Develop

.vscode/settings.json

@@ -0,0 +1,3 @@
+{
+  "rust-analyzer.cargo.features": ["in_memory"]
+}

CHANGELOG.md

@@ -2,6 +2,18 @@
 
 All notable changes to this project will be documented in this file.
 
+## [5.0.0] - 2023-07-21
+
+### Features
+
+- Changed the Search, Upsert and Compact API from mutable to immutable
+- Upsert now returns the set of new keywords added to the Entry Table
+
+### Bug Fixes
+
+- add missing `async` keyword for compact callbacks
+- fix `list_removed_locations` doc
+
 ## [4.0.3] - 2023-07-11
 
 ### Features

Cargo.toml

@@ -1,42 +1,65 @@
 [package]
 name = "cosmian_findex"
-version = "4.0.3"
+version = "5.0.0"
 authors = [
-  "Chloé Hébant <[email protected]>",
-  "Bruno Grieder <[email protected]>",
-  "Célia Corsin <[email protected]>",
-  "Emmanuel Coste <[email protected]>",
-  "Théophile Brézot <[email protected]>",
+    "Chloé Hébant <[email protected]>",
+    "Bruno Grieder <[email protected]>",
+    "Célia Corsin <[email protected]>",
+    "Emmanuel Coste <[email protected]>",
+    "Théophile Brézot <[email protected]>",
 ]
 categories = ["cosmian::crypto"]
 edition = "2021"
 keywords = ["SSE"]
-license = "MIT/Apache-2.0"
+license-file = "LICENSE.md"
 repository = "https://github.com/Cosmian/findex/"
 description = "Symmetric Searchable Encryption"
 
 [lib]
-crate-type = ["cdylib", "rlib", "staticlib"]
+crate-type = ["cdylib", "lib", "staticlib"]
 name = "cosmian_findex"
 path = "src/lib.rs"
 
-[[bench]]
-harness = false
-name = "benches"
-
 [features]
 in_memory = []
-live_compact = []
 
 [dependencies]
 base64 = "0.21.2"
-cosmian_crypto_core = { version = "9.0.0", features = ["aes", "sha3", "ser"], default-features = false }
+cosmian_crypto_core = { version = "9.0.0", features = [
+    "aes",
+    "sha3",
+    "ser",
+], default-features = false }
 rand = "0.8"
 tiny-keccak = { version = "2.0.2", features = ["kmac", "sha3"] }
 zeroize = "1.5"
 
 [dev-dependencies]
 actix-rt = "2.8"
-criterion = { version = "0.4", features = ["html_reports", "async_futures"], default_features = false }
+criterion = { version = "0.4", features = [
+    "html_reports",
+    "async_futures",
+], default_features = false }
 futures = "0.3"
 serde_json = "1.0"
+
+[[bench]]
+harness = false
+name = "benches"
+required-features = ["in_memory"]
+
+[[test]]
+name = "test_in_memory"
+required-features = ["in_memory"]
+
+[[test]]
+name = "non_regression"
+required-features = ["in_memory"]
+
+[[example]]
+name = "search"
+required-features = ["in_memory"]
+
+[[example]]
+name = "upsert"
+required-features = ["in_memory"]

README.md

@@ -2,12 +2,12 @@
 
 ![Build status](https://github.com/Cosmian/findex/actions/workflows/ci.yml/badge.svg)
 ![Build status](https://github.com/Cosmian/findex/actions/workflows/build.yml/badge.svg)
-![latest version](<https://img.shields.io/crates/v/cosmian_findex.svg>)
+![latest version](https://img.shields.io/crates/v/cosmian_findex.svg)
 
 Findex aims to solve the following problem:
 
-> How to securely recover the *location* of an encrypted data matching a given
-> *keyword*?
+> How to securely recover the _location_ of an encrypted data matching a given
+> _keyword_?
 
 It is a cryptographic protocol designed to securely make search queries on an
 untrusted cloud server. Thanks to its encrypted indexes, large databases can
@@ -20,35 +20,33 @@ Findex is part of Cosmian Cloudproof Encryption.
 - [Getting started](#getting-started)
 - [Building and testing](#building-and-testing)
 - [Findex indexes](#findex-indexes)
-  * [Two indexing strategies](#two-indexing-strategies)
+  - [Two indexing strategies](#two-indexing-strategies)
 - [Benchmarks](#benchmarks)
 - [Documentation](#documentation)
 
 <!-- tocstop -->
 
 ## Getting started
 
-Findex allows to index values by keywords. These values can be locations (UIDs
+Findex allows indexing values by keywords. These values can be locations (UIDs
 of an encrypted database, URLs, paths, etc.).
 
 Using Findex API one can:
 
-- index or desindex values by keywords via the `FindexUpsert` trait;
+- index or deindex values by keywords via the `FindexUpsert` trait;
 - search for keywords via the `FindexSearch` trait;
 - compact the indexes via the `FindexCompact` trait.
 
 These traits can be automatically implemented and a macro is provided to help
 with the syntax. The default parameters (the ones used by the macro) are
 defined in [`parameters.rs`](./src/parameters.rs).
 
-Findex delegates to the user the implementation of *callbacks* to manipulate
-the indexes. This makes Findex compatible with any database technology since no
-database specific code is part of it. Implementation is done via the
+Findex delegates to the user the implementation of _callbacks_ to manipulate
+the indexes. This makes Findex compatible with any database technology since no database-specific code is part of it. Implementation is done via the
 `FindexCallbacks` trait. See [`callbacks.md`](./callbacks.md) for details on
 the implementation of the callbacks.
 
-See [`in_memory_example.rs`](./src/in_memory_example.rs) for a example of
-implementation.
+See `in_memory_example.rs` for an example of implementation.
 
 ## Building and testing
 
@@ -72,7 +70,7 @@ cargo bench --all-features
 
 ## Findex indexes
 
-Findex relies on two server side indexes:
+Findex relies on two server-side indexes:
 
 - **Entry Table**: provides the values needed to fetch the correct locations
   from the Chain Table. Each indexing keyword matches a line in the Entry
@@ -84,8 +82,8 @@ Findex relies on two server side indexes:
   stored by blocks of fixed length and the same number of blocks is stored in
   each line (padding is added where necessary).
 
-Findex indexes are key value stores which structure is given in the following
-tables, with $K_{w_i}$ the ephemeral key associated to a keyword $w_i$,
+Findex indexes are key-value stores whose structure is given in the following
+tables, with $K_{w_i}$ the ephemeral key associated with a keyword $w_i$,
 $H_{w_i}$ the hash of $w_i$ and $UID_{last}$ the last UID of the chain of
 indexed values associated to $w_i$.
 
@@ -214,7 +212,7 @@ More client/server interactions are needed for the graph solution: the depth of
 the graph (4 in this example) compared to 1 for the naive solution and 2 for
 the mixed solution.
 
-In the other hand, the graph solution optimizes the size of the Chain Table.
+On the other hand, the graph solution optimizes the size of the Chain Table.
 
 <table>
   <tr>
@@ -292,7 +290,7 @@ In the other hand, the graph solution optimizes the size of the Chain Table.
 
 ## Benchmarks
 
-The benchmarks presented in this section are run on a Intel(R) Xeon(R) Platinum 8171M CPU @ 2.60GHz.
+The benchmarks presented in this section are run on an Intel(R) Xeon(R) Platinum 8171M CPU @ 2.60GHz.
 
 - [Findex in memory (no database)](./benches/BENCHMARKS.md)
 

benches/benches.rs

@@ -1,6 +1,3 @@
-#[cfg(not(feature = "in_memory"))]
-compile_error!("Benches require the `in_memory` feature.");
-
 use std::collections::{HashMap, HashSet};
 
 use cosmian_crypto_core::CsRng;
@@ -48,7 +45,7 @@ fn bench_search(c: &mut Criterion) {
     //
     // Prepare indexes to be search
     //
-    let mut findex = FindexInMemory::default();
+    let findex = FindexInMemory::default();
     block_on(findex.add(&master_key, &label, locations_and_words)).expect("msg");
 
     println!("Entry Table length: {}", findex.entry_table_len());