troggle-unchained/parsers/logbooks.py

import csv
import datetime
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 import DataIssue, Expedition
from troggle.core.models_caves import Cave, LogbookEntry, PersonTrip, GetCaveLookup
from parsers.people import GetPersonExpeditionNameLookup
from utils import save_carefully

# 
# 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)
#

#
# the logbook loading section
#
def GetTripPersons(trippeople, expedition, logtime_underground):
    res = [ ]
    author = None
    round_bracket_regex = re.compile(r"[\(\[].*?[\)\]]")
    for tripperson in re.split(r",|\+|&|&(?!\w+;)| and ", trippeople):
        tripperson = tripperson.strip()
        mul = re.match(r"<u>(.*?)</u>$(?i)", tripperson)
        if mul:
            tripperson = mul.group(1).strip()
        if tripperson and tripperson[0] != '*':
            #assert tripperson in personyearmap, "'%s' << %s\n\n %s" % (tripperson, trippeople, personyearmap)
            tripperson = re.sub(round_bracket_regex, "", tripperson).strip()
            personyear = GetPersonExpeditionNameLookup(expedition).get(tripperson.lower())
            if not personyear:
                message = "No name match for: ||'%s'|| in year '%s'" % (tripperson, expedition.year)
                print(message)
                DataIssue.objects.create(parser='logbooks', message=message)
                logdataissues[expedition.year + "~" + tripperson]=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 GetTripCave(place):  
    try:
        katastNumRes=[]
        katastNumRes=list(Cave.objects.filter(kataster_number=int(place)))
    except ValueError:
        message = " ! - ValueError on finding place " + str(place) + " entered. " + tripdate + " - " + year 
        DataIssue.objects.create(parser='logbooks', message=message)
        logdataissues["author"]=message
    officialNameRes=list(Cave.objects.filter(official_name=place))
    tripCaveRes=officialNameRes+katastNumRes

    if len(tripCaveRes)==1:
        return tripCaveRes[0]
    elif len(tripCaveRes)>1:
        message = " ! - Ambiguous place " + str(place) + " entered. " + tripdate + " - " + year + " " + str(tripCaveRes)
        DataIssue.objects.create(parser='logbooks', message=message)
        logdataissues["author"]=message
        return tripCaveRes[0]
    else:
        print((" " , place))
        message = " ! - No cave found for place:" + str(place) + tripdate + " - " + year 
        DataIssue.objects.create(parser='logbooks', message=message)
        logdataissues["author"]=message
        return None


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 = {}
trips ={}

def EnterLogIntoDbase(date, place, title, text, trippeople, expedition, logtime_underground, entry_type="wiki"):
    """ saves a logbook entry and related persontrips 
    Does NOT save the expeditionday_id  - all NULLs. why?
    """

    trippersons, author = GetTripPersons(trippeople, expedition, logtime_underground)
    if not author:
        print(" ! - Skipping logentry: " + title + " - no 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["title"]=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)
        # message = " ! - '" + lplace + "' place not in noncaveplaces."
        # print(message)
        # DataIssue.objects.create(parser='logbooks', message=message)

    #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
    nonLookupAttribs={'place':place, 'text':text, 'expedition':expedition, 'cave_slug':str(cave), 'slug':slugify(title)[:50], '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  """
    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:
        assert mdatestandard.group(1) == year, (tripdate, year)
        year, month, day = int(mdatestandard.group(1)), int(mdatestandard.group(2)), int(mdatestandard.group(3))
    elif mdategoof:
        assert not mdategoof.group(3) or mdategoof.group(3) == year[:2], mdategoof.groups()
        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
        assert False, tripdate

    return datetime.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
        tripheadp = triphead.split("|")
        assert len(tripheadp) == 3, (tripheadp, triptext)
        tripdate, tripplace, trippeople = tripheadp
        tripsplace = tripplace.split(" - ")
        tripcave = tripsplace[0].strip()

        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)
        
        EnterLogIntoObjStore(year, ldate, tripcave, tripplace, triptext, trippeople, 
                tu, "wiki", tripid, logbook_entry_count)


def EnterLogIntoObjStore(year, date, tripcave, triptitle, text, trippeople, tu, formattype, tripid1, seq):
    # 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 class object in models.py eventually
    global logdataissues

    if tripid1 is None or tripid1 =="":
       tid= "n{}-s{:02d}".format(str(date),seq)
       #print(" - New id ",tid)
    else:
        tid= tripid1
    if tid in trips:
        msg = "   ! DUPLICATE id .{}. {} ~{}~".format(tid, trips[tid][0], trips[tid][1])
        print(msg)
        DataIssue.objects.create(parser='logbooks', message=msg)
        tid= "d{}-s{:02d}".format(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"<hr\s*/>([\s\S]*?)(?=<hr)", txt)
    logbook_entry_count = 0
    for trippara in tripparas:
        logbook_entry_count += 1
        
        s = re.match(r'''(?x)(?:\s*<div\sclass="tripdate"\sid=".*?">.*?</div>\s*<p>)?  # second date
                            \s*(?:<a\s+id="(.*?)"\s*/>\s*</a>)?
                            \s*<div\s+class="tripdate"\s*(?:id="(.*?)")?>(.*?)</div>(?:<p>)?
                            \s*<div\s+class="trippeople">\s*(.*?)</div>
                            \s*<div\s+class="triptitle">\s*(.*?)</div>
                            ([\s\S]*?)
                            \s*(?:<div\s+class="timeug">\s*(.*?)</div>)?
                            \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"</p>", "", triptext)
        ltriptext = re.sub(r"\s*?\n\s*", " ", ltriptext)
        ltriptext = re.sub(r"<p>", "</br></br>", 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")

        EnterLogIntoObjStore(year, ldate, tripcave, triptitle, ltriptext, trippeople, tu, 
                            "html", tripid1, logbook_entry_count)

# 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"<hr[\s/]*>([\s\S]*?)(?=<hr)", txt)
    logbook_entry_count = 0
    for trippara in tripparas:
        logbook_entry_count += 1
        try:
            tripentry = year + "." + str(logbook_entry_count)
            s = re.match("(?s)\s*(?:<p>)?(.*?)</?p>(.*)$(?i)", trippara)
            if not s:
                message = " ! - Skipping logentry on failure to parse header: " + tripentry + trippara[:300] + "..."
                DataIssue.objects.create(parser='logbooks', message=message)
                logdataissues[tripentry]=message
                print(message)
                break
            tripheader, triptext = s.group(1), s.group(2)
            mtripid = re.search(r'<a id="(.*?)"', tripheader)
            tripid = mtripid and mtripid.group(1) or ""
            tripheader = re.sub(r"</?(?:[ab]|span)[^>]*>", "", tripheader)

            tripdate, triptitle, trippeople = tripheader.split("|")
            ldate = ParseDate(tripdate.strip(), year)
        
            mtu = re.search(r'<p[^>]*>(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'(?:<a href="[^"]*">[^<]*</a>|\s|/|-|&amp;|</?p>|\((?:same day|\d+)\))*$', ltriptext)
            if mtail:
                ltriptext = ltriptext[:mtail.start(0)]
            ltriptext = re.sub(r"</p>", "", ltriptext)
            ltriptext = re.sub(r"\s*?\n\s*", " ", ltriptext)
            ltriptext = re.sub(r"<p>|<br>", "\n\n", ltriptext).strip()
            ltriptext = re.sub(r"</?u>", "_", ltriptext)
            ltriptext = re.sub(r"</?i>", "''", ltriptext)
            ltriptext = re.sub(r"</?b>", "'''", ltriptext)

            entrytuple = (ldate, tripcave, triptitle, ltriptext, 
                    trippeople, expedition, tu, "html01", tripid)
            logentries.append(entrytuple)

            EnterLogIntoDbase(date=ldate, place=tripcave, title=triptitle, text=ltriptext,
                              trippeople=trippeople, expedition=expedition, logtime_underground=0,
                              entry_type="html")

            EnterLogIntoObjStore(year, ldate, tripcave, triptitle, ltriptext, trippeople, tu, 
                                "html01", tripid, logbook_entry_count)
        except:
            message = " ! - Skipping logentry due to exception in: " + tripentry 
            DataIssue.objects.create(parser='logbooks', message=message)
            logdataissues[tripentry]=message
            print(message)
            errorcount += 1
            if errorcount >5 :
                message = " !!- TOO MANY ERRORS - aborting logbook: " + year
                DataIssue.objects.create(parser='logbooks', message=message)
                logdataissues[tripentry]=message
                print(message)
                return

# parser for 2003
def Parseloghtml03(year, expedition, txt):
    global logentries
    global logdataissues

    tripparas = re.findall(r"<hr\s*/>([\s\S]*?)(?=<hr)", txt)
    logbook_entry_count = 0
    for trippara in tripparas:
        logbook_entry_count += 1
        
        s = re.match("(?s)\s*<p>(.*?)</p>(.*)$", trippara)
        assert s, trippara
        tripheader, triptext = s.group(1), s.group(2)
        tripheader = re.sub(r"&nbsp;", " ", 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"</p>", "", triptext)
        ltriptext = re.sub(r"\s*?\n\s*", " ", ltriptext)
        ltriptext = re.sub(r"<p>", "\n\n", ltriptext).strip()
        ltriptext = re.sub(r"[^\s0-9a-zA-Z\-.,:;'!&()\[\]<>?=+*%]", "_NONASCII_", ltriptext)

        tid= "n{}-s{:02d}".format(str(ldate),logbook_entry_count)

        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")

        EnterLogIntoObjStore(year, ldate, tripcave, triptitle, ltriptext, trippeople, tu, 
                            "html03", tid, logbook_entry_count)


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
    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
    
    if expedition.year in yearlinks:
        logbookfile = os.path.join(expologbase, yearlinks[expedition.year][0])
        parsefunc   = yearlinks[expedition.year][1]
    else:
        logbookfile = os.path.join(expologbase, expedition.year, settings.DEFAULT_LOGBOOK_FILE)
        parsefunc   = settings.DEFAULT_LOGBOOK_PARSER
    cache_filename = logbookfile + ".cache"

    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("  !- Should be ", expect, " but ", len(logentries), " found in cache")
                raise
        except:
            print("   ! Failed to load corrupt cache. 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)
        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", 
        "1983", "1984", "1985", "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": 1,"1984": 1,"1983": 1,"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:
            if expo.year not in nologbook:
                print((" - Logbook for: " + expo.year))
                numentries = LoadLogbookForExpedition(expo, entries[expo.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'<span\s+class="date">(\d\d\d\d)-(\d\d)-(\d\d)</span>', re.S)
expeditionYearRegex = re.compile(r'<span\s+class="expeditionyear">(.*?)</span>', re.S)
titleRegex = re.compile(r'<H1>(.*?)</H1>', re.S)
reportRegex = re.compile(r'<div\s+class="report">(.*)</div>\s*</body>', re.S)
personRegex = re.compile(r'<div\s+class="person">(.*?)</div>', re.S)
nameAuthorRegex = re.compile(r'<span\s+class="name(,author|)">(.*?)</span>', re.S)
TURegex = re.compile(r'<span\s+class="TU">([0-9]*\.?[0-9]+)</span>', re.S)
locationRegex = re.compile(r'<span\s+class="location">(.*?)</span>', re.S)
caveRegex = re.compile(r'<span\s+class="cave">(.*?)</span>', re.S)

def parseAutoLogBookEntry(filename):
    errors = []
    f = open(filename, "r")
    contents = f.read()
    f.close()

    dateMatch = dateRegex.search(contents)
    if dateMatch:
        year, month, day = [int(x) for x in dateMatch.groups()]
        date = datetime.date(year, month, day)
    else:
        errors.append(" - Date could not be found")

    expeditionYearMatch = expeditionYearRegex.search(contents)
    if expeditionYearMatch:
        try:
            expedition = Expedition.objects.get(year = expeditionYearMatch.groups()[0])
            personExpeditionNameLookup = GetPersonExpeditionNameLookup(expedition)
        except Expedition.DoesNotExist:
            errors.append(" - Expedition not in database")   
    else:
        errors.append(" - Expedition Year could not be parsed")   

    titleMatch = titleRegex.search(contents)
    if titleMatch:
        title, = titleMatch.groups()
        if len(title) > settings.MAX_LOGBOOK_ENTRY_TITLE_LENGTH:
            errors.append(" - Title too long")   
    else:
        errors.append(" - Title could not be found") 

    caveMatch = caveRegex.search(contents)
    if caveMatch:
        caveRef, = caveMatch.groups()
        try:
            # this is a slow and uncertain function:
            cave = getCaveByReference(caveRef)
        except AssertionError:
            cave = None
            errors.append(" - Cave not found in database")   
    else:
        cave = None

    locationMatch = locationRegex.search(contents)
    if locationMatch:
        location, = locationMatch.groups() 
    else:
        location = None
       
    if cave is None and location is None:
        errors.append(" - Location nor cave could not be found") 

    reportMatch = reportRegex.search(contents)
    if reportMatch:
        report, = reportMatch.groups()
    else:
        errors.append(" - Contents could not be found") 
    if errors:
        return errors # Easiest to bail out at this point as we need to make sure that we know which expedition to look for people from.
    people = []
    for personMatch in personRegex.findall(contents):
       nameAuthorMatch = nameAuthorRegex.search(contents)
       if nameAuthorMatch:
           author, name = nameAuthorMatch.groups()
           if name.lower() in personExpeditionNameLookup:
               personExpo = personExpeditionNameLookup[name.lower()]
           else:
               errors.append(" - Person could not be found in database")
           author = bool(author)
       else:
           errors.append(" - Persons name could not be found")
       
       TUMatch = TURegex.search(contents)
       if TUMatch:
           TU, = TUMatch.groups()
       else:
           errors.append(" - TU could not be found")
       if not errors:
           people.append((name, author, TU))
    if errors:
        return errors # Bail out before committing to the database
    logbookEntry = LogbookEntry(date = date, 
                                       expedition  = expedition,
                                       title = title, cave = cave, place = location, 
                                       text = report, slug = slugify(title)[:50],
                                       filename = filename)
    logbookEntry.save()
    for name, author, TU in people:
        PersonTrip(personexpedition = personExpo,  
                          time_underground = TU, 
                          logbook_entry = logbookEntry, 
                          is_logbook_entry_author = author).save()
    print(logbookEntry)