import csv from datetime import datetime, date, time import os import re #import time import pickle import shelve from django.conf import settings from django.template.defaultfilters import slugify from django.utils.timezone import get_current_timezone, make_aware from troggle.core.models.troggle import DataIssue, Expedition from troggle.core.utils import TROG, save_carefully from troggle.core.models.caves import Cave, LogbookEntry, PersonTrip, GetCaveLookup from parsers.people import GetPersonExpeditionNameLookup ''' Parses and imports logbooks in all their wonderful confusion # When we edit logbook entries, allow a "?" after any piece of data to say we've frigged it and # it can be checked up later from the hard-copy if necessary; or it's not possible to determin (name, trip place, etc) ''' todo=''' - Put the object store 'trips' and the 'logdataissues' into TROG global object - refactor everything with some urgency, esp. LoadLogbookForExpedition() - Logbooks log.htm exist for 1983, 84, 85, 87, 88, 89 but have no full-working parser, or it is broken/incomplete and need hand-editing. - import/parse/re-export-as-html the 'artisanal-format' old logbooks so that we keep only a modern HTML05 format. Then we can retiure the old parsers and reduce the volume of code here substantially. - edit LoadLogbooks() to use coroutines to speed up import substantially, but perhaps we had better profile it first? - rewrite to use generators rather than storing everything intermediate in lists - to reduce memory impact. - the object store will need additional functions to replicate the persontrip calculation and storage. For the moment we leave all that to be done in the django db - We should ensure logbook.html is utf-8 and stop this crap: file_in = open(logbookfile,'rb') txt = file_in.read().decode("latin1") - this is a slow and uncertain function: cave = getCaveByReference(caveRef) ''' logentries = [] # the entire logbook for one year is a single object: a list of entries noncaveplaces = [ "QMplaceholder", "Journey", "Loser Plateau", "UNKNOWN", 'plateau', 'base camp', 'basecamp', 'top camp', 'topcamp' ] logdataissues = TROG['issues']['logdataissues'] trips ={} # # the logbook loading section # def set_trip_id(year, seq): tid= f"{year}.s{seq:02d}" return tid def GetTripPersons(trippeople, expedition, logtime_underground, tid=None): res = [ ] author = None round_bracket_regex = re.compile(r"[\(\[].*?[\)\]]") #print(f'# {tid}') for tripperson in re.split(r",|\+|&|&(?!\w+;)| and ", trippeople): tripperson = tripperson.strip() if not tid: tid = expedition.year + "." + tripperson + datetime.now().strftime("%S%f") # no good. Should be getting the tid mul = re.match(r"(.*?)$(?i)", tripperson) if mul: tripperson = mul.group(1).strip() if tripperson and tripperson[0] != '*': tripperson = re.sub(round_bracket_regex, "", tripperson).strip() personyear = GetPersonExpeditionNameLookup(expedition).get(tripperson.lower()) if not personyear: message = f" ! - {expedition.year} No name match for: '{tripperson}' " print(message) DataIssue.objects.create(parser='logbooks', message=message) logdataissues[tid]=message res.append((personyear, logtime_underground)) if mul: author = personyear if not author: if not res: return None, None author = res[-1][0] return res, author def EnterLogIntoDbase(date, place, title, text, trippeople, expedition, logtime_underground, entry_type="wiki", tid=None): """ saves a logbook entry and related persontrips Does NOT save the expeditionday_id - all NULLs. why? """ try: trippersons, author = GetTripPersons(trippeople, expedition, logtime_underground, tid=tid) except: message = f" ! - {expedition.year} Skipping logentry: {title} - GetTripPersons FAIL" DataIssue.objects.create(parser='logbooks', message=message) logdataissues["title"]=message print(message) return if not author: message = f" ! - {expedition.year} Skipping logentry: {title} - - no author for entry in year " DataIssue.objects.create(parser='logbooks', message=message) logdataissues["title"]=message print(message) return # This needs attention. The slug field is derived from 'title' # both GetCaveLookup() and GetTripCave() need to work together better. None of this data is *used* though? #tripCave = GetTripCave(place): lplace = place.lower() cave=None if lplace not in noncaveplaces: cave = GetCaveLookup().get(lplace) #Check for an existing copy of the current entry, and save expeditionday = expedition.get_expedition_day(date) lookupAttribs={'date':date, 'title':title} # 'cave' is converted to a string doing this, which renders as the cave slug. # but it is a db query which we should try to avoid - rewrite this #NEW slug for a logbook entry here! Use the unique id, not the title !!! slug = tid + "." + slugify(title)[:10] nonLookupAttribs={'place':place, 'text':text, 'expedition':expedition, 'cave_slug':str(cave), 'slug': slug, 'entry_type':entry_type} lbo, created=save_carefully(LogbookEntry, lookupAttribs, nonLookupAttribs) for tripperson, time_underground in trippersons: lookupAttribs={'personexpedition':tripperson, 'logbook_entry':lbo} nonLookupAttribs={'time_underground':time_underground, 'is_logbook_entry_author':(tripperson == author)} save_carefully(PersonTrip, lookupAttribs, nonLookupAttribs) # PersonTrip also saved in SetDatesFromLogbookEntries def ParseDate(tripdate, year): """ Interprets dates in the expo logbooks and returns a correct datetime.date object """ try: mdatestandard = re.match(r"(\d\d\d\d)-(\d\d)-(\d\d)", tripdate) mdategoof = re.match(r"(\d\d?)/0?(\d)/(20|19)?(\d\d)", tripdate) if mdatestandard: if not (mdatestandard.group(1) == year): message = " ! - Bad date (year) in logbook: " + tripdate + " - " + year DataIssue.objects.create(parser='logbooks', message=message) logdataissues["tripdate"]=message return datetime.date('1970', '01', '01') else: year, month, day = int(mdatestandard.group(1)), int(mdatestandard.group(2)), int(mdatestandard.group(3)) elif mdategoof: if not (not mdategoof.group(3) or mdategoof.group(3) == year[:2]): message = " ! - Bad date mdategoof.group(3) in logbook: " + tripdate + " - " + mdategoof.group(3) DataIssue.objects.create(parser='logbooks', message=message) logdataissues["tripdate"]=message return date('1970', '01', '01') else: yadd = int(year[:2]) * 100 day, month, year = int(mdategoof.group(1)), int(mdategoof.group(2)), int(mdategoof.group(4)) + yadd else: message = " ! - Bad date in logbook: " + tripdate + " - " + year DataIssue.objects.create(parser='logbooks', message=message) logdataissues["tripdate"]=message return date(year, month, day) except: message = " ! - Failed to parse date in logbook: " + tripdate + " - " + year DataIssue.objects.create(parser='logbooks', message=message) logdataissues["tripdate"]=message return date(year, month, day) # (2006 - not any more), 2008 - 2009 def Parselogwikitxt(year, expedition, txt): global logentries global logdataissues logbook_entry_count = 0 trippara = re.findall(r"===(.*?)===([\s\S]*?)(?====)", txt) for triphead, triptext in trippara: logbook_entry_count += 1 tid = set_trip_id(year,logbook_entry_count) tripheadp = triphead.split("|") if not (len(tripheadp) == 3): message = " ! - Bad no of items in tripdate in logbook: " + tripdate + " - " + tripheadp DataIssue.objects.create(parser='logbooks', message=message) logdataissues["tripdate"]=message tripdate, tripplace, trippeople = tripheadp tripsplace = tripplace.split(" - ") tripcave = tripsplace[0].strip() if len(tripsplace) == 1: tripsplace = tripsplace[0] else: tripsplace = tripsplace[1] #print(f"! LOGBOOK {year} {logbook_entry_count:2} {len(triptext):4} '{tripsplace}'") tul = re.findall(r"T/?U:?\s*(\d+(?:\.\d*)?|unknown)\s*(hrs|hours)?", triptext) if tul: tu = tul[0][0] else: tu = "" ldate = ParseDate(tripdate.strip(), year) tripid ="" entrytuple = (ldate, tripcave, tripsplace, triptext, trippeople, expedition, tu, "wiki", tripid) logentries.append(entrytuple) EnterLogIntoDbase(date = ldate, place = tripcave, title = tripplace, text = triptext, trippeople=trippeople, expedition=expedition, logtime_underground=0, tid=tid) EnterLogIntoObjStore(year, ldate, tripcave, tripplace, triptext, trippeople, tu, "wiki", tripid, logbook_entry_count, tid=tid) def EnterLogIntoObjStore(year, date, tripcave, triptitle, text, trippeople, tu, formattype, tripid1, seq, tid=None): # This will need additional functions to replicate the persontrip calculation and storage. For the # moment we leave all that to be done in the django db global trips # should be a singleton TROG eventually global logdataissues if tid in trips: tyear, tdate, *trest = trips[tid] msg = f" ! DUPLICATE on {tdate} id: '{tid}'" print(msg) DataIssue.objects.create(parser='logbooks', message=msg) tid = set_trip_id(str(date),seq) #print(" - De-dup ",seq, tid) logdataissues[tid]=msg trips[tid] = (year, date, tripcave, triptitle, text, trippeople, tu, formattype) ## copy a lot of checking functionality here from EnterLogIntoDbase() # GetTripPersons is a db query, so this will need to be put in ObjStore before this will work.. # or design a different way to do it. #trippersons, author = GetTripPersons(trippeople, expedition, logtime_underground) # if not author: # print(" ! - Skipping logentry: " + title + " - no RECOGNISED author for entry") # message = " ! - Skipping logentry: %s - no author for entry in year '%s'" % (title, expedition.year) # DataIssue.objects.create(parser='logbooks', message=message) # logdataissues[tid+"author"]=message pass # 2002, 2004, 2005, 2007, 2010 - now # 2006 wiki text is incomplete, but the html all there. So using this parser now. def Parseloghtmltxt(year, expedition, txt): global logentries global logdataissues tripparas = re.findall(r"([\s\S]*?)(?=.*?\s*

)? # second date \s*(?:\s*)? \s*(.*?)(?:

)? \s*\s*(.*?) \s*\s*(.*?) ([\s\S]*?) \s*(?:\s*(.*?))? \s*$ ''', trippara) if not s: if not re.search(r"Rigging Guide", trippara): msg = " !- Logbook. Can't parse: {} entry:{}".format(trippara, logbook_entry_count) print(msg) DataIssue.objects.create(parser='logbooks', message=msg) logdataissues[tid]=msg continue tripid, tripid1, tripdate, trippeople, triptitle, triptext, tu = s.groups() ldate = ParseDate(tripdate.strip(), year) triptitles = triptitle.split(" - ") if len(triptitles) >= 2: tripcave = triptitles[0] else: tripcave = "UNKNOWN" ltriptext = re.sub(r"

", "", triptext) ltriptext = re.sub(r"\s*?\n\s*", " ", ltriptext) ltriptext = re.sub(r"

", "

", ltriptext).strip() entrytuple = (ldate, tripcave, triptitle, ltriptext, trippeople, expedition, tu, "html", tripid1) logentries.append(entrytuple) EnterLogIntoDbase(date = ldate, place = tripcave, title = triptitle, text = ltriptext, trippeople=trippeople, expedition=expedition, logtime_underground=0, entry_type="html", tid=tid) EnterLogIntoObjStore(year, ldate, tripcave, triptitle, ltriptext, trippeople, tu, "html", tripid1, logbook_entry_count, tid=tid) # main parser for 1991 - 2001. simpler because the data has been hacked so much to fit it # trying it out for years 1982 - 1990 too. Some logbook editing required by hand.. def Parseloghtml01(year, expedition, txt): global logentries global logdataissues errorcount = 0 tripparas = re.findall(r"([\s\S]*?)(?=)?(.*?)(.*)$(?i)", trippara) if not s: message = " ! - Skipping logentry {year} failure to parse header: " + tid + trippara[:300] + "..." DataIssue.objects.create(parser='logbooks', message=message) logdataissues[tid]=message print(message) break tripheader, triptext = s.group(1), s.group(2) mtripid = re.search(r']*>", "", tripheader) tripdate, triptitle, trippeople = tripheader.split("|") ldate = ParseDate(tripdate.strip(), year) mtu = re.search(r']*>(T/?U.*)', triptext) if mtu: tu = mtu.group(1) triptext = triptext[:mtu.start(0)] + triptext[mtu.end():] else: tu = "" triptitles = triptitle.split(" - ") tripcave = triptitles[0].strip() ltriptext = triptext mtail = re.search(r'(?:[^<]*|\s|/|-|&||\((?:same day|\d+)\))*$', ltriptext) if mtail: ltriptext = ltriptext[:mtail.start(0)] ltriptext = re.sub(r"

", "", ltriptext) ltriptext = re.sub(r"\s*?\n\s*", " ", ltriptext) ltriptext = re.sub(r"

|
", "\n\n", ltriptext).strip() ltriptext = re.sub(r"", "_", ltriptext) ltriptext = re.sub(r"", "''", ltriptext) ltriptext = re.sub(r"", "'''", ltriptext) if ltriptext == "": message = " ! - Zero content for logbook entry!: " + tid DataIssue.objects.create(parser='logbooks', message=message) logdataissues[tid]=message print(message) entrytuple = (ldate, tripcave, triptitle, ltriptext, trippeople, expedition, tu, "html01", tripid) logentries.append(entrytuple) try: EnterLogIntoDbase(date=ldate, place=tripcave, title=triptitle, text=ltriptext, trippeople=trippeople, expedition=expedition, logtime_underground=0, entry_type="html", tid=tid) except: message = " ! - Enter log entry into database FAIL exception in: " + tid DataIssue.objects.create(parser='logbooks', message=message) logdataissues[tid]=message print(message) try: EnterLogIntoObjStore(year, ldate, tripcave, triptitle, ltriptext, trippeople, tu, "html01", tripid, logbook_entry_count, tid=tid) except: message = " ! - Enter log entry into ObjectStore FAIL exception in: " + tid DataIssue.objects.create(parser='logbooks', message=message) logdataissues[tid]=message print(message) except: message = f" ! - Skipping logentry {year} due to exception in: {tid}" DataIssue.objects.create(parser='logbooks', message=message) logdataissues[tid]=message print(message) errorcount += 1 if errorcount >5 : message = f" !!- TOO MANY ERRORS - aborting at '{tid}' logbook: {year}" DataIssue.objects.create(parser='logbooks', message=message) logdataissues[tid]=message print(message) return # parser for 2003 def Parseloghtml03(year, expedition, txt): global logentries global logdataissues tripparas = re.findall(r"([\s\S]*?)(?=(.*?)

(.*)$", trippara) if not ( s ) : message = " ! - Skipping logentry {year} on failure to parse Parseloghtml03: {} {} {}...".format(tid,s,trippara[:300]) DataIssue.objects.create(parser='logbooks', message=message) logdataissues[tid]=message print(message) break tripheader, triptext = s.group(1), s.group(2) tripheader = re.sub(r" ", " ", tripheader) tripheader = re.sub(r"\s+", " ", tripheader).strip() sheader = tripheader.split(" -- ") tu = "" if re.match("T/U|Time underwater", sheader[-1]): tu = sheader.pop() if len(sheader) != 3: print((" ! Header not three pieces", sheader)) tripdate, triptitle, trippeople = sheader ldate = ParseDate(tripdate.strip(), year) triptitles = triptitle.split(" , ") if len(triptitles) >= 2: tripcave = triptitles[0] else: tripcave = "UNKNOWN" ltriptext = re.sub(r"

", "", triptext) ltriptext = re.sub(r"\s*?\n\s*", " ", ltriptext) ltriptext = re.sub(r"

", "\n\n", ltriptext).strip() ltriptext = re.sub(r"[^\s0-9a-zA-Z\-.,:;'!&()\[\]<>?=+*%]", "_NONASCII_", ltriptext) entrytuple = (ldate, tripcave, triptitle, ltriptext, trippeople, expedition, tu, "html03", tid) logentries.append(entrytuple) EnterLogIntoDbase(date = ldate, place = tripcave, title = triptitle, text = ltriptext, trippeople=trippeople, expedition=expedition, logtime_underground=0, entry_type="html", tid=tid) EnterLogIntoObjStore(year, ldate, tripcave, triptitle, ltriptext, trippeople, tu, "html03", tid, logbook_entry_count, tid=tid) def SetDatesFromLogbookEntries(expedition): """ Sets the date_from and date_to field for an expedition based on persontrips. Then sets the expedition date_from and date_to based on the personexpeditions. """ # Probably a faster way to do this. This uses a lot of db queries, but we have all this # in memory.. for personexpedition in expedition.personexpedition_set.all(): persontrips = personexpedition.persontrip_set.order_by('logbook_entry__date') # sequencing is difficult to do lprevpersontrip = None for persontrip in persontrips: persontrip.persontrip_prev = lprevpersontrip if lprevpersontrip: lprevpersontrip.persontrip_next = persontrip lprevpersontrip.save() persontrip.persontrip_next = None lprevpersontrip = persontrip persontrip.save() # also saved in EnterLogIntoDbase. MERGE these to speed up import. def LoadLogbookForExpedition(expedition, expect): """ Parses all logbook entries for one expedition If a cache is found it uses it. If not found, or fails sanity checks, parses source file. """ # absolutely horrid. REFACTOR THIS (all my fault..) global logentries global logdataissues logbook_parseable = False logbook_cached = False yearlinks = settings.LOGBOOK_PARSER_SETTINGS expologbase = os.path.join(settings.EXPOWEB, "years") logentries=[] def validcache(year,n): if year != expedition: print(" ! year != expedition ",year, expedition ) return False if len(logentries) != n: print(" ! len(logentries) != n ",len(logentries), n ) return False if n != expect: print(" ! n != expect ",n, expect ) return False return True def cleanerrors(year): global logdataissues dataissues = DataIssue.objects.filter(parser='logbooks') for di in dataissues: ph = year if re.search(ph, di.message) is not None: #print(f' - CLEANING dataissue {di.message}') di.delete() #print(f' - CLEAN {year} {len(logdataissues)} {type(logdataissues)} data issues for this year') dellist = [] for key, value in logdataissues.items(): #print(f' - CLEANING logdataissues [{key}]: {value}') if key.startswith(year): #print(f' - CLEANING logdataissues [{key:12}]: {value} ') dellist.append(key) for i in dellist: del logdataissues[i] cleanerrors(expedition.year) if expedition.year in yearlinks: logbookfile = os.path.join(expologbase, yearlinks[expedition.year][0]) expedition.logbookfile = yearlinks[expedition.year][0] parsefunc = yearlinks[expedition.year][1] else: logbookfile = os.path.join(expologbase, expedition.year, settings.DEFAULT_LOGBOOK_FILE) expedition.logbookfile = settings.DEFAULT_LOGBOOK_FILE parsefunc = settings.DEFAULT_LOGBOOK_PARSER cache_filename = logbookfile + ".cache" expedition.save() try: bad_cache = False now = time.time() cache_t = os.path.getmtime(cache_filename) if os.path.getmtime(logbookfile) - cache_t > 2: # at least 2 secs later bad_cache= True if now - cache_t > 30*24*60*60: bad_cache= True if bad_cache: print(" - ! Cache is either stale or more than 30 days old. Deleting it.") os.remove(cache_filename) logentries=[] print(" ! Removed stale or corrupt cache file") raise print(" - Reading cache: " + cache_filename, end='') try: with open(cache_filename, "rb") as f: year,n,logentries = pickle.load(f) if validcache(year,n): print(" -- Loaded ", len(logentries), " log entries") logbook_cached = True else: print(" !- Told to expect ", expect, " but ", len(logentries), " found in cache") raise except: print(" ! Failed to load corrupt cache. (Or I was told to ignore it). Deleting it.") os.remove(cache_filename) logentries=[] raise except : # no cache found #print(" - No cache \"" + cache_filename +"\"") try: file_in = open(logbookfile,'rb') txt = file_in.read().decode("latin1") file_in.close() logbook_parseable = True print((" - Using: " + parsefunc + " to parse " + logbookfile)) except (IOError): logbook_parseable = False print((" ! Couldn't open logbook " + logbookfile)) if logbook_parseable: parser = globals()[parsefunc] parser(expedition.year, expedition, txt) # this launches the parser SetDatesFromLogbookEntries(expedition) if len(logentries) >0: print(" - Cacheing " , len(logentries), " log entries") with open(cache_filename, "wb") as fc: logbk=(expedition,len(logentries),logentries) pickle.dump(logbk, fc, protocol=4) else: print(" ! NO TRIP entries found in logbook, check the syntax.") if logbook_cached: # working on this bit... i=0 for entrytuple in range(len(logentries)): date, tripcave, triptitle, text, trippeople, expedition, logtime_underground, entry_type, tripid1 = logentries[i] EnterLogIntoDbase(date, tripcave, triptitle, text, trippeople, expedition, 0, entry_type) EnterLogIntoObjStore(expedition.year, date, tripcave, triptitle, text, trippeople, logtime_underground, entry_type, tripid1, i) i +=1 SetDatesFromLogbookEntries(expedition) return len(logentries) def LoadLogbooks(): """ This is the master function for parsing all logbooks into the Troggle database. Parser settings appropriate for each year are set in settings.py LOGBOOK_PARSER_SETTINGS. This should be rewritten to use coroutines to load all logbooks from disc in parallel. """ global logdataissues logdataissues = {} DataIssue.objects.filter(parser='logbooks').delete() expos = Expedition.objects.all() if len(expos) <= 1: print(" ! No expeditions found. Load 'people' first.\n") nologbook = ["1976", "1977", "1978", "1979", "1980", "1981", "1987", "1988", "1989", "1986", "2020",] entries = {"2021": 0, "2019": 20, "2018": 74, "2017": 60, "2016": 81, "2015": 79, "2014": 65, "2013": 51, "2012": 75, "2011": 68, "2010": 22, "2009": 52, "2008": 49, "2007": 111, "2006": 60, "2005": 55, "2004": 76, "2003": 40, "2002": 31, "2001": 48, "2000": 54, "1999": 79, "1998": 43, "1997": 53, "1996": 94, "1995": 41, "1994": 32, "1993": 41, "1992": 61, "1991": 38, "1990": 87, "1989": 1,"1988": 1,"1987": 1, "1985": 21,"1984": 19,"1983": 22,"1982": 42,} # Logbooks log.htm exist for 1983, 84, 85, 87, 88, 89 but have no full-working parser, or need hand-editing. try: os.remove("loadlogbk.log") except OSError: pass nlbe={} expd ={} with open("loadlogbk.log", "a") as log: for expo in expos: TROG['pagecache']['expedition'][expo.year] = None # clear cache if expo.year not in nologbook: print((" - Logbook for: " + expo.year)) numentries = LoadLogbookForExpedition(expo, entries[expo.year]) # this actually loads the logbook for one year log.write("{} {:5d} should be {}\n".format(expo.year, numentries, entries[expo.year])) nlbe[expo.year]=numentries expd[expo.year]= 0 print("** total trips in ObjStore:", len(trips)) #for i in logdataissues: # print("{:15s}: {}".format(i, logdataissues[i])) for lbe in trips: year, date, tripcave, triptitle, text, trippeople, tu, formattype = trips[lbe] expd[year] += 1 yt=0 for y in expd: print("{} {}".format(y, expd[y]), nlbe[y]) yt += expd[y] print("total {} log entries in all expeditions".format(yt)) with shelve.open('logbktrips.shelve',writeback=True) as odb: for lbe in trips: odb[lbe]=trips[lbe] odb.sync() odb.close() # dateRegex = re.compile(r'(\d\d\d\d)-(\d\d)-(\d\d)', re.S) # expeditionYearRegex = re.compile(r'(.*?)', re.S) # titleRegex = re.compile(r'

(.*?)

', re.S) # reportRegex = re.compile(r'(.*)\s*', re.S) # personRegex = re.compile(r'(.*?)', re.S) # nameAuthorRegex = re.compile(r'(.*?)', re.S) # TURegex = re.compile(r'([0-9]*\.?[0-9]+)', re.S) # locationRegex = re.compile(r'(.*?)', re.S) # caveRegex = re.compile(r'(.*?)', re.S)