Add archive support, fix #1634

interface/urls.py

@@ -1,21 +1,41 @@
 # Copyright: 2022, ECP, NLnet Labs and the Internet.nl contributors
 # SPDX-License-Identifier: Apache-2.0
+from dateutil.parser import isoparse
+
 from django.conf import settings
-from django.urls import path, re_path
+from django.urls import path, re_path, register_converter
 from django.conf.urls.static import static
 
 from interface import views
 from interface.batch import BATCH_API_MAJOR_VERSION
 from interface.batch import views as batch
 from interface.views import connection, domain, mail, stats
 
+
 regex_tld = r"(?:[a-zA-Z]{2,63}|xn--[a-zA-Z0-9]+)"
 regex_dname = r"(?P<dname>([a-zA-Z0-9]([a-zA-Z0-9\-]{0,61}[a-zA-Z0-9])?\.)+" + regex_tld + ")"
 regex_testid = r"(?P<request_id>[a-zA-Z0-9]{1,35})"
 regex_mailaddr = (
     r"(?P<mailaddr>([a-zA-Z0-9]{0,61}@)?([a-zA-Z0-9]([a-zA-Z0-9\-]{0,61}[a-zA-Z0-9])?\.)+" r"" + regex_tld + ")"
 )
 
+
+class DateConverter:
+    # be as lenient as the isoparse library allows
+    regex = ".+"
+
+    def to_python(self, value):
+        # until python3.10 use dateutil.isoparser
+        # see: https://stackoverflow.com/questions/127803/how-do-i-parse-an-iso-8601-formatted-date-and-time
+        return isoparse(value)
+
+    def to_url(self, value):
+        return value
+
+
+register_converter(DateConverter, "isodatetime")
+
+
 urlpatterns = [
     path("", views.indexpage),
     re_path(r"^statistics/(?P<start_date>[0-9]{8})/(?P<end_date>[0-9]{8})/$", stats.statistics),
@@ -47,6 +67,7 @@
     # these url()s should always be the last in the ^domain/ group
     re_path(r"^(domain|site)/(?P<dname>.*)/$", domain.validate_domain),
     re_path(r"^(domain|site)/(?P<dname>.*)/results$", domain.validate_domain),
+    path(r"archive/site/<str:domain_name>/<isodatetime:date>/", domain.resultsstored_historic),
     path("test-mail/", views.testmailpage),
     path("mail/", mail.index),
     re_path(rf"^mail/{regex_mailaddr}/$", mail.mailprocess),
@@ -58,6 +79,7 @@
     # these url()s should always be the last in the ^mail/ group
     re_path(r"^mail/(?P<mailaddr>.*)/$", mail.validate_domain),
     re_path(r"^mail/(?P<mailaddr>.*)/results$", mail.validate_domain),
+    path(r"archive/mail/<str:domain_name>/<isodatetime:date>/", mail.resultsstored_historic),
     re_path(rf"^clear/{regex_dname}/$", views.clear),
     path("change_language/", views.change_language, name="change_language"),
     path("contact/", views.indexpage),

interface/views/domain.py

@@ -6,7 +6,7 @@
 from django.conf import settings
 from django.core.cache import cache
 from django.core.exceptions import DisallowedRedirect
-from django.http import HttpResponse, HttpResponseRedirect
+from django.http import HttpResponse, HttpResponseRedirect, HttpResponseNotFound
 from django.shortcuts import render
 from django.utils.translation import gettext as _
 
@@ -34,6 +34,7 @@
     redirect_invalid_domain,
     update_report_with_registrar_and_score,
     SafeHttpResponseRedirect,
+    latest_report,
 )
 
 # Entrance after form submission.
@@ -188,6 +189,12 @@ def resultscurrent(request, dname):
     return HttpResponseRedirect(f"/site/{addr}/{report.id}/")
 
 
+def resultsstored_historic(request, domain_name, date):
+    if report_id := latest_report(DomainTestReport, domain_name, date):
+        return HttpResponseRedirect(f"/site/{domain_name}/{report_id}/")
+    return HttpResponseNotFound()
+
+
 # URL: /(site|domain)/<dname>/<reportid>/
 @simple_cache_page
 def resultsstored(request, dname, id):

interface/views/mail.py

@@ -5,7 +5,7 @@
 
 from django.conf import settings
 from django.core.cache import cache
-from django.http import HttpResponse, HttpResponseRedirect
+from django.http import HttpResponse, HttpResponseRedirect, HttpResponseNotFound
 from django.shortcuts import render
 from django.utils.translation import gettext as _
 
@@ -32,6 +32,7 @@
     process,
     redirect_invalid_domain,
     update_report_with_registrar_and_score,
+    latest_report,
 )
 from internetnl import log
 
@@ -182,6 +183,12 @@ def resultscurrent(request, mailaddr):
     return HttpResponseRedirect(f"/mail/{addr}/{report.id}/")
 
 
+def resultsstored_historic(request, domain_name, date):
+    if report_id := latest_report(MailTestReport, domain_name, date):
+        return HttpResponseRedirect(f"/mail/{domain_name}/{report_id}/")
+    return HttpResponseNotFound()
+
+
 # URL: /mail/<dname>/<reportid>/
 def resultsstored(request, dname, id):
     """

interface/views/shared.py

@@ -3,6 +3,7 @@
 import random
 import re
 from datetime import datetime
+from typing import Optional
 from urllib.parse import urlparse
 
 import dns
@@ -27,7 +28,6 @@
 from interface import redis_id
 from internetnl import log
 
-
 # See: https://stackoverflow.com/a/53875771 for a good summary of the various
 # RFCs and other rulings that combine to define what is a valid domain name.
 # Of particular note are xn-- which is used for internationalized TLDs, and
@@ -482,3 +482,25 @@ def __init__(self, redirect_to, *args, **kwargs):
 
         if not settings.DEBUG and not url_has_allowed_host_and_scheme(redirect_to, allowed_hosts=allowed_hosts):
             raise DisallowedRedirect("Unsafe redirect to URL: %s" % redirect_to)
+
+
+def latest_report(model, domain_name: str, date: datetime) -> Optional[int]:
+    # probably the default ordering in the database is on -id or -timestamp. Both will work.
+    # domain does not have an index, timestamp neither. This might be a problem, but it also might just work fine.
+
+    # for caching purposes don't allow users to request domains newer than today:
+    # assume users don't enter timezones, they can, and that will crash here due to non-zone-awareness :)
+    if date > datetime.now():
+        log.debug("Will not hand out reports in the future.")
+        date = datetime.now()
+
+    # if no time is given, the first moment of the day is assumed, at that time the report did not exist yet.
+    # so if people try with just a date, they will not get what they expect: the latest report for that day.
+    if date.hour == 0 and date.minute == 0 and date.second == 0:
+        log.debug("Aiding the user to set the date of the date to the latest report of the day.")
+        date = date.replace(hour=23, minute=59, second=59)  # try to get the latest report of this day.
+
+    report = (
+        model.objects.all().filter(domain=domain_name, timestamp__lte=date).order_by("-timestamp").only("id").first()
+    )
+    return report.id if report else None