232 export files improvements

core/settings.py

@@ -54,13 +54,12 @@
     "django.contrib.humanize",
     "data_registry",
     "markdownx",
-    "exporter",
 ]
 
 MIDDLEWARE = [
     "django.middleware.security.SecurityMiddleware",
     # Add before GZipMiddleware to modify its response.
-    "exporter.middleware.ContentEncodingMiddleware",
+    "data_registry.middleware.ContentEncodingMiddleware",
     # This site is not affected by BREACH.
     # https://docs.djangoproject.com/en/4.2/ref/middleware/#django.middleware.gzip.GZipMiddleware
     "django.middleware.gzip.GZipMiddleware",

core/urls.py

@@ -4,7 +4,6 @@
 
 urlpatterns = [
     path("", include("data_registry.urls"), name="data-registry"),
-    path("", include("exporter.urls"), name="exporter"),
     path("admin/", admin.site.urls),
     path("markdownx/", include("markdownx.urls")),
 ]

data_registry/templates/includes/files.html

@@ -4,13 +4,13 @@
 <ul>
 {% if files.full %}
     <li>
-        <a href="{% url 'download-export' %}?spider={{ collection.source_id }}&job_id={{ job.id }}&full=true&suffix={{ suffix }}" rel="nofollow" download>{% translate "All time" %}</a>
+        <a href="{% url 'download' collection.id %}?name=full.{{ suffix }}" rel="nofollow" download>{% translate "All time" %}</a>
         <span class="text-muted small">{{ files.full|humanfilesize }}</span>
     </li>
 {% endif %}
 {% for file in files.by_year|dictsortreversed:"year" %}
     <li>
-        <a href="{% url 'download-export' %}?spider={{ collection.source_id }}&job_id={{ job.id }}&year={{ file.year }}&suffix={{ suffix }}" rel="nofollow" download>{{ file.year }}</a>
+        <a href="{% url 'download' collection.id %}?name={{ file.year }}.{{ suffix }}" rel="nofollow" download>{{ file.year }}</a>
         <span class="text-muted small">{{ file.size|humanfilesize }}</span>
     </li>
 {% empty %}

data_registry/urls.py

@@ -6,6 +6,7 @@
     path("", views.index, name="index"),
     path("search/", views.search, name="search"),
     path("publication/<int:id>", views.detail, name="detail"),
+    path("publication/<int:id>/download", views.download_export, name="download"),
     # https://code.djangoproject.com/ticket/26556
     path("i18n/setlang/", i18n.set_language, name="set-language"),
     # Uncomment after re-integrating Spoonbill.

data_registry/views.py

@@ -1,4 +1,5 @@
 import logging
+import re
 import string
 from collections import defaultdict
 from datetime import date, datetime, timedelta
@@ -11,20 +12,21 @@
 from django.contrib.postgres.aggregates import ArrayAgg
 from django.db.models import Count, OuterRef, Q, Subquery
 from django.db.models.functions import Substr
-from django.http.response import HttpResponse, JsonResponse
+from django.http.response import FileResponse, HttpResponse, HttpResponseBadRequest, HttpResponseNotFound, JsonResponse
 from django.shortcuts import get_object_or_404, redirect, render
 from django.utils.translation import get_language, get_language_from_request
 from django.utils.translation import gettext as _
 
 from data_registry.models import Collection, Job
 from data_registry.util import collection_queryset
-from exporter.util import Export
+from exporter.util import Export, TaskStatus
 
 logger = logging.getLogger(__name__)
 
 alphabets = defaultdict(lambda: string.ascii_uppercase)
 # https://en.wikipedia.org/wiki/Cyrillic_script_in_Unicode#Basic_Cyrillic_alphabet
 alphabets["ru"] = "АБВГДЕЁЖЗИЙКЛМНОПРСТУФХЦЧШЩЪЫЬЭЮЯ"
+EXPORT_PATTERN = re.compile(r"\A(full|\d{4})\.(jsonl\.gz|csv\.tar\.gz|xlsx)\Z")
 
 
 def index(request):
@@ -174,6 +176,30 @@ def spiders(request):
     return JsonResponse(json.get("spiders"), safe=False)
 
 
+def download_export(request, id):
+    name = request.GET.get("name")
+
+    # Guard against path traversal.
+    if not EXPORT_PATTERN.match(name):
+        return HttpResponseBadRequest("The name query string parameter is invalid")
+
+    collection = get_object_or_404(collection_queryset(request), id=id)
+    active_job = get_object_or_404(collection.job, active=True)
+
+    export = Export(active_job.id, basename=name)
+    if export.status != TaskStatus.COMPLETED:
+        return HttpResponseNotFound("File not found")
+
+    return FileResponse(
+        export.path.open("rb"),
+        as_attachment=True,
+        filename=f"{collection.source_id}_{name}",
+        # Set Content-Encoding to skip GZipMiddleware. (ContentEncodingMiddleware removes the empty header.)
+        # https://docs.djangoproject.com/en/4.2/ref/middleware/#module-django.middleware.gzip
+        headers={"Content-Encoding": ""},
+    )
+
+
 def excel_data(request, job_id, job_range=None):
     job = Job.objects.get(id=job_id)
     export = Export(job_id)

tests/data_registry/test_views.py

@@ -21,7 +21,7 @@ def setUpTestData(cls):
     @patch("exporter.util.Export.get_files")
     def test_detail(self, get_files):
         get_files.return_value = {"jsonl": {"by_year": [{"year": 2022, "size": 1}]}}
-        url = f"/api/download_export?spider=paraguay_dncp_records&job_id={self.job.id}&year=2022&suffix=jsonl.gz"
+        url = f"/en/publication/{self.collection.id}/download?name=2022.jsonl.gz"
 
         with self.assertNumQueries(2):
             response = Client().get(f"/en/publication/{self.collection.id}")

tests/exporter/test_views.py

@@ -3,36 +3,64 @@
 
 from django.test import Client, TestCase, override_settings
 
+from data_registry.models import Collection
+
 
 @override_settings(EXPORTER_DIR=os.path.join("tests", "fixtures"))
 class ViewsTests(TestCase):
+    @classmethod
+    def setUp(cls):
+        cls.collection = Collection.objects.create(
+            id=2,
+            title="Dirección Nacional de Contrataciones Públicas (DNCP)",
+            source_id="abc",
+            public=True,
+        )
+        cls.job = cls.collection.job.create(
+            active=True,
+            id=2,
+        )
+        cls.collection_no_job = Collection.objects.create(
+            id=3,
+            title="Test",
+            source_id="abc",
+            public=True,
+        )
+        cls.collection_no_job.job.create(
+            active=True,
+            id=4,
+        )
+
+    def test_collection_not_found(self):
+        with self.assertNumQueries(1):
+            response = Client().get("/en/publication/10/download?name=2000.jsonl.gz")
+
+        self.assertEqual(response.status_code, 404)
+
     def test_download_export_invalid_suffix(self):
         with self.assertNumQueries(0):
-            response = Client().get("/api/download_export?suffix=invalid")
+            response = Client().get(f"/en/publication/{self.collection.id}/download?name=invalid")
 
         self.assertEqual(response.status_code, 400)
-        self.assertEqual(response.content, b"Suffix not recognized")
+        self.assertEqual(response.content, b"The name query string parameter is invalid")
 
     def test_download_export_empty_parameter(self):
-        for parameter in ("job_id", "year"):
-            with self.subTest(parameter=parameter):
-                with self.assertNumQueries(0):
-                    response = Client().get(f"/api/download_export?suffix=jsonl.gz&{parameter}=")
+        with self.assertNumQueries(0):
+            response = Client().get(f"/en/publication/{self.collection.id}/download?name=")
 
-                self.assertEqual(response.status_code, 404)
-                self.assertEqual(response.content, b"File not found")
+        self.assertEqual(response.status_code, 400)
+        self.assertEqual(response.content, b"The name query string parameter is invalid")
 
     def test_download_export_waiting(self):
-        with self.assertNumQueries(0):
-            response = Client().get("/api/download_export?suffix=jsonl.gz&year=2000&job_id=0")
-
+        with self.assertNumQueries(2):
+            response = Client().get(f"/en/publication/{self.collection_no_job.id}/download?name=2000.jsonl.gz")
         self.assertEqual(response.status_code, 404)
         self.assertEqual(response.content, b"File not found")
 
     @patch("exporter.util.Export.lockfile", new_callable=PropertyMock)
     def test_download_export_running(self, exists):
-        with self.assertNumQueries(0):
-            response = Client().get("/api/download_export?suffix=jsonl.gz&year=2000&job_id=1")
+        with self.assertNumQueries(2):
+            response = Client().get(f"/en/publication/{self.collection.id}/download?name=2000.jsonl.gz")
 
         self.assertEqual(response.status_code, 404)
         self.assertEqual(response.content, b"File not found")
@@ -44,12 +72,11 @@ def test_download_export_completed(self):
             ("xlsx", "application/vnd.openxmlformats-officedocument.spreadsheetml.sheet"),
         ):
             with self.subTest(suffix=suffix):
-                with self.assertNumQueries(0):
+                with self.assertNumQueries(2):
                     response = Client().get(
-                        f"/api/download_export?suffix={suffix}&year=2000&job_id=1&spider=abc",
+                        f"/en/publication/{self.collection.id}/download?name=2000.{suffix}",
                         HTTP_ACCEPT_ENCODING="gzip",
                     )
-
                 self.assertEqual(response.status_code, 200)
                 response.headers.pop("Content-Length")
                 self.assertDictEqual(
@@ -60,10 +87,10 @@ def test_download_export_completed(self):
                         "Content-Type": content_type,
                         "Cross-Origin-Opener-Policy": "same-origin",
                         "Referrer-Policy": "same-origin",
-                        "Vary": "Accept-Language",
+                        "Vary": "Cookie",
                         "X-Content-Type-Options": "nosniff",
                         "X-Frame-Options": "DENY",
                     },
                 )
-                with open(os.path.join("tests", "fixtures", "1", f"2000.{suffix}"), "rb") as f:
+                with open(os.path.join("tests", "fixtures", "2", f"2000.{suffix}"), "rb") as f:
                     self.assertEqual(b"".join(response.streaming_content), f.read())