From d66d52d3ed8aab0b4f126169c2855d00f4053907 Mon Sep 17 00:00:00 2001 From: Andrey Rakhmatullin Date: Tue, 2 Feb 2021 15:21:12 +0500 Subject: [PATCH] Add process_iterable_helper. --- scrapy/utils/middlewares.py | 35 +++++++++++++ tests/test_utils_middlewares.py | 87 +++++++++++++++++++++++++++++++++ 2 files changed, 122 insertions(+) create mode 100644 scrapy/utils/middlewares.py create mode 100644 tests/test_utils_middlewares.py diff --git a/scrapy/utils/middlewares.py b/scrapy/utils/middlewares.py new file mode 100644 index 000000000..da28e0ddf --- /dev/null +++ b/scrapy/utils/middlewares.py @@ -0,0 +1,35 @@ +# coding: utf-8 +import inspect + + +def process_normal_iterable_helper(it, in_predicate=None, out_predicate=None, processor=None): + for o in it: + if in_predicate and not in_predicate(o): + continue + if processor is not None: + o = processor(o) + if out_predicate and not out_predicate(o): + continue + yield o + + +async def process_async_iterable_helper(it, in_predicate=None, out_predicate=None, processor=None): + async for o in it: + if in_predicate and not in_predicate(o): + continue + if processor is not None: + o = processor(o) + if out_predicate and not out_predicate(o): + continue + yield o + + +def process_iterable_helper(it, in_predicate=None, out_predicate=None, processor=None): + """ + For each item in the iterable: skips it if in_predicate is False, applies processor, + skips the result if out_predicate is False, else yields it. + """ + if inspect.isasyncgen(it): + return process_async_iterable_helper(it, in_predicate, out_predicate, processor) + else: + return process_normal_iterable_helper(it, in_predicate, out_predicate, processor) diff --git a/tests/test_utils_middlewares.py b/tests/test_utils_middlewares.py new file mode 100644 index 000000000..d395ba1a9 --- /dev/null +++ b/tests/test_utils_middlewares.py @@ -0,0 +1,87 @@ +import collections + +from twisted.trial import unittest + +from scrapy.utils.asyncgen import as_async_generator, collect_asyncgen +from scrapy.utils.defer import deferred_f_from_coro_f +from scrapy.utils.middlewares import process_iterable_helper + + +def predicate1(o): + return bool(o % 2) + + +def predicate2(o): + return o < 10 + + +def processor(o): + return o * 2 + + +class ProcessIterableHelperNormalTest(unittest.TestCase): + + def test_normal_in_predicate(self): + iterable1 = iter([1, 2, 3]) + iterable2 = process_iterable_helper(iterable1, in_predicate=predicate1) + self.assertIsInstance(iterable2, collections.abc.Iterable) + list2 = list(iterable2) + self.assertEqual(list2, [1, 3]) + + def test_normal_out_predicate(self): + iterable1 = iter([1, 2, 10, 3, 15]) + iterable2 = process_iterable_helper(iterable1, out_predicate=predicate2) + self.assertIsInstance(iterable2, collections.abc.Iterable) + list2 = list(iterable2) + self.assertEqual(list2, [1, 2, 3]) + + def test_normal_processor(self): + iterable1 = iter([1, 2, 3]) + iterable2 = process_iterable_helper(iterable1, processor=processor) + self.assertIsInstance(iterable2, collections.abc.Iterable) + list2 = list(iterable2) + self.assertEqual(list2, [2, 4, 6]) + + def test_normal_combined(self): + iterable1 = iter([1, 2, 10, 3, 6, 18, 5, 15]) + iterable2 = process_iterable_helper(iterable1, in_predicate=predicate1, + out_predicate=predicate2, processor=processor) + self.assertIsInstance(iterable2, collections.abc.Iterable) + list2 = list(iterable2) + self.assertEqual(list2, [2, 6]) + + +class ProcessIterableHelperAsyncTest(unittest.TestCase): + + @deferred_f_from_coro_f + async def test_async_in_predicate(self): + iterable1 = as_async_generator([1, 2, 3]) + iterable2 = process_iterable_helper(iterable1, in_predicate=predicate1) + self.assertIsInstance(iterable2, collections.abc.AsyncIterable) + list2 = await collect_asyncgen(iterable2) + self.assertEqual(list2, [1, 3]) + + @deferred_f_from_coro_f + async def test_async_out_predicate(self): + iterable1 = as_async_generator([1, 2, 10, 3, 15]) + iterable2 = process_iterable_helper(iterable1, out_predicate=predicate2) + self.assertIsInstance(iterable2, collections.abc.AsyncIterable) + list2 = await collect_asyncgen(iterable2) + self.assertEqual(list2, [1, 2, 3]) + + @deferred_f_from_coro_f + async def test_async_processor(self): + iterable1 = as_async_generator([1, 2, 3]) + iterable2 = process_iterable_helper(iterable1, processor=processor) + self.assertIsInstance(iterable2, collections.abc.AsyncIterable) + list2 = await collect_asyncgen(iterable2) + self.assertEqual(list2, [2, 4, 6]) + + @deferred_f_from_coro_f + async def test_async_combined(self): + iterable1 = as_async_generator([1, 2, 10, 3, 6, 18, 5, 15]) + iterable2 = process_iterable_helper(iterable1, in_predicate=predicate1, + out_predicate=predicate2, processor=processor) + self.assertIsInstance(iterable2, collections.abc.AsyncIterable) + list2 = await collect_asyncgen(iterable2) + self.assertEqual(list2, [2, 6])