import csv import os import re import time from random import randint from datetime import datetime, date from pathlib import Path 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 See detailed explanation of the complete process: https://expo.survex.com/handbook/computing/logbooks-parsing.html ''' todo=''' - refactor everything with some urgency, esp. LoadLogbookForExpedition() - remove the TROG and lbo things since we need the database for multiuser access? Or not? - profile the code to find bad repetitive things, of which there are many. - far too many uses of Django field dereferencing to get values, which is SLOW - Logbooks 1987, 1988, 1989 all crash on MySql - but not sqlite - with db constraint fail. Edit logbook to fix. - import/parse/re-export-as-html the 'artisanal-format' old logbooks so that we keep only a modern HTML05 format. Then we can retire the old parsers and reduce the volume of code here substantially. - rewrite to use generators rather than storing everything intermediate in lists - to reduce memory impact. - 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) - use Fixtures https://docs.djangoproject.com/en/4.1/ref/django-admin/#django-admin-loaddata to cache data for old logbooks. New design needed, with a mechanism for flagging fixtures as outdated after edits. ''' MAX_LOGBOOK_ENTRY_TITLE_LENGTH = 200 BLOG_PARSER_SETTINGS = { # "2022": ("ukcavingblog.html", "parser_blog"), # "2019": ("ukcavingblog.html", "parser_blog"), # "2018": ("ukcavingblog.html", "parser_blog"), # "2017": ("ukcavingblog.html", "parser_blog"), # now folded in to logbooks.html } DEFAULT_LOGBOOK_FILE = "logbook.html" DEFAULT_LOGBOOK_PARSER = "parser_html" # All years since 2002 use the default value for Logbook parser LOGBOOK_PARSER_SETTINGS = { # "2009": ("2009logbook.txt", "wiki_parser"), # converted to html # "2008": ("2008logbook.txt", "wiki_parser"), # converted to html # "2006": ("logbook/logbook_06.txt", "wiki_parser"), # converted to html "2002": ("logbook.html", "parser_html"), "2001": ("log.htm", "parser_html_01"), "2000": ("log.htm", "parser_html_01"), "1999": ("log.htm", "parser_html_01"), "1998": ("log.htm", "parser_html_01"), "1997": ("log.htm", "parser_html_01"), "1996": ("log.htm", "parser_html_01"), "1995": ("log.htm", "parser_html_01"), "1994": ("log.htm", "parser_html_01"), "1993": ("log.htm", "parser_html_01"), "1992": ("log.htm", "parser_html_01"), "1991": ("log.htm", "parser_html_01"), "1990": ("log.htm", "parser_html_01"), "1989": ("log.htm", "parser_html_01"), #crashes MySQL "1988": ("log.htm", "parser_html_01"), #crashes MySQL #"1987": ("log.htm", "parser_02"), #crashes MySQL "1987": ("logbook.html", "parser_html"), "1985": ("logbook.html", "parser_html"), "1984": ("logbook.html", "parser_html"), "1983": ("logbook.html", "parser_html"), "1982": ("logbook.html", "parser_html"), } entries = { "2022": 86, "2019": 56, "2018": 100, "2017": 76, "2016": 83, "2015": 79, "2014": 65, "2013": 51, "2012": 75, "2011": 68, "2010": 22, "2009": 53, "2008": 49, "2007": 113, "2006": 60, "2005": 55, "2004": 76, "2003": 42, "2002": 31, "2001": 48, "2000": 54, "1999": 79, "1998": 43, "1997": 53, "1996": 95, "1995": 42, "1994": 32, "1993": 41, "1992": 62, "1991": 39, "1990": 87, "1989": 1,"1988": 1,"1987": 34, "1985": 24, "1984": 32, "1983": 52, "1982": 42,} # Logbooks log.htm exist for 87, 88, 89 but have no full-working parser, or need hand-editing. logentries = [] # the entire logbook for one year is a single object: a list of entries noncaveplaces = [ "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 rx_tripperson = re.compile(r'(?i)(.*?)$') rx_round_bracket = re.compile(r"[\(\[].*?[\)\]]") def GetTripPersons(trippeople, expedition, logtime_underground, tid=None): res = [ ] author = None # print(f'# {tid}') # print(f" - {tid} '{trippeople}' ") for tripperson in re.split(r",|\+|&|&(?!\w+;)| and ", trippeople): tripperson = tripperson.strip() # mul = re.match(r"(?i)(.*?)$", tripperson) mul = rx_tripperson.match(tripperson) if mul: tripperson = mul.group(1).strip() if tripperson and tripperson[0] != '*': tripperson = re.sub(rx_round_bracket, "", tripperson).strip() # these aliases should be moved to people.py GetPersonExpeditionNameLookup(expedition) if tripperson =="Wiggy": tripperson = "Phil Wigglesworth" if tripperson =="Animal": tripperson = "Mike Richardson" if tripperson =="MikeTA": tripperson = "Mike Richardson" if tripperson =="CavingPig": tripperson = "Elaine Oliver" if tripperson =="nobrotson": tripperson = "Rob Watson" if tripperson =="Tinywoman": tripperson = "Nadia" if tripperson =="tcacrossley": tripperson = "Tom Crossley" if tripperson =="Samouse1": tripperson = "Todd Rye" personyear = GetPersonExpeditionNameLookup(expedition).get(tripperson.lower()) if not personyear: message = f" ! - {expedition.year} No name match for: '{tripperson}' in entry {tid=} for this expedition year." 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 "", 0 author = res[-1][0] # the previous valid person and a time of 0 hours #print(f" - {tid} [{author.person}] '{res[0][0].person}'...") return res, author def EnterLogIntoDbase(date, place, title, text, trippeople, expedition, logtime_underground, tid=None): """ saves a logbook entry and related persontrips Does NOT save the expeditionday_id - all NULLs. why? Because we are deprecating expeditionday ! troggle.log shows that we are creating lots of duplicates, which is no no problem with SQL as they just overwrite but we are saving the same thing too many times.. Until 18 Dec.2022, this was overwriting logbook entries for the same date with the same title, because lookupAttribs={'date':date, 'title':title} """ # Nasty hack, must tidy this up.. if logtime_underground: try: logtime_underground = float(logtime_underground) except: # print(f"logtime_underground = {logtime_underground}") tu_match = re.match(r"(T/U:\s*)?(\d+[.]?\d*).*", logtime_underground) if tu_match: # print(f"logtime_underground = {tu_match.group(2)}") logtime_underground = float(tu_match.group(2)) else: logtime_underground = 0 else: logtime_underground = 0 try: trippersons, author = GetTripPersons(trippeople, expedition, logtime_underground, tid=tid) # print(f" - {author} - {logtime_underground}") except: message = f" ! - {expedition.year} Skipping logentry: {title} - GetTripPersons FAIL" DataIssue.objects.create(parser='logbooks', message=message) logdataissues["title"]=message print(message) raise return if not author: message = f" ! - {expedition.year} Warning: logentry: {title} - no expo member author for entry '{tid}'" 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) y = str(date)[:4] text = text.replace(' src="', f' src="/years/{y}/' ) text = text.replace(" src='", f" src='/years/{y}/" ) text = text.replace(f' src="/years/{y}//years/{y}/', f' src="/years/{y}/' ) text = text.replace(f" src='/years/{y}//years/{y}/", f" src='/years/{y}/" ) text = text.replace('\t', '' ) text = text.replace('\n\n\n', '\n\n' ) #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! Unique id + slugified title fragment if tid is not None: slug = tid # slug = tid + "_" + slugify(title)[:10].replace('-','_') else: slug = str(randint(1000,9999)) + "_" + slugify(title)[:10].replace('-','_') nonLookupAttribs={'place':place, 'text':text, 'expedition':expedition, 'time_underground':logtime_underground, 'cave_slug':str(cave), 'slug': slug} # This creates the lbo instance of LogbookEntry lbo, created=save_carefully(LogbookEntry, lookupAttribs, nonLookupAttribs) # for PersonTrip time_underground is float (decimal hours) for tripperson, time_underground in trippersons: # print(f" - {tid} '{tripperson}' author:{tripperson == author}") lookupAttribs={'personexpedition':tripperson, 'logbook_entry':lbo} nonLookupAttribs={'time_underground':time_underground, 'is_logbook_entry_author':(tripperson == author)} # this creates the PersonTrip instance. save_carefully(PersonTrip, lookupAttribs, nonLookupAttribs) def ParseDate(tripdate, year): """ Interprets dates in the expo logbooks and returns a correct datetime.date object """ dummydate = date(1970, 1, 1) month = 1 day = 1 # message = f" ! - Trying to parse date in logbook: {tripdate} - {year}" # print(message) 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 = f" ! - Bad date (year) in logbook: {tripdate} - {year}" DataIssue.objects.create(parser='logbooks', message=message) logdataissues["tripdate"]=message return dummydate 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 dummydate else: yadd = int(year[:2]) * 100 day, month, year = int(mdategoof.group(1)), int(mdategoof.group(2)), int(mdategoof.group(4)) + yadd else: year = 1970 message = f" ! - Bad date in logbook: {tripdate} - {year}" DataIssue.objects.create(parser='logbooks', message=message) logdataissues["tripdate"]=message return date(year, month, day) except: message = f" ! - Failed to parse date in logbook: {tripdate} - {year}" DataIssue.objects.create(parser='logbooks', message=message) logdataissues["tripdate"]=message return datetime.date(1970, 1, 1) # 2002, 2004 - now def parser_html(year, expedition, txt, seq=""): '''This uses some of the more obscure capabilities of regular expressions, see https://docs.python.org/3/library/re.html ''' global logentries global logdataissues # extract front material and stash for later use when rebuilding from list of entries headmatch = re.match(r"(?i)(?s).*
]*>(.*?))? # second date \s*(?:\s*)? \s*
)? \s*
)? # second date \s*(?:\s*)? \s*
)? \s*
", "
", ltriptext).strip()
triptitle = triptitle.strip()
entrytuple = (ldate, tripcave, triptitle, ltriptext,
trippeople, expedition, tu, tripid1)
logentries.append(entrytuple)
# 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.. place
def parser_html_01(year, expedition, txt, seq=""):
global logentries
global logdataissues
errorcount = 0
tripparas = re.findall(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|/|-|&|?p>|\((?:same day|\d+)\))*$', 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"?u>", "_", ltriptext) ltriptext = re.sub(r"?i>", "''", ltriptext) ltriptext = re.sub(r"?b>", "'''", ltriptext) ltriptext = re.sub(r"", "
", ltriptext).strip()
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, tid)
logentries.append(entrytuple)
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
raise
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
# variant parser for 1987
def parser_02(year, expedition, txt, seq=""):
'''This uses some of the more obscure capabilities of regular expressions,
see https://docs.python.org/3/library/re.html
'''
global logentries
global logdataissues
errorcount = 0
# extract front material and stash for later use when rebuilding from list of entries
headmatch = re.match(r"(?i)(?s).*
]*>(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|/|-|&|?p>|\((?:same day|\d+)\))*$', 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"?u>", "_", ltriptext) ltriptext = re.sub(r"?i>", "''", ltriptext) ltriptext = re.sub(r"?b>", "'''", ltriptext) ltriptext = re.sub(r"", "
", ltriptext).strip()
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, tid)
logentries.append(entrytuple)
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
raise
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
def parser_blog(year, expedition, txt, sq=""):
'''Parses the format of web pages collected as 'Save As HTML" from the UK Caving blog website.
Note that the entries have dates and authors, but no titles.
See detailed explanation of the complete process:
https://expo.survex.com/handbook/computing/logbooks-parsing.html
https://expo.survex.com/handbook/computing/log-blog-parsing.html
This uses some of the more obscure capabilities of regular expressions,
see https://docs.python.org/3/library/re.html
BLOG entries have this structure: