update

Cargo.toml

@@ -16,7 +16,7 @@ tokio = { version = "1.35.1", features = ["full"] }
 uuid = { version = "1.8.0", features = ["v4", "serde"] }
 sqlx = { version = "0.8", features = ["postgres", "runtime-tokio-native-tls", "derive", "uuid", "chrono", "json"] }
 maud = "0.26.0"
-based = { git = "https://git.hydrar.de/jmarya/based", features = ["htmx"] }
+based = { git = "https://git.hydrar.de/jmarya/based", branch = "ui" }
 url = "2.5.4"
 reqwest = "0.12.11"
 ollama-rs = "0.2.2"

config.toml

@@ -17,6 +17,12 @@ BLACKLIST_DOMAINS = [
 # The domain the config applies to
 domain = "example.com"
 
+# Consider a refresh after interval (in days)
+outdated = 30
+
+# Keep last n versions
+keep_n = 5
+
 # Blacklisted Path (Regexes)
 blacklist_paths = ["/.*"]
 
@@ -70,4 +76,8 @@ no_video = true
 [[websites.domains]]
 domain = "docs.flutter.dev"
 no_javascript = true
-no_video = true
+no_video = true
+
+[[websites.domains]]
+domain = "home-assistant.io"
+no_javascript = true

src/ai.rs

@@ -1,6 +1,6 @@
 use std::collections::{HashMap, VecDeque};
 
-use based::{get_pg, request::api::ToAPI, result::LogNoneAndPass};
+use based::{get_pg, request::api::ToAPI, result::LogNoneAndPass, ui::components::Shell};
 use ollama_rs::generation::embeddings::request::{EmbeddingsInput, GenerateEmbeddingsRequest};
 use serde::Serialize;
 use serde_json::json;
@@ -110,6 +110,7 @@ pub trait Embedding {
     fn embedding(
         &self,
         ver: Option<String>,
+        shell: &Shell,
     ) -> impl std::future::Future<Output = Option<Vec<Vec<f32>>>>;
 }
 
@@ -130,7 +131,7 @@ pub fn remove_data_urls(input: &str) -> String {
 }
 
 impl Embedding for Document {
-    async fn embedding(&self, ver: Option<String>) -> Option<Vec<Vec<f32>>> {
+    async fn embedding(&self, ver: Option<String>, shell: &Shell) -> Option<Vec<Vec<f32>>> {
         let latest = "latest".to_string();
         log::info!(
             "Generating Vector embeddings for {} / {} @ {}",
@@ -139,7 +140,7 @@ impl Embedding for Document {
             ver.as_ref().unwrap_or(&latest)
         );
 
-        let content_html = self.render_local(ver.clone()).await?;
+        let content_html = self.render_local(ver.clone(), shell).await?;
         let content = remove_data_urls(&html2md::parse_html(&content_html));
 
         let mut embeddings = Vec::new();
@@ -219,9 +220,9 @@ impl EmbedStore {
             .unwrap()
     }
 
-    pub async fn embed_document(doc: &Document, ver: &str) {
+    pub async fn embed_document(doc: &Document, ver: &str, shell: &Shell) {
         if let Some(embed) = doc
-            .embedding(Some(ver.to_string()))
+            .embedding(Some(ver.to_string()), shell)
             .await
             .log_warn_none_and_pass(|| {
                 format!(
@@ -253,10 +254,10 @@ impl EmbedStore {
         }
     }
 
-    pub async fn ensure_embedding(doc: &Document) {
+    pub async fn ensure_embedding(doc: &Document, shell: &Shell) {
         for ver in doc.versions() {
             if Self::get_embedding(doc, Some(ver.as_str())).await.is_none() {
-                Self::embed_document(doc, &ver).await;
+                Self::embed_document(doc, &ver, shell).await;
             }
         }
     }
@@ -299,25 +300,25 @@ impl EmbedStore {
         flat
     }
 
-    pub async fn generate_embeddings_for(arc: &WebsiteArchive) {
+    pub async fn generate_embeddings_for(arc: &WebsiteArchive, shell: &Shell) {
         log::info!("Generating embeddings");
 
         for dom in arc.domains() {
             let dom = arc.get_domain(&dom);
-            embed_path(&dom, "/").await;
+            embed_path(&dom, "/", shell).await;
         }
 
         log::info!("Done generating embeddings");
     }
 }
 
-pub async fn embed_path(dom: &Domain, path: &str) {
+pub async fn embed_path(dom: &Domain, path: &str, shell: &Shell) {
     let (paths, is_doc) = dom.paths(path);
 
     // If the path is a document, process the root path.
     if is_doc {
         let doc = dom.path("/");
-        EmbedStore::ensure_embedding(&doc).await;
+        EmbedStore::ensure_embedding(&doc, shell).await;
     }
 
     // Create a queue to process paths iteratively
@@ -331,7 +332,7 @@ pub async fn embed_path(dom: &Domain, path: &str) {
 
         if is_doc {
             let doc = dom.path(next_path.path());
-            EmbedStore::ensure_embedding(&doc).await;
+            EmbedStore::ensure_embedding(&doc, shell).await;
         }
 
         queue.extend(next_paths);

src/archive/document.rs

@@ -1,6 +1,6 @@
 use std::{io::Read, path::PathBuf};
 
-use based::request::RequestContext;
+use based::{request::RequestContext, ui::components::Shell};
 use maud::html;
 
 use crate::{blacklist::check_blacklist, conf::get_config, render_page};
@@ -52,12 +52,17 @@ impl Document {
     ///
     /// # Returns
     /// An `Option` containing the rendered content as a string, or `None` if nothing could be rendered.
-    pub async fn render_local(&self, version: Option<String>) -> Option<String> {
+    pub async fn render_local(&self, version: Option<String>, shell: &Shell) -> Option<String> {
         if check_blacklist(&self.domain) {
             let content = html! {
                 h3 { "This site is blacklisted" };
             };
-            return Some(render_page(content, RequestContext::default()).await.1 .1);
+            return Some(
+                render_page(content, RequestContext::default(), shell)
+                    .await
+                    .1
+                     .1,
+            );
         }
 
         let mut file_path = self.doc_dir();
@@ -100,6 +105,20 @@ impl Document {
         file_path
     }
 
+    fn latest_time_since(&self) -> Option<usize> {
+        if let Some(t_str) = self.versions().first() {
+            let given_date = chrono::NaiveDate::parse_from_str(t_str, "%Y-%m-%d")
+                .expect("Invalid date format. Expected yyyy-mm-dd.");
+
+            let today = chrono::Local::now().date_naive();
+
+            let duration = today.signed_duration_since(given_date);
+            return Some(duration.num_days() as usize);
+        }
+
+        None
+    }
+
     /// Retrieves available versions of the document.
     ///
     /// # Returns

src/archive/mod.rs

@@ -28,6 +28,7 @@ pub fn read_dir(dir: &PathBuf) -> Vec<String> {
 
 /// Rewrite all URLs in `input` to the format `/s/<domain>/<path..>`
 fn internalize_urls(input: &str, base: &str) -> String {
+    // todo : fix regex, domains without path are not captured
     let url_pattern = r#"(\ |"|')(?:(<?)(https?:\/\/([a-zA-Z0-9.-]+))?(\/[\w./-]*))"#;
     let re = regex::Regex::new(url_pattern).unwrap();
 

src/conf.rs

@@ -67,6 +67,8 @@ pub struct WebsiteConfig {
 #[derive(Debug, Clone, Deserialize)]
 pub struct DomainConfig {
     pub domain: String,
+    pub outdated: Option<usize>,
+    pub keep_n: Option<usize>,
     pub blacklist_paths: Option<Vec<String>>,
     pub no_audio: Option<bool>,
     pub no_video: Option<bool>,
@@ -83,6 +85,8 @@ impl Default for DomainConfig {
         Self {
             domain: String::new(),
             blacklist_paths: None,
+            outdated: None,
+            keep_n: None,
             no_audio: Some(false),
             no_video: Some(false),
             no_image: Some(false),

src/lib.rs

@@ -1,7 +1,5 @@
-use based::{
-    page::Shell,
-    request::{RequestContext, StringResponse},
-};
+use based::request::{RequestContext, StringResponse};
+use based::ui::components::Shell;
 use maud::{html, PreEscaped};
 
 pub mod ai;
@@ -10,20 +8,10 @@ pub mod blacklist;
 pub mod conf;
 pub mod favicon;
 
-pub async fn render_page(content: PreEscaped<String>, ctx: RequestContext) -> StringResponse {
-    based::page::render_page(
-        content,
-        "Website Archive",
-        ctx,
-        &Shell::new(
-            html! {
-                script src="https://cdn.tailwindcss.com" {};
-                meta name="viewport" content="width=device-width, initial-scale=1.0" {};
-                script src="/assets/htmx.min.js" {};
-            },
-            html! {},
-            Some("bg-zinc-950 text-white min-h-screen flex pt-8 justify-center".to_string()),
-        ),
-    )
-    .await
+pub async fn render_page(
+    content: PreEscaped<String>,
+    ctx: RequestContext,
+    shell: &Shell,
+) -> StringResponse {
+    shell.render_page(content, "Website Archive", ctx).await
 }

src/main.rs

@@ -1,5 +1,9 @@
+use based::asset::AssetRoutes;
 use based::get_pg;
+use based::ui::components::Shell;
+use based::ui::prelude::*;
 use rocket::routes;
+use rocket::time::format_description::modifier::Padding;
 use webarc::ai::EmbedStore;
 use webarc::archive::WebsiteArchive;
 use webarc::conf::{get_config, load_config, load_default_config};
@@ -15,6 +19,8 @@ async fn main() {
 
     let archive_dir: &String = args.get_one("dir").unwrap();
 
+    let shell = get_shell();
+
     match args.subcommand() {
         Some(("serve", serve_args)) => {
             let config: &String = serve_args.get_one("config").unwrap();
@@ -30,7 +36,7 @@ async fn main() {
             let archive = arc.clone();
             if get_config().ai.is_some() {
                 tokio::spawn(async move {
-                    EmbedStore::generate_embeddings_for(&archive).await;
+                    EmbedStore::generate_embeddings_for(&archive, &get_shell()).await;
                 });
             }
 
@@ -40,10 +46,10 @@ async fn main() {
             });
 
             rocket::build()
+                .mount_assets()
                 .mount(
                     "/",
                     routes![
-                        based::htmx::htmx_script_route,
                         pages::index,
                         pages::render_website,
                         pages::domain_info_route,
@@ -53,6 +59,7 @@ async fn main() {
                     ],
                 )
                 .manage(arc)
+                .manage(shell)
                 .launch()
                 .await
                 .unwrap();
@@ -146,7 +153,7 @@ async fn main() {
 
                     let md = get_args.get_flag("md");
 
-                    let content = doc.render_local(Some(ver)).await;
+                    let content = doc.render_local(Some(ver), &shell).await;
 
                     if content.is_none() {
                         println!("No document found");
@@ -168,3 +175,20 @@ async fn main() {
         None => {}
     }
 }
+
+pub fn get_shell() -> Shell {
+    Shell::new(
+        Nothing(),
+        Nothing(),
+        Background(MinHeight(
+            ScreenValue::screen,
+            Padding(Text("").white()).top(ScreenValue::_8),
+        ))
+        .color(Zinc::_950),
+    )
+    .use_ui()
+}
+
+// TODO : redownload after threshold
+// TODO : keep n versions
+// TODO : archive cleanup code

src/pages/component.rs

@@ -1,4 +1,5 @@
-use maud::{html, PreEscaped};
+use based::ui::prelude::*;
+use maud::{html, PreEscaped, Render};
 
 /// Generates an SVG arrow icon with the specified color.
 ///
@@ -20,9 +21,9 @@ pub fn arrow_icon(color: &str) -> PreEscaped<String> {
 /// # Returns
 /// A `PreEscaped<String>` containing the HTML markup for a slash separator.
 pub fn slash_seperator() -> PreEscaped<String> {
-    html! {
-        p class="font-bold p-2 text-gray-400" { " / " };
-    }
+    Padding(Text("/").bold().color(&Gray::_400))
+        .all(ScreenValue::_2)
+        .render()
 }
 
 /// Generates a hyperlink for a specific path within a domain.
@@ -42,9 +43,12 @@ pub fn gen_path_link(
     domain: &str,
 ) -> PreEscaped<String> {
     let upto: Vec<&str> = path_seperations.iter().take(index + 1).cloned().collect();
-    html! {
-        a href=(format!("/d/{}/{}", domain, upto.join("/"))) { (path)}
-    }
+    Link(
+        &format!("/d/{}/{}", domain, upto.join("/")),
+        path.to_string(),
+    )
+    .use_htmx()
+    .render()
 }
 
 /// Generates a breadcrumb-like header for a path within a domain.

src/pages/mod.rs

@@ -1,10 +1,11 @@
-use std::{io::Read, path::PathBuf};
+use std::{io::Read, path::PathBuf, sync::Arc};
 
+use based::ui::prelude::*;
 use based::{
-    page::search::Search,
     request::{
         api::GeneratedPager, assets::DataResponse, respond_json, RequestContext, StringResponse,
     },
+    ui::components::{Search, Shell},
 };
 use maud::{html, PreEscaped};
 use rocket::{get, request::FromSegments, State};
@@ -20,6 +21,9 @@ use webarc::{
     render_page,
 };
 
+// TODO : Implement archive timeline page (chrono sorted documents)
+// TODO : impl archive index to db
+
 const SEARCH_BAR_STYLE: &str = "w-full px-4 mb-4 py-2 text-white bg-black border-2 border-neon-blue placeholder-neon-blue focus:ring-2 focus:ring-neon-pink focus:outline-none font-mono text-lg";
 
 /// Get the favicon of a domain
@@ -42,7 +46,11 @@ pub async fn favicon_route(domain: &str) -> Option<DataResponse> {
 
 /// Websites Overview
 #[get("/")]
-pub async fn index(ctx: RequestContext, arc: &State<WebsiteArchive>) -> StringResponse {
+pub async fn index(
+    ctx: RequestContext,
+    arc: &State<WebsiteArchive>,
+    shell: &State<Shell>,
+) -> StringResponse {
     let websites = arc.domains();
 
     let content = html! {
@@ -71,7 +79,7 @@ pub async fn index(ctx: RequestContext, arc: &State<WebsiteArchive>) -> StringRe
         }
     };
 
-    render_page(content, ctx).await
+    render_page(content, ctx, &shell).await
 }
 
 /// Overview on <domain> / <path>
@@ -81,6 +89,7 @@ pub async fn domain_info_route(
     domain: &str,
     paths: PathBuf,
     arc: &State<WebsiteArchive>,
+    shell: &State<Shell>,
 ) -> StringResponse {
     let domain = arc.get_domain(domain);
     let document = domain.path(paths.to_str().unwrap());
@@ -88,10 +97,15 @@ pub async fn domain_info_route(
     let (path_entries, is_doc) = domain.paths(paths.to_str().unwrap());
     let path_seperations: Vec<&str> = paths.to_str().unwrap().split('/').collect();
 
-    let domains = extract_domains(&document.render_local(None).await.unwrap_or_default());
+    let domains = extract_domains(
+        &document
+            .render_local(None, &shell)
+            .await
+            .unwrap_or_default(),
+    );
 
     let content = html! {
-        h2 class="text-xl font-bold mb-4 flex items-center" {
+        h2 class="text-xl font-bold mb-4 flex items-center w-fit mx-auto" {
             img class="p-2" src=(format!("/favicon/{}", &domain.name)) {};
             a href=(format!("/d/{}", &domain.name)) { (domain.name) };
             (slash_seperator())
@@ -150,7 +164,7 @@ pub async fn domain_info_route(
     };
     };
 
-    render_page(content, ctx).await
+    render_page(content, ctx, &shell).await
 }
 
 #[get("/txt/<domain>/<path..>?<time>&<no_data_urls>")]
@@ -160,11 +174,12 @@ pub async fn render_txt_website(
     time: Option<&str>,
     no_data_urls: Option<&str>,
     arc: &State<WebsiteArchive>,
+    shell: &State<Shell>,
 ) -> Option<String> {
     let document = arc.get_domain(domain).path(path.to_str().unwrap());
 
     let mut content = document
-        .render_local(time.map(|time| time.to_string()))
+        .render_local(time.map(|time| time.to_string()), &shell)
         .await?;
 
     if no_data_urls.is_some() {
@@ -215,11 +230,12 @@ pub async fn render_website(
     path: PathSegment,
     time: Option<&str>,
     arc: &State<WebsiteArchive>,
+    shell: &State<Shell>,
 ) -> Option<DataResponse> {
     let document = arc.get_domain(domain).path(&path.to_str());
 
     let content = document
-        .render_local(time.map(|time| time.to_string()))
+        .render_local(time.map(|time| time.to_string()), &shell)
         .await;
 
     if let Some(content) = content {
@@ -232,7 +248,7 @@ pub async fn render_website(
         arc.archive_url(&format!("https://{domain}/{}", path.to_str()))
             .await;
 
-        let content = document.render_local(None).await?;
+        let content = document.render_local(None, &shell).await?;
 
         return Some(DataResponse::new(
             content.as_bytes().to_vec(),
@@ -264,6 +280,7 @@ pub async fn vector_search(
     query: Option<&str>,
     page: Option<i64>,
     ctx: RequestContext,
+    shell: &State<Shell>,
 ) -> Option<StringResponse> {
     get_config().ai.as_ref()?;
 
@@ -278,10 +295,12 @@ pub async fn vector_search(
         let real_query = query.trim_end_matches(".json");
 
         // Search Results
-        let vector = pgvector::Vector::from(generate_embedding(real_query.to_string()).await?);
+        let vector = Arc::new(pgvector::Vector::from(
+            generate_embedding(real_query.to_string()).await?,
+        ));
 
         let results = GeneratedPager::new(
-            |input, offset, limit| {
+            |input: Arc<pgvector::Vector>, offset, limit| {
                 Box::pin(async move {
                     EmbedStore::search_vector(&input, limit as i64, offset as i64).await
                 })
@@ -299,10 +318,10 @@ pub async fn vector_search(
 
         let content = search.build_response(&ctx, results, page, real_query, gen_search_element);
 
-        return Some(render_page(content, ctx).await);
+        return Some(render_page(content, ctx, &shell).await);
     }
 
     // Return new search site
     let content = search.build("", html! {});
-    Some(render_page(content, ctx).await)
+    Some(render_page(content, ctx, &shell).await)
 }