Versioned vector indexing & search for ProllyTree

.github/workflows/python.yml

@@ -67,7 +67,7 @@ jobs:
         PYO3_CROSS_PYTHON_VERSION: "3.11"
       with:
         target: ${{ matrix.target }}
-        args: --release --out dist --features python,git,sql,rocksdb_storage --interpreter python3.11
+        args: --release --out dist --features python,git,sql,rocksdb_storage,proximity,proximity_text --interpreter python3.11
         sccache: 'true'
         manylinux: ${{ matrix.manylinux }}
         rust-toolchain: stable

.github/workflows/python_examples.yml

@@ -0,0 +1,94 @@
+name: Python Examples
+
+# Builds the wheel with the full feature set (matching what PyPI ships) and
+# runs every example under `python/examples/`. No path filter — examples are
+# a user-facing contract and any PR that breaks them should fail this gate.
+
+on:
+  pull_request:
+    types: [opened, ready_for_review, synchronize]
+  push:
+    branches:
+      - main
+
+jobs:
+  examples:
+    name: Run Python examples
+    runs-on: ubuntu-latest
+
+    steps:
+    - uses: actions/checkout@v4
+
+    - name: Setup Python
+      uses: actions/setup-python@v5
+      with:
+        python-version: '3.11'
+
+    # Cache the MiniLM-L6-v2 model weights between runs so the text_index
+    # example doesn't pay the ~90 MB download on every PR. The cache key is
+    # stable (model id + revision); restore-keys lets future schema bumps fall
+    # back to an older cache if needed.
+    - name: Cache MiniLM embedder weights
+      uses: actions/cache@v4
+      with:
+        path: ~/.cache/prollytree/embedders
+        key: minilm-l6-v2-main-v1
+        restore-keys: |
+          minilm-l6-v2-
+
+    - name: Build wheel (with all features)
+      uses: PyO3/maturin-action@v1
+      env:
+        PYO3_CROSS_PYTHON_VERSION: "3.11"
+      with:
+        target: x86_64
+        args: --release --out dist --features python,git,sql,rocksdb_storage,proximity,proximity_text --interpreter python3.11
+        sccache: 'true'
+        manylinux: 2_28
+        rust-toolchain: stable
+        before-script-linux: |
+          # librocksdb-sys needs clang for bindgen + a C++ toolchain. Mirrors
+          # the install in python.yml.
+          if command -v yum >/dev/null 2>&1; then
+            yum install -y clang gcc gcc-c++ glibc-devel
+          elif command -v apt-get >/dev/null 2>&1; then
+            apt-get update && apt-get install -y --no-install-recommends clang g++ libc6-dev
+          fi
+
+    - name: Configure git identity
+      run: |
+        git config --global user.name  "CI Examples"
+        git config --global user.email "ci-examples@example.com"
+        git config --global init.defaultBranch main
+
+    - name: Install wheel
+      run: |
+        python -m pip install --upgrade pip
+        WHEEL=$(ls dist/prollytree-*-cp38-abi3-manylinux*_x86_64.whl | head -1)
+        echo "Installing $WHEEL"
+        python -m pip install --force-reinstall "$WHEEL"
+
+    - name: Verify install + feature flags
+      run: |
+        python -c "import prollytree as p; \
+          print('proximity_available     =', p.proximity_available); \
+          print('proximity_text_available=', p.proximity_text_available); \
+          print('sql_available           =', p.sql_available); \
+          print('git_available           =', p.git_available); \
+          print('namespaced_available    =', p.namespaced_available)"
+
+    # Run each example by file. Fail fast (`set -e`) so the first failure
+    # surfaces immediately in the PR check. text_index_example.py downloads
+    # MiniLM weights into ~/.cache/prollytree/embedders on first run; the
+    # cache step above persists them between PR runs.
+    - name: Run examples
+      run: |
+        set -e
+        # Run from /tmp so examples don't pollute the workspace and don't pick
+        # up the repo's pyproject.toml as their CWD (matches local convention).
+        cd /tmp
+        python "$GITHUB_WORKSPACE/python/examples/basic_usage.py"
+        python "$GITHUB_WORKSPACE/python/examples/sql_example.py"
+        python "$GITHUB_WORKSPACE/python/examples/merge_example.py"
+        python "$GITHUB_WORKSPACE/python/examples/namespaced_example.py"
+        python "$GITHUB_WORKSPACE/python/examples/text_index_example.py"

.github/workflows/release.yml

@@ -117,7 +117,7 @@ jobs:
           - os: ubuntu-latest
             target: x86_64
             manylinux: 2_28
-            features: python,git,sql,rocksdb_storage
+            features: python,git,sql,rocksdb_storage,proximity,proximity_text
           # Linux aarch64 — cross-compile via rust-cross/manylinux_2_28-cross.
           # rocksdb_storage is OFF here: librocksdb-sys's bindgen invokes clang
           # targeting aarch64-unknown-linux-gnu and can't resolve the aarch64
@@ -126,17 +126,17 @@ jobs:
           - os: ubuntu-latest
             target: aarch64
             manylinux: 2_28
-            features: python,git,sql
+            features: python,git,sql,proximity,proximity_text
           # Windows x86_64
           - os: windows-latest
             target: x64
             manylinux: false
-            features: python,git,sql,rocksdb_storage
+            features: python,git,sql,rocksdb_storage,proximity,proximity_text
           # macOS ARM64 (Apple Silicon) — arm64 wheel works on Intel Macs via Rosetta
           - os: macos-14
             target: aarch64
             manylinux: false
-            features: python,git,sql,rocksdb_storage
+            features: python,git,sql,rocksdb_storage,proximity,proximity_text
 
     steps:
       - uses: actions/checkout@v4

.github/workflows/rust_examples.yml

@@ -0,0 +1,60 @@
+name: Rust Examples
+
+# Runs every example under `examples/` to catch regressions that the unit/
+# integration test suite doesn't surface (CLI plumbing, demo dual-write
+# patterns, multi-backend storage paths). No path filter — examples are a
+# user-facing contract and any PR that breaks them should fail this gate.
+
+on:
+  pull_request:
+    types: [opened, ready_for_review, synchronize]
+  push:
+    branches:
+      - main
+
+jobs:
+  examples:
+    name: Run examples
+    runs-on: ubuntu-latest
+
+    steps:
+    - uses: actions/checkout@v4
+
+    - name: Check Cargo.lock is up to date
+      run: cargo check --locked
+
+    # librocksdb-sys (pulled in by the `rocksdb_storage` feature, which the
+    # `storage` example needs) requires clang for bindgen and a C++ toolchain
+    # to compile RocksDB itself.
+    - name: Install clang + C++ toolchain for librocksdb-sys
+      run: |
+        sudo apt-get update
+        sudo apt-get install -y --no-install-recommends clang g++ libc6-dev
+
+    # Examples that touch GitVersionedKvStore / GitNamespacedKvStore run
+    # `git init` against temp dirs and need a configured identity.
+    - name: Configure git identity
+      run: |
+        git config --global user.name  "CI Examples"
+        git config --global user.email "ci-examples@example.com"
+        git config --global init.defaultBranch main
+
+    # Build every example once with the union feature set; each example's
+    # `required-features` is satisfied as long as it is a subset.
+    - name: Build all examples
+      run: |
+        cargo build --locked --examples \
+          --features "git sql proximity rocksdb_storage" --verbose
+
+    # Run each example by name. Fail fast (`set -e`) so the first failure
+    # surfaces immediately in the PR check.
+    - name: Run examples
+      run: |
+        set -e
+        cargo run --locked --example proof
+        cargo run --locked --example sql --features sql
+        cargo run --locked --example versioning
+        cargo run --locked --example worktree --features "git sql"
+        cargo run --locked --example storage --features rocksdb_storage
+        cargo run --locked --example namespaced --features git
+        cargo run --locked --example text_index --features "git proximity"