Show both database and documents database sizes

public/script.js

@@ -63,5 +63,8 @@ $('#docs-count').text(function(index, text) {
 
 // Make the database a little bit easier to read
 $('#db-size').text(function(index, text) {
-  return filesize(parseInt(text))
+  let arr = text.split("+");
+  let database_size = filesize(parseInt(arr[0]));
+  let documents_size = filesize(parseInt(arr[1]));
+  return `${database_size} + ${documents_size}`
 });

src/bin/serve.rs

@@ -62,6 +62,7 @@ fn highlight_string(string: &str, words: &HashSet<String>) -> String {
 struct IndexTemplate {
     db_name: String,
     db_size: usize,
+    docs_size: usize,
     docs_count: usize,
 }
 
@@ -88,6 +89,7 @@ async fn main() -> anyhow::Result<()> {
     // the disk file size and the number of documents in the database.
     let db_name = opt.database.file_stem().and_then(|s| s.to_str()).unwrap_or("").to_string();
     let db_size = File::open(opt.database.join("data.mdb"))?.metadata()?.len() as usize;
+    let docs_size = File::open(opt.database.join("documents.mtbl"))?.metadata()?.len() as usize;
     let docs_count = index.number_of_documents();
 
     // We run and wait on the HTTP server
@@ -99,6 +101,7 @@ async fn main() -> anyhow::Result<()> {
             IndexTemplate {
                 db_name: db_name.clone(),
                 db_size,
+                docs_size,
                 docs_count: docs_count as usize,
             }
         });

src/lib.rs

@@ -185,9 +185,9 @@ impl Index {
             let mut stream = fst.search(&dfa).into_stream();
             while let Some(word) = stream.next() {
                 let word = std::str::from_utf8(word)?;
-                if let Some(right) = self.word_positions.get(rtxn, word)? {
+                if let Some(positions) = self.word_positions.get(rtxn, word)? {
-                    union_positions.union_with(&right);
+                    union_positions.union_with(&positions);
-                    derived_words.push((word.as_bytes().to_vec(), right));
+                    derived_words.push((word.as_bytes().to_vec(), positions));
                     count += 1;
                 }
             }
@@ -206,7 +206,7 @@ impl Index {
         let number_of_attributes = self.number_of_attributes(rtxn)?.map_or(0, |n| n as u32);
         for (i, derived_words) in words.iter().enumerate() {
             let mut union_docids = RoaringBitmap::new();
-            for (word, _) in derived_words {
+            for (word, _positions) in derived_words {
                 for attr in 0..number_of_attributes {
                     let mut key = word.to_vec();
                     key.extend_from_slice(&attr.to_be_bytes());

src/query_tokens.rs

@@ -81,11 +81,10 @@ impl<'a> Iterator for QueryTokens<'a> {
 #[cfg(test)]
 mod tests {
     use super::*;
+    use QueryToken::{Quoted, Free};
 
     #[test]
     fn one_quoted_string() {
-        use QueryToken::Quoted;
-
         let mut iter = QueryTokens::new("\"hello\"");
         assert_eq!(iter.next(), Some(Quoted("hello")));
         assert_eq!(iter.next(), None);
@@ -93,8 +92,6 @@ mod tests {
 
     #[test]
     fn one_pending_quoted_string() {
-        use QueryToken::Quoted;
-
         let mut iter = QueryTokens::new("\"hello");
         assert_eq!(iter.next(), Some(Quoted("hello")));
         assert_eq!(iter.next(), None);
@@ -102,8 +99,6 @@ mod tests {
 
     #[test]
     fn one_non_quoted_string() {
-        use QueryToken::Free;
-
         let mut iter = QueryTokens::new("hello");
         assert_eq!(iter.next(), Some(Free("hello")));
         assert_eq!(iter.next(), None);
@@ -111,8 +106,6 @@ mod tests {
 
     #[test]
     fn quoted_directly_followed_by_free_strings() {
-        use QueryToken::{Quoted, Free};
-
         let mut iter = QueryTokens::new("\"hello\"world");
         assert_eq!(iter.next(), Some(Quoted("hello")));
         assert_eq!(iter.next(), Some(Free("world")));
@@ -121,8 +114,6 @@ mod tests {
 
     #[test]
     fn free_directly_followed_by_quoted_strings() {
-        use QueryToken::{Quoted, Free};
-
         let mut iter = QueryTokens::new("hello\"world\"");
         assert_eq!(iter.next(), Some(Free("hello")));
         assert_eq!(iter.next(), Some(Quoted("world")));
@@ -131,8 +122,6 @@ mod tests {
 
     #[test]
     fn free_followed_by_quoted_strings() {
-        use QueryToken::{Quoted, Free};
-
         let mut iter = QueryTokens::new("hello \"world\"");
         assert_eq!(iter.next(), Some(Free("hello")));
         assert_eq!(iter.next(), Some(Quoted("world")));
@@ -141,8 +130,6 @@ mod tests {
 
     #[test]
     fn multiple_spaces_separated_strings() {
-        use QueryToken::Free;
-
         let mut iter = QueryTokens::new("hello    world   ");
         assert_eq!(iter.next(), Some(Free("hello")));
         assert_eq!(iter.next(), Some(Free("world")));
@@ -151,8 +138,6 @@ mod tests {
 
     #[test]
     fn multi_interleaved_quoted_free_strings() {
-        use QueryToken::{Quoted, Free};
-
         let mut iter = QueryTokens::new("hello \"world\" coucou \"monde\"");
         assert_eq!(iter.next(), Some(Free("hello")));
         assert_eq!(iter.next(), Some(Quoted("world")));
@@ -160,4 +145,13 @@ mod tests {
         assert_eq!(iter.next(), Some(Quoted("monde")));
         assert_eq!(iter.next(), None);
     }
+
+    #[test]
+    fn multi_quoted_strings() {
+        let mut iter = QueryTokens::new("\"hello world\" coucou \"monde est beau\"");
+        assert_eq!(iter.next(), Some(Quoted("hello world")));
+        assert_eq!(iter.next(), Some(Free("coucou")));
+        assert_eq!(iter.next(), Some(Quoted("monde est beau")));
+        assert_eq!(iter.next(), None);
+    }
 }

templates/index.html

@@ -34,7 +34,7 @@
           <div class="level-item has-text-centered">
             <div>
               <p class="heading">Database Size</p>
-              <p class="title" id="db-size">{{ db_size }}</p>
+              <p class="title" id="db-size">{{ db_size }} + {{ docs_size }}</p>
             </div>
           </div>
           <div class="level-item has-text-centered">