truncate_cells_html now affects URLs too, refs #1805

datasette/utils/__init__.py

@@ -1167,3 +1167,13 @@ def resolve_routes(routes, path):
         if match is not None:
             return match, view
     return None, None
+
+
+def truncate_url(url, length):
+    if (not length) or (len(url) <= length):
+        return url
+    bits = url.rsplit(".", 1)
+    if len(bits) == 2 and 1 <= len(bits[1]) <= 4 and "/" not in bits[1]:
+        rest, ext = bits
+        return rest[: length - 1 - len(ext)] + "…." + ext
+    return url[: length - 1] + "…"

datasette/views/database.py

@@ -20,6 +20,7 @@ from datasette.utils import (
     path_with_format,
     path_with_removed_args,
     sqlite3,
+    truncate_url,
     InvalidSql,
 )
 from datasette.utils.asgi import AsgiFileDownload, NotFound, Response, Forbidden
@@ -371,6 +372,7 @@ class QueryView(DataView):
 
         async def extra_template():
             display_rows = []
+            truncate_cells = self.ds.setting("truncate_cells_html")
             for row in results.rows if results else []:
                 display_row = []
                 for column, value in zip(results.columns, row):
@@ -396,9 +398,12 @@ class QueryView(DataView):
                         if value in ("", None):
                             display_value = Markup(" ")
                         elif is_url(str(display_value).strip()):
-                            display_value = Markup(
+                            display_value = markupsafe.Markup(
-                                '<a href="{url}">{url}</a>'.format(
+                                '<a href="{url}">{truncated_url}</a>'.format(
-                                    url=escape(value.strip())
+                                    url=markupsafe.escape(value.strip()),
+                                    truncated_url=markupsafe.escape(
+                                        truncate_url(value.strip(), truncate_cells)
+                                    ),
                                 )
                             )
                         elif isinstance(display_value, bytes):

datasette/views/table.py

@@ -24,6 +24,7 @@ from datasette.utils import (
     path_with_removed_args,
     path_with_replaced_args,
     to_css_class,
+    truncate_url,
     urlsafe_components,
     value_as_boolean,
 )
@@ -966,8 +967,11 @@ async def display_columns_and_rows(
                 display_value = markupsafe.Markup(" ")
             elif is_url(str(value).strip()):
                 display_value = markupsafe.Markup(
-                    '<a href="{url}">{url}</a>'.format(
+                    '<a href="{url}">{truncated_url}</a>'.format(
-                        url=markupsafe.escape(value.strip())
+                        url=markupsafe.escape(value.strip()),
+                        truncated_url=markupsafe.escape(
+                            truncate_url(value.strip(), truncate_cells)
+                        ),
                     )
                 )
             elif column in table_metadata.get("units", {}) and value != "":

tests/fixtures.py

@@ -598,23 +598,24 @@ CREATE TABLE roadside_attractions (
     pk integer primary key,
     name text,
     address text,
+    url text,
     latitude real,
     longitude real
 );
 INSERT INTO roadside_attractions VALUES (
-    1, "The Mystery Spot", "465 Mystery Spot Road, Santa Cruz, CA 95065",
+    1, "The Mystery Spot", "465 Mystery Spot Road, Santa Cruz, CA 95065", "https://www.mysteryspot.com/",
     37.0167, -122.0024
 );
 INSERT INTO roadside_attractions VALUES (
-    2, "Winchester Mystery House", "525 South Winchester Boulevard, San Jose, CA 95128",
+    2, "Winchester Mystery House", "525 South Winchester Boulevard, San Jose, CA 95128", "https://winchestermysteryhouse.com/",
     37.3184, -121.9511
 );
 INSERT INTO roadside_attractions VALUES (
-    3, "Burlingame Museum of PEZ Memorabilia", "214 California Drive, Burlingame, CA 94010",
+    3, "Burlingame Museum of PEZ Memorabilia", "214 California Drive, Burlingame, CA 94010", null,
     37.5793, -122.3442
 );
 INSERT INTO roadside_attractions VALUES (
-    4, "Bigfoot Discovery Museum", "5497 Highway 9, Felton, CA 95018",
+    4, "Bigfoot Discovery Museum", "5497 Highway 9, Felton, CA 95018", "https://www.bigfootdiscoveryproject.com/",
     37.0414, -122.0725
 );
 

tests/test_api.py

@@ -339,7 +339,7 @@ def test_database_page(app_client):
         },
         {
             "name": "roadside_attractions",
-            "columns": ["pk", "name", "address", "latitude", "longitude"],
+            "columns": ["pk", "name", "address", "url", "latitude", "longitude"],
             "primary_keys": ["pk"],
             "count": 4,
             "hidden": False,

tests/test_table_api.py

@@ -615,11 +615,12 @@ def test_table_through(app_client):
     response = app_client.get(
         '/fixtures/roadside_attractions.json?_through={"table":"roadside_attraction_characteristics","column":"characteristic_id","value":"1"}'
     )
-    assert [
+    assert response.json["rows"] == [
         [
             3,
             "Burlingame Museum of PEZ Memorabilia",
             "214 California Drive, Burlingame, CA 94010",
+            None,
             37.5793,
             -122.3442,
         ],
@@ -627,13 +628,15 @@ def test_table_through(app_client):
             4,
             "Bigfoot Discovery Museum",
             "5497 Highway 9, Felton, CA 95018",
+            "https://www.bigfootdiscoveryproject.com/",
             37.0414,
             -122.0725,
         ],
-    ] == response.json["rows"]
+    ]
+
     assert (
-        'where roadside_attraction_characteristics.characteristic_id = "1"'
+        response.json["human_description_en"]
-        == response.json["human_description_en"]
+        == 'where roadside_attraction_characteristics.characteristic_id = "1"'
     )
 
 

tests/test_table_html.py

@@ -69,6 +69,17 @@ def test_table_cell_truncation():
             td.string
             for td in table.findAll("td", {"class": "col-neighborhood-b352a7"})
         ]
+        # URLs should be truncated too
+        response2 = client.get("/fixtures/roadside_attractions")
+        assert response2.status == 200
+        table = Soup(response2.body, "html.parser").find("table")
+        tds = table.findAll("td", {"class": "col-url"})
+        assert [str(td) for td in tds] == [
+            '<td class="col-url type-str"><a href="https://www.mysteryspot.com/">http…</a></td>',
+            '<td class="col-url type-str"><a href="https://winchestermysteryhouse.com/">http…</a></td>',
+            '<td class="col-url type-none">\xa0</td>',
+            '<td class="col-url type-str"><a href="https://www.bigfootdiscoveryproject.com/">http…</a></td>',
+        ]
 
 
 def test_add_filter_redirects(app_client):

tests/test_utils.py

@@ -626,3 +626,23 @@ def test_tilde_encoding(original, expected):
     assert actual == expected
     # And test round-trip
     assert original == utils.tilde_decode(actual)
+
+
+@pytest.mark.parametrize(
+    "url,length,expected",
+    (
+        ("https://example.com/", 5, "http…"),
+        ("https://example.com/foo/bar", 15, "https://exampl…"),
+        ("https://example.com/foo/bar/baz.jpg", 30, "https://example.com/foo/ba….jpg"),
+        # Extensions longer than 4 characters are not treated specially:
+        ("https://example.com/foo/bar/baz.jpeg2", 30, "https://example.com/foo/bar/b…"),
+        (
+            "https://example.com/foo/bar/baz.jpeg2",
+            None,
+            "https://example.com/foo/bar/baz.jpeg2",
+        ),
+    ),
+)
+def test_truncate_url(url, length, expected):
+    actual = utils.truncate_url(url, length)
+    assert actual == expected