracking down - Learn to use Django .select_related() and .prefetch_related() to speed things up https://zerotobyte.com/how-to-use-django-select-related-and-prefetch-related/ - LoadSurvexFile() Creates a new current survexfile and valid .survexdirectory The survexblock passed-in is not necessarily the parent. FIX THIS. - Finish writing the parse_one_file() function for survexfiles edited online. Perhaps easier if this is a completely new file rather than an existing file.. nasty. - When Olly implements LEG in the 'dump3d --legs' utility, then we can use that to get the length of all the legs in a survex block instead of adding them up oursleves. Which means that we can ignore all the Units and offset stuff, that troggle will work with survex files with backsights, repeated readings from distox etc.. Not actually useful for pre 2022 survey data, but good future-proofing. Also it will be a tiny bit more accurate as these leg lengths are after loop closure fixup. """ survexblockroot = None survexomitsroot = None ROOTBLOCK = "rootblock" OMITBLOCK = "omitblock" METRESINFEET = 3.28084 UNSEENS = "_unseens.svx" stop_dup_warning = False dup_includes = 0 debugprint = False # Turns on debug printout for just one *include file debugprinttrigger = "!" dataissues = [] class SurvexLeg: """No longer a models.Model subclass, so no longer a database table""" tape = 0.0 compass = 0.0 clino = 0.0 def IdentifyCave(cavepath): """Given a file path for a survex file, or a survex-block path, return the cave object """ caveslist = GetCaveLookup() if cavepath.lower() in caveslist: return caveslist[cavepath.lower()] # TO DO - this predates the big revision to Gcavelookup so look at this again carefully path_match = LoadingSurvex.rx_cave.search(cavepath) # use as Class method if path_match: sluggy = f"{path_match.group(1)}-{path_match.group(2)}" guesses = [sluggy.lower(), path_match.group(2).lower()] for g in guesses: if g in caveslist: caveslist[cavepath] = caveslist[g] return caveslist[g] print(f" ! Failed to find cave for {cavepath.lower()}") else: # not a cave, but that is fine. # print(f' ! No regex(standard identifier) cave match for {cavepath.lower()}') return None def datewallet(w, earliest): """Gets the date of the youngest survexblock associated with the wallet REFACTOR this to do the whole date-getting task Currently there is only one SurvexBlock, but this is in anticipation of chnaging the schema to allow many. """ first = earliest blocks = SurvexBlock.objects.filter(scanswallet=w) # only ONE I think ?! for b in blocks: if b.date: if b.date < first: first = b.date if first == earliest: # no date found w.date = None else: w.date = first.isoformat() return w.date def set_walletdate(w): earliest = datetime.now().date() if not w.date(): # sets .walletdate as a side-effect if it gets it from JSON d = datewallet(w, earliest) # Not in JSON, so checks all the survex blocks w.walletdate = d w.save() def stash_data_issue(parser=None, message=None, url=None, sb=None): """Avoid hitting the database for error messages until the end of the import""" global dataissues dataissues.append((parser, message, url, sb)) def store_data_issues(): """Take the stash and store it permanently in the database instead use BULK creation here !""" global dataissues print(f" - Storing {len(dataissues)} Data Issues into database") # make a list of objects, but don't commit to database yet di_list = [] for issue in dataissues: parser, message, url, sb = issue if url is None: if sb is not None: url = get_offending_filename(sb) di_list.append(DataIssue(parser=parser, message=message, url=url)) # Now commit to db DataIssue.objects.bulk_create(di_list) dataissues = [] # in database now, so empty cache def get_offending_filename(path): """Used to provide the URL for a line in the DataErrors page which reports problems on importing data into troggle """ return "/survexfile/" + path + ".svx" trip_people_cache = {} # per survexblock, so robust wrt PUSH/POP begin/end def get_team_on_trip(survexblock): """Uses a cache to avoid a database query if it doesn't need to. Only used for complete team.""" global trip_people_cache if survexblock in trip_people_cache: return trip_people_cache[survexblock] qpeople = SurvexPersonRole.objects.filter(survexblock=survexblock) # not very good Django style trip_people_cache[survexblock] = qpeople # this is a query list return qpeople def get_people_on_trip(survexblock): """Gets the displayable names of the people on a survexbock trip. Only used for complete team.""" qpeople = get_team_on_trip(survexblock) people = [] for p in qpeople: people.append(f"{p.personname}") return list(set(people)) trip_person_record = {} # per survexblock, so robust wrt PUSH/POP begin/end trip_team_cache = {} # per survexblock, so robust wrt PUSH/POP begin/end def put_person_on_trip(survexblock, personexpedition, tm): """Uses a cache to avoid a database query if it doesn't need to. Only used for a single person""" global trip_person_record global trip_team_cache if (survexblock, personexpedition) in trip_person_record: return True try: personrole = SurvexPersonRole( # does not commit to db yet survexblock=survexblock, person = personexpedition.person, personexpedition=personexpedition, personname=tm ) except: message = f"! *team '{tm}' FAIL, already created {survexblock.survexfile.path} ({survexblock}) " print(self.insp + message) stash_data_issue( parser="survex", message=message, url=None, sb=(survexblock.survexfile.path) ) if survexblock not in trip_team_cache: trip_team_cache[survexblock] = [] trip_team_cache[survexblock].append(personrole) trip_person_record[(survexblock, personexpedition)] = 1 return False def confirm_team_on_trip(survexblock): global trip_team_cache if survexblock not in trip_team_cache: return # Now commit to db SurvexPersonRole.objects.bulk_create(trip_team_cache[survexblock]) trip_team_cache[survexblock] = [] # in database now, so empty cache def check_team_cache(): global trip_team_cache message = f"! *team CACHEFAIL check_team_cache() called " print(message) for block in trip_team_cache: message = f"! *team CACHEFAIL, already created {block.survexfile.path} ({block}) " print(message) person_pending_cache = {} # per survexblock, so robust wrt PUSH/POP begin/end def add_to_pending(survexblock, tm): """Collects team names before we have a date so cannot validate against expo attendance yet""" global person_pending_cache if survexblock not in person_pending_cache: person_pending_cache[survexblock] = set() person_pending_cache[survexblock].add(tm) def get_team_pending(survexblock): """A set of *team names before we get to the *date line in a survexblock """ global person_pending_cache if survexblock in person_pending_cache: teamnames = person_pending_cache[survexblock] # a set of names person_pending_cache[survexblock] = () return teamnames return 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". """ # python regex flags (?i) means case-insentitive, (?s) means . matches newline too # see https://docs.python.org/3/library/re.html rx_begin = re.compile(r"(?i)begin") rx_begin2 = re.compile("(?i)begin$") rx_end = re.compile(r"(?i)end$") rx_end2 = re.compile("(?i)end$") rx_title = re.compile(r"(?i)title$") rx_title2 = re.compile("(?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\-+.]+$") instruments = "(bitch|bodger|bolt|bolter|bolting|book|clino|comp|compass|consultant|disto|distox|distox2|dog|dogsbody|drawing|drill|gps|helper|inst|instr|instrument|monkey|nagging|nail|nail_polish|nail_polish_bitch|nail_polish_monkey|nail_varnish|nail_varnish_bitch|note|paint|photo|pic|point|polish|powerdrill|rig|rigger|rigging|shoot|sketch|slacker|something|tape|topodroid|unknown|useless|varnish|waiting_patiently)" rx_teammem = re.compile(r"(?i)" + instruments + "?(?:es|s)?\s+(.*)$") rx_teamold = re.compile(r"(?i)(.*)\s+" + instruments + "?(?:es|s)?$") rx_teamabs = re.compile(r"(?i)^\s*(" + instruments + ")?(?:es|s)?\s*$") rx_person = re.compile(r"(?i) and |/| / |, | , |&| & | \+ |^both$|^none$") rx_qm = re.compile( # r"(?i)^\s*QM(\d+)\s+(.+)\s+([\w\-\_]+)\.([\w\.\-]+)\s+(([\w\-]+)\.([\w\.\-]+)|\-)\s+(.+)$" r"(?i)^\s*QM(\d+)\s+([^\s]+)\s+([^\s]+)\s+([^\s]+)\s+(.+)$" ) # does not recognise non numeric suffix survey point ids rx_qm0 = re.compile(r"(?i)^\s*QM(\d+)\s+(.+)$") rx_qm_tick = re.compile(r"(?i)^\s*QM(\d+)\s+TICK\s([\d\-]+)\s(.*)$") # 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].*)$") rx_include2 = re.compile("(?i)include$") rx_commref = re.compile(r"(?i)^\s*ref(?:erence)?[\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+.*?)$") rx_argsref = re.compile(r"(?i)^[\s.:]*((?:19[6789]\d)|(?:20[012345]\d))\s*#?\s*(X)?\s*(.*?\d+.*?)$") rx_badmerge = re.compile(r"(?i).*(\>\>\>\>\>)|(\=\=\=\=\=)|(\<\<\<\<\<).*$") rx_ref2 = re.compile(r"(?i)\s*ref[.;]?") rx_commteam = re.compile(r"(?i)\s*(Messteam|Zeichner)\s*[:]?(.*)") rx_quotedtitle = re.compile(r'(?i)^"(.*)"$') # 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 = [] teaminheritstack = [] teamcurrentstack = [] dateinheritstack = [] datecurrentstack = [] stackbegin = [] flagsstack = [] datastack = [] includestack = [] stacksvxfiles = [] svxfileslist = [] svxdirs = {} uniquefile = {} # each key is a survex path, and its value is a list of parent files 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", "deprecated"] ignorenoncave = [ "caves-1623", "caves-1623/2007-NEU", "caves-1626", "caves-1624", "caves-1627", "fixedpts/gps/gps00raw", "", ] TREE = "tree" ODDS = "oddments" svxpass = TREE includedfilename = "" currentsurvexblock = None currentsurvexfile = None currentcave = None caverndate = None currentteam = set() inheritteam = set() currentdate = None inheritdate = None pending = [] adhocload = False 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 = ( f"! Warning. Unparsed [*{cmd}]: '{line}' {survexblock.survexfile.path} - not an error (probably)" ) print(self.insp + message) stash_data_issue( parser="survex", message=message, url=None, sb=(survexblock.survexfile.path) ) else: message = ( f"! Bad unrecognised svx command: [*{cmd}] {line} ({survexblock}) {survexblock.survexfile.path}" ) print(self.insp + message) stash_data_issue( parser="survex", message=message, url=None, sb=(survexblock.survexfile.path) ) def get_team_inherited(self, survexblock): # survexblock only used for debug mesgs """See get_team_pending(survexblock) which gets called at the same time, when we see a *date line""" global person_pending_cache if self.inheritteam: message = ( f"- no *team INHERITING ({survexblock.parent})>({survexblock}) {survexblock.survexfile.path} '{self.inheritteam}'" ) print(self.insp + message) # stash_data_issue( # parser="survex", message=message, url=None, sb=(survexblock.survexfile.path) # ) return self.inheritteam def fix_undated(self, survexblock): """Called when we reach *end of a block OR when a QM is seen. Checks to see if the block has no *date, in which case it uses the inherited date. This is fine if the inherited date is from the same SurvexFile, but inheriting dates across *include files is almost certainly NOT expected behaviour, even though it is syntactically "correct", so triggers a Warning. """ if survexblock.parent.name == "troggle_unseens": # Bolluxed up if we try to inherit from this random junk, so don't. return if self.currentdate: # already set if not survexblock.date: # error message = ( f"! no survexblock.date but currentdate is set. ({survexblock})-{survexblock.survexfile.path} {self.currentdate=}" ) print(self.insp + message) stash_data_issue( parser="survex", message=message, url=None, sb=(survexblock.survexfile.path) ) return if self.inheritdate: survexblock.date = self.inheritdate self.currentdate = self.inheritdate # unecessary duplication # Not an error, so not put in DataIssues, but is printed to debug output message = ( f"- No *date. INHERITING date '{self.inheritdate:%Y-%m-%d}' from ({survexblock.parent})-{survexblock.parent.survexfile.path} to ({survexblock})-{survexblock.survexfile.path} {self.inheritdate:%Y-%m-%d}" ) print(self.insp + message) # stash_data_issue( # parser="survex", message=message, url=None, sb=(survexblock.survexfile.path) # child # ) if survexblock.survexfile != survexblock.parent.survexfile: # This is noteworthy, however. if survexblock.parent.name == "rootblock": # Not a sensible thing to inherit a date from, even if a date exists, which it shouldn't... message = ( f"- No *date. But not sensible to inherit from rootblock. From ({survexblock.parent})-{survexblock.parent.survexfile.path} to ({survexblock})-{survexblock.survexfile.path} {self.inheritdate:%Y-%m-%d}" ) print(self.insp + message) # stash_data_issue( # parser="survex", message=message, url=None, sb=(survexblock.survexfile.path) # ) return else: message = ( f"- Warning *date '{self.inheritdate:%Y-%m-%d}' INHERITED from DIFFERENT file:\n ({survexblock.parent})-{survexblock.parent.survexfile.path} to ({survexblock})-{survexblock.survexfile.path} {self.inheritdate:%Y-%m-%d}\n {self.stackbegin} {self.inheritdate:%Y-%m-%d}" ) print(self.insp + message) stash_data_issue( parser="survex", message=message, url=None, sb=(survexblock.parent.survexfile.path) # PARENT ) return self.inheritdate else: # This is not an error in the Expo dataset. # Many files just holding *include lines do not have dates. # Hardly _any_ of the ARGE survex files have dates ! pass # message = f" ! No survexblock.date inheritable in '{survexblock}' in '{survexblock.survexfile.path}', setting to 1976" # print(self.insp + message) # stash_data_issue( # parser="survex", message=message, url=None, sb=(survexblock.survexfile.path) # ) # expoyear = "1976" return def fix_anonymous(self, survexblock): """Called when we reach *end of a block Checks to see if the block has no team attached, in which case it uses the inherited team. This is fine if the inherited team is from the same SurvexFile, but inheriting team across *include files is almost certainly NOT expected behaviour, even though it is syntactically "correct". If the block has no date, then it is assumed to be an abstract container, with no relevant team, and anyway we can't attach a PersonExpedition without knowing the year. Unless its parent has an identified expo""" if survexblock.parent.name == "troggle_unseens": # Bolluxed up if we try to inherit from this random junk, so don't. return expo = survexblock.expedition # may be None if no *date yet if not expo: expo = survexblock.parent.expedition # immediate parent works mostly if not expo: return if not self.currentteam: # i.e. if it is a dated block and has no team if teamnames := self.get_team_inherited(survexblock):# WALRUS for tm in teamnames: personexpedition = GetPersonExpeditionNameLookup(expo).get(tm.lower()) if personexpedition: put_person_on_trip(survexblock, personexpedition, tm) return 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 personrole is used to record that a person was on a survex trip, NOT the role they played. (NB PersonLogEntry is a logbook thing, not a survex thing. ) """ def record_team_member(tm, survexblock): tm = tm.strip("\"'").strip() # Refactor. The dict GetPersonExpeditionNameLookup(expo) indexes by name and has values of personexpedition # This is convoluted # we need the current expedition, but if there has been no date yet in the survex file, we don't know which one it is. # so we can't validate whether the person was on expo or not. # we will have to attach them to the survexblock anyway, and then do a # later check on whether they are valid when we get the date. self.currentteam.add(tm) # used in push/pop block code expo = survexblock.expedition # may be None if no *date yet if expo: personexpedition = GetPersonExpeditionNameLookup(expo).get(tm.lower()) if personexpedition: put_person_on_trip(survexblock, personexpedition, tm) elif known_foreigner(tm): # note, not using .lower() message = f"- *team {expo.year} '{tm}' known foreigner on *team {survexblock.survexfile.path} ({survexblock}) in '{line}'" print(self.insp + message) # stash_data_issue(parser='survex', message=message, url=None, sb=(survexblock.survexfile.path)) else: # we know the date and expo, but can't find the person message = f"! *team {expo.year} '{tm}' FAIL personexpedition lookup on *team {survexblock.survexfile.path} ({survexblock}) in '{line}'" print(self.insp + message) stash_data_issue( parser="survex", message=message, url=None, sb=(survexblock.survexfile.path) ) else: add_to_pending(survexblock, tm) # don't know the date yet, so cannot query the table about validity. # assume the person is valid. It will get picked up with the *date appears # There are hundreds of these.. message = ( f"- Team before Date: {line} ({survexblock}) {survexblock.survexfile.path}" ) # print(self.insp + message) # stash_data_issue( # parser="survex team", message=message, url=None, sb=(survexblock.survexfile.path) # ) mteammember = self.rx_teammem.match(line) # matches the role at the beginning if not mteammember: moldstyle = self.rx_teamold.match(line) # matches the role at the the end of the string if moldstyle: for tm in self.rx_person.split(moldstyle.group(1)): if tm: record_team_member(tm, survexblock) # seems to be working # msg = "! OLD tm='{}' line: '{}' ({}) {}".format(tm, line, survexblock, survexblock.survexfile.path) # print(msg, file=sys.stderr) else: message = f"! *team {survexblock.survexfile.path} ({survexblock}) Weird '{mteammember.group(1)}' oldstyle line: '{line}'" print(self.insp + message) stash_data_issue( parser="survex", message=message, url=None, sb=(survexblock.survexfile.path) ) else: nullmember = self.rx_teamabs.match(line) # matches empty role line. Ignore these. if not nullmember: message = f"! *team {survexblock.survexfile.path} ({survexblock}) Bad line: '{line}'" print(self.insp + message) stash_data_issue( parser="survex", message=message, url=None, sb=(survexblock.survexfile.path) ) else: for tm in self.rx_person.split(mteammember.group(2)): if tm: record_team_member(tm, survexblock) else: if mteammember.group(2).lower() not in ("none", "both"): message = f"! Weird *team '{mteammember.group(2)}' newstyle line: '{line}' ({survexblock}) {survexblock.survexfile.path}" print(self.insp + message) stash_data_issue( parser="survex", message=message, url=None, sb=(survexblock.survexfile.path) ) def LoadSurvexEntrance(self, survexblock, line): # Not using this yet pass def LoadSurvexAlias(self, survexblock, line): # *alias station - .. splayalias = re.match("(?i)\s*station\s*\-\s*\.\.\s*$", line) if splayalias: self.flagsstar["splayalias"] = True print(line) else: message = f"! Bad *ALIAS: '{line}' ({survexblock}) {survexblock.survexfile.path}" print(self.insp + message) stash_data_issue(parser="survex", message=message) def LoadSurvexUnits(self, survexblock, line): # all for 4 survex files with measurements in feet. bugger. # Won't need this once we move to using cavern or d3dump output for lengths tapeunits = self.rx_tapelng.match(line) # tape|length if not tapeunits: return convert = re.match("(?i)(\w*)\s*([\.\d]+)\s*(\w*)", line) if convert: factor = convert.groups()[1] self.unitsfactor = float(factor) if debugprint: message = ( f"! *UNITS NUMERICAL conversion [{factor}x] '{line}' ({survexblock}) {survexblock.survexfile.path}" ) print(self.insp + message) stash_data_issue(parser="survexunits", message=message) feet = re.match("(?i).*feet$", line) metres = re.match("(?i).*(METRIC|METRES|METERS)$", line) if feet: self.units = "feet" elif metres: self.units = "metres" else: message = f"! *UNITS in YARDS!? - not converted '{line}' ({survexblock}) {survexblock.survexfile.path}" print(self.insp + message) stash_data_issue(parser="survexunits", message=message) def get_expo_from_year(self, year, line, survexblock): # cacheing to save DB query on every block if year in self.expos: expo = self.expos[year] else: expeditions = Expedition.objects.filter(year=year) if len(expeditions) != 1: message = ( f"! More than one expedition in year {year} '{line}' ({survexblock}) {survexblock.survexfile.path}" ) print(self.insp + message) stash_data_issue( parser="survex", message=message, url=None, sb=(survexblock.survexfile.path) ) if expeditions: expo = expeditions[0] self.expos[year] = expo else: expo = Expedition.objects.get(year="1976") message = f"! DATE INCORRECT. There is no expedition for the year {year}. {survexblock.survexfile.path} ({survexblock}) - set to 1976." print(self.insp + message) stash_data_issue(parser='survex', message=message, url=None, sb=(survexblock.survexfile.path)) return expo def LoadSurvexDate(self, survexblock, line): """We now have a valid date for this survexblock, so we now know the expo it relates to and can use GetPersonExpeditionNameLookup(expo) to check whether the people are correct. Note that a *team line can come before AND after a *date line""" def setdate_on_survexblock(year): """Either *date comes before any *team, in which case there are no prior PersonRoles attached, or *team came before this *date, in which case the names are only in 'pending'""" global trip_person_record expo = self.get_expo_from_year(year, line, survexblock) survexblock.expedition = expo team = get_team_on_trip(survexblock) # should be empty, should only be in 'pending' # team = SurvexPersonRole.objects.filter(survexblock=survexblock) if len(team) > 0: message = f"! *team {expo.year} Multiple *date in one block? Already someone on team when *date seen. {survexblock.survexfile.path} ({survexblock}) in '{line}'" print(self.insp + message) stash_data_issue(parser='survex', message=message, url=None, sb=(survexblock.survexfile.path)) if teamnames := get_team_pending(survexblock): # WALRUS https://docs.python.org/3/whatsnew/3.8.html#assignment-expressions for tm in teamnames: if known_foreigner(tm): message = f"- *team {expo.year} '{tm}' known foreigner *date (misordered) {survexblock.survexfile.path} ({survexblock}) in '{line}'" print(self.insp + message) # stash_data_issue(parser='survex', message=message, url=None, sb=(survexblock.survexfile.path)) else: pe = GetPersonExpeditionNameLookup(expo).get(tm.lower()) if pe: put_person_on_trip(survexblock, pe, tm) else: message = f"! *team {year} '{tm}' FAIL personexpedition lookup on *date {survexblock.survexfile.path} ({survexblock}) " print(self.insp + message) stash_data_issue( parser="survex", message=message, url=None, sb=(survexblock.survexfile.path), ) oline = line perps = get_people_on_trip(survexblock) # What, you don't know Judge Dredd slang ? if len(line) > 10: message = "! DATE Warning LONG DATE '{}' ({}) {}".format(oline, survexblock, survexblock.survexfile.path) print(self.insp+message) stash_data_issue(parser='svxdate', message=message, url=None, sb=(survexblock.survexfile.path)) if line[10] == "-": # ie a range, just look at first date 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(line.replace(".", "-"), "%Y-%m-%d") elif len(line) == 7: year = line[:4] message = f"! DATE Warning only accurate to the month, setting to 1st '{oline}' ({survexblock}) {survexblock.survexfile.path} {perps}" print(self.insp + message) stash_data_issue( parser="svxdate", message=message, url=None, sb=(survexblock.survexfile.path) ) survexblock.date = datetime.strptime(line.replace(".", "-"), "%Y-%m") # sets to first of month elif len(line) == 4: year = line[:4] message = f"! DATE WARNING only accurate to the YEAR, setting to 1st January '{oline}' ({survexblock}) {survexblock.survexfile.path} {perps}" print(self.insp + message) stash_data_issue( parser="svxdate", message=message, url=None, sb=(survexblock.survexfile.path) ) survexblock.date = datetime.strptime(line, "%Y") # sets to January 1st elif len(line) == 9 or len(line) == 8: year = line[:4] message = f"! DATE format WARNING, single digit day or month number,'{oline}' [{line[-5]}][{line[-2]}] ({survexblock}) {survexblock.survexfile.path}" print(self.insp + message) stash_data_issue( parser="svxdate", message=message, url=None, sb=(survexblock.survexfile.path) ) if line[-2] == "-" or line[-2] == ".": line = line[:-1] + '0' + line[-1] survexblock.date = datetime.strptime(line.replace(".", "-"), "%Y-%m-%d") print(f"! DATE -2 '{line}' '{survexblock.date}'") elif line[-5] == "-" or line[-5] == ".": line = line[:-4] + '0' + line[-4:] survexblock.date = datetime.strptime(line.replace(".", "-"), "%Y-%m-%d") print(f"! DATE -5 '{line}' '{survexblock.date}'") else: year = line[:4] message = ( f"! DATE Error SHORT LINE '{line}' '{oline}-{survexblock}' ({type(survexblock)}) {survexblock.survexfile.path}" ) print(self.insp + message) stash_data_issue( parser="svxdate", message=message, url=None, sb=(survexblock.survexfile.path) ) else: # these errors are reporting the wrong survexblock, which is actually a SurvexFile (!) # see To Do notes on how to trigger this. Still needs investigating.. message = ( f"! DATE Error unrecognised '{oline}-{survexblock}' ({type(survexblock)}) {survexblock.survexfile.path}" ) print(self.insp + message) stash_data_issue( parser="survex", message=message, url=None, sb=(survexblock.survexfile.path) ) print(f" {type(survexblock)=}") # survexblock.parent fails as a SurvexFile has no .parent ...ugh. print(f" {survexblock.survexpath=}") print(f" {survexblock.survexfile=}") # Not setting 'year' crashes entire import on databaseReset. year = line[:4] perps = get_people_on_trip(survexblock) # raise try: setdate_on_survexblock(year) except NameError: print(f">> why is year not set ?! {survexblock.survexfile.path}") setdate_on_survexblock("1976") if survexblock.date: # do not actually need a distict variable 'currentdate' but it makes the code clearer self.currentdate = survexblock.date survexblock.save() def LoadSurvexLeg(self, survexblock, sline, comment, svxline): """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. REPLACE ALL THIS by reading the .log output of cavern for the file. But we need the lengths per Block, not by File. dump3d will do lengths per block. """ 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(f"! LEG datastar type:{self.datastar['type'].upper()}++{survexblock.survexfile.path}\n{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( f" !! LEG data lineno:{self.lineno}\n !! sline:'{sline}'\n !! datastar['tape']: {self.datastar['tape']}" ) if self.datastar["type"] != "normal": return ls = sline.lower().split() # NORMAL, so there should be 5 fields # from the content, this is clearly reading fixedpts/gps/gps00raw.svx, but not reporting it by that name if len(ls) < 5: print("! Fewer than 5 fields in NORMAL in ", survexblock.survexfile.path, survexfile, survexfile.parent) print(" datastar NORMAL:", self.datastar) print(f" Line (split): {ls}, comment: {comment}") print(f" Line: {sline}\nsvxline: {svxline}") message = f" ! Not 5 fields in line '{sline.lower()}' {self.datastar=} {ls=} in\n{survexblock}\n{survexblock.survexfile}\n{survexblock.survexfile.path}" stash_data_issue( parser="survexleg", message=message, url=None, sb=(survexblock.survexfile.path) ) datastar = self.datastar # shallow copy: alias but the things inside are the same things survexleg = SurvexLeg() # skip all splay legs try: if "splayalias" in self.flagsstar: if ls[datastar["from"]] == "-" or ls[datastar["to"]] == "-": if debugprint: print("Aliased splay in ", survexblock.survexfile.path) return 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 ls[datastar["to"]] == "-": message = f" ! Suspected splay, not declared, in line {ls} in {survexblock.survexfile.path}" print(self.insp + message) stash_data_issue( parser="survexleg", message=message, url=None, sb=(survexblock.survexfile.path) ) return except: message = f" ! datastar parsing from/to incorrect in line {ls} in {survexblock.survexfile.path}" print(self.insp + message) stash_data_issue( parser="survexleg", message=message, url=None, sb=(survexblock.survexfile.path) ) return try: tape = ls[datastar["tape"]] except: message = f" ! datastar parsing incorrect in line {ls} in {survexblock.survexfile.path}" print(self.insp + message) stash_data_issue( parser="survexleg", message=message, url=None, sb=(survexblock.survexfile.path) ) 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 = f" ! Units: Length scaled {tape}m '{ls}' in ({survexblock.survexfile.path}) units:{self.units} factor:{self.unitsfactor}x" print(self.insp + message) stash_data_issue( parser="survexleg", message=message, url=None, sb=(survexblock.survexfile.path) ) if self.units == "feet": tape = float(tape) / METRESINFEET if debugprint: message = f" ! Units: converted to {tape:.3f}m from {self.units} '{ls}' in ({survexblock.survexfile.path})" print(self.insp + message) stash_data_issue( parser="survexleg", message=message, url=None, sb=(survexblock.survexfile.path) ) survexleg.tape = float(tape) self.legsnumber += 1 except ValueError: message = f" ! Value Error: Tape misread in line'{ls}' in {survexblock.survexfile.path} units:{self.units}" print(self.insp + message) stash_data_issue( parser="survexleg", message=message, url=None, sb=(survexblock.survexfile.path) ) survexleg.tape = invalid_tape try: survexblock.legslength += survexleg.tape self.slength += survexleg.tape except ValueError: message = ( f" ! Value Error: Tape length not added '{ls}' in {survexblock.survexfile.path} units:{self.units}" ) print(self.insp + message) stash_data_issue( parser="survexleg", message=message, url=None, sb=(survexblock.survexfile.path) ) try: lcompass = ls[datastar["compass"]] except: message = f" ! Value Error: Compass not found in line {ls} in {survexblock.survexfile.path}" print(self.insp + message) stash_data_issue( parser="survexleg", message=message, url=None, sb=(survexblock.survexfile.path) ) lcompass = invalid_compass try: lclino = ls[datastar["clino"]] except: print(("! Clino misread in", survexblock.survexfile.path)) print((" datastar:", datastar)) print((" Line:", ls)) message = f" ! Value Error: Clino misread in line '{sline.lower()}' {datastar=} {self.datastar=} {ls=} in\n{survexblock}\n{survexblock.survexfile}\n{survexblock.survexfile.path}" stash_data_issue( parser="survexleg", message=message, url=None, sb=(survexblock.survexfile.path) ) 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) stash_data_issue( parser="survexleg", message=message, url=None, sb=(survexblock.survexfile.path) ) survexleg.compass = invalid_compass # delete the object to save memory survexleg = None def LoadSurvexRef(self, survexblock, args): """Interpret the *ref record, and all the many variants""" # print(self.insp+ "*REF ---- '"+ args +"'") url = get_offending_filename(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(f"{self.insp} *REF quoted text so ignored:{args} in {survexblock.survexfile.path}") return if len(args) < 4: message = f" ! Empty or BAD *REF statement '{args}' in '{survexblock.survexfile.path}'" print(self.insp + message) stash_data_issue(parser="survex", message=message, url=url) return argsgps = self.rx_argsref.match(args) if argsgps: yr, letterx, wallet = argsgps.groups() else: perps = get_people_on_trip(survexblock) message = f" ! Wallet *REF bad in '{survexblock.survexfile.path}' malformed id '{args}' {perps}" print(self.insp + message) stash_data_issue(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) < 2050): message = " ! Wallet year out of bounds {yr} '{refscan}' {survexblock.survexfile.path}" print(self.insp + message) stash_data_issue(parser="survex", message=message, url=url) refscan = f"{yr}#{letterx}{wallet}" try: if int(wallet) > 99: message = f" ! Wallet *REF {refscan} - very big (more than 99) so probably wrong in '{survexblock.survexfile.path}'" print(self.insp + message) stash_data_issue(parser="survex", message=message, url=url) except: message = f" ! Wallet *REF {refscan} - not numeric in '{survexblock.survexfile.path}'" print(self.insp + message) stash_data_issue(parser="survex", message=message, url=url) manywallets = Wallet.objects.filter( walletname=refscan ) # assumes all wallets found in earlier pass of data import if manywallets: if len(manywallets) > 1: message = f" ! Wallet *REF {refscan} - more than one found {len(manywallets)} wallets in db with same id {survexblock.survexfile.path}" print(self.insp + message) stash_data_issue(parser="survex", message=message, url=url) if survexblock.scanswallet: if survexblock.scanswallet.walletname != refscan: message = f" ! Wallet *REF {refscan} in {survexblock.survexfile.path} - Already a DIFFERENT wallet is set for this block '{survexblock.scanswallet.walletname}'" print(self.insp + message) stash_data_issue(parser="survex", message=message, url=url) else: survexblock.scanswallet = manywallets[0] # this is a ForeignKey field survexblock.save() # This is where we should check that the wallet JSON contains a link to the survexfile # and that the JSON date and walletdate are set correctly to the survexblock date. set_walletdate(survexblock.scanswallet) else: perps = get_people_on_trip(survexblock) message = f" ! Wallet *REF bad in '{survexblock.survexfile.path}' '{refscan}' NOT in database i.e. wallet does not exist {perps}." print(self.insp + message) stash_data_issue(parser="survex", message=message, url=url) 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 = ( f" ! - Unrecognised *data normal statement '{args}' {survexblock.name}|{survexblock.survexpath}" ) print(message) print(message, file=sys.stderr) stash_data_issue( parser="survex", message=message, url=None, sb=(survexblock.survexfile.path) ) 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].lower() == "newline": message = f" ! - ABORT *data statement has NEWLINE in it in {survexblock.survexfile.path}. Not parsed by troggle. '{args}'" print(message) print(message, file=sys.stderr) stash_data_issue( parser="survex", message=message, url=None, sb=(survexblock.survexfile.path) ) return False 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] == "passage" or ls[0] == "nosurvey" or ls[0] == "diving" or ls[0] == "cylpolar": # message = " ! - *data {} blocks ignored. {}|{}" '{}' .format(ls[0].upper(), survexblock.name, survexblock.survexpath, args) # print(message) # print(message,file=sys.stderr) # stash_data_issue(parser='survex', message=message) self.datastar["type"] = ls[0] elif ls[0] == "cartesian": # We should not ignore this ?! Default for Germans ? # message = " ! - *data {} blocks ignored. {}|{}" '{}' .format(ls[0].upper(), survexblock.name, survexblock.survexpath, args) # print(message) # print(message,file=sys.stderr) # stash_data_issue(parser='survex', message=message) self.datastar["type"] = ls[0] else: message = f" ! - Unrecognised *data statement '{args}' {survexblock.name}|{survexblock.survexpath}" print(message) print(message, file=sys.stderr) stash_data_issue( parser="survex", message=message, url=None, sb=(survexblock.survexfile.path) ) self.datastar["type"] = ls[0] 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( f" ^ flagslist:{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"] is True or self.flagsstar["splay"] is True: self.flagsstar["skiplegs"] = True if debugprint: print( f" $ flagslist:{flags}", ) 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, depth): """Ignore surface, kataser and gpx *include survex files""" if not self.pending: self.pending = set() fpending = Path(settings.CAVEDESCRIPTIONS, "pendingcaves.txt") if fpending.is_file(): with open(fpending, "r") as fo: cids = fo.readlines() for cid in cids: id = cid.strip().rstrip("\n").upper() if cid.startswith("162"): self.pending.add(id) else: self.pending.add("1623-" + id) if headpath in self.ignorenoncave: message = f" - {headpath} is (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 (while creating '{includelabel}' sfile & sdirectory)" ) # print("\n"+message) # print("\n"+message,file=sys.stderr) return caveid = f"{headpath[6:10]}-{headpath[11:]}".upper() if caveid in self.pending: # Yes we didn't find this cave, but we know it is a pending one. So not an error. # print(f'! ALREADY PENDING {caveid}',file=sys.stderr) return id = caveid[5:] if id in self.pending: print(f"! ALREADY PENDING {id}", file=sys.stderr) return # It is too late to add it to the pending caves list here, they were already # processed in parsers/caves.py So we have to do a bespoke creation. cave = create_new_cave(includelabel) message = f" ! Warning: cave identifier '{caveid}'or {id} (guessed from file path) is not a known cave. Need to add to expoweb/cave_data/pending.txt ? In '{includelabel}.svx' at depth:[{len(depth)}]." print("\n" + message) print("\n" + message, file=sys.stderr) print(f"{self.pending}", end="", file=sys.stderr) stash_data_issue(parser="survex", message=message, url=None, sb=(includelabel)) # It is too late to add it to pending caves here, they were already processed in parsers/caves.py # and something else is creating them... # cave = create_new_cave(includelabel) 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 Inspects the parent folder of the survexfile and uses that to decide if this is a cave we know. If we see a duplicate cave, this is TOO LATE. It has already been included into the long linear file. We prevent duplication when the long linear file is created, so if we see a duplicate here, it is a serious error. The survexblock passed-in is not necessarily the parent. FIX THIS. """ global dup_includes if debugprint: print(f" # datastack in LoadSurvexFile:{svxid} 'type':", end="") for dict in self.datastack: print(f"'{dict['type'].upper()}' ", end="") print("") depth = " " * self.depthbegin print("{:2}{} - NEW survexfile:'{}'".format(self.depthbegin, depth, svxid)) headpath = os.path.dirname(svxid) newfile, created = SurvexFile.objects.update_or_create(path=svxid) if not created: dup_includes += 1 message = f" ! DUPLICATE SurvexFile '{svxid}' create attempt in LoadSurvexFile()" print(message) # print(message, file=sys.stderr) stash_data_issue(parser="survex", message=message, url=f"/survexfile/{svxid}") self.currentsurvexfile = newfile return # abort as everything already done for object creation 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) if not newdirectory: message = f" ! 'None' SurvexDirectory returned from GetSurvexDirectory({headpath})" print(message) print(message, file=sys.stderr) stash_data_issue(parser="survex", message=message, url=f"/survexfile/{svxid}") cave = IdentifyCave(headpath) # cave already exists in db if not cave: # probably a surface survey, or a cave in a new area # e.g. 1624 not previously managed, and not in the pending list self.ReportNonCaveIncludes(headpath, svxid, depth) #try again cave = IdentifyCave(headpath) if cave: newdirectory.cave = cave newfile.cave = cave # print(f"\n - New directory '{newdirectory}' for cave '{cave}'",file=sys.stderr) if not newfile.survexdirectory: message = f" ! SurvexDirectory NOT SET in new SurvexFile {svxid} " print(message) print(message, file=sys.stderr) stash_data_issue(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(f" # datastack end LoadSurvexFile:{svxid} 'type':", end="") for dict in self.datastack: print(f"'{dict['type'].upper()}' ", end="") print("") def ProcessIncludeLine(self, included): """As we read the long linear file, we come across lines telling us that the content from this point on is from a particular included file """ 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): """As we read the long linear file, we come across lines telling us that the we are about to pop back out of the contents of an included file Saves the current survexfile object in the db to include the data parsed from it""" global debugprint svxid = edulcni.groups()[0] if debugprint: depth = " " * self.depthbegin print(f"{self.depthbegin:2}{depth} - Edulcni survexfile:'{svxid}'") if svxid.lower() == debugprinttrigger.lower(): debugprint = False self.currentsurvexfile.save() self.currentsurvexfile = self.stacksvxfiles.pop() def TickSurvexQM(self, survexblock, qmtick): """Interpret the specially formatted comment which is a QM TICKED statement""" # Now we need to find the correct QM object. It will be in the same block and have the same number. try: # could try to search on blockname instead? # but the QMn TICK has to be in the same block anyway qm = QM.objects.filter(block=survexblock, number=int(qmtick.group(1))) except: # raise message = f' ! QM TICK find FAIL QM{qmtick.group(1)} date:"{qmtick.group(2)}" qmlist:"{qm}" in "{survexblock.survexfile.path}" + completion_description:"{qmtick.group(3)}" ' print(message) stash_data_issue( parser="survex", message=message, url=None, sb=(survexblock.survexfile.path) ) if len(qm) > 1: message = f' ! QM TICK MULTIPLE found FAIL QM{qmtick.group(1)} date:"{qmtick.group(2)}" in "{survexblock.survexfile.path}" + completion_description:"{qmtick.group(3)}" ' print(message) stash_data_issue( parser="survex", message=message, url=None, sb=(survexblock.survexfile.path) ) qm[0].ticked = True # qm[0].ticked_date = qmtick.group(2) # not in data model yet qm[0].completion_description = qmtick.group(3) qm[0].save() def LoadSurvexQM(self, survexblock, qmline): """Interpret the specially formatted comment which is a QM definition""" # r"(?i)^\s*QM(\d+)\s+(.+)\s+([\w\-\_]+)\.([\w\.\-]+)\s+(([\w\-]+)\.([\w\.\-]+)|\-)\s+(.+)$" # r"(?i)^\s*QM(\d+)\s+([^\s]+)\s+([^\s]+)\s+([^\s]+)\s+(.+)$" # rx_qm_tick QMnn TICK date comment # (r"(?i)^\s*QM(\d+)\s+TICK\s([\d\-]+)\s(.*)$") insp = self.insp # create a short, hopefully-unique name for this block to be used in the QM id if len(survexblock.name) < 7: blockname = survexblock.name else: blockname = survexblock.name[:6] + survexblock.name[-1:] # logslug = f'D{int(qmyear)}_{blockname}_{int(qm_no):03d}' qm_ticked = False # default qm_no = qmline.group(1) # this is NOT unique across multiple survex files qm_grade = qmline.group(2).strip().upper() # TICK or [a-dA-DvVxX?] if qm_grade == "TICK": self.TickSurvexQM(survexblock, qmline) return if qm_grade not in ["A", "B", "C", "D", "X"]: # "V", "?" not allowed in survex file QMs message = f" ! QM{qm_no} INVALID code '{qm_grade}' [{blockname}] '{survexblock.survexfile.path}'" print(insp + message) stash_data_issue( parser="survex", message=message, url=None, sb=(survexblock.survexfile.path) ) qm_nearest = qmline.group(3) # if qmline.group(3): # usual closest survey station # qm_nearest = qmline.group(3) # if qmline.group(4): # qm_nearest = qm_nearest + "." + qmline.group(4) resolution_station_name = qmline.group(4) if (resolution_station_name=="-"): pass else: qm_ticked = True # print(f"{survexblock.survexfile.cave} {survexblock}:{qm_no}{qm_grade} {qmline.group(4)}", file=sys.stderr) qm_notes = qmline.group(5) # qm_notes = qmline.group(8) # Spec of QM in SVX files: # ;Serial number grade(A/B/C/D/V/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 #;QM1 A B6 - see plan drawing there is definitely a QM # 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. if survexblock.survexfile.cave: survexblock.survexfile.cave.slug() self.fix_undated(survexblock) # null-op if already set try: expoyear = str(survexblock.date.year) except: print(f">> why is survexblock not set ?! in LoadSurvexQM()/n {survexblock.survexfile.path}") expoyear = "1970" try: qm = QM.objects.create( number=qm_no, # nearest_station=a_survex_station_object, # can be null resolution_station_name=resolution_station_name, nearest_station_name=qm_nearest, ticked=qm_ticked, grade=qm_grade.upper(), location_description=qm_notes, block=survexblock, # only set for survex-imported QMs blockname=blockname, # only set for survex-imported QMs expoyear=expoyear, cave=survexblock.survexfile.cave, ) qm.save except: qms = QM.objects.filter( number=qm_no, # nearest_station=a_survex_station_object, # can be null resolution_station_name=resolution_station_name, nearest_station_name=qm_nearest, ticked=qm_ticked, grade=qm_grade.upper(), location_description=qm_notes, block=survexblock, # only set for survex-imported QMs blockname=blockname, # only set for survex-imported QMs expoyear=expoyear, cave=survexblock.survexfile.cave, ) message = f" ! QM{qm_no} FAIL to create {qm_nearest} in'{survexblock.survexfile.path}' found {len(qms)}:{qms}" print(insp + message) stash_data_issue( parser="survex", message=message, url=None, sb=(survexblock.survexfile.path) ) def ProcessQM(self, survexblock, qml, comment): """Process the line beginning ;QM which is a QM new declaration or a QM TICK closing declaration. It _should_ recognise a non-numeric survey station ID, but currently doesn't. Valid QM types are [a-dA-DvVxX?] A-D, V for Vertical, X for horrible and ? for unknown """ # rx_qm : r"(?i)^\s*QM(\d+)\s+?(.+)\s+([\w\-\_]+)(\.([\w\.\-]+)?)\s+(([\w\-]+)\.([\w\.\-]+)|\-)\s+(.+)$) qmline = self.rx_qm.match(comment) if qmline: self.LoadSurvexQM(survexblock, qmline) else: message = f' ! QM Unrecognised as valid in "{survexblock.survexfile.path}" QM{qml.group(1)} "{qml.group(2)}" : regex failure typo?' print(message) stash_data_issue( parser="survex", message=message, url=None, sb=(survexblock.survexfile.path) ) 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[.;]?') # This _should_ also check that the QM survey point exists in the block depth = " " * self.depthbegin refline = self.rx_commref.match(comment) if refline: comment = self.rx_ref2.sub("", comment.strip()) print(f"{self.depthbegin:2}{depth} - rx_ref2 -- {comment=} in {survexblock.survexfile.path} :: {survexblock}") self.LoadSurvexRef(survexblock, comment) # handle # ; Messteam: Jörg Haussmann, Robert Eckardt, Thilo Müller # ; Zeichner: Thilo Müller # But none of these will be valid teammembers because they are not actually on our expo team = self.rx_commteam.match(comment) if team: # print(f'rx_commteam -- {comment=} in {survexblock.survexfile.path} :: {survexblock}') pass # rx_qm0 = re.compile(r"(?i)^\s*QM(\d+)\s+(.+)$") qml = self.rx_qm0.match(comment) if qml: self.ProcessQM(survexblock, qml, comment) included = self.rx_comminc.match(comment) # ;|*include means 'we have been included'; whereas *include means 'proceed to include' # No test here to check that this file has not already been included. Ouch. 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 get_cave(self, path): """Read the file path to a survex file and guesses the cave """ path_match = re.search(r"caves-(\d\d\d\d)/(\d+|\d\d\d\d-?\w+-\d+)/", path) if path_match: pos_cave = f"{path_match.group(1)}-{path_match.group(2)}" cave = getCaveByReference(pos_cave) return cave return None def LinearLoad(self, survexblock, path, collatefilename): """Loads a single survex file. Usually used to import all the survex files which have been collated into a single file: either the tree or the _unseens. Also used for loading a single file which has been edited online. Loads the begin/end blocks using a stack for labels. Uses the python generator idiom to avoid loading the whole file (21MB) into memory. """ blkid = None pathlist = None args = None oldflags = None blockcount = 0 self.lineno = 0 slengthtotal = 0.0 nlegstotal = 0 self.relativefilename = path 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 % 40 == 0: print(".", file=sys.stderr, end="") if blockcount % 1600 == 0: print("\n", file=sys.stderr, end="") mem = get_process_memory() print(f" - MEM: {mem:7.2f} MB in use", 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(f"{self.depthbegin:2}{depth} - Begin for :'{blkid}'") pathlist = "" for id in self.stackbegin: if len(id) > 0: pathlist += "." + id def printend(): nonlocal args depth = " " * self.depthbegin if debugprint: print(f"{self.depthbegin:2}{depth} - End from:'{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(f" # datastack at 1 *begin {blkid} 'type':", end="") for dict in self.datastack: print(f"'{dict['type'].upper()}' ", end="") print("") print(f"'{self.datastar['type'].upper()}' self.datastar ") # ------------ * DATA self.datastack.append(copy.deepcopy(self.datastar)) # ------------ * DATA if debugprint: print(f" # datastack at 2 *begin {blkid} 'type':", end="") for dict in self.datastack: print(f"'{dict['type'].upper()}' ", end="") print("") print(f"'{self.datastar['type'].upper()}' self.datastar ") # ------------ * FLAGS self.flagsstack.append(copy.deepcopy(self.flagsstar)) # ------------ * FLAGS pass def popblock(): nonlocal blkid nonlocal oldflags if debugprint: print(f" # datastack at *end '{blkid} 'type':", end="") for dict in self.datastack: print(f"'{dict['type'].upper()}' ", end="") print("") print(f"'{self.datastar['type'].upper()}' self.datastar ") # ------------ * DATA self.datastar = copy.deepcopy(self.datastack.pop()) # ------------ * DATA if debugprint: print(f" # datastack after *end '{blkid} 'type':", end="") for dict in self.datastack: print(f"'{dict['type'].upper()}' ", end="") print("") print(f"'{self.datastar['type'].upper()}' self.datastar ") # ------------ * FLAGS self.flagsstar = copy.deepcopy(self.flagsstack.pop()) # ------------ * FLAGS if debugprint: if oldflags["skiplegs"] != self.flagsstar["skiplegs"]: print(f" # POP 'any' flag now:'{self.flagsstar['skiplegs']}' was:{oldflags['skiplegs']} ") def starstatement(star): """Interprets a survex comamnd where * is the first character on the line, e.g. *begin""" 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.depthbegin += 1 self.stackbegin.append(blkid) self.unitsstack.append((self.units, self.unitsfactor)) self.legsnumberstack.append(self.legsnumber) self.slengthstack.append(self.slength) self.teaminheritstack.append(self.inheritteam) self.teamcurrentstack.append(self.currentteam) self.dateinheritstack.append(self.inheritdate) self.datecurrentstack.append(self.currentdate) pushblock() # PUSH state ++++++++++++++ self.legsnumber = 0 self.slength = 0.0 self.units = "metres" self.inheritteam = self.currentteam self.currentteam = set() # zero the current team when we start a new block self.inheritdate = self.currentdate self.currentdate = None # zero the current date when we start a new block printbegin() newsurvexblock = SurvexBlock( name=blkid, parent=survexblock, survexpath=pathlist, survexfile=self.currentsurvexfile, legsall=0, legslength=0.0, ) newsurvexblock.save() print(f"SB: #{newsurvexblock.id} '{newsurvexblock}' parent:{newsurvexblock.parent} f:{newsurvexblock.survexfile}") 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 self.fix_undated(survexblock) self.fix_anonymous(survexblock) 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 confirm_team_on_trip(survexblock) # POP state ++++++++++++++ popblock() self.inheritteam = self.teaminheritstack.pop() self.currentteam = self.teamcurrentstack.pop() self.inheritdate = self.dateinheritstack.pop() self.currentdate = self.datecurrentstack.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 = self.rx_quotedtitle.match(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(f" # CHANGE 'any' flag now:'{self.flagsstar['skiplegs']}' was:{oldflags['skiplegs']} ") elif self.rx_data.match(cmd): if self.LoadSurvexDataNormal(survexblock, args): pass else: # Abort, we do not cope with this *data format return 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 = f" ! -ERROR *include command not expected here {path}. Re-run a full Survex import." print(message) print(message, file=sys.stderr) stash_data_issue( parser="survex", message=message, ) else: self.LoadSurvexFallThrough(survexblock, args, cmd) # this is a python generator idiom. # see https://realpython.com/introduction-to-python-generators/ # this is the first use of generators in troggle (Oct.2022) and save 21 MB of memory with open(collatefilename, "r") as fcollate: for svxline in fcollate: 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) stash_data_issue(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, svxline) self.legsnumber = nlegstotal self.slength = slengthtotal def PushdownStackScan(self, survexblock, path, finname, flinear, fcollate): """Follows the *include links in all the survex files from the root file (usually 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. """ global stop_dup_warning def process_line(svxline): self.lineno += 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) stash_data_issue(parser="survex", message=message, url=None, sb=(path)) return # skip this survex file and all things *included in it includestmt = self.rx_include.match(svxline) if not includestmt: fcollate.write(f"{svxline.strip()}\n") 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 self.rx_include2.match(cmd): # rx_include2 = re.compile("(?i)include$") # if re.match("(?i)include$", cmd): includepath = os.path.normpath(os.path.join(os.path.split(path)[0], re.sub(r"\.svx$", "", args))) # normalises path syntax if self.never_seen(includepath, path): fullpath = os.path.join(settings.SURVEX_DATA, includepath + ".svx") self.RunSurvexIfNeeded(os.path.join(settings.SURVEX_DATA, includepath), path) self.check_unique_name(os.path.join(settings.SURVEX_DATA, includepath)) if os.path.isfile(fullpath): # -------------------------------------------------------- self.depthinclude += 1 # fininclude = open(fullpath,'r') finincludename = fullpath fcollate.write(f";|*include {includepath}\n") flinear.write(f"{self.depthinclude:2} {indent} *include {includepath}\n") push = includepath.lower() self.includestack.append(push) # ----------------- self.PushdownStackScan(survexblock, includepath, finincludename, 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) stash_data_issue(parser="survex", message=message, url=None, sb=(path)) flinear.write(f"{self.depthinclude:2} {indent} *edulcni {pop}\n") fcollate.write(f";|*edulcni {pop}\n") # fininclude.close() self.depthinclude -= 1 # -------------------------------------------------------- else: message = f" ! ERROR *include file '{includepath}' not found, listed in '{fin.name}'" print(message) print(message, file=sys.stderr) stash_data_issue(parser="survex", message=message, url=None, sb=(path)) elif self.rx_begin2.match(cmd): #elif re.match("(?i)begin$", cmd): self.depthbegin += 1 depth = " " * self.depthbegin if args: pushargs = args else: pushargs = " " self.stackbegin.append(pushargs.lower()) flinear.write(f" {self.depthbegin:2} {depth} *begin {args}\n") pass elif self.rx_end2.match(cmd): # elif re.match("(?i)end$", cmd): depth = " " * self.depthbegin flinear.write(f" {self.depthbegin:2} {depth} *end {args}\n") if not args: args = " " popargs = self.stackbegin.pop() if popargs != args.lower(): message = ( f"!! ERROR mismatch in BEGIN/END labels pop!=push '{popargs}'!='{args}'\n{self.stackbegin}" ) print(message) print(message, file=flinear) print(message, file=sys.stderr) stash_data_issue(parser="survex", message=message, url=None, sb=(path)) self.depthbegin -= 1 pass elif self.rx_title2.match(cmd): # elif re.match("(?i)title$", cmd): depth = " " * self.depthbegin flinear.write(f" {self.depthbegin:2} {depth} *title {args}\n") pass 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: # We have already used os.normpath() so this is OK. "/../" and "//" have been simplified already. if stop_dup_warning: # print("D",end="", file=sys.stderr) pass else: message = f" * Warning. Duplicate detected. We have already seen this *include '{path}' from another survex file. Detected at callcount:{self.callcount} depth:{self.depthinclude}" print(message) print(message, file=flinear) # print(message,file=sys.stderr) stash_data_issue(parser="survex", message=message, url=None, sb=(path)) if self.svxfileslist.count(path) > 2: message = f" ! ERROR. Should have been caught before this. Survex file already *included 2x. Probably an infinite loop so fix your *include statements that include this. Aborting. {path}" print(message) print(message, file=flinear) # print(message,file=sys.stderr) stash_data_issue(parser="survex", message=message, url=None, sb=(path)) return return try: # python generator idiom again. Not important here as these are small files with open(finname, "r") as fin: for svxline in fin: process_line(svxline) self.svxfileslist.append(path) except UnicodeDecodeError: # some bugger put an umlaut in a non-UTF survex file ?! message = f" ! ERROR *include file '{path}' in '{survexblock}' has UnicodeDecodeError. Omitted." print(message) print(message, file=sys.stderr) stash_data_issue(parser="survex", message=message, url=None, sb=(path)) return # skip this survex file and all things *included in it except: message = f" ! ERROR *include file '{path}' in '{survexblock}' has unexpected error on opening file. OMITTED!" print(message) print(message, file=sys.stderr) stash_data_issue(parser="survex", message=message, url=None, sb=(path)) raise return # skip this survex file and all things *included in it def never_seen(self, incpath, parent): """The _unseen files may include survex files we have already seen, and we do not want to process them again. For the _unseens this is not an error, but for the main *include tree it is an error. """ if incpath in self.uniquefile: self.uniquefile[incpath].append(parent) if self.svxpass == self.TREE: message = ( f" DUP: skipping non-unique survex filepath, '{incpath}' - #{len(self.uniquefile[incpath])} '{self.uniquefile[incpath]}'" ) print(message) stash_data_issue(parser='survex', message=message) for p in self.uniquefile[incpath]: if p in self.uniquefile: print(f"{p} <- {self.uniquefile[p]}") return False else: self.uniquefile[incpath] = [parent] return True def check_unique_name(self, fullpath): """This only checks whether the last bit of the name of the survex file is unique, e.g. "bigpitch", not whether the whole path of the survexfile has been seen before. We don't care about this any more. """ return def RunSurvexIfNeeded(self, fullpath, calledpath): 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. """ try: 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(str(settings.SURVEX_DATA), "") stash_data_issue(parser="xEntrances", message=message, url=url) print(message) print( "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 except: message = f' ! FAIL running cavern on survex file "{fullpath}" specified in *include in {calledpath} ' stash_data_issue(parser="survex", message=message) print(message) svxpath = Path(fullpath + ".svx") logpath = Path(fullpath + ".log") Path(svxpath).parent if not svxpath.is_file(): message = f' ! BAD. "{fullpath}" is not a file, specified in *include in {calledpath} ' stash_data_issue(parser="survex", message=message) print(message) return 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", f"{settings.CAVERN}"], capture_output=True, check=False, text=True) if sp.returncode != 0: message = f' ! Error running "which" on {settings.CAVERN}' stash_data_issue(parser="survex", message=message) print(message) print( "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 survex files This proceeds in 3 phases: 1. The root survex file is read and all the *include files are found, using PushdownStackScan() 2. All the other survex files in the :loser: repo are found, and their *includes found, using another PushdownStackScan() [duplicates omitted] 3. The combined expanded file containing all the survex data is parsed as a single file, using LinearLoad()""" global stop_dup_warning print(" - redirecting stdout to svxblks.log...") stdout_orig = sys.stdout # Redirect sys.stdout to the file sys.stdout = open("svxblks.log", "w") print(f" - Scanning Survex Blocks tree from {settings.SURVEX_TOPNAME}.svx ...", 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) print(f" - RunSurvexIfNeeded cavern on '{fullpathtotop}'", file=sys.stderr) svx_scan.RunSurvexIfNeeded(fullpathtotop, fullpathtotop) svx_scan.check_unique_name(fullpathtotop) svx_scan.uniquefile[str(survexfileroot)] = ["0"] indent = "" fcollate = open(collatefilename, "w") mem0 = get_process_memory() print(f" - MEM:{mem0:7.2f} MB START '{survexfileroot}'", file=sys.stderr) flinear = open("svxlinear.log", "w") flinear.write(f" - MEM:{mem0:7.2f} MB START '{survexfileroot.path}'\n") print(" ", file=sys.stderr, end="") finrootname = Path(settings.SURVEX_DATA, survexfileroot.path + ".svx") fcollate.write(f";*include {survexfileroot.path}\n") flinear.write(f"{svx_scan.depthinclude:2} {indent} *include {survexfileroot.path}\n") import cProfile import pstats from pstats import SortKey pr = cProfile.Profile() pr.enable() svx_scan.svxpass = svx_scan.TREE # ---------------------------------------------------------------- svx_scan.PushdownStackScan(survexblockroot, survexfileroot.path, finrootname, flinear, fcollate) # ---------------------------------------------------------------- svx_scan.svxpass = "" 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(f"{svx_scan.depthinclude:2} {indent} *edulcni {survexfileroot.path}\n") fcollate.write(f";*edulcni {survexfileroot.path}\n") mem1 = get_process_memory() flinear.write(f"\n - MEM:{mem1:.2f} MB STOP {survexfileroot.path}\n") flinear.write(f" - MEM:{mem1 - mem0:.3f} MB ADDITIONALLY USED\n") flinear.write(f" - {len(svx_scan.svxfileslist):,} survex files in linear include list \n") flinear.write(f" - {len(svx_scan.uniquefile):,} unique survex files in linear include list \n") for j in svx_scan.svxfileslist: if j not in svx_scan.uniquefile: flinear.write(f" - '{j}' {type(j)} not in unique list \n") for f in svx_scan.uniquefile: # flinear.write(f" - '{f}' {type(f)} {svx_scan.uniquefile[f]} \n") if len(svx_scan.uniquefile[f]) > 1: flinear.write(f" - '{f}' {type(f)} {svx_scan.uniquefile[f]} dup survex files \n") print(f"\n - {svx_scan.caverncount:,} runs of survex 'cavern' refreshing .3d files", file=sys.stderr) print(f" - {len(svx_scan.svxfileslist):,} survex files from tree in linear include list", file=sys.stderr) print(f" - {len(svx_scan.uniquefile):,} unique survex files from tree in linear include list", file=sys.stderr) mem1 = get_process_memory() print(f" - MEM:{mem1:7.2f} MB END ", file=sys.stderr) print(f" - MEM:{mem1 - mem0:7.3f} MB ADDITIONALLY USED", file=sys.stderr) # # Process all the omitted files in :loser: with some exceptions unseens = set() b = [] for p in Path(settings.SURVEX_DATA).rglob("*.svx"): if p.is_file(): po = p.relative_to(Path(settings.SURVEX_DATA)) pox = po.with_suffix("") if str(pox) not in svx_scan.svxfileslist: # print(f"[{pox}]", file=sys.stderr) unseens.add(pox) else: b.append(pox) if len(b) != len(svx_scan.svxfileslist): print( f" ! Mismatch. {len(b)} survex files found which should be {len(svx_scan.svxfileslist)} in main tree)", file=sys.stderr, ) unseensroot = re.sub(r"\.svx$", "", UNSEENS) excpts = ["surface/terrain", "kataster/kataster-boundaries", "template", "docs", "deprecated", "subsections", unseensroot] removals = [] for x in unseens: for o in excpts: if str(x).strip().startswith(o): removals.append(x) # special fix for .svx file not actually in survex format unseens.remove(Path("fixedpts/gps/gps00raw")) for x in removals: unseens.remove(x) print( f"\n - {len(unseens)} survex files found which were not included in main tree. ({len(svx_scan.svxfileslist)} in main tree)", file=sys.stderr, ) check_team_cache() print(f" -- Now loading the previously-omitted survex files.", file=sys.stderr) print(f" - (except: {excpts})", file=sys.stderr) with open(Path(settings.SURVEX_DATA, UNSEENS), "w") as u: u.write( f"; {len(unseens):,} survex files not *included by {settings.SURVEX_TOPNAME} (which are {len(svx_scan.svxfileslist):,} files)\n" ) u.write(f"; autogenerated by parser/survex.py from databasereset.py on '{datetime.now(timezone.utc)}'\n") u.write(f"; omitting any file beginning with {excpts}\n\n") u.write("*begin troggle_unseens\n") u.write("*title \"Collated unseen and unlinked survex files\"\n") for x in sorted(unseens): u.write(f" *include {x}\n") u.write("*end troggle_unseens\n") survexfileroot = survexblockroot.survexfile # i.e. SURVEX_TOPNAME only omit_scan = LoadingSurvex() omit_scan.callcount = 0 omit_scan.depthinclude = 0 fullpathtotop = os.path.join(survexfileroot.survexdirectory.path, UNSEENS) # copy the list to prime the next pass through the files omit_scan.svxfileslist = svx_scan.svxfileslist[:] svx_scan.svxfileslist = [] # free memory svx_scan = None # Hmm. Does this actually delete all the instance variables if they are lists, dicts etc.? omit_scan.check_unique_name(fullpathtotop) omit_scan.uniquefile[unseensroot] = ["0"] mem0 = get_process_memory() print(f" - MEM:{mem0:7.2f} MB START '{unseensroot}'", file=sys.stderr) # flinear = open('svxlinear.log', 'w') flinear.write(f" - MEM:{mem0:7.2f} MB START '{unseensroot}'\n") print(" ", file=sys.stderr, end="") # this is a bit tricky as some unseen files will *include files we have already seen, which # we should not process again. finrootname = fullpathtotop fcollate.write(f";*include {UNSEENS}\n") flinear.write(f"{omit_scan.depthinclude:2} {indent} *include {unseensroot}\n") omit_scan.svxpass = omit_scan.ODDS # stop_dup_warning = True # ---------------------------------------------------------------- omit_scan.PushdownStackScan(survexblockroot, unseensroot, finrootname, flinear, fcollate) # ---------------------------------------------------------------- # stop_dup_warning = False omit_scan.svxpass = "" flinear.write(f"{omit_scan.depthinclude:2} {indent} *edulcni {unseensroot}\n") fcollate.write(f";*edulcni {UNSEENS}\n") check_team_cache() mem1 = get_process_memory() flinear.write(f"\n - MEM:{mem1:.2f} MB STOP {UNSEENS} Unseen Oddments\n") flinear.write(f" - MEM:{mem1 - mem0:.3f} MB ADDITIONALLY USED Unseen Oddments\n") flinear.write(f" - {len(omit_scan.svxfileslist):,} survex files in linear include list Unseen Oddments \n") flinear.close() fcollate.close() print( f"\n - {omit_scan.caverncount:,} runs of survex 'cavern' refreshing .3d files in the unseen list", file=sys.stderr, ) print( f" - {len(omit_scan.svxfileslist):,} survex files in linear include list including previously unseen ones \n", file=sys.stderr, ) omit_scan = None # Hmm. Does this actually delete all the instance variables if they are lists, dicts etc.? mem1 = get_process_memory() print(f" - MEM:{mem1:7.2f} MB END ", file=sys.stderr) print(f" - MEM:{mem1 - mem0:7.3f} MB ADDITIONALLY USED", file=sys.stderr) # 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 # pr2 = cProfile.Profile() # pr2.enable() print(" ", file=sys.stderr, end="") # ---------------------------------------------------------------- svx_load.LinearLoad(survexblockroot, survexfileroot.path, collatefilename) # ---------------------------------------------------------------- # pr2.disable() # with open('LinearLoad.prof', 'w') as f: # ps = pstats.Stats(pr2, stream=f) # ps.sort_stats(SortKey.CUMULATIVE) # ps.print_stats() mem1 = get_process_memory() print(f"\n - MEM:{mem1:7.2f} MB STOP", file=sys.stderr) print(f" - MEM:{mem1 - mem0:7.3f} MB ADDITIONALLY USED", 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(f" - Number of SurvexDirectories: {len(svx_load.survexdict):,}") tf = 0 for d in svx_load.survexdict: tf += len(svx_load.survexdict[d]) print(f" - Number of SurvexFiles: {tf:,}") print(f" - Number of Survex legs: {legsnumber:,}") svx_load = None return legsnumber def display_contents(blocks): for b in blocks: print(f"B {b} {b.parent=} {b.expedition=}") sfs = SurvexFile.objects.filter(survexblock=b) for sf in sfs: print(f" SF {sf}") print(f" SD {sf.survexdirectory} {sf.survexdirectory.cave}") ws = Wallet.objects.filter(survexblock=b) for w in ws: print(f" W {w}") sfs = QM.objects.filter(block=b) for sf in sfs: print(f" QM {sf}") sfs = SurvexStation.objects.filter(block=b) for sf in sfs: print(f" SS {sf}") def parse_one_file(fpath): # --------------------------------------in progress------------------- """Parse just one file. Use when re-loading after editing. NOTE: *include lines are ignored. In the initial file parsing in databaseReset, the *include expansion is done in an earlier stange than LinearLoad(). By the time LinearLoad() is called, all the *include expansion has happened. """ def parse_new_svx(fpath, svx_load, svxfileroot=None): """We need a dummy survex block which has the survexfile being parsed as its .survexfile field. But it is used in two ways, it is also set as the parent block for the new blocks being created. This has to be fixed later. This all needs refactoring. We also need to re-plumb the fileroot after importing, so that the new survexfile appears in the survexdirectory lists """ if svxfileroot == None: svxfileroot = MakeFileRoot(fpath) svxfileroot.save() # It is vital that the block has attached the survexfile object which is being parsed. block_dummy = SurvexBlock( name="dummy", survexpath="", survexfile=svxfileroot, legsall=0, legslength=0.0 ) svxfileroot.save() block_dummy.save() newname = f"#{block_dummy.id}_" + str(Path(str(svxfileroot)).name) block_dummy.name = newname block_dummy.save() print(f" - block_dummy now '{block_dummy}' {type(block_dummy)} id={block_dummy.id} f:{block_dummy.survexfile}") svx_load.survexdict[svxfileroot.survexdirectory] = [] svx_load.survexdict[svxfileroot.survexdirectory].append(svxfileroot) svx_load.svxdirs[""] = svxfileroot.survexdirectory # ---------------------------------------------------------------- svx_load.LinearLoad(block_dummy, svxfileroot.path, fname) # ---------------------------------------------------------------- # Now we don't need or want the dummy any more block_dummy.delete() #svxfile.survexdirectory = global svx_load print(f"\n - Loading One Survex file '{fpath}'", file=sys.stderr) svx_load = LoadingSurvex() svx_load.survexdict = {} fname = Path(settings.SURVEX_DATA, (fpath + ".svx")) svxs = SurvexFile.objects.filter(path=fpath) if svxs: if len(svxs)>1: print(f" ! Mistake? More than one survex file object in database with the same file-path {svxs}") print(f" - Aborting file parsing & import into database.") return False print(f" - Pre-existing survexfile {svxs}.") existingsvx = SurvexFile.objects.get(path=fpath) existingcave = existingsvx.cave print(f" - survexfile id={existingsvx.id} {existingsvx} {existingcave}") sbs = existingsvx.survexblock_set.all() existingparent = None parents =set() if sbs: for sb in sbs: # print(f" - {sb.id} checking survex block {sb=}") try: if sb.parent: parents.add(sb.parent) # print(f" - adding {sb.parent=}") except: print(f" ! FAILURE to access sb.parent {sb=}\n ! {sb.parent_id=} ")# \n{dir(sb)} # even though the parent_id exists.. hmm. for sb in sbs: # print(f" - {sb.id} {sb.pk} {sb}") sb_keep = sb if sb not in parents: # print(f" - {sb.id} Deleting survex block {sb=}") sb.delete() if parents: # print(f" - parents get {parents}") if len(parents) > 1: print(f" - WARNING more than one parent survex block!") existingparent = parents.pop() # removes it parents.add(existingparent) # restores it print(f" - Reloading and parsing this survexfile '{fpath}' Loading...") # Logic is that we need an SB which links to the survexfile we are parsing for the parser # to work, but we delete all those before we start parsing. Urk. #=========== parse_new_svx(fpath, svx_load, svxfileroot=existingsvx) #=========== print(f" - survexfile id={existingsvx.id} update ") if parents: print(f" - parents set {parents}") sbs = existingsvx.survexblock_set.all() if len(sbs)<1: print(f" ! No survex blocks found. Parser failure...") for sb in sbs: print(f" - {sb.id} re-setting survex block parent{sb=}") sb.parent = existingparent # should be all the same sb.save() else: print(f" - Not seen this survexfile before '{fpath}' Loading. ..") #=========== parse_new_svx(fpath,svx_load) #=========== # print(f" - Number of SurvexDirectories: {len(svx_load.survexdict):,}") # tf = 0 # for d in svx_load.survexdict: # print(f" - SD: {d}") # tf += len(svx_load.survexdict[d]) # print(f" - Number of SurvexFiles: {tf:,}") # print(f" - Number of Survex legs: {svx_load.legsnumber:,}") # print(f" - Length of Survex legs: {svx_load.slength:.2f} m") svx_load = None return True def MakeSurvexFileRoot(): """Returns a file_object.path = SURVEX_TOPNAME associated with directory_object.path = SURVEX_DATA""" # find a cave, any cave.. smk = Cave.objects.filter(kataster_number="000") # returns a list, a QuerySet fileroot = SurvexFile(path=settings.SURVEX_TOPNAME, cave=None) fileroot.save() directoryroot = SurvexDirectory(path=settings.SURVEX_DATA, cave=smk[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 MakeFileRoot(fn): """Returns a file_object.path = _unseens.svx associated with directory_object.path = SURVEX_DATA or tries to find the primary survex file for this cave """ cave = IdentifyCave(fn) if not cave: if fn != UNSEENS: cave = create_new_cave(fn) print(f" - Making/finding a new root survexfile for this import: {fn}") fileroot = SurvexFile(path=fn, cave=cave) try: fileroot.survexdirectory = SurvexDirectory.objects.get(id=1) # default except: fileroot.survexdirectory = None if cave: # But setting the SurvexDirectory does work ! # The fluffy stuff is because of errors in the original setting of survex directories # which needs to be cleaned up.. for sd in cave.survexdirectory_set.filter(cave=cave): if f"{sd.primarysurvexfile}".replace("caves-","").startswith(f"{sd.cave}"[:4]): fileroot.survexdirectory = sd fileroot.save() fileroot.cave = cave print(f" - new fileroot {type(fileroot)} for {fn} with cave {cave} - {fileroot}") return fileroot def LoadSurvexBlocks(): global dup_includes mem1 = get_process_memory() print(f" - MEM:{mem1:7.2f} MB now ", file=sys.stderr) start = time.time() print(" - Flushing All Survex Blocks...") # why does this increase memory use by 20 MB ?! # We have foreign keys, Django needs to load the related objects # in order to resolve how the relation should handle the deletion: # https://docs.djangoproject.com/en/dev/ref/models/fields/#django.db.models.ForeignKey.on_delete SurvexBlock.objects.all().delete() SurvexFile.objects.all().delete() SurvexDirectory.objects.all().delete() SurvexPersonRole.objects.all().delete() SurvexStation.objects.all().delete() mem1 = get_process_memory() print(f" - MEM:{mem1:7.2f} MB now. Foreign key objects loaded on deletion. ", file=sys.stderr) print(" - Flushing survex Data Issues ") global dataissues dataissues = [] DataIssue.objects.filter(parser="survex").delete() DataIssue.objects.filter(parser="svxdate").delete() DataIssue.objects.filter(parser="survexleg").delete() DataIssue.objects.filter(parser="survexunits").delete() DataIssue.objects.filter(parser="survex team").delete() DataIssue.objects.filter(parser="entrances").delete() DataIssue.objects.filter(parser="xEntrances").delete() print(" - survex Data Issues flushed") mem1 = get_process_memory() print(f" - MEM:{mem1:7.2f} MB now ", file=sys.stderr) survexfileroot = MakeSurvexFileRoot() # this next makes a block_object assciated with a file_object.path = SURVEX_TOPNAME survexblockroot = SurvexBlock( name=ROOTBLOCK, survexpath="", survexfile=survexfileroot, legsall=0, legslength=0.0 ) # crashes here sometimes on MariaDB complaining that cave_id should not be null. But it should be. # django.db.utils.IntegrityError: (1048, "Column 'cave_id' cannot be null") # fix by restarting db on server # sudo service mariadb stop # sudo service mariadb start survexblockroot.save() omitsfileroot = MakeFileRoot(UNSEENS) survexomitsroot = SurvexBlock( name=OMITBLOCK, survexpath="", survexfile=omitsfileroot, legsall=0, legslength=0.0 ) survexomitsroot.save() print(" - Loading Survex Blocks...") memstart = get_process_memory() # ---------------------------------------------------------------- FindAndLoadSurvex(survexblockroot) # ---------------------------------------------------------------- memend = get_process_memory() print(f" - MEMORY start:{memstart:.3f} MB end:{memend:.3f} MB increase={memend - memstart:.3f} MB") survexblockroot.save() global person_pending_cache for sb in person_pending_cache: if len(person_pending_cache[sb]) > 0: print(f" ") message = f" ! PENDING team list not emptied {sb.survexfile.path} {len(person_pending_cache[sb])} people: {person_pending_cache[sb]}" stash_data_issue(parser="survex", message=message, url=None, sb=(sb.survexfile.path)) print(message) # duration = time.time() - start # print(f" - TIME: {duration:7.2f} s", file=sys.stderr) store_data_issues() # duration = time.time() - start # print(f" - TIME: {duration:7.2f} s", file=sys.stderr) if dup_includes > 0: print(f" - ERROR: There are {dup_includes} duplicate *includes in the final list") print(" - Loaded All Survex Blocks.")