From 27d6359e5a56972fa315128203c7ad32a33ca40a Mon Sep 17 00:00:00 2001
From: Bruno Rocha <rochacbruno@gmail.com>
Date: Sun, 3 Nov 2024 16:26:26 +0000
Subject: [PATCH] feat: extract card image from html if not defined

---
 ...10-19-what-is-marmite-static-site-generator.md |  6 ++++++
 example/content/customizing-templates.md          |  1 +
 example/templates/content.html                    |  4 ++--
 src/content.rs                                    |  1 +
 src/markdown.rs                                   | 15 +++++++++++++++
 src/site.rs                                       |  1 +
 src/tera_functions.rs                             |  1 +
 7 files changed, 27 insertions(+), 2 deletions(-)

diff --git a/example/content/2024-10-19-what-is-marmite-static-site-generator.md b/example/content/2024-10-19-what-is-marmite-static-site-generator.md
index e56aa55..3e5c6d6 100644
--- a/example/content/2024-10-19-what-is-marmite-static-site-generator.md
+++ b/example/content/2024-10-19-what-is-marmite-static-site-generator.md
@@ -157,6 +157,12 @@ Content
       - tag2
     ```
   **default** empty
+  
+**card_image**
+
+  : Image url to use as social card image `og:image`  
+  **format**  `./media/file.png` or `https://path/to/img.jpg`  
+  **default** first image extracted from html content, config card_image, or None.
 
 **extra**
 
diff --git a/example/content/customizing-templates.md b/example/content/customizing-templates.md
index a32a339..2bba553 100644
--- a/example/content/customizing-templates.md
+++ b/example/content/customizing-templates.md
@@ -68,6 +68,7 @@ date: DateTimeObject or None
 extra: {key: value}
 links_to: [str] or None
 back_links: [Content] or []
+card_image: str or None
 ```
 
 There are 6 templates inside the `templates` folder, each adds more data to context.
diff --git a/example/templates/content.html b/example/templates/content.html
index e3c273c..74eb0a8 100644
--- a/example/templates/content.html
+++ b/example/templates/content.html
@@ -7,8 +7,8 @@
 {% else %}
 <meta property="og:description" content="{{ content.html | striptags | trim_start_matches(pat=content.title) | truncate(length=100, end="...") }}">
 {%- endif %}
-{% if content.extra.card_image %}
-<meta property="og:image" content="{{url_for(path=content.extra.card_image, abs=true)}}">
+{% if content.card_image %}
+<meta property="og:image" content="{{url_for(path=content.card_image, abs=true)}}">
 {% elif site.card_image %}
 <meta property="og:image" content="{{url_for(path=site.card_image, abs=true)}}">
 {% endif %}
diff --git a/src/content.rs b/src/content.rs
index c565e8c..356018c 100644
--- a/src/content.rs
+++ b/src/content.rs
@@ -20,6 +20,7 @@ pub struct Content {
     pub extra: Option<Value>,
     pub links_to: Option<Vec<String>>,
     pub back_links: Vec<Self>,
+    pub card_image: Option<String>,
 }
 
 pub fn get_title<'a>(frontmatter: &'a Frontmatter, html: &'a str) -> String {
diff --git a/src/markdown.rs b/src/markdown.rs
index 4e056d3..658fb6b 100644
--- a/src/markdown.rs
+++ b/src/markdown.rs
@@ -29,6 +29,7 @@ pub fn get_content(path: &Path) -> Result<Content, String> {
     let extra = frontmatter.get("extra").map(std::borrow::ToOwned::to_owned);
     let links_to = get_links_to(&html);
     let back_links = Vec::new(); // will be mutated later
+    let card_image = get_card_image(&frontmatter, &html);
     let content = Content {
         title,
         description,
@@ -39,10 +40,24 @@ pub fn get_content(path: &Path) -> Result<Content, String> {
         extra,
         links_to,
         back_links,
+        card_image,
     };
     Ok(content)
 }
 
+/// Capture `card_image` from frontmatter, then if not defined
+/// take the first img src found in the post content
+fn get_card_image(frontmatter: &Frontmatter, html: &str) -> Option<String> {
+    if let Some(card_image) = frontmatter.get("card_image") {
+        return Some(card_image.to_string());
+    }
+    // first <img> src attribute
+    let img_regex = Regex::new(r#"<img[^>]*src="([^"]+)""#).unwrap();
+    img_regex
+        .captures(html)
+        .and_then(|caps| caps.get(1).map(|m| m.as_str().to_string()))
+}
+
 fn get_links_to(html: &str) -> Option<Vec<String>> {
     let mut result = Vec::new();
     let re = Regex::new(r#"href="\./(.*?)\.html""#).unwrap();
diff --git a/src/site.rs b/src/site.rs
index 80ae8bd..d62a7b5 100644
--- a/src/site.rs
+++ b/src/site.rs
@@ -564,6 +564,7 @@ fn handle_404(
         tags: vec![],
         links_to: None,
         back_links: vec![],
+        card_image: None,
     };
     if input_404_path.exists() {
         let custom_content = get_content(&input_404_path)?;
diff --git a/src/tera_functions.rs b/src/tera_functions.rs
index cf91165..03dd30f 100644
--- a/src/tera_functions.rs
+++ b/src/tera_functions.rs
@@ -14,6 +14,7 @@ impl Function for UrlFor {
             .get("path")
             .and_then(Value::as_str)
             .ok_or_else(|| tera::Error::msg("Missing `path` argument"))?
+            .trim_start_matches("./")
             .to_string();
 
         let abs_prefixes = ["http", "https", "mailto"];