From 9481d2a725c0354cf8a3c7eca8b4d87076bf9c8c Mon Sep 17 00:00:00 2001
From: Stephen Rosen <sirosen@globus.org>
Date: Sat, 29 Jul 2023 22:22:09 -0500
Subject: [PATCH] Introduce remote $ref tests and fix URL join logic

URL joining needs to be done between any relative ref paths and the
base URI. This was not correct, as revealed by a new testcase.
After fixing, this revealed that the local file loading was somewhat
indirect in a way that it no longer needs to be, so local file
handling in the ref resolver has been refactored as well.
---
 .../schema_loader/resolver.py                 | 26 +++----
 .../acceptance/test_remote_ref_resolution.py  | 76 +++++++++++++++++++
 2 files changed, 87 insertions(+), 15 deletions(-)
 create mode 100644 tests/acceptance/test_remote_ref_resolution.py

diff --git a/src/check_jsonschema/schema_loader/resolver.py b/src/check_jsonschema/schema_loader/resolver.py
index 6207b30cf..5d658eadc 100644
--- a/src/check_jsonschema/schema_loader/resolver.py
+++ b/src/check_jsonschema/schema_loader/resolver.py
@@ -1,6 +1,5 @@
 from __future__ import annotations
 
-import pathlib
 import typing as t
 import urllib.parse
 
@@ -38,25 +37,22 @@ def create_retrieve_callable(
     parser_set: ParserSet, schema_uri: str | None
 ) -> t.Callable[[str], referencing.Resource[Schema]]:
     def get_local_file(uri: str) -> t.Any:
-        path = pathlib.Path(uri)
-        if not path.is_absolute():
-            if schema_uri is None:
-                raise referencing.exceptions.Unretrievable(
-                    f"Cannot retrieve schema reference data for '{uri}' from "
-                    "local filesystem. "
-                    "The path appears relative, but there is no known local base path."
-                )
-            schema_path = filename2path(schema_uri)
-            path = schema_path.parent / path
+        path = filename2path(uri)
         return parser_set.parse_file(path, "json")
 
     def retrieve_reference(uri: str) -> referencing.Resource[Schema]:
         scheme = urllib.parse.urlsplit(uri).scheme
-        if scheme in ("http", "https"):
-            data = requests.get(uri, stream=True)
-            parsed_object = parser_set.parse_data_with_path(data.raw, uri, "json")
+        if scheme == "" and schema_uri is not None:
+            full_uri = urllib.parse.urljoin(schema_uri, uri)
         else:
-            parsed_object = get_local_file(uri)
+            full_uri = uri
+
+        full_uri_scheme = urllib.parse.urlsplit(full_uri).scheme
+        if full_uri_scheme in ("http", "https"):
+            data = requests.get(full_uri, stream=True)
+            parsed_object = parser_set.parse_data_with_path(data.raw, full_uri, "json")
+        else:
+            parsed_object = get_local_file(full_uri)
 
         return referencing.Resource.from_contents(
             parsed_object, default_specification=DRAFT202012
diff --git a/tests/acceptance/test_remote_ref_resolution.py b/tests/acceptance/test_remote_ref_resolution.py
new file mode 100644
index 000000000..626b93bf9
--- /dev/null
+++ b/tests/acceptance/test_remote_ref_resolution.py
@@ -0,0 +1,76 @@
+import json
+
+import pytest
+import responses
+
+from check_jsonschema import cachedownloader
+
+CASES = {
+    "case1": {
+        "main_schema": {
+            "$schema": "http://json-schema.org/draft-07/schema",
+            "properties": {
+                "title": {"$ref": "./title_schema.json"},
+            },
+            "additionalProperties": False,
+        },
+        "other_schemas": {"title_schema": {"type": "string"}},
+        "passing_document": {"title": "doc one"},
+        "failing_document": {"title": 2},
+    },
+    "case2": {
+        "main_schema": {
+            "$schema": "http://json-schema.org/draft-07/schema",
+            "type": "object",
+            "required": ["test"],
+            "properties": {"test": {"$ref": "./values.json#/$defs/test"}},
+        },
+        "other_schemas": {
+            "values": {
+                "$schema": "http://json-schema.org/draft-07/schema",
+                "$defs": {"test": {"type": "string"}},
+            }
+        },
+        "passing_document": {"test": "some data"},
+        "failing_document": {"test": {"foo": "bar"}},
+    },
+}
+
+
+@pytest.mark.parametrize("check_passes", (True, False))
+@pytest.mark.parametrize("casename", ("case1", "case2"))
+def test_remote_ref_resolution_simple_case(
+    run_line, check_passes, casename, tmp_path, monkeypatch
+):
+    def _fake_compute_default_cache_dir(self):
+        return str(tmp_path)
+
+    monkeypatch.setattr(
+        cachedownloader.CacheDownloader,
+        "_compute_default_cache_dir",
+        _fake_compute_default_cache_dir,
+    )
+
+    main_schema_loc = "https://example.com/main.json"
+    responses.add("GET", main_schema_loc, json=CASES[casename]["main_schema"])
+    for name, subschema in CASES[casename]["other_schemas"].items():
+        other_schema_loc = f"https://example.com/{name}.json"
+        responses.add("GET", other_schema_loc, json=subschema)
+
+    instance_path = tmp_path / "instance.json"
+    instance_path.write_text(
+        json.dumps(
+            CASES[casename]["passing_document"]
+            if check_passes
+            else CASES[casename]["failing_document"]
+        )
+    )
+
+    result = run_line(
+        ["check-jsonschema", "--schemafile", main_schema_loc, str(instance_path)]
+    )
+    output = f"\nstdout:\n{result.stdout}\n\nstderr:\n{result.stderr}"
+    if check_passes:
+        assert result.exit_code == 0, output
+    else:
+        assert result.exit_code == 1, output