.*?\s*)? # second date
\s*(?:\s*)?
\s*(.*?)(?:)?
\s*\s*(.*?)
\s*\s*(.*?)
([\s\S]*?)
\s*(?:\s*(.*?))?
\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[:55]}'...'{trippara}'"
print(msg)
DataIssue.objects.create(parser="logbooks", message=msg)
s2 = re.match(
r"""(?x)(?:\s*.*?\s*)? # second date
\s*(?:\s*)?
\s*(.*?)(?:)?
\s*\s*(.*?)
\s*\s*(.*?)
([\s\S]*?)
\s*(?:\s*(.*?))?
\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[:55]}'...'{trippara}'"
print(msg)
DataIssue.objects.create(parser="logbooks", message=msg)
continue
ldate = parser_date(tripdate.strip(), year)
triptitles = triptitle.split(" - ")
if len(triptitles) >= 2:
place = triptitles[0]
else:
place = "UNKNOWN"
tripcontent = re.sub(r"
", "", triptext)
tripcontent = re.sub(r"", "
", tripcontent).strip()
triptitle = triptitle.strip()
# triptitle must be unique for a given date. We fix this here.
check = (ldate, triptitle)
if check in dupl:
dupl[check] += 1
triptitle = f"{triptitle} #{dupl[check]}"
print(f" - {triptitle} -- {ldate}")
else:
dupl[check] = 1
tu = tidy_time_underground(tu)
trippersons, author = tidy_trip_persons(trippeople, triptitle, expedition, tu, tid)
tripcave = tidy_trip_cave(place)
tripcontent = tidy_trip_image_urls(tripcontent, ldate)
tid = tidy_tid(tid, triptitle)
entrytuple = (ldate, place, tripcave, triptitle, tripcontent, trippersons, author, expedition, tu, tid)
logentries.append(entrytuple)
return logentries
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:
So the content is nested inside the header. Attachments (images) come after the content.
It's a bugger, but it's out of our control.
"""
logentries = []
tripheads = re.findall(
r"\s*([\s\S]*?)(]*>)([\s\S]*?)(?=[\s\S]*?(?=)", "", attach)
attach = re.sub(r")", "", 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)
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)
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)
print(message)
# fallback, ignore the timestamp bits:
tripdate = datetime.fromisoformat(datestamp[0:10])
# print(f" - tid: {tid} '{trippeople}' '{tripdate}'")
# triptitle must have the location then a hyphen at the beginning as it is ignored by export function. We can't know what this is, so we set it as 'Expo' and 'Unknown'.
place = "Unknown"
# triptitle must be unique for a given date. We can enforce this here.
triptitle = f"Expo - UK Caving Blog{sq} post {logbook_entry_count}"
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"", "", tripcontent)
tripcontent = f"\n\n\nBlog Author: {trippeople}" + tripcontent
trippersons, author = tidy_trip_persons(trippeople, expedition, logtime_underground, tid)
tripcave = tidy_trip_cave(place)
tripcontent = tidy_trip_image_urls(tripcontent, date)
tid = tidy_tid(tid, triptitle)
entrytuple = (tripdate, place, tripcave, triptitle, tripcontent, trippersons, author, expedition, tu, tid)
logentries.append(entrytuple)
return logentries
def clean_all_logbooks():
DataIssue.objects.filter(parser="logbooks").delete()
LogbookEntry.objects.all().delete()
def clean_logbook_for_expedition(expedition):
"""Only used when loading a single logbook. Deletes database LogBookEntries and
DataIssues for this expedition year.
"""
lbes = LogbookEntry.objects.filter(expedition=expedition).delete()
dataissues = DataIssue.objects.filter(parser="logbooks")
for di in dataissues:
ph = expedition.year
if re.search(ph, di.message) is not None: # SLOW just to delete issues for one year
# print(f' - CLEANING dataissue {di.message}')
di.delete()
def parse_logbook_for_expedition(expedition, blog=False):
"""Parses all logbook entries for one expedition
"""
global ENTRIES
logentries = []
logbook_parseable = False
expologbase = Path(settings.EXPOWEB, LOGBOOKS_DIR)
year = expedition.year
expect = ENTRIES[year]
# print(" - Logbook for: " + year)
if year in LOGBOOK_PARSER_SETTINGS:
yearfile, parsefunc = LOGBOOK_PARSER_SETTINGS[year]
expedition.logbookfile = yearfile # don't change this if a blog
else:
yearfile = DEFAULT_LOGBOOK_FILE
expedition.logbookfile = DEFAULT_LOGBOOK_FILE # don't change this if a blog
parsefunc = DEFAULT_LOGBOOK_PARSER
if blog:
print(f" - BLOG file {yearfile} using parser {parsefunc}")
if year not in BLOG_PARSER_SETTINGS:
message = f" ! - Expecting blog parser buut none specified for {year}"
DataIssue.objects.create(parser="logbooks", message=message)
print(message)
else:
yearfile, parsefunc = BLOG_PARSER_SETTINGS[year]
logbookpath = Path(yearfile)
# print(f" - Logbook file {yearfile} using parser {parsefunc}")
# expedition.save()
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}")
print(" .", end="")
logentries = parser(year, expedition, txt, sq) # this launches the right parser
# --------------------
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 logentries
def LoadLogbook(year):
"""One off logbook for testing purposes, and also reloadable on '/expedition/2022?reload'
This is inside an atomic transaction"""
expo = Expedition.objects.get(year=year)
year = expo.year # some type funny
clean_logbook_for_expedition(expo)
logentries = []
logentries = parse_logbook_for_expedition(expo) # this actually loads the logbook for one expo
if year in BLOG_PARSER_SETTINGS:
logentries += parse_logbook_for_expedition(expo, blog=True) # this loads the blog logbook
else:
print(
f" - Not a year with extant blog entries to import: '{year}' not in BLOG_PARSER_SETTINGS {BLOG_PARSER_SETTINGS}"
)
for entrytuple in logentries:
date, place, tripcave, triptitle, text, trippersons, author, expedition, tu, tid = entrytuple
if expo == expedition: # unneeded check, we zeroed it before filling it
#print(f" - {triptitle}")
store_entry_into_database(date, place, tripcave, triptitle, text, trippersons, author, expedition, tu, tid)
else:
print(f" ! unexpected log entry labelled as '{expedition}' {tid}" )
expo.save() # to save logbook name property
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.
This is inside an atomic transaction. Maybe it shouldn't be..
"""
global ENTRIES
global logentries
allentries = []
mem1 = get_process_memory()
print(f" - MEM:{mem1:7.2f} MB now ", file=sys.stderr)
start = time.time()
clean_all_logbooks()
expos = Expedition.objects.all()
if len(expos) <= 1:
message = " ! - No expeditions found. Attempting to 'people' first"
DataIssue.objects.create(parser="logbooks", message=message)
print(message)
load_people_expos()
expos = Expedition.objects.all()
if len(expos) <= 1:
message = " ! - No expeditions found, even after attempting to load 'people'. Abort."
DataIssue.objects.create(parser="logbooks", message=message)
print(message)
return
noexpo = [
"1986",
"2020",
"2021",
] # no expo
lostlogbook = ["1976", "1977", "1979", "1980", "1981"]
sqlfail = [""] # breaks mysql with db constraint fail - all now fixed.]
nologbook = noexpo + lostlogbook + sqlfail
nlbe = {}
loglist = []
bloglist = []
for expo in expos:
year = expo.year
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)
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:
logentries = parse_logbook_for_expedition(ex) # this loads the logbook for one expo
allentries += logentries
for b in bloglist:
print(f" - BLOG: {b}")
logentries = parse_logbook_for_expedition(b, blog=True) # loads the blog logbook for one expo
allentries += logentries
print(f"\n - {len(allentries):,} log entries parsed in all expeditions")
mem = get_process_memory()
print(f" - MEM: {mem:7.2f} MB in use, {mem-mem1:7.2f} MB more", file=sys.stderr)
duration = time.time() - start
print(f" - TIME: {duration:7.2f} s", file=sys.stderr)
# Now we serially store the parsed data in the database, updating 3 types of object:
# - Expedition (the 'logbook.html' value)
# - LogBookEntry (text, who when etc.)
# - PersonLogEntry (who was on that specific trip mentione din the logbook entry)
for entrytuple in allentries:
date, place, tripcave, triptitle, text, trippersons, author, expedition, tu, tid = entrytuple
store_entry_into_database(date, place, tripcave, triptitle, text, trippersons, author, expedition, tu, tid)
for expo in expos:
expo.save() # to save logbook name property
mem = get_process_memory()
print(f" - {len(allentries):,} log entries saved into database")
print(f" - MEM: {mem:7.2f} MB in use, {mem-mem1:7.2f} MB more", file=sys.stderr)
duration = time.time() - start
print(f" - TIME: {duration:7.2f} s", file=sys.stderr)
# 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*