feat(seach-index): build search index

Author: fiji-flo

Diff for: crates/rari-doc/src/lib.rs

@@ -9,6 +9,7 @@ pub mod percent;
 pub mod reader;
 pub mod redirects;
 pub mod resolve;
+pub mod search_index;
 pub mod sidebars;
 pub mod specs;
 pub mod templ;

Diff for: crates/rari-doc/src/search_index.rs

@@ -0,0 +1,63 @@
+use std::cmp::Ordering;
+use std::collections::HashMap;
+use std::fs::File;
+use std::io::BufWriter;
+
+use rari_types::globals::{build_out_root, content_root};
+use rari_types::locale::Locale;
+use rari_types::Popularities;
+use rari_utils::io::read_to_string;
+use serde::Serialize;
+
+use crate::error::DocError;
+use crate::pages::page::{Page, PageLike};
+
+#[derive(Debug, Serialize)]
+pub struct SearchItem<'a> {
+    title: &'a str,
+    url: &'a str,
+}
+
+pub fn build_search_index(docs: &[Page]) -> Result<(), DocError> {
+    let in_file = content_root().join("en-US").join("popularities.json");
+    let json_str = read_to_string(in_file)?;
+    let popularities: Popularities = serde_json::from_str(&json_str)?;
+
+    let mut all_indices: HashMap<Locale, Vec<(&Page, f64)>> = HashMap::new();
+
+    for doc in docs {
+        let entry = all_indices.entry(doc.locale()).or_default();
+        entry.push((
+            doc,
+            popularities
+                .popularities
+                .get(doc.url())
+                .cloned()
+                .unwrap_or_default(),
+        ));
+    }
+
+    for (locale, mut index) in all_indices.into_iter() {
+        if !index.is_empty() {
+            index.sort_by(|(da, a), (db, b)| match b.partial_cmp(a) {
+                None | Some(Ordering::Equal) => da.title().cmp(db.title()),
+                Some(ord) => ord,
+            });
+            let out = index
+                .into_iter()
+                .map(|(doc, _)| SearchItem {
+                    title: doc.title(),
+                    url: doc.url(),
+                })
+                .collect::<Vec<_>>();
+            let out_file = build_out_root()?
+                .join(locale.as_folder_str())
+                .join("search-index.json");
+            let file = File::create(out_file)?;
+            let buffed = BufWriter::new(file);
+
+            serde_json::to_writer(buffed, &out)?;
+        }
+    }
+    Ok(())
+}

Diff for: src/main.rs

@@ -14,6 +14,7 @@ use rari_doc::build::{
 use rari_doc::cached_readers::{read_and_cache_doc_pages, CACHED_DOC_PAGE_FILES};
 use rari_doc::pages::types::doc::Doc;
 use rari_doc::reader::read_docs_parallel;
+use rari_doc::search_index::build_search_index;
 use rari_doc::utils::TEMPL_RECORDER_SENDER;
 use rari_tools::history::gather_history;
 use rari_tools::popularities::update_popularities;
@@ -70,6 +71,8 @@ struct BuildArgs {
     #[arg(long)]
     skip_contributors: bool,
     #[arg(long)]
+    skip_search_index: bool,
+    #[arg(long)]
     skip_blog: bool,
     #[arg(long)]
     skip_curriculum: bool,
@@ -189,6 +192,11 @@ fn main() -> Result<(), anyhow::Error> {
                 urls.extend(build_docs(&docs)?);
                 println!("Took: {: >10.3?} to build content", start.elapsed());
             }
+            if !args.skip_search_index && args.files.is_empty() {
+                let start = std::time::Instant::now();
+                build_search_index(&docs)?;
+                println!("Took: {: >10.3?} to build search index", start.elapsed());
+            }
             if !args.skip_curriculum && args.files.is_empty() {
                 let start = std::time::Instant::now();
                 urls.extend(build_curriculum_pages()?);