C++ beam search and edit distance

.github/workflows/build_wheels.yml

@@ -0,0 +1,38 @@
+name: Build wheels
+
+on:
+  workflow_dispatch:
+  push:
+    tags:
+      - "v*"
+
+jobs:
+  build_wheels:
+    name: Build wheels on ${{ matrix.os }}
+    runs-on: ${{ matrix.os }}
+    strategy:
+      matrix:
+        os: [ubuntu-latest, macos-latest, windows-latest]
+
+    env:
+      CIBW_ARCHS_LINUX: "x86_64"
+
+    steps:
+      - uses: actions/checkout@v4
+
+      - name: Install Python
+        uses: actions/setup-python@v5
+        with:
+          python-version: "3.12"
+
+      - name: Install cibuildwheel
+        run: pip install cibuildwheel==2.22.0
+
+      - name: Build wheels
+        run: cibuildwheel --output-dir wheelhouse
+
+      - name: Upload wheels as artifacts
+        uses: actions/upload-artifact@v4
+        with:
+          name: wheels-${{ matrix.os }}
+          path: wheelhouse/*.whl

.github/workflows/tests_and_coverage.yml

@@ -26,10 +26,11 @@ jobs:
         run: |
           python -m pip install --upgrade pip
           pip install poetry
-          poetry install --with test
+          poetry install --with test,dev
+          poetry run pip install .
 
       - name: Run tests with coverage
-        run: poetry run pytest
+        run: poetry run pytest --cov=error_align --cov-report=xml
 
       - name: Upload coverage to Codecov
         if: matrix.python-version == '3.12'

.pre-commit-config.yaml

@@ -4,6 +4,7 @@ repos:
     rev: v6.0.0
     hooks:
     -   id: check-added-large-files  # Check for large files.
+        exclude: ^src/error_align/baselines/power/cmudict.rep.json$
     -   id: check-ast  # Check Python file syntax using the ast module.
     -   id: check-json  # Check JSON file syntax.
     -   id: check-merge-conflict  # Check for merge conflict strings.
@@ -18,6 +19,7 @@ repos:
     hooks:
       - id: ruff-format
       - id: ruff-check
+        args: ["--fix"]
 -   repo: https://github.com/asottile/yesqa
     rev: v1.5.0
     hooks:

CMakeLists.txt

@@ -0,0 +1,20 @@
+cmake_minimum_required(VERSION 3.18)
+project(error_align LANGUAGES CXX)
+
+set(CMAKE_CXX_STANDARD 17)
+set(CMAKE_CXX_STANDARD_REQUIRED ON)
+
+find_package(pybind11 CONFIG REQUIRED)
+
+# Build the module
+pybind11_add_module(_cpp_edit_distance cpp/edit_distance.cpp)
+pybind11_add_module(_cpp_beam_search cpp/beam_search.cpp)
+
+# Install it under the Python package directory (src/error_align)
+install(
+    TARGETS
+        _cpp_edit_distance
+        _cpp_beam_search
+    LIBRARY DESTINATION error_align
+    RUNTIME DESTINATION error_align
+)

README.md

@@ -1,5 +1,5 @@
 <p align="center">
-  <img src="https://raw.githubusercontent.com/corticph/error-align/refs/heads/main/.github/assets/logo.svg" alt="ErrorAlign Logo" width="85%"/>
+  <img src=".github/assets/logo.svg" alt="ErrorAlign Logo" width="100%"/>
 </p>
 
 <p align="center">
@@ -14,9 +14,11 @@
 
 **Text-to-text alignment algorithm for speech recognition error analysis.** ErrorAlign helps you dig deeper into your speech recognition projects by accurately aligning each word in a reference transcript with the model-generated transcript. Unlike traditional methods, such as Levenshtein-based alignment, it is not restricted to simple one-to-one alignment, but can map a single reference word to multiple words or subwords in the model output. This enables quick and reliable identification of error patterns in rare words, names, or domain-specific terms that matter most for your application.
 
+→ **Update [2025-12-10]:** As of version `0.1.0b5`, `error-align` will include a word-level pass to efficiently identify unambiguous matches, along with C++ extensions to accelerate beam search and backtrace construction. The combined speedup is ~15× over the pure-Python implementation ⚡
 
-__Contents__ | [Installation](#installation) | [Quickstart](#quickstart) | [Work-in-Progress](#wip) | [Citation and Research](#citation) |
+[//]: <> (https://raw.githubusercontent.com/corticph/error-align/refs/heads/main/.github/assets/logo_gpt.svg)
 
+__Contents__ | [Installation](#installation) | [Quickstart](#quickstart) | [Citation and Research](#citation) |
 
 
 
@@ -51,15 +53,6 @@ Alignment(SUBSTITUTE: "noting" -> "nothing"),
 Alignment(INSERT: "period")
 ```
 
-<a name="wip">
-
-## Work-in-Progress
-
-- Optimization for longform text.
-- Efficient word-level first-pass.
-- C++ version with Python bindings.
-
-
 
 <a name="citation">
 

codecov.yml

@@ -1,2 +1,11 @@
 ignore:
-  - "src/error_align/baselines/*"
+  - "src/error_align/baselines/*"
+
+coverage:
+  status:
+    project:
+      default:
+        informational: true
+    patch:
+      default:
+        informational: true