Fix AuthorizationException with AWSV4SignerAsyncAuth when the doc ID has special characters. (#848)

* Lifecycle integration tests. Signed-off-by: dblock <[email protected]> * Added a test that makes sure the slash is properly encoded. Signed-off-by: dblock <[email protected]> * Added more tests for signer and _make_path. Signed-off-by: Nathalie Jonathan <[email protected]> * Prevent AIOHttpConnection from encoding the url a second time. Signed-off-by: Nathalie Jonathan <[email protected]> --------- Signed-off-by: dblock <[email protected]> Signed-off-by: Nathalie Jonathan <[email protected]> Co-authored-by: dblock <[email protected]>
2024-11-27 17:50:22 -05:00
parent bf9add4eed
commit b9e48dc847
13 changed files with 445 additions and 9 deletions
@@ -29,6 +29,7 @@ from typing import Any
 from unittest import mock

 import pytest
+import yarl
 from multidict import CIMultiDict

 from opensearchpy._async._extra_imports import aiohttp  # type: ignore
@@ -91,7 +92,7 @@ class TestAsyncHttpConnection:
        await c.perform_request("post", "/test")
        mock_request.assert_called_with(
            "post",
-            "http://localhost:9200/test",
+            yarl.URL("http://localhost:9200/test", encoded=True),
            data=None,
            auth=c._http_auth,
            headers={},
@@ -120,7 +121,7 @@ class TestAsyncHttpConnection:

        mock_request.assert_called_with(
            "post",
-            "http://localhost:9200/test",
+            yarl.URL("http://localhost:9200/test", encoded=True),
            data=None,
            auth=None,
            headers={
@@ -30,10 +30,70 @@ from typing import Any
 import pytest
 from _pytest.mark.structures import MarkDecorator

+from opensearchpy.exceptions import RequestError
+
 pytestmark: MarkDecorator = pytest.mark.asyncio


+class TestSpecialCharacters:
+    async def test_index_with_slash(self, async_client: Any) -> None:
+        index_name = "movies/shmovies"
+        with pytest.raises(RequestError) as e:
+            await async_client.indices.create(index=index_name)
+        assert (
+            str(e.value)
+            == "RequestError(400, 'invalid_index_name_exception', 'Invalid index name [movies/shmovies], must not contain the following characters [ , \", *, \\\\, <, |, ,, >, /, ?]')"
+        )
+
+
 class TestUnicode:
+    async def test_indices_lifecycle_english(self, async_client: Any) -> None:
+        index_name = "movies"
+
+        index_create_result = await async_client.indices.create(index=index_name)
+        assert index_create_result["acknowledged"] is True
+        assert index_name == index_create_result["index"]
+
+        document = {"name": "Solaris", "director": "Andrei Tartakovsky", "year": "2011"}
+        id = "solaris@2011"
+        doc_insert_result = await async_client.index(
+            index=index_name, body=document, id=id, refresh=True
+        )
+        assert "created" == doc_insert_result["result"]
+        assert index_name == doc_insert_result["_index"]
+        assert id == doc_insert_result["_id"]
+
+        doc_delete_result = await async_client.delete(index=index_name, id=id)
+        assert "deleted" == doc_delete_result["result"]
+        assert index_name == doc_delete_result["_index"]
+        assert id == doc_delete_result["_id"]
+
+        index_delete_result = await async_client.indices.delete(index=index_name)
+        assert index_delete_result["acknowledged"] is True
+
+    async def test_indices_lifecycle_russian(self, async_client: Any) -> None:
+        index_name = "кино"
+        index_create_result = await async_client.indices.create(index=index_name)
+        assert index_create_result["acknowledged"] is True
+        assert index_name == index_create_result["index"]
+
+        document = {"название": "Солярис", "автор": "Андрей Тарковский", "год": "2011"}
+        id = "соларис@2011"
+        doc_insert_result = await async_client.index(
+            index=index_name, body=document, id=id, refresh=True
+        )
+        assert "created" == doc_insert_result["result"]
+        assert index_name == doc_insert_result["_index"]
+        assert id == doc_insert_result["_id"]
+
+        doc_delete_result = await async_client.delete(index=index_name, id=id)
+        assert "deleted" == doc_delete_result["result"]
+        assert index_name == doc_delete_result["_index"]
+        assert id == doc_delete_result["_id"]
+
+        index_delete_result = await async_client.indices.delete(index=index_name)
+        assert index_delete_result["acknowledged"] is True
+
    async def test_indices_analyze(self, async_client: Any) -> None:
        await async_client.indices.analyze(body='{"text": "привет"}')

@@ -8,6 +8,7 @@
 # GitHub history for details.

 import uuid
+from typing import Any, Collection, Dict, Mapping, Optional, Tuple, Union
 from unittest.mock import Mock

 import pytest
@@ -103,3 +104,75 @@ class TestAsyncSignerWithFrozenCredentials(TestAsyncSigner):
        assert "X-Amz-Date" in headers
        assert "X-Amz-Security-Token" in headers
        assert len(mock_session.get_frozen_credentials.mock_calls) == 1
+
+
+class TestAsyncSignerWithSpecialCharacters:
+    def mock_session(self) -> Mock:
+        access_key = uuid.uuid4().hex
+        secret_key = uuid.uuid4().hex
+        token = uuid.uuid4().hex
+        dummy_session = Mock()
+        dummy_session.access_key = access_key
+        dummy_session.secret_key = secret_key
+        dummy_session.token = token
+
+        del dummy_session.get_frozen_credentials
+
+        return dummy_session
+
+    async def test_aws_signer_async_consitent_url(self) -> None:
+        region = "us-west-2"
+
+        from opensearchpy import AsyncOpenSearch
+        from opensearchpy.connection.http_async import AsyncHttpConnection
+        from opensearchpy.helpers.asyncsigner import AWSV4SignerAsyncAuth
+
+        # Store URLs for comparison
+        signed_url = None
+        sent_url = None
+
+        doc_id = "doc_id:with!special*chars%3A"
+        quoted_doc_id = "doc_id%3Awith%21special*chars%253A"
+        url = f"https://search-domain.region.es.amazonaws.com:9200/index/_doc/{quoted_doc_id}"
+
+        # Create a mock signer class to capture the signed URL
+        class MockSigner(AWSV4SignerAsyncAuth):
+            def _sign_request(
+                self,
+                method: str,
+                url: str,
+                query_string: Optional[str] = None,
+                body: Optional[Union[str, bytes]] = None,
+            ) -> Dict[str, str]:
+                nonlocal signed_url
+                signed_url = url
+                return {}
+
+        # Create a mock connection class to capture the sent URL
+        class MockConnection(AsyncHttpConnection):
+            async def perform_request(
+                self: "MockConnection",
+                method: str,
+                url: str,
+                params: Optional[Mapping[str, Any]] = None,
+                body: Optional[Any] = None,
+                timeout: Optional[Union[int, float]] = None,
+                ignore: Collection[int] = (),
+                headers: Optional[Mapping[str, str]] = None,
+            ) -> Tuple[int, Mapping[str, str], str]:
+                nonlocal sent_url
+                sent_url = f"{self.host}{url}"
+                return 200, {}, "{}"
+
+        auth = MockSigner(self.mock_session(), region)
+        auth("GET", url)
+
+        client = AsyncOpenSearch(
+            hosts=[{"host": "search-domain.region.es.amazonaws.com"}],
+            http_auth=auth,
+            use_ssl=True,
+            verify_certs=True,
+            connection_class=MockConnection,
+        )
+        await client.index("index", {"test": "data"}, id=doc_id)
+        assert signed_url == sent_url, "URLs don't match"