From fac7a4d5ac4c5acafda9f97851807e35d57c4154 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Dawid=20Ci=C4=99=C5=BCarkiewicz?= <dpc@dpc.pw>
Date: Sun, 10 May 2020 23:59:43 -0700
Subject: [PATCH] Compact results

---
 src/index.rs | 56 +++++++++++++++++++++++++++++++++++++++++++++++++---
 src/main.rs  | 16 ++++++++++++---
 src/page.rs  |  5 +++--
 3 files changed, 69 insertions(+), 8 deletions(-)

diff --git a/src/index.rs b/src/index.rs
index 9f0d57d..5154bcb 100644
--- a/src/index.rs
+++ b/src/index.rs
@@ -6,24 +6,30 @@ use async_trait::async_trait;
 use log::info;
 use std::collections::{HashMap, HashSet};
 
+/// Indexing wrapper over `page::Store`
+///
+/// `Index` keeps track of page data neccessary
+/// to quickly look them up by a tag query.
 #[derive(Default)]
 pub struct Index<T> {
     page_ids_by_tag: HashMap<String, HashSet<Id>>,
-    tags_by_page_id: HashMap<Id, Vec<Tag>>,
+    tags_by_page_id: HashMap<Id, HashSet<Tag>>,
     title_by_page_id: HashMap<Id, String>,
     store: T,
 }
 
+/// Basic page info
 #[derive(Debug, Clone)]
 pub struct PageInfo {
     pub id: Id,
     pub title: String,
 }
 
+/// Results of tag query lookup
 #[derive(Default, Debug, Clone)]
 pub struct FindResults {
     pub matching_pages: Vec<PageInfo>,
-    pub matching_tags: Vec<page::Tag>,
+    pub matching_tags: Vec<Tag>,
 }
 
 impl FindResults {
@@ -32,6 +38,14 @@ impl FindResults {
     }
 }
 
+/// More compact (post-processed) `FindResults`
+pub struct CompactResults {
+    // all tags that were not already filtered on
+    pub tags: Vec<(Tag, usize)>,
+    // all pages that can't be reached by one of the `tags`
+    pub pages: Vec<PageInfo>,
+}
+
 impl<T> Index<T>
 where
     T: page::StoreMut,
@@ -49,6 +63,7 @@ where
         Ok(s)
     }
 
+    /// Index the inner `Store`
     async fn index_inner(&mut self) -> Result<()> {
         let mut count = 0;
         let ids = self.store.iter().await?.collect::<Vec<page::Id>>();
@@ -60,9 +75,44 @@ where
         info!("Indexed {} pages", count);
         Ok(())
     }
+
+    /// Compact the results to a shorter form
+    pub fn compact_results(&self, results: FindResults) -> CompactResults {
+        let matching_tags: HashSet<String> = results.matching_tags.iter().cloned().collect();
+        let mut unmatched_tags: HashMap<Tag, usize> = Default::default();
+        for page_info in &results.matching_pages {
+            for page_tag in &self.tags_by_page_id[&page_info.id] {
+                if !matching_tags.contains(page_tag.as_str()) {
+                    *unmatched_tags.entry(page_tag.to_owned()).or_default() += 1;
+                }
+            }
+        }
+
+        let unmatched_tags_set: HashSet<Tag> = unmatched_tags.keys().cloned().collect();
+
+        let mut pages: Vec<PageInfo> = results
+            .matching_pages
+            .into_iter()
+            .filter(|page_info| {
+                unmatched_tags_set
+                    .intersection(&self.tags_by_page_id[&page_info.id])
+                    .next()
+                    .is_none()
+            })
+            .collect();
+
+        pages.sort_by(|a, b| a.title.cmp(&b.title));
+
+        let mut tags: Vec<_> = unmatched_tags.into_iter().collect();
+
+        tags.sort_by(|a, b| a.1.cmp(&b.1).reverse().then_with(|| a.0.cmp(&b.0)));
+
+        CompactResults { tags, pages }
+    }
 }
 
 impl<T> Index<T> {
+    /// Lookup pages with a list of tags
     pub fn find(&self, tags: &[TagRef]) -> FindResults {
         let mut matching_pages: Vec<PageInfo> = vec![];
         let mut matching_tags: Vec<String> = vec![];
@@ -146,7 +196,7 @@ impl<T> Index<T> {
             .tags_by_page_id
             .get(&id)
             .cloned()
-            .unwrap_or_else(|| vec![])
+            .unwrap_or_else(|| HashSet::new())
         {
             self.page_ids_by_tag
                 .get_mut(&tag)
diff --git a/src/main.rs b/src/main.rs
index 209562c..a872348 100644
--- a/src/main.rs
+++ b/src/main.rs
@@ -8,7 +8,7 @@ use warp::{path::FullPath, Filter};
 
 use serde_derive::Deserialize;
 
-use page::StoreMut;
+use page::{StoreMut, Tag};
 
 /// Command line options
 mod cli;
@@ -157,7 +157,10 @@ fn render_page_view(page: &page::Parsed) -> impl RenderOnce {
     }
 }
 
-fn render_post_list(posts: impl Iterator<Item = index::PageInfo> + 'static) -> impl RenderOnce {
+fn render_post_list(
+    unmatched_tags: impl Iterator<Item = (Tag, usize)>,
+    posts: impl Iterator<Item = index::PageInfo> + 'static,
+) -> impl RenderOnce {
     owned_html! {
         div(class="pure-menu pure-menu-horizontal") {
             form(action="..", method="get", class="pure-menu-item") {
@@ -173,6 +176,11 @@ fn render_post_list(posts: impl Iterator<Item = index::PageInfo> + 'static) -> i
             }
         }
         ul {
+            @ for tag in unmatched_tags {
+                li {
+                    a(href=format!("./{}", tag.0)) : format!("{} ({})", tag.0, tag.1)
+                }
+            }
             @ for post in posts {
                 li {
                     a(href=format!("?id={}", post.id)) : post.title
@@ -358,8 +366,10 @@ async fn handle_get(
             query.edit.is_some(),
         ))))
     } else {
+        let compact_results = read.compact_results(results);
         Ok(warp_reply_from_render(render_html_page(render_post_list(
-            results.matching_pages.into_iter(),
+            compact_results.tags.into_iter(),
+            compact_results.pages.into_iter(),
         ))))
     }
 }
diff --git a/src/page.rs b/src/page.rs
index 61c9b82..36a15d5 100644
--- a/src/page.rs
+++ b/src/page.rs
@@ -3,6 +3,7 @@ pub mod store;
 #[allow(unused)]
 use anyhow::Result;
 use lazy_static::lazy_static;
+use std::collections::HashSet;
 pub use store::{InMemoryStore, Store, StoreMut};
 
 use digest::Digest;
@@ -23,7 +24,7 @@ pub struct Parsed {
     pub source_body: String,
     pub html: String,
     pub headers: Headers,
-    pub tags: Vec<Tag>,
+    pub tags: HashSet<Tag>,
     pub title: String,
 }
 
@@ -165,7 +166,7 @@ impl Parsed {
             html: html_output,
             source_body: body,
             source: Source(source),
-            tags,
+            tags: tags.into_iter().collect(),
             title,
         }
     }