Merge pull request #167 from ashvardanian/main

Disaggregating vector storage from the HNSW index

.github/workflows/prerelease.yml

@@ -17,8 +17,8 @@ permissions:
 
 jobs:
 
-  build_test:
-    name: Test USearch
+  test_cpp:
+    name: Test C++
     runs-on: ubuntu-latest
     steps:
     - uses: actions/checkout@v3
@@ -163,7 +163,7 @@ jobs:
     - name: Build USearch by Emscripten
       run: |
         ./emsdk/emsdk activate latest && source ./emsdk/emsdk_env.sh
-        emcmake cmake -B ./build -DCMAKE_CXX_FLAGS="${CMAKE_CXX_FLAGS} -s TOTAL_MEMORY=64MB" && emmake make -C ./build
+        emcmake cmake -DUSEARCH_BUILD_BENCHMARK=0 -DUSEARCH_BUILD_WASM=1 -B ./build -DCMAKE_CXX_FLAGS="${CMAKE_CXX_FLAGS} -s TOTAL_MEMORY=64MB" && emmake make -C ./build
     - name: Run tests
       run: node ./build/test.js
 

.vscode/launch.json

@@ -22,16 +22,16 @@
             ],
             "stopAtEntry": false,
             "linux": {
-                "preLaunchTask": "Linux Build C++ Bench Debug",
+                "preLaunchTask": "Linux Build C++ Debug",
                 "MIMode": "gdb"
             },
             "osx": {
-                "preLaunchTask": "MacOS Build C++ Bench Debug",
+                "preLaunchTask": "MacOS Build C++ Debug",
                 "MIMode": "lldb"
             }
         },
         {
-            "name": "Debug C++ Bench on Wiki 1M @Cos",
+            "name": "Debug C++ on Wiki 1M @Cos",
             "type": "cppdbg",
             "request": "launch",
             "program": "${workspaceFolder}/build_debug/bench",
@@ -53,7 +53,7 @@
             ],
             "stopAtEntry": false,
             "linux": {
-                "preLaunchTask": "Linux Build C++ Bench Debug",
+                "preLaunchTask": "Linux Build C++ Debug",
                 "MIMode": "gdb",
                 "environment": [
                     {
@@ -63,12 +63,12 @@
                 ]
             },
             "osx": {
-                "preLaunchTask": "MacOS Build C++ Bench Debug",
+                "preLaunchTask": "MacOS Build C++ Debug",
                 "MIMode": "lldb"
             }
         },
         {
-            "name": "Debug C++ Bench on Text-to-Image 1B @Cos",
+            "name": "Debug C++ on Text-to-Image 1B @Cos",
             "type": "cppdbg",
             "request": "launch",
             "program": "${workspaceFolder}/build_debug/bench",
@@ -90,16 +90,16 @@
             ],
             "stopAtEntry": false,
             "linux": {
-                "preLaunchTask": "Linux Build C++ Bench Debug",
+                "preLaunchTask": "Linux Build C++ Debug",
                 "MIMode": "gdb"
             },
             "osx": {
-                "preLaunchTask": "MacOS Build C++ Bench Debug",
+                "preLaunchTask": "MacOS Build C++ Debug",
                 "MIMode": "lldb"
             }
         },
         {
-            "name": "Run C++ Bench on Text-to-Image 1M @Cos",
+            "name": "Run C++ on Text-to-Image 1M @Cos",
             "type": "cppdbg",
             "request": "launch",
             "program": "${workspaceFolder}/build_release/bench",
@@ -113,7 +113,7 @@
                 "datasets/t2i_1M/groundtruth.public.100K.ibin",
             ],
             "stopAtEntry": false,
-            "preLaunchTask": "Linux Build C++ Bench Release",
+            "preLaunchTask": "Linux Build C++ Release",
             "MIMode": "gdb"
         },
         {

.vscode/settings.json

@@ -153,6 +153,9 @@
     "python.testing.pytestArgs": [
         "python"
     ],
+    "python.analysis.diagnosticSeverityOverrides": {
+        "reportMissingImports": "none"
+    },
     "python.testing.unittestEnabled": false,
     "python.testing.pytestEnabled": true
 }

.vscode/tasks.json

@@ -2,7 +2,7 @@
     "version": "2.0.0",
     "tasks": [
         {
-            "label": "Linux Build C++ Bench Debug",
+            "label": "Linux Build C++ Debug",
             "command": "cmake -DCMAKE_CXX_COMPILER=gcc-12 -DCMAKE_CXX_COMPILER=g++-12 -DCMAKE_BUILD_TYPE=Debug -B ./build_debug && make -C ./build_debug",
             "args": [],
             "type": "shell",
@@ -11,7 +11,7 @@
             ]
         },
         {
-            "label": "Linux Build C++ Bench Release",
+            "label": "Linux Build C++ Release",
             "command": "cmake -DCMAKE_CXX_COMPILER=gcc-12 -DCMAKE_CXX_COMPILER=g++-12 -DCMAKE_BUILD_TYPE=RelWithDebInfo -B ./build_release && make -C ./build_release",
             "args": [],
             "type": "shell",
@@ -20,13 +20,13 @@
             ]
         },
         {
-            "label": "MacOS Build C++ Bench Debug",
+            "label": "MacOS Build C++ Debug",
             "command": "cmake -DCMAKE_C_COMPILER=/opt/homebrew/opt/llvm/bin/clang -DCMAKE_CXX_COMPILER=/opt/homebrew/opt/llvm/bin/clang++ -DCMAKE_BUILD_TYPE=Debug -B ./build_debug && make -C ./build_debug",
             "args": [],
             "type": "shell",
         },
         {
-            "label": "MacOS Build C++ Bench Release",
+            "label": "MacOS Build C++ Release",
             "command": "cmake -DCMAKE_C_COMPILER=/opt/homebrew/opt/llvm/bin/clang -DCMAKE_CXX_COMPILER=/opt/homebrew/opt/llvm/bin/clang++ -DCMAKE_BUILD_TYPE=RelWithDebInfo -B ./build_release && make -C ./build_release",
             "args": [],
             "type": "shell"

Cargo.toml

@@ -1,6 +1,6 @@
 [package]
 name = "usearch"
-version = "0.22.2"
+version = "0.22.3"
 authors = ["Ash Vardanian <[email protected]>"]
 description = "Smaller & Faster Single-File Vector Search Engine from Unum"
 edition = "2021"

README.md

@@ -45,7 +45,7 @@ Linux • MacOS • Windows • Docker • WebAssembly
 - ✅ [View large indexes from disk](#disk-based-indexes) without loading into RAM.
 - ✅ Space-efficient point-clouds with `uint40_t`, accommodating 4B+ size.
 - ✅ Compatible with OpenMP and custom "executors", for fine-grained control over CPU utilization.
-- ✅ On-the-fly deletions.
+- ✅ Heterogeneous lookups, renaming/relabeleing, and on-the-fly deletions.
 - ✅ [Semantic Search](#usearch--ai--multi-modal-semantic-search) and [Joins](#joins).
 
 [usearch-header]: https://github.com/unum-cloud/usearch/blob/main/include/usearch/index.hpp
@@ -91,7 +91,7 @@ matches: Matches = index.search(vector, 10)
 
 assert len(index) == 1
 assert len(matches) == 1
-assert matches[0].label == 42
+assert matches[0].key == 42
 assert matches[0].distance <= 0.001
 assert np.allclose(index[42], vector)
 ```
@@ -214,17 +214,17 @@ model = uform.get_model('unum-cloud/uform-vl-multilingual')
 index = usearch.index.Index(ndim=256)
 
 @server
-def add(label: int, photo: pil.Image.Image):
+def add(key: int, photo: pil.Image.Image):
     image = model.preprocess_image(photo)
     vector = model.encode_image(image).detach().numpy()
-    index.add(label, vector.flatten(), copy=True)
+    index.add(key, vector.flatten(), copy=True)
 
 @server
 def search(query: str) -> np.ndarray:
     tokens = model.preprocess_text(query)
     vector = model.encode_text(tokens).detach().numpy()
     matches = index.search(vector.flatten(), 3)
-    return matches.labels
+    return matches.keys
 
 server.run()
 ```
@@ -268,9 +268,9 @@ fingerprints = np.vstack([encoder.GetFingerprint(x) for x in molecules])
 fingerprints = np.packbits(fingerprints, axis=1)
 
 index = Index(ndim=2048, metric=MetricKind.Tanimoto)
-labels = np.arange(len(molecules))
+keys = np.arange(len(molecules))
 
-index.add(labels, fingerprints)
+index.add(keys, fingerprints)
 matches = index.search(fingerprints, 10)
 ```
 

VERSION

@@ -1 +1 @@
-0.22.2
+0.22.3

build.rs

@@ -13,7 +13,7 @@ fn main() {
     println!("cargo:rerun-if-changed=rust/lib.rs");
     println!("cargo:rerun-if-changed=rust/lib.cpp");
     println!("cargo:rerun-if-changed=rust/lib.hpp");
-    println!("cargo:rerun-if-changed=include/index_punned_helpers.hpp");
-    println!("cargo:rerun-if-changed=include/index_punned_dense.hpp");
+    println!("cargo:rerun-if-changed=include/index_plugins.hpp");
+    println!("cargo:rerun-if-changed=include/index_dense.hpp");
     println!("cargo:rerun-if-changed=include/usearch/index.hpp");
 }