From b7ba24b567b21138bccff179fa8500d3f699ad1e Mon Sep 17 00:00:00 2001
From: Edward Betts <edward@4angle.com>
Date: Sat, 22 Aug 2026 21:36:03 +0100
Subject: [PATCH] Handle Wikipedia category API failures

---
 main.py                       | 16 +++++++
 templates/category.html       |  2 +-
 tests/test_category_search.py | 85 +++++++++++++++++++++++++++++++++++
 3 files changed, 102 insertions(+), 1 deletion(-)
 create mode 100644 tests/test_category_search.py

diff --git a/main.py b/main.py
index 111c7bc..ee6fee4 100755
--- a/main.py
+++ b/main.py
@@ -354,6 +354,7 @@ class CategoryResult:
 
     articles: list[ArticleWithoutImage]
     gcmcontinue: str | None
+    error: str | None = None
 
 
 # Common non-content images to ignore when checking if an article has images
@@ -435,6 +436,7 @@ def get_articles_without_images(
     articles_without_images: list[ArticleWithoutImage] = []
     seen_pageids: set[int] = set()
     next_gcmcontinue: str | None = None
+    error: str | None = None
 
     # Build initial continue params from the external pagination token
     continue_params: dict[str, str] = {}
@@ -451,8 +453,20 @@ def get_articles_without_images(
             )
             response.raise_for_status()
             data = response.json()
+            if api_error := data.get("error"):
+                raise requests.RequestException(
+                    f"{api_error.get('code', 'unknown')}: "
+                    f"{api_error.get('info', 'unknown Wikipedia API error')}"
+                )
         except (requests.RequestException, json.JSONDecodeError) as e:
             print(f"Wikipedia API error: {e}")
+            if articles_without_images:
+                error = (
+                    "Wikipedia stopped responding before the whole category could "
+                    "be checked. Showing partial results; please try again."
+                )
+            else:
+                error = "Unable to load articles from Wikipedia. Please try again."
             break
 
         pages = data.get("query", {}).get("pages", {})
@@ -491,6 +505,7 @@ def get_articles_without_images(
     return CategoryResult(
         articles=articles_without_images,
         gcmcontinue=next_gcmcontinue,
+        error=error,
     )
 
 
@@ -891,6 +906,7 @@ def category_search() -> str:
         category_name=category_name,
         articles=result.articles,
         gcmcontinue=result.gcmcontinue,
+        error=result.error,
     )
 
 
diff --git a/templates/category.html b/templates/category.html
index 62fcb1a..2fe8df8 100644
--- a/templates/category.html
+++ b/templates/category.html
@@ -50,7 +50,7 @@
       </div>
       {% endif %}
 
-      {% else %}
+      {% elif not error %}
       <div class="alert alert-success mt-3">
         All articles in this category have images!
       </div>
diff --git a/tests/test_category_search.py b/tests/test_category_search.py
new file mode 100644
index 0000000..8a22e1f
--- /dev/null
+++ b/tests/test_category_search.py
@@ -0,0 +1,85 @@
+"""Tests for Wikipedia category searches."""
+
+from unittest.mock import patch
+
+import requests
+
+from main import CategoryResult, app, get_articles_without_images
+
+
+def test_api_failure_is_not_reported_as_an_empty_success():
+    with patch("main.requests.get", side_effect=requests.ConnectionError("offline")):
+        result = get_articles_without_images("Category:Example")
+
+    assert result.articles == []
+    assert result.error == "Unable to load articles from Wikipedia. Please try again."
+
+
+def test_api_error_response_is_not_reported_as_an_empty_success():
+    error_response = type(
+        "Response",
+        (),
+        {
+            "raise_for_status": lambda self: None,
+            "json": lambda self: {
+                "error": {"code": "readonly", "info": "Wikipedia is read-only"}
+            },
+        },
+    )()
+
+    with patch("main.requests.get", return_value=error_response):
+        result = get_articles_without_images("Category:Example")
+
+    assert result.articles == []
+    assert result.error == "Unable to load articles from Wikipedia. Please try again."
+
+
+def test_api_failure_after_results_marks_them_as_partial():
+    first_response = type(
+        "Response",
+        (),
+        {
+            "raise_for_status": lambda self: None,
+            "json": lambda self: {
+                "query": {
+                    "pages": {
+                        "1": {"pageid": 1, "title": "No image", "images": []}
+                    }
+                },
+                "continue": {
+                    "gcmcontinue": "next-page",
+                    "continue": "gcmcontinue||",
+                },
+            },
+        },
+    )()
+
+    with patch(
+        "main.requests.get",
+        side_effect=[first_response, requests.ConnectionError("offline")],
+    ):
+        result = get_articles_without_images("Category:Example")
+
+    assert [article.title for article in result.articles] == ["No image"]
+    assert result.error is not None
+    assert "partial results" in result.error
+
+
+def test_category_page_does_not_claim_success_after_api_failure():
+    result = CategoryResult(
+        articles=[],
+        gcmcontinue=None,
+        error="Unable to load articles from Wikipedia. Please try again.",
+    )
+
+    with (
+        patch("main.get_articles_without_images", return_value=result),
+        patch("main.log_interaction"),
+        app.test_client() as client,
+    ):
+        response = client.get("/category?cat=Category:Example")
+
+    page = response.get_data(as_text=True)
+    assert response.status_code == 200
+    assert "Unable to load articles from Wikipedia" in page
+    assert "All articles in this category have images" not in page