mirror of
https://github.com/simonw/datasette.git
synced 2025-12-10 16:51:24 +01:00
I've run the black code formatting tool against everything:
black tests datasette setup.py
I also added a new unit test, in tests/test_black.py, which will fail if the code does not
conform to black's exacting standards.
This unit test only runs on Python 3.6 or higher, because black itself doesn't run on 3.5.
101 lines
3.4 KiB
Python
101 lines
3.4 KiB
Python
from .fixtures import ( # noqa
|
|
app_client,
|
|
app_client_csv_max_mb_one,
|
|
app_client_with_cors,
|
|
)
|
|
|
|
EXPECTED_TABLE_CSV = """id,content
|
|
1,hello
|
|
2,world
|
|
3,
|
|
4,RENDER_CELL_DEMO
|
|
""".replace(
|
|
"\n", "\r\n"
|
|
)
|
|
|
|
EXPECTED_CUSTOM_CSV = """content
|
|
hello
|
|
world
|
|
""".replace(
|
|
"\n", "\r\n"
|
|
)
|
|
|
|
EXPECTED_TABLE_WITH_LABELS_CSV = """
|
|
pk,planet_int,on_earth,state,city_id,city_id_label,neighborhood,tags
|
|
1,1,1,CA,1,San Francisco,Mission,"[""tag1"", ""tag2""]"
|
|
2,1,1,CA,1,San Francisco,Dogpatch,"[""tag1"", ""tag3""]"
|
|
3,1,1,CA,1,San Francisco,SOMA,[]
|
|
4,1,1,CA,1,San Francisco,Tenderloin,[]
|
|
5,1,1,CA,1,San Francisco,Bernal Heights,[]
|
|
6,1,1,CA,1,San Francisco,Hayes Valley,[]
|
|
7,1,1,CA,2,Los Angeles,Hollywood,[]
|
|
8,1,1,CA,2,Los Angeles,Downtown,[]
|
|
9,1,1,CA,2,Los Angeles,Los Feliz,[]
|
|
10,1,1,CA,2,Los Angeles,Koreatown,[]
|
|
11,1,1,MI,3,Detroit,Downtown,[]
|
|
12,1,1,MI,3,Detroit,Greektown,[]
|
|
13,1,1,MI,3,Detroit,Corktown,[]
|
|
14,1,1,MI,3,Detroit,Mexicantown,[]
|
|
15,2,0,MC,4,Memnonia,Arcadia Planitia,[]
|
|
""".lstrip().replace(
|
|
"\n", "\r\n"
|
|
)
|
|
|
|
|
|
def test_table_csv(app_client):
|
|
response = app_client.get("/fixtures/simple_primary_key.csv")
|
|
assert response.status == 200
|
|
assert not response.headers.get("Access-Control-Allow-Origin")
|
|
assert "text/plain; charset=utf-8" == response.headers["Content-Type"]
|
|
assert EXPECTED_TABLE_CSV == response.text
|
|
|
|
|
|
def test_table_csv_cors_headers(app_client_with_cors):
|
|
response = app_client_with_cors.get("/fixtures/simple_primary_key.csv")
|
|
assert response.status == 200
|
|
assert "*" == response.headers["Access-Control-Allow-Origin"]
|
|
|
|
|
|
def test_table_csv_with_labels(app_client):
|
|
response = app_client.get("/fixtures/facetable.csv?_labels=1")
|
|
assert response.status == 200
|
|
assert "text/plain; charset=utf-8" == response.headers["Content-Type"]
|
|
assert EXPECTED_TABLE_WITH_LABELS_CSV == response.text
|
|
|
|
|
|
def test_custom_sql_csv(app_client):
|
|
response = app_client.get(
|
|
"/fixtures.csv?sql=select+content+from+simple_primary_key+limit+2"
|
|
)
|
|
assert response.status == 200
|
|
assert "text/plain; charset=utf-8" == response.headers["Content-Type"]
|
|
assert EXPECTED_CUSTOM_CSV == response.text
|
|
|
|
|
|
def test_table_csv_download(app_client):
|
|
response = app_client.get("/fixtures/simple_primary_key.csv?_dl=1")
|
|
assert response.status == 200
|
|
assert "text/csv; charset=utf-8" == response.headers["Content-Type"]
|
|
expected_disposition = 'attachment; filename="simple_primary_key.csv"'
|
|
assert expected_disposition == response.headers["Content-Disposition"]
|
|
|
|
|
|
def test_max_csv_mb(app_client_csv_max_mb_one):
|
|
response = app_client_csv_max_mb_one.get(
|
|
"/fixtures.csv?sql=select+randomblob(10000)+"
|
|
"from+compound_three_primary_keys&_stream=1&_size=max"
|
|
)
|
|
# It's a 200 because we started streaming before we knew the error
|
|
assert response.status == 200
|
|
# Last line should be an error message
|
|
last_line = [line for line in response.body.split(b"\r\n") if line][-1]
|
|
assert last_line.startswith(b"CSV contains more than")
|
|
|
|
|
|
def test_table_csv_stream(app_client):
|
|
# Without _stream should return header + 100 rows:
|
|
response = app_client.get("/fixtures/compound_three_primary_keys.csv?_size=max")
|
|
assert 101 == len([b for b in response.body.split(b"\r\n") if b])
|
|
# With _stream=1 should return header + 1001 rows
|
|
response = app_client.get("/fixtures/compound_three_primary_keys.csv?_stream=1")
|
|
assert 1002 == len([b for b in response.body.split(b"\r\n") if b])
|