now with configuration

tests/test_config.py

@@ -0,0 +1,131 @@
+from pathlib import Path
+
+from repub.config import (
+    FeedConfig,
+    RepublisherConfig,
+    build_base_settings,
+    build_feed_settings,
+    load_config,
+)
+
+
+def test_load_config_resolves_relative_out_dir_against_config_path(
+    tmp_path: Path,
+) -> None:
+    config_path = tmp_path / "configs" / "repub.toml"
+    config_path.parent.mkdir(parents=True)
+    config_path.write_text(
+        """
+out_dir = "../mirror"
+
+[[feeds]]
+name = "gp-pod"
+url = "https://guardianproject.info/podcast/podcast.xml"
+
+[[feeds]]
+name = "nasa"
+url = "https://www.nasa.gov/rss/dyn/breaking_news.rss"
+""".strip()
+        + "\n",
+        encoding="utf-8",
+    )
+
+    config = load_config(config_path)
+
+    assert config.out_dir == (tmp_path / "mirror").resolve()
+    assert config.feeds == (
+        FeedConfig(
+            name="gp-pod",
+            url="https://guardianproject.info/podcast/podcast.xml",
+        ),
+        FeedConfig(
+            name="nasa",
+            url="https://www.nasa.gov/rss/dyn/breaking_news.rss",
+        ),
+    )
+
+
+def test_load_config_preserves_absolute_out_dir(tmp_path: Path) -> None:
+    absolute_out_dir = (tmp_path / "absolute-out").resolve()
+    config_path = tmp_path / "repub.toml"
+    config_path.write_text(
+        f"""
+out_dir = "{absolute_out_dir}"
+
+[[feeds]]
+name = "nasa"
+url = "https://www.nasa.gov/rss/dyn/breaking_news.rss"
+""".strip()
+        + "\n",
+        encoding="utf-8",
+    )
+
+    config = load_config(config_path)
+
+    assert config.out_dir == absolute_out_dir
+
+
+def test_build_feed_settings_derives_output_paths_from_out_dir(tmp_path: Path) -> None:
+    out_dir = (tmp_path / "mirror").resolve()
+    config = RepublisherConfig(
+        config_path=tmp_path / "repub.toml",
+        out_dir=out_dir,
+        feeds=(
+            FeedConfig(
+                name="nasa",
+                url="https://www.nasa.gov/rss/dyn/breaking_news.rss",
+            ),
+        ),
+        scrapy_settings={"LOG_LEVEL": "DEBUG"},
+    )
+
+    base_settings = build_base_settings(config)
+    feed_settings = build_feed_settings(
+        base_settings, out_dir=out_dir, feed_name="nasa"
+    )
+
+    assert base_settings["LOG_LEVEL"] == "DEBUG"
+    assert feed_settings["REPUBLISHER_OUT_DIR"] == str(out_dir)
+    assert feed_settings["LOG_FILE"] == str(out_dir / "logs" / "nasa.log")
+    assert feed_settings["HTTPCACHE_DIR"] == str(out_dir / "httpcache")
+    assert feed_settings["IMAGES_STORE"] == str(out_dir / "nasa" / "images")
+    assert feed_settings["AUDIO_STORE"] == str(out_dir / "nasa" / "audio")
+    assert feed_settings["VIDEO_STORE"] == str(out_dir / "nasa" / "video")
+    assert feed_settings["FILES_STORE"] == str(out_dir / "nasa" / "files")
+    assert feed_settings["FEEDS"] == {
+        str(out_dir / "nasa.rss"): {
+            "format": "rss",
+            "postprocessing": [],
+            "feed_name": "nasa",
+        }
+    }
+
+
+def test_build_feed_settings_uses_runtime_media_dir_overrides(tmp_path: Path) -> None:
+    out_dir = (tmp_path / "mirror").resolve()
+    config = RepublisherConfig(
+        config_path=tmp_path / "repub.toml",
+        out_dir=out_dir,
+        feeds=(
+            FeedConfig(
+                name="gp-pod",
+                url="https://guardianproject.info/podcast/podcast.xml",
+            ),
+        ),
+        scrapy_settings={
+            "REPUBLISHER_VIDEO_DIR": "videos-custom",
+            "REPUBLISHER_AUDIO_DIR": "audio-custom",
+        },
+    )
+
+    base_settings = build_base_settings(config)
+    feed_settings = build_feed_settings(
+        base_settings,
+        out_dir=out_dir,
+        feed_name="gp-pod",
+    )
+
+    assert feed_settings["REPUBLISHER_VIDEO_DIR"] == "videos-custom"
+    assert feed_settings["REPUBLISHER_AUDIO_DIR"] == "audio-custom"
+    assert feed_settings["VIDEO_STORE"] == str(out_dir / "gp-pod" / "videos-custom")
+    assert feed_settings["AUDIO_STORE"] == str(out_dir / "gp-pod" / "audio-custom")