[FEATURE] Introduce Config API

eliashaeussler · Feb 22, 2024 · 7e8f626 · 7e8f626
1 parent 0442a60
commit 7e8f626
Show file tree

Hide file tree

Showing 17 changed files with 1,228 additions and 58 deletions.
diff --git a/README.md b/README.md
@@ -86,6 +86,7 @@ The following input parameters are available:
 |-------------------------------------------------|------------------------------------------------------------------------|
 | [`sitemaps`](#sitemaps)                         | URLs or local filenames of XML sitemaps to be warmed up                |
 | [`--allow-failures`](#--allow-failures)         | Allow failures during URL crawling and exit with zero                  |
+| [`--config`](#--config)                         | Path to configuration file                                             |
 | [`--crawler-options`, `-o`](#--crawler-options) | JSON-encoded string of additional config for configurable crawlers     |
 | [`--crawler`, `-c`](#--crawler)                 | FQCN of the crawler to use for cache warmup                            |
 | [`--exclude`, `-e`](#--exclude)                 | Patterns of URLs to be excluded from cache warmup                      |
@@ -130,6 +131,31 @@ $ cache-warmup -u "https://www.example.org/" -u "https://www.example.org/de/"
 | Multiple values allowed | ✅                                                                            |
 | Default                 | **–**                                                                        |
 
+#### `--config`
+
+Path to a configuration file. Read more in the
+[Configuration file](#configuration-file) section below.
+
+At the moment, the following file formats are available:
+
+* [`json`][25]
+* [`php`][26]
+* [`yaml`/`yml`][25]
+
+> [!NOTE]
+> Configuration options provided by command options take
+> precedence over those provided by configuration files.
+
+```bash
+$ cache-warmup --config cache-warmup.yaml
+```
+
+| Shorthand               | –     |
+|:------------------------|:------|
+| Required                | **–** |
+| Multiple values allowed | **–** |
+| Default                 | **–** |
+
 #### `--exclude`
 
 Patterns of URLs to be excluded from cache warmup.
@@ -533,6 +559,41 @@ object. It includes the following properties:
 
 💡 The complete JSON structure can be found in the provided [JSON schema][22].
 
+### Configuration file
+
+Instead of passing all available configuration options as
+command parameters, it is also possible to provide a
+configuration file to be used for cache warmup. At the moment,
+the following file formats are supported:
+
+* [`json`][25]
+* [`php`][26]
+* [`yaml`/`yml`][25]
+
+For JSON and YAML files, the name of each configuration option
+can be derived from the list of available
+[command parameters](#command-line-usage) and must be written in
+camel case, e.g. `crawler-options` is configured as `crawlerOptions`.
+
+Example:
+
+```yaml
+sitemaps:
+  - https://www.example.org/sitemap.xml
+
+# Crawler options must be written in object notation
+# instead of a JSON-encoded string
+crawlerOptions:
+  concurrency: 3
+  request_options:
+    delay: 3000
+```
+
+> [!TIP]
+> For a full list of supported configuration options, have a look at the
+> [`EliasHaeussler\CacheWarmup\Config\CacheWarmupConfig`](src/Config/CacheWarmupConfig.php)
+> class.
+
 ## 🧑‍💻 Contributing
 
 Please have a look at [`CONTRIBUTING.md`](CONTRIBUTING.md).
@@ -565,3 +626,5 @@ This project is licensed under [GNU General Public License 3.0 (or later)](LICEN
 [22]: res/cache-warmup-result.schema.json
 [23]: src/Crawler/StoppableCrawlerInterface.php
 [24]: src/Crawler/LoggingCrawlerInterface.php
+[25]: src/Config/Adapter/FileConfigAdapter.php
+[26]: src/Config/Adapter/PhpConfigAdapter.php
diff --git a/composer.json b/composer.json
@@ -26,7 +26,8 @@
 		"psr/http-message": "^1.0 || ^2.0",
 		"psr/log": "^2.0 || ^3.0",
 		"symfony/console": "^5.4 || ^6.0 || ^7.0",
-		"symfony/filesystem": "^5.4 || ^6.0 || ^7.0"
+		"symfony/filesystem": "^5.4 || ^6.0 || ^7.0",
+		"symfony/yaml": "^5.4 || ^6.0 || ^7.0"
 	},
 	"require-dev": {
 		"armin/editorconfig-cli": "^1.8 || ^2.0",

diff --git a/composer.lock b/composer.lock
diff --git a/src/CacheWarmer.php b/src/CacheWarmer.php
@@ -26,8 +26,6 @@
 use GuzzleHttp\Client;
 use GuzzleHttp\ClientInterface;
 use GuzzleHttp\Exception\GuzzleException;
-use GuzzleHttp\Psr7;
-use Psr\Http\Message;
 
 use function array_key_exists;
 use function array_values;
@@ -36,7 +34,6 @@
 use function is_array;
 use function is_string;
 use function preg_match;
-use function str_contains;
 use function str_starts_with;
 
 /**
@@ -124,8 +121,7 @@ public function addSitemaps(array|string|Sitemap\Sitemap $sitemaps): self
         foreach ($sitemaps as $sitemap) {
             // Parse sitemap URL to valid sitemap object
             if (is_string($sitemap)) {
-                $sitemapUri = $this->resolveSitemapUri($sitemap);
-                $sitemap = new Sitemap\Sitemap($sitemapUri);
+                $sitemap = Sitemap\Sitemap::createFromString($sitemap);
             }
 
             // Throw exception if sitemap is invalid
@@ -196,19 +192,6 @@ public function addUrl(string|Sitemap\Url $url): self
         return $this;
     }
 
-    private function resolveSitemapUri(string $sitemap): Message\UriInterface
-    {
-        // Sitemap is a remote URL
-        if (str_contains($sitemap, '://')) {
-            return new Psr7\Uri($sitemap);
-        }
-
-        // Sitemap is a local file
-        $file = Helper\FilesystemHelper::resolveRelativePath($sitemap);
-
-        return new Psr7\Uri('file://'.$file);
-    }
-
     private function exceededLimit(): bool
     {
         return $this->limit > 0 && count($this->urls) >= $this->limit;