Improve BibTeX completion performance

CHANGELOG.md

@@ -5,6 +5,14 @@ All notable changes to this project will be documented in this file.
 The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.0.0/),
 and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html).
 
+## [Unreleased]
+
+### Fixed
+
+- Improve performance when completing BibTeX entries ([#493](https://github.com/latex-lsp/texlab/issues/493))
+- Don't report unused entries for very large bibliographies
+- Avoid redundant reparses after saving documents
+
 ## [5.9.0] - 2023-08-06
 
 ### Added

crates/base-db/Cargo.toml

@@ -21,6 +21,7 @@ rustc-hash = "1.1.0"
 syntax = { path = "../syntax" }
 text-size = "1.1.1"
 url = "=2.3.1"
+bibtex-utils = { path = "../bibtex-utils" }
 
 [lib]
 doctest = false

crates/base-db/src/semantics/bib.rs

@@ -1,7 +1,11 @@
+use bibtex_utils::field::text::TextFieldData;
+use itertools::Itertools;
 use rowan::ast::AstNode;
-use syntax::bibtex::{self, HasName};
+use syntax::bibtex::{self, HasName, HasType, HasValue};
 use text_size::TextRange;
 
+use crate::data::{BibtexEntryType, BibtexEntryTypeCategory};
+
 use super::Span;
 
 #[derive(Debug, Clone, Default)]
@@ -23,12 +27,27 @@ impl Semantics {
 
     fn process_entry(&mut self, entry: bibtex::Entry) {
         if let Some(name) = entry.name_token() {
+            let type_token = entry.type_token().unwrap();
+            let category = BibtexEntryType::find(&type_token.text()[1..])
+                .map_or(BibtexEntryTypeCategory::Misc, |ty| ty.category);
+
+            let field_values = entry
+                .fields()
+                .filter_map(|field| Some(TextFieldData::parse(&field.value()?)?.text));
+
+            let keywords = [name.text().into(), type_token.text().into()]
+                .into_iter()
+                .chain(field_values)
+                .join(" ");
+
             self.entries.push(Entry {
                 name: Span {
                     range: name.text_range(),
                     text: name.text().into(),
                 },
                 full_range: entry.syntax().text_range(),
+                category,
+                keywords,
             });
         }
     }
@@ -50,6 +69,8 @@ impl Semantics {
 pub struct Entry {
     pub name: Span,
     pub full_range: TextRange,
+    pub keywords: String,
+    pub category: BibtexEntryTypeCategory,
 }
 
 #[derive(Debug, Clone)]

crates/base-db/src/workspace.rs

@@ -75,6 +75,12 @@ impl Workspace {
             Cow::Owned(text) => text,
         };
 
+        if let Some(document) = self.lookup_path(path) {
+            if document.text == text {
+                return Ok(());
+            }
+        }
+
         self.open(uri, text, language, owner, LineCol { line: 0, col: 0 });
         Ok(())
     }

crates/bibtex-utils/Cargo.toml

@@ -0,0 +1,23 @@
+[package]
+name = "bibtex-utils"
+version = "0.0.0"
+license.workspace = true
+authors.workspace = true
+edition.workspace = true
+rust-version.workspace = true
+
+[dependencies]
+chrono = { version = "0.4.26", default-features = false, features = ["std"] }
+human_name = "2.0.2"
+itertools = "0.11.0"
+rowan = "0.15.11"
+rustc-hash = "1.1.0"
+syntax = { path = "../syntax" }
+unicode-normalization = "0.1.22"
+
+[lib]
+doctest = false
+
+[dev-dependencies]
+expect-test = "1.4.1"
+parser = { path = "../parser" }

crates/bibtex-utils/src/lib.rs

@@ -0,0 +1 @@
+pub mod field;

crates/citeproc/Cargo.toml

@@ -7,8 +7,7 @@ edition.workspace = true
 rust-version.workspace = true
 
 [dependencies]
-chrono = { version = "0.4.26", default-features = false, features = ["std"] }
-human_name = "2.0.2"
+bibtex-utils = { path = "../bibtex-utils" }
 isocountry = "0.3.2"
 itertools = "0.11.0"
 rowan = "0.15.11"

crates/citeproc/src/driver.rs

@@ -1,3 +1,9 @@
+use bibtex_utils::field::{
+    author::AuthorField,
+    date::DateField,
+    number::{NumberField, NumberFieldData},
+    text::TextField,
+};
 use isocountry::CountryCode;
 use itertools::Itertools;
 use syntax::bibtex;
@@ -6,12 +12,6 @@ use url::Url;
 
 use super::{
     entry::{EntryData, EntryKind},
-    field::{
-        author::AuthorField,
-        date::DateField,
-        number::{NumberField, NumberFieldData},
-        text::TextField,
-    },
     output::{Inline, InlineBuilder, Punct},
 };
 

crates/citeproc/src/entry.rs

@@ -1,12 +1,11 @@
-use rustc_hash::FxHashMap;
-use syntax::bibtex::{Entry, Field, HasName, HasType, HasValue, Value};
-
-use super::field::{
+use bibtex_utils::field::{
     author::{AuthorField, AuthorFieldData},
     date::{DateField, DateFieldData},
     number::{NumberField, NumberFieldData},
     text::{TextField, TextFieldData},
 };
+use rustc_hash::FxHashMap;
+use syntax::bibtex::{Entry, Field, HasName, HasType, HasValue, Value};
 
 #[derive(Debug, PartialEq, Eq, PartialOrd, Ord, Clone, Copy, Hash)]
 pub enum EntryKind {

crates/citeproc/src/lib.rs

@@ -1,6 +1,5 @@
 mod driver;
 mod entry;
-pub mod field;
 mod output;
 
 use syntax::bibtex;

crates/completion-data/src/lib.rs

@@ -107,7 +107,7 @@ const JSON_GZ: &[u8] = include_bytes!("../data/completion.json.gz");
 
 pub static DATABASE: Lazy<Database<'static>> = Lazy::new(|| {
     let mut decoder = GzDecoder::new(JSON_GZ);
-    let json = Box::leak(Box::new(String::new()));
+    let json = Box::leak(Box::default());
     decoder.read_to_string(json).unwrap();
     let mut db: Database = serde_json::from_str(json).unwrap();
     db.lookup_packages = db

crates/definition/src/citation.rs

@@ -7,7 +7,7 @@ use crate::DefinitionContext;
 
 use super::DefinitionResult;
 
-pub(super) fn goto_definition<'db>(context: &mut DefinitionContext<'db>) -> Option<()> {
+pub(super) fn goto_definition(context: &mut DefinitionContext) -> Option<()> {
     let data = context.params.document.data.as_tex()?;
 
     let citation = queries::object_at_cursor(

crates/definition/src/command.rs

@@ -6,7 +6,7 @@ use crate::DefinitionContext;
 
 use super::DefinitionResult;
 
-pub(super) fn goto_definition<'db>(context: &mut DefinitionContext<'db>) -> Option<()> {
+pub(super) fn goto_definition(context: &mut DefinitionContext) -> Option<()> {
     let data = context.params.document.data.as_tex()?;
     let root = data.root_node();
     let name = root

crates/definition/src/include.rs

@@ -4,7 +4,7 @@ use crate::DefinitionContext;
 
 use super::DefinitionResult;
 
-pub(super) fn goto_definition<'db>(context: &mut DefinitionContext<'db>) -> Option<()> {
+pub(super) fn goto_definition(context: &mut DefinitionContext) -> Option<()> {
     let start = context.params.document;
     let parents = context.params.workspace.parents(start);
     let results = parents

crates/definition/src/label.rs

@@ -10,7 +10,7 @@ use crate::DefinitionContext;
 
 use super::DefinitionResult;
 
-pub(super) fn goto_definition<'db>(context: &mut DefinitionContext<'db>) -> Option<()> {
+pub(super) fn goto_definition(context: &mut DefinitionContext) -> Option<()> {
     let data = context.params.document.data.as_tex()?;
     let reference = queries::object_at_cursor(
         &data.semantics.labels,

crates/definition/src/lib.rs

@@ -29,7 +29,7 @@ struct DefinitionContext<'db> {
     results: Vec<DefinitionResult<'db>>,
 }
 
-pub fn goto_definition<'db>(params: DefinitionParams<'db>) -> Vec<DefinitionResult<'db>> {
+pub fn goto_definition(params: DefinitionParams) -> Vec<DefinitionResult> {
     let project = params.workspace.project(params.document);
     let mut context = DefinitionContext {
         params,

crates/definition/src/string_ref.rs

@@ -5,7 +5,7 @@ use crate::DefinitionContext;
 
 use super::DefinitionResult;
 
-pub(super) fn goto_definition<'db>(context: &mut DefinitionContext<'db>) -> Option<()> {
+pub(super) fn goto_definition(context: &mut DefinitionContext) -> Option<()> {
     let data = context.params.document.data.as_bib()?;
     let root = data.root_node();
     let name = root

crates/diagnostics/src/build_log.rs

@@ -25,19 +25,24 @@ impl DiagnosticSource for BuildErrors {
     fn update(&mut self, workspace: &Workspace, log_document: &Document) {
         let mut errors: FxHashMap<Url, Vec<Diagnostic>> = FxHashMap::default();
 
-        let Some(data) = log_document.data.as_log() else { return };
+        let Some(data) = log_document.data.as_log() else {
+            return;
+        };
 
         let parents = workspace.parents(log_document);
-        let Some(root_document) = parents.iter().next() else { return };
+        let Some(root_document) = parents.iter().next() else {
+            return;
+        };
 
-        let Some(base_path) = root_document
-            .path
-            .as_deref()
-            .and_then(|path| path.parent()) else { return };
+        let Some(base_path) = root_document.path.as_deref().and_then(|path| path.parent()) else {
+            return;
+        };
 
         for error in &data.errors {
             let full_path = base_path.join(&error.relative_path);
-            let Ok(full_path_uri) = Url::from_file_path(&full_path) else { continue };
+            let Ok(full_path_uri) = Url::from_file_path(&full_path) else {
+                continue;
+            };
             let tex_document = workspace.lookup(&full_path_uri).unwrap_or(root_document);
 
             let range = find_range_of_hint(tex_document, error).unwrap_or_else(|| {
@@ -74,9 +79,12 @@ impl DiagnosticSource for BuildErrors {
         self.logs.retain(|uri, _| workspace.lookup(uri).is_some());
 
         for document in workspace.iter() {
-            let Some(log) = self.logs.get(&document.uri) else { continue };
+            let Some(log) = self.logs.get(&document.uri) else {
+                continue;
+            };
+
             for (uri, errors) in &log.errors {
-                builder.push_many(&uri, errors.iter().map(Cow::Borrowed));
+                builder.push_many(uri, errors.iter().map(Cow::Borrowed));
             }
         }
     }