troggle/urls.py

from django.conf import settings
from django.contrib import admin
from django.urls import include, path, re_path

from troggle.core.views import statistics, survex
from troggle.core.views.auth import expologin, expologout
from troggle.core.views.caves import (cave3d, caveEntrance, caveindex,
                                      cavepage, caveQMs, edit_cave,
                                      edit_entrance, get_entrances, qm)
from troggle.core.views.drawings import dwgallfiles, dwgfilesingle
from troggle.core.views.editor_helpers import image_selector, new_image_form
from troggle.core.views.expo import (editexpopage, expofiles_redirect,
                                     expofilessingle, expopage, map, mapfile,
                                     mediapage, spider)
from troggle.core.views.logbooks import (QMs_jsonListView, Expeditions_jsonListView,
                                         Expeditions_tsvListView, expedition,
                                         get_logbook_entries, get_people,
                                         logbookentry, notablepersons, person,
                                         personexpedition)
from troggle.core.views.other import (controlpanel, exportlogbook, frontpage,
                                      todos)
from troggle.core.views.prospect import prospecting
from troggle.core.views.scans import (allscans, cavewallets, scansingle,
                                      walletslistperson, walletslistyear)
from troggle.core.views.uploads import dwgupload, photoupload
from troggle.core.views.wallets_edit import walletedit
"""This sets the actualurlpatterns[] and urlpatterns[] lists which django uses 
to resolve urls - in both directions as these are declarative. 

HOW THIS WORKS
This is a "url dispatcher" - something needed by every web framework.
re_path( <regular expression that matches the thing in the web browser>,
<reference to python function in 'core' folder>, <optional name>)

Django also provides the reverse function: given an an object, provide the URL
which is vital to writing code for the webapp. So the URL dispatch is declarative.

The API urls return TSV or JSON and are new in July 2020.
"""

todo = '''
- Replace more re_path() with modern and simpler path(). Careful: some have to stay as re_path()

- The admin and logout paths need to stay using re_path() as they 
  have to be locked to the start.

- The final _edit and CATCHALL also have to use re_path().

- Test VERY CAREFULLY for each change. It is fragile.
'''

# Many of these patterns do not work because troggle spent many years broken and we have
# not yet restored all the functions. Some may have never been fully implemented in
# the first place and what they were intended to provide is obscure.

if settings.EXPOFILESREMOTE:
    expofilesurls = [
        path('<path:filepath>',    expofiles_redirect, name="expofiles_redirect"), # to http://expo.survex.com/expofiles
        path('',                    expofiles_redirect, {'filepath': ""}, name="expofiles_redirect"),
        ]
else:
    expofilesurls = [
        path('',                 expofilessingle, {'filepath': ""}, name="single"), 
        path('<path:filepath>', expofilessingle, name="single"), # local copy of EXPOFILES
        ]

# see https://docs.djangoproject.com/en/dev/topics/auth/default/tiny
# The URLs provided by include('django.contrib.auth.urls') are:
# 
# accounts/login/ [name='login']
# accounts/logout/ [name='logout']
# accounts/password_change/ [name='password_change']
# accounts/password_change/done/ [name='password_change_done']
# accounts/password_reset/ [name='password_reset']
# accounts/password_reset/done/ [name='password_reset_done']
# accounts/reset/<uidb64>/<token>/ [name='password_reset_confirm']
# accounts/reset/done/ [name='password_reset_complete']

# BUT many of these are set up by opinionated Django even if 'django.contrib.auth.urls' is NOT included.
# Some overlap with 'admin.site.urls' needs to be investigated.

# admin.site.urls is urls() which maps to get_urls() which is a function declared  
# in django/contrib/admin/sites.py which for me is 
# /home/philip/p11d5/lib/python3.11/site-packages/django/contrib/admin/sites.py

trogglepatterns = [
    path('expofiles/', include(expofilesurls)), # intercepted by Apache, if it is running.
    path('expofiles',  include(expofilesurls)), # curious interaction with the include() here, not just a slash problem.

    re_path(r'^(.*)_edit_edit$',    spider, name="spider"), # web spider funny

    re_path(r'^caves$',      caveindex,      name="caveindex"),
    re_path(r'^indxal.htm$', caveindex,      name="caveindex"), # ~420 hrefs to this url in expoweb files
    re_path(r'^people/?$',   notablepersons, name="notablepersons"),

    re_path(r'^admin/doc/',  include('django.contrib.admindocs.urls')), # needs docutils Python module (http://docutils.sf.net/).
    re_path(r'^admin/',      admin.site.urls), # includes admin login & logout urls  & /admin/jsi18n/
    
# Uploads - uploading a file
    path('walletedit/',                walletedit,  name='walletedit'), 
    path('walletedit/<path:path>',     walletedit,  name='walletedit'), # path=2020#01
    path('photoupload/',                 photoupload, name='photoupload'), # restricted to current year
    path('photoupload/<path:folder>',    photoupload, name='photoupload'), # restricted to current year
    path('dwgupload/<path:folder>',      dwgupload,   name='dwgupload'), 
    path('dwgupload/',                   dwgupload,   name='dwgupload'), 
    path('dwguploadnogit/',              dwgupload,  {'gitdisable': 'yes'}, name='dwguploadnogit'), # used in testing
    path('dwguploadnogit/<path:folder>', dwgupload,  {'gitdisable': 'yes'}, name='dwguploadnogit'), # used in testing

# setting LOGIN_URL = '/accounts/login/' is default. 
#   NB setting url pattern name to 'login' instea dof 'expologin' with override Django, see https://docs.djangoproject.com/en/dev/topics/http/urls/#naming-url-patterns
    path('accounts/logout/', expologout, name='expologout'), # same as in django.contrib.auth.urls
    path('accounts/login/',  expologin,  name='expologin'),  # same as in django.contrib.auth.urls
    #re_path(r'^accounts/', include('django.contrib.auth.urls')),  # see site-packages\registration\auth_urls_classes.py

#   Persons - nasty surname recognition logic fails for 19 people!
    re_path(r'^person/(?P<first_name>[A-Z]*[a-z\-\'&;]*)[^a-zA-Z]*(?P<last_name>[a-z\-\']*[^a-zA-Z]*[\-]*[A-Z]*[a-zA-Z\-&;]*)/?', person, name="person"),
    re_path(r'^personexpedition/(?P<first_name>[A-Z]*[a-z&;]*)[^a-zA-Z]*(?P<last_name>[A-Z]*[a-zA-Z&;]*)/(?P<year>\d+)/?$', personexpedition, name="personexpedition"),

#   Expedition master page & API exports
    re_path(r'^expedition/(\d+)$',  expedition,  name="expedition"),
    re_path(r'^api/expeditions_tsv$', Expeditions_tsvListView.as_view()),
    re_path(r'^api/expeditions_json$', Expeditions_jsonListView.as_view()),
    re_path(r'^api/QMs_json$', QMs_jsonListView.as_view()),

#   Logbook entries
    re_path(r'^logbookentry/(?P<date>.*)/(?P<slug>.*)/?$', logbookentry,name="logbookentry"),
    re_path(r'^logbook(?P<year>\d\d\d\d)\.(?P<extension>.*)/?$', exportlogbook, name='exportlogbook'), # e.g. /logbook2019.html # working but old CSS in     
    re_path(r'^logbook$', exportlogbook, name='exportlogbook'), 
   
#   Internal. editfile.html template uses these internally
    re_path(r'^getPeople/(?P<expeditionslug>.*)', get_people, name = "get_people"), 
    re_path(r'^getLogBookEntries/(?P<expeditionslug>.*)', get_logbook_entries, name = "get_logbook_entries"), 
    re_path(r'^getEntrances/(?P<caveslug>.*)', get_entrances, name = "get_entrances"), # used internally ?

#   Cave description pages
    re_path(r'^newcave/$', edit_cave, name="newcave"),
    re_path(r'^cave/3d/(?P<cave_id>[^/]+)$', cave3d, name="cave3d"),

    #re_path(r'^cave/description/([^/]+)/?$', caves.caveDescription), #!!!BAD, local links fail..
    #re_path(r'^cave/(?P<cave_id>[^/]+)/?$', caves.cave, name="cave"), # used only in testing !? XXXXXXXXXXXXXXXXXXXXXXXXXX
    #re_path(r'^cave/(?P<cave_id>[^/]+)/?(?P<ent_letter>[^/])$', ent), #!!!BAD, local links fail..# view_caves.ent
    
#   Edit caves and entrances
    re_path(r'^(?P<path>.*)/(?P<slug>[^/]+)_cave_edit/$',           edit_cave, name="edit_cave"), # edit_cave needed by cave.html template for url matching
    re_path(r'^(?P<path>.*)/(?P<caveslug>[^/]+):(?P<slug>[^:]+)_entrance_edit', edit_entrance, name = "editentrance"), #edit existing entrance
    re_path(r'^(?P<path>.*)/(?P<caveslug>[^/]+)_entrance_new$', edit_entrance, name = "newentrance"), # new entrance for a cave

    re_path(r'^(.*)_edit$',         editexpopage, name="editexpopage"),
    re_path(r'^(?P<karea>\d\d\d\d)(?P<subpath>.*)$',   cavepage,  name="cavepage"), # shorthand /1623/264 or 1623/161/top.htm
    # Note that urls eg '/1623/161/l/rl89a.htm' are handled by cavepage which redirects them to 'expopage'    
    # Note that _edit$ for a cave description page in a subfolder e.g. /1623/204/204.html_edit  gets caught here and breaks with 404
    
    # These re-enable archaic URLs which are in old web pages which are still public on other servers
    path('smkridge/<path:subpath>',  cavepage,  {'karea': "1623/"}, name="cavepage"), 
    path('expo/smkridge/<path:subpath>',  cavepage,  {'karea': "1623/"}, name="cavepage"), 

    # Archaic, kept. This /expo/ prefix only works for expoweb HTML pages not troggle pages   
    path('expo/<path:path>',              expopage,     name="expopage"),  
    
#   Entrances
    re_path(r'^cave/entrance/([^/]+)/?$',    caveEntrance), # lists all entrances !!!BAD, local links fail

   
# System admin and monitoring
    path('statistics',        statistics.stats,       name="stats"),
    path('stats',             statistics.stats,       name="stats"),
    path('pathsreport',       statistics.pathsreport, name="pathsreport"),
    path('dataissues',        statistics.dataissues,  name="dataissues"),
    path('therionissues',     statistics.therionissues,  name="therionissues"),
    path('surveximport',      statistics.surveximport,  name="surveximport"),
    path('survexdebug',       statistics.survexdebug, name="survexdebug"),
    path('eastings',          statistics.eastings,    name="eastings"),
    path('aliases/<int:year>',statistics.aliases,     name="aliases"),

    path('troggle',            frontpage,     name="frontpage"), # control panel. Shows recent actions.
    path('todo/<path:module>', todos,         name="todos"),
    path('controlpanel',       controlpanel,  name="controlpanel"),

# The survexfile pages
    path('survexfile',                    survex.survexcavesingle, {'survex_cave': ''},   name="survexcavessingle"),
    path('survexfile/',                   survex.survexcavesingle, {'survex_cave': ''},   name="survexcavessingle"),
    path('survexfile/caves',              survex.survexcaveslist,     name="survexcaveslist"),
    path('survexfile/caves/',             survex.survexcaveslist,     name="survexcaveslist"), # auto slash not working

    path('survexfile/<path:survex_file>.svx', survex.svx,        name="svx"), 
    path('survexfile/<path:survex_file>.3d',  survex.threed,     name="threed"), 
    path('survexfile/<path:survex_file>.log', survex.svxlog,     name="svxlog"),  
    path('survexfile/<path:survex_file>.err', survex.err,        name="err"), 
    path('survexfile/<path:survex_cave>',     survex.survexcavesingle,    name="survexcavessingle"),


# The survey scans in the wallets. This short-cuts SCANS_URL which is not used anymore and is defunct
    path('survey_scans/',                    allscans,    name="allscans"), # all the scans in all wallets
    path('survey_scans/<path:path>/',        walletedit,  name="singlewallet"), # replaced singlewallet()
    path('survey_scans/<path:path>/<file>',  scansingle,  name="scansingle"), # works, but html href goes direct to /expofiles/ too
    path('cave/scans/<slug:caveid>',         cavewallets, name="cavewallets"), # like allscans, but for just one cave
    
# The data about the wallets themselves, not the scans inside tehm
    path('wallets/year/<int:year>',       walletslistyear,  name="walletslistyear"), # wallets that are for a specific year, as an integer '1985'
    re_path('wallets/person/(?P<first_name>[A-Z]*[a-z\-\'&;]*)[^a-zA-Z]*(?P<last_name>[a-z\-\']*[^a-zA-Z]*[\-]*[A-Z]*[a-zA-Z\-&;]*)/?',   walletslistperson,  name="walletslistperson"), 


# The tunnel and therion drawings files pageswalletslistcave
    path('dwgfiles',                      dwgallfiles,       name="dwgallfiles"), 
    path('dwgfiles/',                     dwgallfiles,       name="dwgallfiles"),
    path('dwgdataraw/<path:path>',        dwgfilesingle,     name="dwgfilesingle"), 

# QMs pages - must precede other /caves pages?
    re_path(r'^cave/qms/([^/]+)/?$', caveQMs, name="caveQMs"),
    re_path(r'^cave/qms/(?P<cave_id>[^/]+)/(?P<year>\d\d\d\d)-(?P<qm_id>\d*)(?P<grade>[ABCDXV\?]?)-?(?P<blockname>[a-zA-Z]+.*)?$', qm, name="qm"), # Dogs breakfast
    # the  resolution of a QM uses several fields together, there is no clean slug field. Artefact of history.

#   Prospecting Guide document
    re_path(r'^prospecting_guide/$', prospecting),    # disabled. Bad links, incompatible image package use and very, very out of date.

# This next set are all intercepted by Apache, if it is running.
    re_path(r'^photos/(?P<subpath>.*)$',      mediapage, {'doc_root': settings.PHOTOS_ROOT}, name="mediapage"), # photo galleries 
    re_path(r'^site_media/(?P<subpath>.*)$',  mediapage, {'doc_root': settings.MEDIA_ROOT},  name="mediapage"), # MEDIA_ROOT: CSS and JS 
    re_path(r'^static/(?P<subpath>.*)$',      mediapage, {'doc_root': settings.MEDIA_ROOT},  name="mediapage"), # STATIC only used by admin pages
    path('javascript/<path:subpath>',  mediapage, {'doc_root': settings.JSLIB_ROOT},  name="mediapage"), # JSLIB_URL 
    
    re_path(r'^/loser/(?P<subpath>.*)$',      mediapage, {'doc_root': settings.SURVEX_DATA},  name="mediapage"),  #  Oddly not working !?
    re_path(r'^map/map.html', map,     name="map"), # Redirects to OpenStreetMap JavaScript
    re_path(r'^map/(?P<path>.*)$',     mapfile, name="mapfile"),  # css, js, gpx
    
# Helpers to edit HTML
    re_path(r'^image_selector/(?P<path>.*)', image_selector, name = 'image_selector'),
    re_path(r'^new_image_form/(?P<path>.*)', new_image_form, name = 'new_image_form'),

# Final catchall which also serves expoweb handbook pages and imagestiny
    re_path(r'^(.*)$',              expopage,     name="expopage"),     # CATCHALL assumed relative to EXPOWEB
]

# do NOT allow DIR_ROOT prefix to all urls
urlpatterns = [
    re_path('', include(trogglepatterns))
]

# When apache is running these prempt Django so Django never sees them.
# NB apache has its own ideas about mimetypes, so behaviour may not be identical for .xml files by troggle

# NEW apache configurations suggested as of 2 April 2021: 
# Alias /site-media/ /home/expo/troggle/media/
# Alias /robots.txt  /home/expo/troggle/media/robots.txt   # does not exist!
# Alias /favicon.ico /home/expo/troggle/media/favicon.ico  # comes from /expoweb/* when running runserver
# Alias /javascript  /home/expo/troggle/media/jslib        # empty
# Alias /search       ?        # the Xapian text search thinggy
# Alias /kanboard     ?        # the Kanban Trello-clone thinggy

# Copy of old standard apache configurations:
# Alias /expofiles   /home/expo/expofiles
# Alias /photos      /home/expo/webphotos
# Alias /map         /home/expo/expoweb/map
# Alias /javascript  /usr/share/javascript              # to be changed, see above
# Alias /robots.txt  /home/expo/static/robots.txt       # to be changed, see above
# Alias /favicon.ico /home/expo/static/favicon.ico      # to be deleted. favicon.ico now in expoweb/
# Alias /static/     /home/expo/static/                 # only used by Django admin, tinymce

# ScriptAlias /repositories  /home/expo/config/apache/services/hgweb/hgweb.cgi # UPDATE this for git
# ScriptAlias /boe           /home/expo/boe/boc/boc.pl
# ScriptAlias /boe-lastyear  /home/expo/boe/boc-previous/boc.pl