troggle-unchained/parsers/caves.py

import os
import re


from pathlib import Path

from django.conf import settings
from django.db import transaction

from troggle.core.models.caves import Area, Cave, CaveAndEntrance, Entrance, GetCaveLookup  
from troggle.core.models.logbooks import CaveSlug
from troggle.core.models.troggle import DataIssue
from troggle.settings import CAVEDESCRIPTIONS, ENTRANCEDESCRIPTIONS, EXPOWEB, SURVEX_DATA

"""Reads all the cave description data and entrance description data 
by parsing the xml files stored as e.g. 
:EXPOWEB:/cave_data/1623-161.html 
or
:EXPOWEB:/entrance_data/1623-161g.html 

and creating the various Cave, Entrance and necessary Area objects.

This is the first import that happens after the database is reinitialised. 
So is the first thing that creates tables.

"""

todo = """  
- Cannot use Edit This Page for pendingcaves.txt_edit as Edit This Page is expecting an html file.
   So we will need a separate file-editing capability just for this configuration file ?!
   
- we want to overwrite a PENDING cave if we are now importing the 1623-xxx.html file for it

- rewrite archaic regex
  re.findall("<%(itemname)s>(.*?)</%(itemname)s>" % {"itemname": itemname}, text, re.S)
  in modern form and pre-compile it. 

- Semi-automagically import all the 1627- pending caves and create HTML files for them to be
  edited individually. (These are caves we only know about because we have German survex files.)
   
- crashes on MariaDB in databasereset.py on server when deleting Caves and complains Area needs a 
  non null parent, But this is not true.   The only solution we have found is to let it crash, then 
  stop and restart MariaDB (requires a logon able to sudo)   and then restart the databasereset.py 
  again. (status as of July 2022)
"""
entrances_xslug = {}
caves_xslug = {}
areas_xslug = {}


def dummy_entrance(k, slug, msg="DUMMY"):
    """Returns an empty entrance object for either a PENDING cave or a DUMMY entrance if
    user forgot to provide one when creating the cave
    """
    ent = Entrance.objects.create( # creates object and saves into db
        name=k, slug=k, filename = k + ".html",
        entrance_description="Dummy entrance: auto-created when registering a new cave "
        + "and you forgot to create an entrance for it. Click on 'Edit' to enter the correct data, then 'Submit'.",
        marking="?",
    )
    if ent:
        return ent
    else:
        message = f" ! {k:11s}  {msg}-{slug} {k} entrance create failure"
        DataIssue.objects.create(parser="entrances", message=message, url=f"{slug}")
        print(message)
        raise # caught and handled by calling routine.



def set_dummy_entrance(id, slug, cave, msg="DUMMY"):
    """Called only when reading the cave and entrance html files
    
    Called when the Entrance field in a cave_data file is either missing or 
    holds a null string instead of a filename.
    
    Previously, the lack of an entrance where an entrance was expected, caused troggle to crash in several places. 
    But it is more robust now, so this is not necessary... we hope.
    
    Also, Cave and Entrance editing now expects there to be a real file (since April 2023), so creating this
    dummy is actually harmful. So this is commented out, pending removal after further experience.
    
    global variable entrances_xslug is simply a cache of references to Entrance objects
    to speed things up when parsing a lot of caves and entrances. All DB actions are time-consuming
    so 
    """
    global entrances_xslug
    message = f" - Note: Missing Entrance for entrance '{id}' on cave '{cave}' - Is this a problem?"
    DataIssue.objects.create(parser="entrances", message=message, url=f"{cave.url}")
    
    # try:        
        # entrance = dummy_entrance(id, slug, msg="DUMMY")
        # entrances_xslug[slug] = entrance
        # CaveAndEntrance.objects.update_or_create(cave=cave, entrance_letter="", entrance=entrance)
        # pass
    # except:
        # message = f' ! Entrance Dummy setting failure, slug:"{slug}" cave id :"{id}" '
        # DataIssue.objects.create(parser="entrances", message=message, url=f"{cave.url}")
        # print(message)

def make_areas():
    print(" - Creating Areas 1623, 1624, 1627 and 1626")
    # This crashes on the server with MariaDB even though a null parent is explicitly allowed.
    area_1623 = Area.objects.create(short_name="1623", super=None)
    area_1623.save()
    area_1624 = Area.objects.create(short_name="1624", super=None)
    area_1624.save()
    area_1626 = Area.objects.create(short_name="1626", super=None)
    area_1626.save()
    area_1627 = Area.objects.create(short_name="1627", super=None)
    area_1627.save()
        
def get_area(areanum):
    """Given the number as a string, return the area object
    """
    a = Area.objects.all()
    if len(a) == 0:
        make_areas()
        
    area = Area.objects.get(short_name="1623") # default

    if areanum == "1623":
        area = Area.objects.get(short_name="1623")
    if areanum == "1624":
        area = Area.objects.get(short_name="1624")
    if areanum == "1626":
        area = Area.objects.get(short_name="1626")
    if areanum == "1627":
        area = Area.objects.get(short_name="1627")
    return area
    
def create_new_cave(svxpath):
    """This is called only when a new survex file is edited online which has a path on the
    :loser: repo which is not recognised as a known cave.
    """
    # e.g. svxpath = "caves-1623/666/antig"
    print(f"Create new cave at {svxpath}")
    #
    survex_file = svxpath + ".svx"
    parts = svxpath.split("/")
    a = parts[0][-4:]
    caveid = parts[1]
    print(f"parts {parts}, {a}, {caveid}")
    # double check
    if a[0:3] == "162":
        areanum = a[0:4]
        url = f"{areanum}/{caveid}.html"  # Note we are appending the .html as we are believe in backwards compatability.
        #url = f"{areanum}/{a[5:]}.html"  # This is original code, but a above is only defined as being 4 characters long, so it did not make sense and produced non unique urls
    else: 
        print(f"WARNING: parsers/caves/create_new_cave called with svxpath '{svxpath}'.  Surely it should start 'caves-162*'?")
        areanum = "1623"
        url = f"1623/{caveid}.html"
        #url = f"1623/{k}.html" # This is original code, but a above is only defined as being 4 characters long, so it did not make sense and produced non unique urls
        
    k = f"{areanum}-{caveid}"
    area = get_area(areanum)

    caves = Cave.objects.filter(unofficial_number=caveid) 
    if caves:
        message = f" ! Already exists, caveid:{k} in area {areanum} {caves}"
        DataIssue.objects.create(parser="caves", message=message)
        print(message)
        return caves[0]

    try:
        cave = do_pending_cave(k, caveid, url, area)
    except:
        message = f" ! Error. Cannot create pending cave and entrance, pending-id:{k} in area {areanum}"
        DataIssue.objects.create(parser="caves", message=message)
        print(message)
        raise    
    
    # we know what the survex file is, we don't need to use the guess
    cave.survex_file=survex_file
    cave.save()
    return cave

def do_pending_cave(k, caveid, url, area):
    """
    default for a PENDING cave, should be overwritten in the db later if a real cave of the same name exists
    in expoweb/cave_data/1623-"k".html

    Note that at this point in importing the data we have not yet seen the survex files, so we can't
    look inside the relevant survex file to find the year and so we con't provide helpful links.
    """

    def get_survex_file(k):
        """Guesses at and finds a survex file for this pending cave.
        Convoluted. Needs rewriting
        """
        if k[0:3] == "162":
            id = Path(k[5:])
        else:
            id = Path(k)

        survex_file = f"caves-{area.short_name}/{id}/{id}.svx"
        if Path(settings.SURVEX_DATA, survex_file).is_file():
            return survex_file
        else:
            survex_file = f"caves-{area.short_name}/{id}.svx"
            if Path(settings.SURVEX_DATA, survex_file).is_file():
                return survex_file

        survex_file = ""
        d = Path(settings.SURVEX_DATA, f"caves-{area.short_name}/{id}")
        if d.is_dir():
            prime_suspect = ""
            dir = d.iterdir()
            for f in dir:
                if f.suffix == ".svx":
                    survex_file = f.relative_to(settings.SURVEX_DATA)
                    chk = min(5, len(f.name) - 1)
                    if str(f.name)[:chk].lower() == str(id.name)[:chk].lower():  # bodge which mostly works
                        prime_suspect = survex_file
            if prime_suspect:
                survex_file = prime_suspect
            # message = f" ! {k:14} Found a survex file which might be the right one: {survex_file}"
            # DataIssue.objects.create(parser='caves', message=message, url=url)
            # print(message)
        return survex_file

    slug = k

    g = GetCaveLookup()
    with transaction.atomic():
        if slug in g:
            message = f" ! {k:18} cave listed in pendingcaves.txt already exists."
            DataIssue.objects.create(parser="caves", message=message, url=url)
            print(message)
            return

        default_note = "A reference has been found to this cave id in a survex file in the loser repo, or in a wallet metadata"
        default_note += " in a JSON file in the drawings repo, but no Cave Description exists in expoweb (in /cave_data/)<br><br><br>\n"
        default_note += "INSTRUCTIONS: First open 'This survex file' (link above the CaveView panel) to find the date and info. Then "
        default_note += '<br><br>\n\n - (0) look in the <a href="/noinfo/cave-number-index">cave number index</a> for notes on this cave,  '
        default_note += "<br><br>\n\n - (1) search in the survex file for the *ref to find a "
        default_note += "relevant wallet, e.g.<a href='/survey_scans/2009%252311/'>2009#11</a> and read the notes image files <br>\n - "
        default_note += (
            "<br><br>\n\n - (2) search in the Expo for that year e.g. <a href='/expedition/2009'>2009</a> to find a "
        )
        default_note += "relevant logbook entry, remember that the date may have been recorded incorrectly, "
        default_note += (
            "so check for trips i.e. logbook entries involving the same people as were listed in the survex file, "
        )
        default_note += (
            "and you should also check the scanned copy of the logbook (linked from each logbook entry page) "
        )
        default_note += "just in case a vital trip was not transcribed, then <br>\n - "
        default_note += (
            "click on 'Edit this cave' and copy the information you find in the survex file and the logbook"
        )
        default_note += "and delete all the text in the 'Notes' section - which is the text you are reading now."
        default_note += "<br><br>\n\n - Only two fields on this  form are essential. "
        default_note += "Documentation of all the fields on 'Edit this cave' form is in <a href='/handbook/survey/caveentryfields.html'>handbook/survey/caveentryfields</a>"
        default_note += "<br><br>\n\n - "
        default_note += "You will also need to create a new entrance from the 'Edit this cave' page. Ignore the existing dummy one, it will evaporate on the next full import."
        default_note += "<br><br>\n\n - "
        default_note += "When you Submit it will create a new file in expoweb/cave_data/ "
        default_note += (
            "<br><br>\n\n - Now you can edit the entrance info: click on Edit below for the dummy entrance. "
        )
        default_note += "and then Submit to save it (if you forget to do this, a dummy entrance will be created for your new cave description)."
        default_note += "<br><br>\n\n - Finally, you need to find a nerd to edit the file '<var>expoweb/cave_data/pending.txt</var>' "
        default_note += (
            f"to remove the line <br><var>{slug}</var><br> as it is no longer 'pending' but 'done. Well Done."
        )

        survex_file = get_survex_file(k)

        cave = Cave(
            unofficial_number=caveid,
            underground_description="Pending cave write-up - No cave description created yet.",
            survex_file=survex_file,
            url=url,
            notes=default_note,
        )
        if cave:
            cave.save()  # must save to have id before foreign keys work. This is also a ManyToMany key.
            cave.area.add(area)
            cave.save()
            message = f" ! {k:18} {cave.underground_description} url: {url}"
            DataIssue.objects.create(parser="caves", message=message, url=url)
            print(message)

            try:  # Now create a cave slug ID
                CaveSlug.objects.update_or_create(cave=cave, slug=slug, primary=False)
            except:
                message = f" ! {k:11s}  PENDING CaveSLUG {slug} create failure"
                DataIssue.objects.create(parser="caves", message=message)
                print(message)
        else:
            message = f" ! {k:11s}  PENDING cave create failure"
            DataIssue.objects.create(parser="caves", message=message)
            print(message)

        try:
            # troggle is more robust against missing entrances now, not needed.
            # set_dummy_entrance(k, slug, cave, msg="PENDING")
            pass
        except:
            message = f" ! {k:11s}  PENDING entrance + cave UNION create failure  '{cave}' [{slug}] {k}"
            # message = f" ! {k:11s}  PENDING entrance + cave UNION create failure  '{cave}' [{ent}]"
            DataIssue.objects.create(parser="caves", message=message)
            print(message)
    return cave

def getXML(text, itemname, minItems=1, maxItems=None, context=""):
    """Reads a single XML tag
    Should throw exception rather than producing error message here,
    then handle exception in calling routine where it has the context.
    
    This always succeeds, but it produices error message on the terminal and in the
    DatIssues log.
    """
    items = re.findall("<%(itemname)s>(.*?)</%(itemname)s>" % {"itemname": itemname}, text, re.S)
    if len(items) < minItems:
        message = (
            " ! %(count)i x %(itemname)s found, at least %(min)i expected. Load ABORT. "
            % {"count": len(items), "itemname": itemname, "min": minItems}
            + " in file "
            + context
        )
        DataIssue.objects.create(parser="caves", message=message, url="" + context)
        print(message)

    if maxItems is not None and len(items) > maxItems:
        message = (
            " ! %(count)i x %(itemname)s found, no more than %(max)i expected in this XML unit. Load ABORT. "
            % {"count": len(items), "itemname": itemname, "max": maxItems}
            + " in file "
            + context
        )
        DataIssue.objects.create(parser="caves", message=message, url="" + context)
        print(message)
    if minItems == 0:
        if not items:
            items = [""]
    return items


def boolify(boolstrs):
            return {
            "True": True,
            "False": False,
            "true": True,
            "false": False}[boolstrs[0]]
            
def validate_station(station):
    """It is possible to break troggle entirely by getting this wrong.
    These station identifiers are matched against other statsions using .endswith()
    in parsers/locations.py
    so a simple number here will match hundreds of SUrvexStation objects
    It should be, e.g. "1623.p240"
    """
    if station == "":
        return True
    dot = station.find(".")
    if dot == -1:
        # no full stop found. Bad station identifier.
        raise
    else:
        return True

def read_entrance(filename, ent=None):
    """Reads an entrance description from the .html file.

      If not called as part of initial import, then the global lists will not be correct
      but this is OK, a search will find them in the db.

      Args:
        filename: The name of the .html file.
        ent: The entrance object, if it already exists.

      Returns:
        The entrance object, or a new entrance object if `ent` is None.
      """
    def getXMLmax1(field):
        return getXML(entrancecontents, field, maxItems=1, context=context)
        
    global entrances_xslug
    global caves_xslug
    global areas_xslug

    # Note: these are HTML files in the EXPOWEB repo, not from the loser repo.
    with open(os.path.join(ENTRANCEDESCRIPTIONS, filename)) as f:
        contents = f.read()
    context = filename
    
    entrancecontentslist = getXML(contents, "entrance", maxItems=1, context=context)
    if len(entrancecontentslist) != 1:
        message = f'! BAD ENTRANCE DATA in "{filename}". More than one entrance. Edit file manually, click.'
        DataIssue.objects.create(parser="entrances", message=message, url=f"/entrance_data/{filename}_edit")
        print(message)
        return None

    entrancecontents = entrancecontentslist[0]
    slugs = getXML(entrancecontents, "slug", context=context)
    slug = slugs[0]

    if len(slugs) >1:
        # Only ever one of these per entrance in the expo dataset
        message = f" ! - More than one slug for an entrance: {entrance}, slugs: {slugs}. Ignoring all except first."
        DataIssue.objects.create(parser="entrances", message=message, url=f"/cave/{slug}/edit/")
        print(message)
        
    lastvisit = getXML(entrancecontents, "lastvisit", maxItems=1, minItems=0, context=context)

    alt = getXMLmax1("alt")
    approach = getXMLmax1("approach")
    bearings = getXMLmax1("bearings")
    easting = getXMLmax1("easting")
    entrance_description = getXMLmax1("entrance_description")
    exact_station = getXMLmax1("exact_station")
    explorers = getXMLmax1("explorers")
    findability = getXMLmax1("findability")
    findability_description = getXMLmax1("findability_description")
    location_description = getXMLmax1("location_description")
    map_description = getXMLmax1("map_description")
    marking = getXMLmax1("marking")
    marking_comment = getXMLmax1("marking_comment")
    name = getXMLmax1("name")
    non_public = getXMLmax1("non_public")
    northing = getXMLmax1("northing")
    other_description = getXMLmax1("other_description")
    other_station = getXMLmax1("other_station")
    photo = getXMLmax1("photo")
    tag_station = getXMLmax1("tag_station")
    underground_description = getXMLmax1("underground_description")
    url = getXMLmax1("url")

    if not ent:
        ent, state = Entrance.objects.update_or_create(slug=slug)
        
    ent.name=name[0]
    ent.non_public=boolify(non_public)
    ent.alt=alt[0]
    ent.approach=approach[0]
    ent.bearings=bearings[0]
    ent.easting=easting[0]
    ent.entrance_description=entrance_description[0]
    ent.exact_station=exact_station[0]
    ent.explorers=explorers[0]
    ent.filename=filename
    ent.findability=findability[0]
    ent.findability_description=findability_description[0]
    ent.lastvisit=lastvisit[0]
    ent.location_description=location_description[0]
    ent.map_description=map_description[0]
    ent.marking=marking[0]
    ent.marking_comment=marking_comment[0]
    ent.northing=northing[0]
    ent.other_description=other_description[0]
    ent.other_station=other_station[0]
    ent.photo=photo[0]
    # ent.slug=slugs[0]
    ent.tag_station=tag_station[0]
    ent.underground_description=underground_description[0]
    ent.url=url[0]
    
    for st in [ent.exact_station, ent.other_station, ent.tag_station]:
        try:
            validate_station(st)
        except:
            message = f" ! BAD ENTRANCE TAG '{st}' in '{filename}'. Must format like '1623.p204'. Edit file manually, click."
            #http://localhost:8000/1623/2023-EBH-01/1623-2023-EBH-01:1623-2023-EBH-01_entrance_edit
            DataIssue.objects.create(parser="entrances", message=message, url=f"/1623/{slug}/{slug}:{slug}_entrance_edit")
            print(message)
            # ent_issues = DataIssue.objects.filter(parser="entrances")
            # print(f".. We now have  {len(ent_issues)} entrance DataIssues")
            return None
    ent.save()
    return ent

def read_cave(filename, cave=None):
    """Reads an entrance description from the .html file
    Convoluted. Sorry. Needs rewriting
    Assumes any area it hasn't seen before is a subarea of 1623
    
    If not called as part of initial import, then the global lists will not be correct
    but this is OK, a search will find them in the db.
    
    Attempted to use standard python3.11 xml library but fails on HTML entities (2023-04-23)
    import xml.etree.ElementTree as ET
    tree = ET.parse(fn)
    xml_root = tree.getroot()
    for t in ["html", "head", "body", "cave","non_public", "caveslug", "official_name","entrance"]:
        elements = xml_root.findall(t)

    """
    def getXMLmax1(field):
        return getXML(cavecontents, field, maxItems=1, context=context)
        
    def do_entrances():
        """For both bulk import and individual re-reading of cave_data file,
        fix the entrances
        What is Class CaveAndEntrance for?
        """
        for e in entrances:
 
            eslug = getXML(e, "entranceslug", maxItems=1, context=context)[0]
            # if eslug.endswith(('a','b','c','d','e','f')):
                # print(f"! Entrance {eslug}")
            if eslug.endswith('a b'):
                message = f' - Entrance has weird name slug:"{eslug}" cave:"{cave}" caveslug:"{slug}" filename:"cave_data/{filename}"'
                DataIssue.objects.create(parser="xEntrances", message=message, url=f"{cave.area}/{cave.area}-{cave.url}_cave_edit/")
                print(message)        

            letter = getXML(e, "letter", maxItems=1, context=context)[0]
            if len(entrances) == 1 and not eslug:  # may be empty: <entranceslug></entranceslug>
                msg="DUMMY: no entrance slug read from file, so assume textually same as cave slug"
                set_dummy_entrance(slug[5:], slug, c, msg=msg)
                print(f"! {msg}\n- {slug} {c}")
            else:
                if eslug in entrances_xslug:
                    # print(f"eslug {eslug} found eslug in xslug cache ")
                    entrance = entrances_xslug[eslug]
                else:
                    # print(f"eslug {eslug} looking up entrance ")
                    try:
                        entrance = Entrance.objects.get(slug=eslug)
                        entrances_xslug[eslug] = entrance
                    except:
                        message = f"! eslug {eslug} Abort entrance loading. Failed to find entrance in db"         
                        DataIssue.objects.create(parser="xEntrances", message=message, url=f"{cave.area}/{cave.area}-{cave.url}_cave_edit/")
                        print(message)   
                        return                        

                if eslug != f"{entrance}":
                    message = f"eslug {eslug} using different entrance {entrance} to set CaveAndEntrance"         
                    DataIssue.objects.create(parser="xEntrances", message=message, url=f"{cave.area}/{cave.area}-{cave.url}_cave_edit/")
                    print(message)        
                try:    
                    CaveAndEntrance.objects.update_or_create(
                        cave=cave, entrance_letter=letter, entrance=entrance
                    )
                except:
                    print(f"! Entrance setting failure {slug}")
                    message = f' ! Entrance setting failure, slug:"{slug}" #entrances:{len(entrances)} letter:"{letter}" cave:"{cave}" filename:"cave_data/{filename}"\n{e}'
                    DataIssue.objects.create(parser="entrances", message=message, url=f"{cave.url}_cave_edit/")
                    print(message)        
    def reload_entrances():
        """For individual re-reading of a cave_data file when editing,
        also re-read the entrance_data files
        """
        for eslug in entrances_xslug:
            entrance = entrances_xslug[eslug]
            read_entrance(entrance.filename, ent=entrance)
            entrance.save()

    def do_caveslugstuff():
        """This may be a fossil. We only have one slug per cave in troggle.
        Pending destruction of this whole concept and Class CaveSlug
        What is Class CaveSlug for?
        """
        primary = True # this sets the first thing we find to be primary=True and all the others =False
        for slug in slugs:
            if slug in caves_xslug:
                cs = caves_xslug[slug]
            else:
               try:  
                    cs = CaveSlug.objects.update_or_create(cave=cave, slug=slug, primary=primary)
                    caves_xslug[slug] = cs
               except Exception as ex:
                    #raise
                    # This fails to do an update! It just crashes.. to be fixed
                    message = f" ! CaveSlug update/create failure : {slug}, skipping file cave_data/{context} with exception\nException: {ex.__class__}"
                    DataIssue.objects.create(parser="caves", message=message, url=f"{cave.url}_edit/")
                    print(message)
            primary = False
            
    global entrances_xslug
    global caves_xslug
    global areas_xslug

    # Note: we want to overwrite a PENDING cave if we are now importing the 1623-xxx.html file for it
    
    # Note: these are HTML files in the EXPOWEB repo, not from the loser repo.
    fn = settings.CAVEDESCRIPTIONS / filename
    context = filename
    
    # print(f" - Reading Cave from cave descriptions file {fn}")
    if not fn.exists():
        message = f" ! Cave_data file reading problem filename:'cave_data/{filename}'"
        DataIssue.objects.create(parser="caves", message=message, url=f"/cave_data/{filename}_edit")
        print(message)
        return None

    with open(fn) as f:
        contents = f.read()
    cavecontentslist = getXML(contents, "cave", maxItems=1, context=context)

    if len(cavecontentslist) != 1:
        message = f'! BAD CAVE DATA in "{filename}". More than one cave. Edit file manually, click.'
        DataIssue.objects.create(parser="caves", message=message, url=f"/cave_data/{filename}_edit")
        print(message)
        return None
        
    cavecontents = cavecontentslist[0]
    slugs = getXML(cavecontents, "caveslug", maxItems=1, context=context)
    if len(slugs) > 1:
        message = f" ! - More than one slug for a cave: {cave}, slugs: {slugs}. Ignoring all except first."
        DataIssue.objects.create(parser="caves", message=message, url=f"{cave.url}_edit/")
        print(message)
    slug = slugs[0]

    non_public = getXMLmax1("non_public")
    official_name = getXMLmax1("official_name")
    kataster_code = getXMLmax1("kataster_code")
    kataster_number = getXMLmax1("kataster_number")
    unofficial_number = getXMLmax1("unofficial_number")
    explorers = getXMLmax1("explorers")
    underground_description = getXMLmax1("underground_description")
    equipment = getXMLmax1("equipment")
    references = getXMLmax1("references")
    survey = getXMLmax1("survey")
    kataster_status = getXMLmax1("kataster_status")
    underground_centre_line = getXMLmax1("underground_centre_line")
    notes = getXMLmax1("notes")
    length = getXMLmax1("length")
    depth = getXMLmax1("depth")
    extent = getXMLmax1("extent")
    survex_file = getXMLmax1("survex_file")
    description_file = getXMLmax1("description_file")
    url = getXMLmax1("url")
    
    manual_edit = True
    if not cave:
        manual_edit = False
        try:
            cave, state = Cave.objects.update_or_create(filename=filename) # replace with slug when CaveSlug tidied up
        except:
            print(" ! FAILED to get only one CAVE in db when updating using: " + filename)
            kaves = Cave.objects.all().filter(filename=filename) # replace with slug when CaveSlug tidied up
            for k in kaves:
                message = " ! - DUPLICATES in db. kataster:" + str(k.kataster_number) + ", slug:" + str(k.slug())
                DataIssue.objects.create(parser="caves", message=message)
                print(message)
            for k in kaves:
                if k.slug() is not None:
                    print(" ! - OVERWRITING this one: slug:" + str(k.slug()))
                    k.notes = "DUPLICATE kataster number found on import. Please fix\n" + k.notes
                    cave = k

    # From here on the code applies to both edited and newly-imported caves (mostly!)
    do_caveslugstuff() # needs cave!=None
    
    cave.non_public=boolify(non_public)
    cave.official_name=official_name[0]
    cave.kataster_code=kataster_code[0]
    cave.kataster_number=kataster_number[0]
    cave.unofficial_number=unofficial_number[0]
    cave.explorers=explorers[0]
    cave.underground_description=underground_description[0]
    cave.equipment=equipment[0]
    cave.references=references[0]
    cave.survey=survey[0]
    cave.kataster_status=kataster_status[0]
    cave.underground_centre_line=underground_centre_line[0]
    cave.notes=notes[0]
    cave.length=length[0]
    cave.depth=depth[0]
    cave.extent=extent[0]
    cave.survex_file=survex_file[0]
    cave.description_file=description_file[0]
    cave.url=url[0]
            
    areas = getXML(cavecontents, "area", context=context)
    cave.area.clear() # Deletes all links to areas in db
    for area_slug in areas:
        if area_slug in areas_xslug:
            newArea = areas_xslug[area_slug]
        else:
            areas_new = Area.objects.filter(short_name=area_slug)
            if areas_new:
                newArea = areas_new[0] # just the first one we find, but we are going to clean up Areas anyway
            else:
                # Area not seen before. SHould not happen with manual edit
                if manual_edit:
                    message = f" ! Cave edit failure due to unrecognised Area: {area_slug[0]}, skipping this field edit. "
                    DataIssue.objects.create(parser="caves", message=message)
                    print(message)                    
                # super value is highly dodgy
                newArea = Area(short_name=area_slug, super=Area.objects.get(short_name="1623"))
                newArea.save()
            areas_xslug[area_slug] = newArea
        cave.area.add(newArea)

    entrances = getXML(cavecontents, "entrance", context=context)
    do_entrances()
    # print(f"- {entrances_xslug=}")
    if not entrances or len(entrances) < 1:
        # missing entrance link in cave_data/1623-* .html file
        set_dummy_entrance(slug[5:], slug, cave, msg="DUMMY: no entrances")
    else:
        do_entrances()
    if manual_edit:
        reload_entrances()
        
    if survex_file[0]:
        if not (Path(SURVEX_DATA) / survex_file[0]).is_file():
            message = f' ! {slug:12} survex filename does not exist :LOSER:"{survex_file[0]}" in "{filename}"'
            DataIssue.objects.create(parser="caves", message=message, url=f"/{slug[0:4]}/{slug}_cave_edit/")
            print(message)

    if description_file[0]:  # if not an empty string
        message = f' - {slug:12} Note (not an error): complex description filename "{description_file[0]}" inside "cave_data/{filename}"'
        DataIssue.objects.create(parser="caves ok", message=message, url=f"/{slug}_cave_edit/")
        print(message)

        if not (Path(EXPOWEB) / description_file[0]).is_file():
            message = f' ! {slug:12} description filename  "{EXPOWEB}/{description_file[0]}" does not refer to a real file'
            DataIssue.objects.create(parser="caves", message=message, url=f"/{slug}_cave_edit/")
            print(message)

    cave.save()
    return cave

def add_cave_to_pending_list(id):
    fpending = Path(CAVEDESCRIPTIONS, "pendingcaves.txt")
    if fpending.is_file():
        with open(fpending, "a") as pend:
            pend.write(f"{id}\n")
 
def readcaves():
    """Called from databaseReset mass importer.
    Reads the xml-format HTML 'cave' files in the EXPOWEB repo, the survex files from the loser repo.
    """
    # Pending is for those caves which do not have cave_data/1623-xxx.html XML files even though 
    # they exist and have surveys. 
    with transaction.atomic():
        pending = set()
        fpending = Path(CAVEDESCRIPTIONS, "pendingcaves.txt")
        if fpending.is_file():
            with open(fpending, "r") as fo:
                cids = fo.readlines()
            for cid in cids:
                pending.add(cid.strip().rstrip("\n").upper())

    with transaction.atomic():
        print(" - Deleting Caves and Entrances")
        # attempting to avoid MariaDB crash when doing this
        try:
            Area.objects.all().delete()
        except:
            pass
        try:
            Cave.objects.all().delete()
        except:
            pass
        try:
            Entrance.objects.all().delete()
        except:
            pass
            
    # Clear the cave data issues and the caves as we are reloading
    DataIssue.objects.filter(parser="areas").delete()
    DataIssue.objects.filter(parser="caves").delete()
    DataIssue.objects.filter(parser="caves ok").delete()
    #DataIssue.objects.filter(parser="entrances").delete()
    #DataIssue.objects.filter(parser="xEntrances").delete()

    with transaction.atomic():
        area = get_area("1623")

        print(" - Reading Entrances from entrance descriptions xml files")
        for filename in next(os.walk(ENTRANCEDESCRIPTIONS))[2]:  # Should be a better way of getting a list of files
            read_entrance(filename)

    # Why is this needed ? Without it, we lose these DataIssues!
    ent_issues = DataIssue.objects.filter(parser="entrances")
    print(f" _ We now have  {len(ent_issues)} entrance DataIssues")

    with transaction.atomic():
        print(" - Reading Caves from cave descriptions xml files")
        for filename in next(os.walk(CAVEDESCRIPTIONS))[2]:  # Should be a better way of getting a list of files
            if filename.endswith(".html"):
                read_cave(filename)

    print(" - Setting up all the variously useful alias names")
    GetCaveLookup()

    print(" - Setting pending caves")
    # Do this last, so we can detect if they are created and no longer 'pending'

    with transaction.atomic():
        for k in pending:

            if k[0:3] == "162":
                areanum = k[0:4]
                number = k[5:]
                url = f"{areanum}/{k[5:]}.html"  # Note we are appending the .htm to allow for offline websites
            else:
                areanum = "1623"
                number = k
                url = f"1623/{k}"

            area = get_area(areanum)
            try:
                do_pending_cave(k, number, url, area)
            except:
                message = f" ! Error. Cannot create pending cave, pending-id:{k} in area {areanum}"
                DataIssue.objects.create(parser="caves", message=message)
                print(message)
                raise