expo/troggle
2
0
Fork 1
mirror of https://expo.survex.com/repositories/troggle/.git synced 2024-11-22 23:31:52 +00:00
Code Issues 19 Projects 3 Activity
9169abdb05
troggle/parsers/logbooks.py
Philip Sargent 9169abdb05 2019 blog edits
2022-12-20 15:18:07 +00:00

797 lines
34 KiB
Python
Raw Blame History

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 New design needed, with a mechanism for flagging fixtures as outdated after edits.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_html_01"), #crashes MySQL
"1985": ("log.htm", "parser_html_01"),
"1984": ("log.htm", "parser_html_01"),
"1983": ("log.htm", "parser_html_01"),
"1982": ("log.htm", "parser_html_01"),
}
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": 1,
"1985": 24, "1984": 32, "1983": 52, "1982": 42,}
# Logbooks log.htm exist for 1983, 84, 85, 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)<u>(.*?)</u>$')
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",|\+|&amp;|&(?!\w+;)| and ", trippeople):
tripperson = tripperson.strip()
# mul = re.match(r"(?i)<u>(.*?)</u>$", 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).*<body[^>]*>(.*?)<hr.*", txt)
headpara = headmatch.groups()[0].strip()
# print(f" - headpara:\n'{headpara}'")
if(len(headpara)>0):
frontpath = Path(settings.EXPOWEB, "years", year, "frontmatter.html")
with open(frontpath,"w") as front:
front.write(headpara+"\n")
tripparas = re.findall(r"<hr\s*/>([\s\S]*?)(?=<hr)", txt)
logbook_entry_count = 0
for trippara in tripparas:
logbook_entry_count += 1
tid = set_trip_id(year,logbook_entry_count)
# print(f' - new tid:{tid} lbe count: {logbook_entry_count}')
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 s:
tripid, tripid1, tripdate, trippeople, triptitle, triptext, tu = s.groups()
else: # allow title and people to be swapped in order
msg = f" !- {year} Can't parse:{logbook_entry_count} '{trippara[:50]}'..."
print(msg)
DataIssue.objects.create(parser='logbooks', message=msg)
logdataissues[tid]=msg
s2 = 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="triptitle">\s*(.*?)</div>
\s*<div\s+class="trippeople">\s*(.*?)</div>
([\s\S]*?)
\s*(?:<div\s+class="timeug">\s*(.*?)</div>)?
\s*$
''', trippara)
if s2:
tripid, tripid1, tripdate, triptitle, trippeople, triptext, tu = s2.groups()
else:
# if not re.search(r"Rigging Guide", trippara):
msg = f" !- Logbook. Can't parse entry on 2nd pass:{logbook_entry_count} '{trippara[:50]}'..."
print(msg)
DataIssue.objects.create(parser='logbooks', message=msg)
logdataissues[tid]=msg
continue
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()
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=""):
'''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).*<body[^>]*>(.*?)<hr.*", txt)
headpara = headmatch.groups()[0].strip()
# print(f" - headpara:\n'{headpara}'")
if(len(headpara)>0):
frontpath = Path(settings.EXPOWEB, "years", year, "frontmatter.html")
with open(frontpath,"w") as front:
front.write(headpara+"\n")
tripparas = re.findall(r"<hr[\s/]*>([\s\S]*?)(?=<hr)", txt)
logbook_entry_count = 0
for trippara in tripparas:
logbook_entry_count += 1
tid = set_trip_id(year,logbook_entry_count)
# print(f" #0 - tid: {tid}")
try:
#print(f" #1 - tid: {tid}")
s = re.match(r"(?i)(?s)\s*(?:<p>)?(.*?)</?p>(.*)$", 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
try:
tripheader, triptext = s.group(1), s.group(2)
except:
message = f" ! - Fail to set tripheader, triptext. trip:<{tid}> s:'{s}'"
DataIssue.objects.create(parser='logbooks', message=message)
logdataissues[tid]=message
print(message)
# mtripid = re.search(r'<a id="(.*?)"', tripheader)
# if not mtripid:
# message = f" ! - A tag id not found. Never mind. Not needed. trip:<{tid}> header:'{tripheader}'"
# DataIssue.objects.create(parser='logbooks', message=message)
# logdataissues[tid]=message
# print(message)
# tripid = mtripid and mtripid.group(1) or ""
# print(f" # - mtripid: {mtripid}")
tripheader = re.sub(r"</?(?:[ab]|span)[^>]*>", "", tripheader)
#print(f" #2 - tid: {tid}")
try:
tripdate, triptitle, trippeople = tripheader.split("|")
except:
message = f" ! - Fail 3 to split out date|title|people. trip:<{tid}> '{tripheader.split('|')}'"
DataIssue.objects.create(parser='logbooks', message=message)
logdataissues[tid]=message
print(message)
try:
tripdate, triptitle = tripheader.split("|")
trippeople = "GUESS ANON"
except:
message = f" ! - Skipping logentry {year} Fail 2 to split out date|title (anon). trip:<{tid}> '{tripheader.split('|')}' CRASHES MySQL !"
DataIssue.objects.create(parser='logbooks', message=message)
logdataissues[tid]=message
print(message)
break
#print(f" #3 - tid: {tid}")
triptitle = triptitle.strip()
ldate = ParseDate(tripdate.strip(), year)
#print(f" # - tid: {tid} <{tripdate}> <{triptitle}> <{trippeople}>")
#print(f" #4 - tid: {tid}")
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"</?u>", "_", ltriptext)
ltriptext = re.sub(r"</?i>", "''", ltriptext)
ltriptext = re.sub(r"</?b>", "'''", ltriptext)
ltriptext = re.sub(r"<p>", "<br /><br />", 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:
<article ... data-author="Tinywoman" data-content="post-298780" id="js-post-298780">
<article class="message-body js-selectToQuote">
</article>
</article>
So the content is nested inside the header. Attachments (images) come after the content.
'''
global logentries
global logdataissues
errorcount = 0
tripheads = re.findall(r"<article class=\"message message--post js-post js-inlineModContainer\s*\"\s*([\s\S]*?)(?=</article)", txt)
if not ( tripheads ) :
message = f" ! - Skipping on failure to parse article header: {txt[:500]}"
print(message)
# (?= is a non-consuming match, see https://docs.python.org/3/library/re.html
tripparas = re.findall(r"<article class=\"message-body js-selectToQuote\"\>\s*([\s\S]*?)(</article[^>]*>)([\s\S]*?)(?=</article)", txt)
if not ( tripparas ) :
message = f" ! - Skipping on failure to parse article content: {txt[:500]}"
print(message)
if (len(tripheads) !=len(tripparas)):
print(f"{len(tripheads)} != {len(tripparas)}")
print(f"{len(tripheads)} - {len(tripparas)}")
location = "Plateau" # best guess, fix manually later
tu = 0
logbook_entry_count = 0
for i in range(0, len(tripparas)):
tripstuff = tripparas[i]
attach = tripstuff[2]
# note use on non-greedy *? regex idiom here
attach = re.sub(r"<div class=\"file-content\">[\s\S]*?(?=</li>)","",attach)
attach = re.sub(r"<footer[\s\S]*(</footer>)","",attach)
tripcontent = tripstuff[0] + attach
#print(f"{i} - {len(tripstuff)} - {tripstuff[1]}")
triphead = tripheads[i]
logbook_entry_count += 1
tid = set_trip_id(year,logbook_entry_count) +"_blog" + sq
# print(f" - tid: {tid}")
# data-author="tcacrossley"
match_author = re.search(r".*data-author=\"([^\"]*)\" data-content=.*", triphead)
if not ( match_author ) :
message = f" ! - Skipping logentry {year}:{logbook_entry_count} on failure to parse data-author {tid} {triphead[:400]}..."
DataIssue.objects.create(parser='logbooks', message=message)
logdataissues[tid]=message
print(message)
break
trippeople = match_author.group(1)
# print(f" - tid: {tid} {trippeople}")
# datetime="2019-07-11T13:16:18+0100"
match_datetime = re.search(r".*datetime=\"([^\"]*)\" data-time=.*", triphead)
if not ( match_datetime ) :
message = f" ! - Skipping logentry {year}:{logbook_entry_count} on failure to parse datetime {tid} {triphead[:400]}..."
DataIssue.objects.create(parser='logbooks', message=message)
logdataissues[tid]=message
print(message)
break
datestamp = match_datetime.group(1)
try:
tripdate = datetime.fromisoformat(datestamp)
except:
message = f" ! - FROMISOFORMAT fail logentry {year}:{logbook_entry_count} {tid} '{datestamp}'"
DataIssue.objects.create(parser='logbooks', message=message)
logdataissues[tid]=message
print(message)
# fallback, ignore the timestamp bits:
tripdate = datetime.fromisoformat(datestamp[0:10])
# print(f" - tid: {tid} '{trippeople}' '{tripdate}'")
# tripname must have the location then a hyphen at the beginning as it is ignored by export function
location = "Unknown"
tripname = f"Expo - UK Caving Blog{sq} post {logbook_entry_count}" # must be unique for a given date
tripcontent = re.sub(r"(width=\"\d+\")","",tripcontent)
tripcontent = re.sub(r"height=\"\d+\"","",tripcontent)
tripcontent = re.sub(r"width: \d+px","",tripcontent)
tripcontent = re.sub(r"\n\n+","\n\n",tripcontent)
tripcontent = re.sub(r"<hr\s*>","",tripcontent)
tripcontent = f"\n\n<!-- Content parsed from UK Caving Blog -->\nBlog Author: {trippeople}" + tripcontent
entrytuple = (tripdate, location, tripname, tripcontent,
trippeople, expedition, tu, tid)
logentries.append(entrytuple)
def LoadLogbookForExpedition(expedition, clean=True):
""" Parses all logbook entries for one expedition
if clean==True then it deletes all entries for this year first.
"""
global logentries
# absolutely horrid. REFACTOR THIS (all my fault..)
global logdataissues
global entries
logbook_parseable = False
yearlinks = LOGBOOK_PARSER_SETTINGS
expologbase = os.path.join(settings.EXPOWEB, "years")
logentries=[]
year = expedition.year
expect = entries[year]
# print(" - Logbook for: " + year)
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]
if (clean):
cleanerrors(year)
if year in yearlinks:
yearfile, yearparser = yearlinks[year]
logbookpath = Path(yearfile)
expedition.logbookfile = yearfile
parsefunc = yearparser
# print(f" - Logbook file {yearfile} using parser {yearparser}")
else:
logbookpath = Path(DEFAULT_LOGBOOK_FILE)
expedition.logbookfile = DEFAULT_LOGBOOK_FILE
parsefunc = DEFAULT_LOGBOOK_PARSER
expedition.save()
lbes = LogbookEntry.objects.filter(expedition=expedition)
if (clean):
for lbe in lbes:
lbe.delete()
for sq in ["", "2", "3", "4"]: # cope with blog saved as many separate files
lb = Path(expologbase, year, logbookpath.stem + sq + logbookpath.suffix)
if not (lb.is_file()):
# print(f" ! End of blog. Next blog file in sequence not there:{lb}")
break
try:
with open(lb,'rb') as file_in:
txt = file_in.read().decode("utf-8")
logbook_parseable = True
except (IOError):
logbook_parseable = False
print(f" ! Couldn't open logbook as UTF-8 {lb}")
except:
logbook_parseable = False
print(f" ! Very Bad Error opening {lb}")
if logbook_parseable:
# --------------------
parser = globals()[parsefunc]
print(f' - {year} parsing with {parsefunc} - {lb}')
parser(year, expedition, txt, sq) # this launches the right parser for this year
# --------------------
dupl = {}
for entrytuple in logentries:
date, tripcave, triptitle, text, trippeople, expedition, logtime_underground, tripid1 = entrytuple
check = (date, triptitle)
if check in dupl:
dupl[check] += 1
triptitle = f"{triptitle} #{dupl[check]}"
print(f' - {triptitle} -- {date}')
else:
dupl[check] = 1
EnterLogIntoDbase(date, tripcave, triptitle, text, trippeople, expedition, logtime_underground,
tripid1)
if len(logentries) == expect:
# print(f"OK {year} {len(logentries):5d} is {expect}\n")
pass
else:
print(f"Mismatch in number of log entries: {year} {len(logentries):5d} is not {expect}\n")
return len(logentries)
def LoadLogbook(year):
'''One off logbook for testing purposes
'''
global LOGBOOK_PARSER_SETTINGS
nlbe={}
TROG['pagecache']['expedition'][year] = None # clear cache
expo = Expedition.objects.get(year=year)
year = expo.year # some type funny
nlbe[expo] = LoadLogbookForExpedition(expo) # this actually loads the logbook for one expo
if year in BLOG_PARSER_SETTINGS:
print("BLOG parsing")
LOGBOOK_PARSER_SETTINGS[year] = BLOG_PARSER_SETTINGS[year]
nlbe[expo] = LoadLogbookForExpedition(expo, clean=False) # this loads the blog logbook for one expo
else:
print(f"Not a year with extant blog entries to import: '{year}' not in BLOG_PARSER_SETTINGS {BLOG_PARSER_SETTINGS}")
def LoadLogbooks():
""" This is the master function for parsing all logbooks into the Troggle database.
This should be rewritten to use coroutines to load all logbooks from disc in parallel,
but must be serialised to write to database as sqlite is single-user.
"""
global logdataissues
global entries
logdataissues = {}
DataIssue.objects.filter(parser='logbooks').delete()
expos = Expedition.objects.all()
if len(expos) <= 1:
message = f" ! - No expeditions found. Load 'people' first"
DataIssue.objects.create(parser='logbooks', message=message)
logdataissues[f"sqlfail 0000"]=message
print(message)
return
noexpo = ["1986", "2020", "2021",] #no expo
lostlogbook = ["1976", "1977", "1978", "1979", "1980", "1981"]
sqlfail = ["1987", "1988", "1989"] # breaks mysql with db constraint fail - debug locally first]
nologbook = noexpo + lostlogbook + sqlfail
nlbe={}
expd ={}
loglist = []
bloglist = []
for expo in expos: # pointless as we explicitly know the years in this code.
year = expo.year
TROG['pagecache']['expedition'][year] = None # clear cache
if year in sqlfail:
print(" - Logbook for: " + year + " NO parsing attempted - known sql failures")
message = f" ! - Not even attempting to parse logbook for {year} until code fixed"
DataIssue.objects.create(parser='logbooks', message=message)
logdataissues[f"sqlfail {year}"]=message
print(message)
if year not in nologbook:
if year in entries:
loglist.append(expo)
else:
print(" - No Logbook yet for: " + year) # catch case when preparing for next expo
if year in BLOG_PARSER_SETTINGS:
bloglist.append(expo)
for ex in loglist:
nlbe[ex] = LoadLogbookForExpedition(ex) # this loads the logbook for one expo
for b in bloglist:
if str(b) in LOGBOOK_PARSER_SETTINGS:
orig = LOGBOOK_PARSER_SETTINGS[str(b)]
else:
orig = (DEFAULT_LOGBOOK_FILE, DEFAULT_LOGBOOK_PARSER)
LOGBOOK_PARSER_SETTINGS[str(b)] = BLOG_PARSER_SETTINGS[str(b)]
print(f" - BLOG: {b}")
nlbe[b] = LoadLogbookForExpedition(b, clean=False) # this loads the blog logbook for one expo
LOGBOOK_PARSER_SETTINGS[str(b)] = orig
# tried to use map with concurrent threads - but sqlite database is not concurrent, so failed with database lock
# yt = 0
# for r in map(LoadLogbookForExpedition, loglist):
# yt = r
yt = 0
for e in nlbe:
yt += nlbe[e]
print(f"total {yt:,} log entries parsed in all expeditions")
# 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)
Reference in New Issue View Git Blame Copy Permalink