UPDATE: Adds bincode to serializer/deserializer

Cargo.lock

@@ -38,6 +38,25 @@ version = "1.1.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "d468802bab17cbc0cc575e9b053f41e72aa36bfa6b7f55e3529ffa43161b97fa"
 
+[[package]]
+name = "bincode"
+version = "2.0.0-rc.3"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "f11ea1a0346b94ef188834a65c068a03aec181c94896d481d7a0a40d85b0ce95"
+dependencies = [
+ "bincode_derive",
+ "serde",
+]
+
+[[package]]
+name = "bincode_derive"
+version = "2.0.0-rc.3"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "7e30759b3b99a1b802a7a3aa21c85c3ded5c28e1c83170d82d70f08bbf7f3e4c"
+dependencies = [
+ "virtue",
+]
+
 [[package]]
 name = "bumpalo"
 version = "3.14.0"
@@ -191,6 +210,7 @@ dependencies = [
 name = "ironcalc"
 version = "0.1.3"
 dependencies = [
+ "bincode",
  "chrono",
  "ironcalc_base",
  "itertools",
@@ -206,6 +226,7 @@ dependencies = [
 name = "ironcalc_base"
 version = "0.1.2"
 dependencies = [
+ "bincode",
  "chrono",
  "chrono-tz",
  "js-sys",
@@ -541,6 +562,12 @@ dependencies = [
  "serde",
 ]
 
+[[package]]
+name = "virtue"
+version = "0.0.13"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "9dcc60c0624df774c82a0ef104151231d37da4962957d691c011c852b2473314"
+
 [[package]]
 name = "wasi"
 version = "0.10.0+wasi-snapshot-preview1"

README.md

@@ -117,55 +117,7 @@ See more examples in the `examples` folder of the xlsx crate.
 
 # ROADMAP
 
-> [!WARNING]  
+See https://github.com/ironcalc
-> This is work-in-progress. IronCalc in developed in the open. Expect things to be broken and change quickly until version 0.5
-
-Major milestones:
-
-* MVP, version 0.5.0: We intend to have a working version by mid March 2024 (version 0.5, MVP)
-* Stable, version 1.0.0 will come later in December 2024
-
-MVP stands for _Minimum Viable Product_
-
-## Version 0.5 or MVP (early 2024)
-
-Version 0.5 includes the engine, javascript and nodejs bindings and a web application
-
-Features of the engine include:
-
-* Read and write xlsx files
-* API to set and read values from cells
-* Implemented 192 Excel functions
-* Time functions with timezones
-* Prepared for i18n but will only support English
-* Wide test coverage
-
-UI features of the web application (backed by the engine):
-
-* Enter values and formulas. Browse mode
-* Italics, bold, underline, horizontal alignment
-* Number formatting
-* Add/remove/rename sheets
-* Copy/Paste extend values
-* Keyboard navigation
-* Delete/Add rows and columns
-* Resize rows and columns
-* Correct scrolling and navigation
-
-## Version 1.0 or Stable (December 2024)
-
-Minor milestones in the ROADMAD for version 1.0.0 (engine and UI):
-
-* Implementation of arrays and array formulas
-* Formula documentation and context help
-* Merge cells
-* Pivot tables
-* Define name manager (mostly UI)
-* Update main evaluation algorithm with a support graph
-* Dynamic arrays (SORT, UNIQUE, ..)
-* Full i18n support with different locales and languages
-* Python bindings
-* Full test coverage
 
 # Early testing
 

base/Cargo.toml

@@ -19,6 +19,7 @@ chrono = "0.4"
 chrono-tz = "0.7.0"
 regex = "1.0"
 once_cell = "1.16.0"
+bincode = "=2.0.0-rc.3"
 
 [target.'cfg(target_arch = "wasm32")'.dependencies]
 js-sys = { version = "0.3.60" }

base/src/expressions/token.rs

@@ -1,5 +1,6 @@
 use std::fmt;
 
+use bincode::{Decode, Encode};
 use serde::{Deserialize, Serialize};
 use serde_repr::{Deserialize_repr, Serialize_repr};
 
@@ -80,7 +81,7 @@ impl fmt::Display for OpProduct {
 ///  * "#ERROR!" means there was an error processing the formula (for instance "=A1+")
 ///  * "#N/IMPL!" means the formula or feature in Excel but has not been implemented in IronCalc
 /// Note that they are serialized/deserialized by index
-#[derive(Serialize_repr, Deserialize_repr, Debug, PartialEq, Eq, Clone)]
+#[derive(Serialize_repr, Deserialize_repr, Decode, Encode, Debug, PartialEq, Eq, Clone)]
 #[repr(u8)]
 pub enum Error {
     REF,

base/src/garbage_collector.rs

@@ -1,12 +0,0 @@
-use crate::model::Model;
-
-// The gc (Garbage Collector) cleans up leftover elements in the workbook:
-// * Strings that are no longe reachable
-// * Styles that are no longer reachable
-// * ...
-
-impl Model {
-    pub(crate) fn gc(&mut self) -> Result<(), String> {
-        todo!()
-    }
-}

base/src/lib.rs

@@ -50,7 +50,6 @@ mod implicit_intersection;
 mod units;
 mod utils;
 mod workbook;
-mod garbage_collector;
 
 #[cfg(test)]
 mod test;

base/src/model.rs

@@ -7,6 +7,7 @@
 //! IronCalc is row first. A cell is referenced by (`sheet`, `row`, `column`)
 //!
 
+use bincode::config;
 use serde_json::json;
 
 use std::collections::HashMap;
@@ -1718,6 +1719,12 @@ impl Model {
         }
     }
 
+    /// bin
+    pub fn to_binary_str(&self) -> Vec<u8> {
+        let config = config::standard();
+        bincode::encode_to_vec(&self.workbook, config).expect("")
+    }
+
     /// Returns markup representation of the given `sheet`.
     pub fn sheet_markup(&self, sheet: u32) -> Result<String, String> {
         let worksheet = self.workbook.worksheet(sheet)?;

base/src/test/mod.rs

@@ -53,4 +53,3 @@ mod test_frozen_rows_and_columns;
 mod test_get_cell_content;
 mod test_percentage;
 mod test_today;
-mod test_gc;

base/src/test/test_gc.rs

@@ -1,22 +0,0 @@
-#![allow(clippy::unwrap_used)]
-
-use crate::test::util::new_empty_model;
-
-#[test]
-fn test_empty_model() {
-    let mut model = new_empty_model();
-    // set a string
-    model._set("A1", "Hello");
-    assert_eq!(model.shared_strings.len(), 1);
-    // calling the gc doesn't do anything
-    model.gc().unwrap();
-    assert_eq!(model.shared_strings.len(), 1);
-
-    // If we delete the cell the string is still in the list
-    model.delete_cell(0, 1, 1).unwrap();
-    assert_eq!(model.shared_strings.len(), 1);
-
-    // after the gc the string is no longer present
-    model.gc().unwrap();
-    assert_eq!(model.shared_strings.len(), 0);
-}

base/src/types.rs

@@ -1,3 +1,4 @@
+use bincode::{Decode, Encode};
 use serde::{Deserialize, Serialize};
 use std::{collections::HashMap, fmt::Display};
 
@@ -33,7 +34,7 @@ fn hashmap_is_empty(h: &HashMap<String, Table>) -> bool {
     h.values().len() == 0
 }
 
-#[derive(Serialize, Deserialize, Debug, PartialEq, Eq, Clone)]
+#[derive(Serialize, Deserialize, Decode, Encode, Debug, PartialEq, Eq, Clone)]
 pub struct Metadata {
     pub application: String,
     pub app_version: String,
@@ -43,13 +44,13 @@ pub struct Metadata {
     pub last_modified: String, //"2020-11-20T16:24:35"
 }
 
-#[derive(Serialize, Deserialize, Debug, PartialEq, Eq, Clone)]
+#[derive(Serialize, Deserialize, Decode, Encode, Debug, PartialEq, Eq, Clone)]
 pub struct WorkbookSettings {
     pub tz: String,
     pub locale: String,
 }
 /// An internal representation of an IronCalc Workbook
-#[derive(Serialize, Deserialize, Debug, PartialEq, Clone)]
+#[derive(Serialize, Deserialize, Decode, Encode, Debug, PartialEq, Clone)]
 #[serde(deny_unknown_fields)]
 pub struct Workbook {
     pub shared_strings: Vec<String>,
@@ -65,7 +66,7 @@ pub struct Workbook {
 }
 
 /// A defined name. The `sheet_id` is the sheet index in case the name is local
-#[derive(Serialize, Deserialize, Debug, PartialEq, Eq, Clone)]
+#[derive(Serialize, Deserialize, Decode, Encode, Debug, PartialEq, Eq, Clone)]
 pub struct DefinedName {
     pub name: String,
     pub formula: String,
@@ -79,7 +80,7 @@ pub struct DefinedName {
 /// * state:
 ///    18.18.68 ST_SheetState (Sheet Visibility Types)
 ///    hidden, veryHidden, visible
-#[derive(Serialize, Deserialize, Debug, PartialEq, Eq, Clone)]
+#[derive(Serialize, Deserialize, Decode, Encode, Debug, PartialEq, Eq, Clone)]
 #[serde(rename_all = "lowercase")]
 pub enum SheetState {
     Visible,
@@ -98,7 +99,7 @@ impl Display for SheetState {
 }
 
 /// Internal representation of a worksheet Excel object
-#[derive(Serialize, Deserialize, Debug, PartialEq, Clone)]
+#[derive(Serialize, Deserialize, Decode, Encode, Debug, PartialEq, Clone)]
 pub struct Worksheet {
     pub dimension: String,
     pub cols: Vec<Col>,
@@ -125,7 +126,7 @@ pub struct Worksheet {
 pub type SheetData = HashMap<i32, HashMap<i32, Cell>>;
 
 // ECMA-376-1:2016 section 18.3.1.73
-#[derive(Serialize, Deserialize, Debug, PartialEq, Clone)]
+#[derive(Serialize, Deserialize, Decode, Encode, Debug, PartialEq, Clone)]
 pub struct Row {
     /// Row index
     pub r: i32,
@@ -139,7 +140,7 @@ pub struct Row {
 }
 
 // ECMA-376-1:2016 section 18.3.1.13
-#[derive(Serialize, Deserialize, Debug, PartialEq, Clone)]
+#[derive(Serialize, Deserialize, Decode, Encode, Debug, PartialEq, Clone)]
 pub struct Col {
     // Column definitions are defined on ranges, unlike rows which store unique, per-row entries.
     /// First column affected by this record. Settings apply to column in \[min, max\] range.
@@ -164,7 +165,7 @@ pub enum CellType {
     CompoundData = 128,
 }
 
-#[derive(Serialize, Deserialize, Debug, Clone, PartialEq)]
+#[derive(Serialize, Deserialize, Decode, Encode, Debug, Clone, PartialEq)]
 #[serde(tag = "t", deny_unknown_fields)]
 pub enum Cell {
     #[serde(rename = "empty")]
@@ -208,7 +209,7 @@ impl Default for Cell {
     }
 }
 
-#[derive(Serialize, Deserialize, Debug, PartialEq, Eq, Clone)]
+#[derive(Serialize, Deserialize, Decode, Encode, Debug, PartialEq, Eq, Clone)]
 pub struct Comment {
     pub text: String,
     pub author_name: String,
@@ -218,7 +219,7 @@ pub struct Comment {
 }
 
 // ECMA-376-1:2016 section 18.5.1.2
-#[derive(Serialize, Deserialize, Debug, PartialEq, Eq, Clone)]
+#[derive(Serialize, Deserialize, Decode, Encode, Debug, PartialEq, Eq, Clone)]
 pub struct Table {
     pub name: String,
     pub display_name: String,
@@ -241,7 +242,7 @@ pub struct Table {
 
 // totals_row_label vs totals_row_function might be mutually exclusive. Use an enum?
 // the totals_row_function is an enum not String methinks
-#[derive(Serialize, Deserialize, Debug, PartialEq, Eq, Clone)]
+#[derive(Serialize, Deserialize, Decode, Encode, Debug, PartialEq, Eq, Clone)]
 pub struct TableColumn {
     pub id: u32,
     pub name: String,
@@ -271,7 +272,7 @@ impl Default for TableColumn {
     }
 }
 
-#[derive(Serialize, Deserialize, Debug, PartialEq, Eq, Clone, Default)]
+#[derive(Serialize, Deserialize, Decode, Encode, Debug, PartialEq, Eq, Clone, Default)]
 pub struct TableStyleInfo {
     #[serde(skip_serializing_if = "Option::is_none")]
     pub name: Option<String>,
@@ -289,7 +290,7 @@ pub struct TableStyleInfo {
     pub show_column_stripes: bool,
 }
 
-#[derive(Serialize, Deserialize, Debug, PartialEq, Eq, Clone)]
+#[derive(Serialize, Deserialize, Decode, Encode, Debug, PartialEq, Eq, Clone)]
 pub struct Styles {
     pub num_fmts: Vec<NumFmt>,
     pub fonts: Vec<Font>,
@@ -314,7 +315,7 @@ impl Default for Styles {
     }
 }
 
-#[derive(Serialize, Deserialize, Debug, PartialEq, Eq, Clone)]
+#[derive(Serialize, Deserialize, Decode, Encode, Debug, PartialEq, Eq, Clone)]
 pub struct Style {
     pub alignment: Option<Alignment>,
     pub num_fmt: String,
@@ -324,7 +325,7 @@ pub struct Style {
     pub quote_prefix: bool,
 }
 
-#[derive(Serialize, Deserialize, Debug, PartialEq, Eq, Clone)]
+#[derive(Serialize, Deserialize, Decode, Encode, Debug, PartialEq, Eq, Clone)]
 pub struct NumFmt {
     pub num_fmt_id: i32,
     pub format_code: String,
@@ -342,7 +343,7 @@ impl Default for NumFmt {
 // ST_FontScheme simple type (§18.18.33).
 // Usually major fonts are used for styles like headings,
 // and minor fonts are used for body and paragraph text.
-#[derive(Serialize, Deserialize, Debug, PartialEq, Eq, Clone)]
+#[derive(Serialize, Deserialize, Decode, Encode, Debug, PartialEq, Eq, Clone)]
 #[serde(rename_all = "lowercase")]
 #[derive(Default)]
 pub enum FontScheme {
@@ -362,7 +363,7 @@ impl Display for FontScheme {
     }
 }
 
-#[derive(Serialize, Deserialize, Debug, PartialEq, Eq, Clone)]
+#[derive(Serialize, Deserialize, Decode, Encode, Debug, PartialEq, Eq, Clone)]
 pub struct Font {
     #[serde(default = "default_as_false")]
     #[serde(skip_serializing_if = "is_false")]
@@ -405,7 +406,7 @@ impl Default for Font {
 }
 
 // TODO: Maybe use an enum for the pattern_type values here?
-#[derive(Serialize, Deserialize, Debug, PartialEq, Eq, Clone)]
+#[derive(Serialize, Deserialize, Decode, Encode, Debug, PartialEq, Eq, Clone)]
 pub struct Fill {
     pub pattern_type: String,
     #[serde(skip_serializing_if = "Option::is_none")]
@@ -424,7 +425,7 @@ impl Default for Fill {
     }
 }
 
-#[derive(Serialize, Deserialize, Debug, PartialEq, Eq, Clone)]
+#[derive(Serialize, Deserialize, Decode, Encode, Debug, PartialEq, Eq, Clone)]
 #[serde(rename_all = "lowercase")]
 pub enum HorizontalAlignment {
     Center,
@@ -466,7 +467,7 @@ impl Display for HorizontalAlignment {
     }
 }
 
-#[derive(Serialize, Deserialize, Debug, PartialEq, Eq, Clone)]
+#[derive(Serialize, Deserialize, Decode, Encode, Debug, PartialEq, Eq, Clone)]
 #[serde(rename_all = "lowercase")]
 pub enum VerticalAlignment {
     Bottom,
@@ -501,7 +502,7 @@ impl Display for VerticalAlignment {
 }
 
 // 1762
-#[derive(Serialize, Deserialize, Debug, PartialEq, Eq, Clone, Default)]
+#[derive(Serialize, Deserialize, Decode, Encode, Debug, PartialEq, Eq, Clone, Default)]
 pub struct Alignment {
     #[serde(default)]
     #[serde(skip_serializing_if = "HorizontalAlignment::is_default")]
@@ -514,7 +515,7 @@ pub struct Alignment {
     pub wrap_text: bool,
 }
 
-#[derive(Serialize, Deserialize, Debug, PartialEq, Eq, Clone)]
+#[derive(Serialize, Deserialize, Decode, Encode, Debug, PartialEq, Eq, Clone)]
 pub struct CellStyleXfs {
     pub num_fmt_id: i32,
     pub font_id: i32,
@@ -557,7 +558,7 @@ impl Default for CellStyleXfs {
     }
 }
 
-#[derive(Serialize, Deserialize, Debug, PartialEq, Eq, Clone, Default)]
+#[derive(Serialize, Deserialize, Decode, Encode, Debug, PartialEq, Eq, Clone, Default)]
 pub struct CellXfs {
     pub xf_id: i32,
     pub num_fmt_id: i32,
@@ -589,7 +590,7 @@ pub struct CellXfs {
     pub alignment: Option<Alignment>,
 }
 
-#[derive(Serialize, Deserialize, Debug, PartialEq, Eq, Clone)]
+#[derive(Serialize, Deserialize, Decode, Encode, Debug, PartialEq, Eq, Clone)]
 pub struct CellStyles {
     pub name: String,
     pub xf_id: i32,
@@ -606,7 +607,7 @@ impl Default for CellStyles {
     }
 }
 
-#[derive(Serialize, Deserialize, Debug, PartialEq, Eq, Clone)]
+#[derive(Serialize, Deserialize, Decode, Encode, Debug, PartialEq, Eq, Clone)]
 #[serde(rename_all = "lowercase")]
 pub enum BorderStyle {
     Thin,
@@ -636,13 +637,13 @@ impl Display for BorderStyle {
     }
 }
 
-#[derive(Serialize, Deserialize, Debug, PartialEq, Eq, Clone)]
+#[derive(Serialize, Deserialize, Decode, Encode, Debug, PartialEq, Eq, Clone)]
 pub struct BorderItem {
     pub style: BorderStyle,
     pub color: Option<String>,
 }
 
-#[derive(Serialize, Deserialize, Debug, PartialEq, Eq, Clone, Default)]
+#[derive(Serialize, Deserialize, Decode, Encode, Debug, PartialEq, Eq, Clone, Default)]
 pub struct Border {
     #[serde(default = "default_as_false")]
     #[serde(skip_serializing_if = "is_false")]
@@ -664,7 +665,7 @@ pub struct Border {
 
 /// Information need to show a sheet tab in the UI
 /// The color is serialized only if it is not Color::None
-#[derive(Serialize, Deserialize, Debug, PartialEq)]
+#[derive(Serialize, Deserialize, Decode, Encode, Debug, PartialEq)]
 pub struct SheetInfo {
     pub name: String,
     pub state: String,

xlsx/Cargo.toml

@@ -23,6 +23,7 @@ thiserror = "1.0"
 ironcalc_base = { path = "../base", version = "0.1.0" }
 itertools = "0.10.5"
 chrono = "0.4"
+bincode = "=2.0.0-rc.3"
 
 [dev-dependencies]
 uuid = { version = "1.2.2", features = ["serde", "v4"] }
@@ -34,3 +35,8 @@ path = "src/lib.rs"
 [[bin]]
 name = "test"
 path = "src/bin/test.rs"
+
+[[bin]]
+name = "bench"
+path = "src/bin/bench.rs"
+

xlsx/src/bin/bench.rs

@@ -0,0 +1,59 @@
+use std::{fs, io::Write, time::Instant};
+
+use bincode::config;
+use ironcalc::import::load_model_from_xlsx;
+use ironcalc_base::{types::Workbook, model::Model};
+
+
+fn main() {
+    let args: Vec<_> = std::env::args().collect();
+    if args.len() != 2 {
+        panic!("Usage: {} <file.xlsx>", args[0]);
+    }
+    let file_name = &args[1];
+    let now = Instant::now();
+    let mut model = load_model_from_xlsx(file_name, "en", "UTC").unwrap();
+    let elapsed_time = now.elapsed();
+    println!("Loaded model from xlsx: {:?}", elapsed_time);
+    model.evaluate();
+    let now = Instant::now();
+    let s = model.to_json_str();
+    let elapsed_time = now.elapsed();
+    println!("Stringify json: {:?}", elapsed_time);
+    {
+        let now = Instant::now();
+        let decoded : Workbook = serde_json::from_str(&s).unwrap();
+        let elapsed_time = now.elapsed();
+        println!("Parse from json: {:?} and name {}", elapsed_time, decoded.name);
+    }
+    let file_name_json = format!("{}.json", file_name);
+    let file_path = std::path::Path::new(&file_name_json);
+    let mut file = fs::File::create(file_path).unwrap();
+    file.write_all(s.as_bytes()).unwrap();
+    let now = Instant::now();
+    let s = model.to_binary_str();
+    let elapsed_time = now.elapsed();
+    println!("stringify to binary: {:?}", elapsed_time);
+    {
+        let config = config::standard();
+        let now = Instant::now();
+        let (decoded, _): (Workbook, usize) = bincode::decode_from_slice(&s[..], config).unwrap();
+        let elapsed_time = now.elapsed();
+        println!("Parse from binary: {:?} and {}", elapsed_time, decoded.name);
+    }
+    let file_name_binary = format!("{}.binary", file_name);
+    let file_path = std::path::Path::new(&file_name_binary);
+    let mut file = fs::File::create(file_path).unwrap();
+    file.write_all(&s).unwrap();
+    {
+        let config = config::standard();
+        let now = Instant::now();
+        let s = &fs::read(file_name_binary).unwrap();
+        let (decoded, _): (Workbook, usize) = bincode::decode_from_slice(&s[..], config).unwrap();
+        let model = Model::from_workbook(decoded).unwrap();
+        let elapsed_time = now.elapsed();
+        println!("Loaded from binary file: {:?} and {}", elapsed_time, model.workbook.name);
+
+    }
+}
+