pixie-py 0.1.0__tar.gz

pixie_py-0.1.0/.clang-format

@@ -0,0 +1 @@
+BasedOnStyle: Chromium

pixie_py-0.1.0/.github/workflows/release.yml

@@ -0,0 +1,75 @@
+name: release
+
+on:
+  push:
+    tags: ['v*']
+
+permissions:
+  contents: read
+
+jobs:
+  build:
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@v4
+      - uses: actions/setup-python@v5
+        with:
+          python-version: '3.12'
+      - name: Check release version
+        run: |
+          python - <<'PY'
+          import os, tomllib
+          with open("pyproject.toml", "rb") as f:
+              version = tomllib.load(f)["project"]["version"]
+          assert os.environ["GITHUB_REF_NAME"] == f"v{version}"
+          PY
+      - run: python -m pip install build twine
+      - run: python -m build --sdist
+      - run: python -m twine check --strict dist/*
+      - uses: actions/upload-artifact@v4
+        with:
+          name: release-distributions
+          path: dist/*.tar.gz
+          if-no-files-found: error
+
+  test:
+    needs: build
+    runs-on: ubuntu-latest
+    strategy:
+      matrix:
+        python: ['3.10', '3.12']
+    steps:
+      - uses: actions/checkout@v4
+      - uses: actions/setup-python@v5
+        with:
+          python-version: ${{ matrix.python }}
+      - uses: actions/download-artifact@v4
+        with:
+          name: release-distributions
+          path: dist
+      - name: Install the release archive
+        run: python -m pip install dist/*.tar.gz pytest -Cbuild.tool-args=-j2
+      - run: python -m pytest
+      - run: python examples/louds.py
+      - name: Check core revision
+        run: |
+          python - <<'PY'
+          import pathlib, re, pixie
+          expected = re.search(r'set\(PIXIE_REVISION ([0-9a-f]+)\)', pathlib.Path('CMakeLists.txt').read_text()).group(1)
+          assert pixie.core_revision == expected, pixie.core_revision
+          PY
+
+  publish:
+    needs: [build, test]
+    runs-on: ubuntu-latest
+    environment:
+      name: pypi
+      url: https://pypi.org/project/pixie-py/
+    permissions:
+      id-token: write
+    steps:
+      - uses: actions/download-artifact@v4
+        with:
+          name: release-distributions
+          path: dist
+      - uses: pypa/gh-action-pypi-publish@release/v1

pixie_py-0.1.0/.github/workflows/tests.yml

@@ -0,0 +1,16 @@
+name: tests
+on: [push, pull_request]
+jobs:
+  python:
+    runs-on: ubuntu-latest
+    strategy:
+      matrix:
+        python: ['3.10', '3.12']
+    steps:
+      - uses: actions/checkout@v4
+      - uses: actions/setup-python@v5
+        with:
+          python-version: ${{ matrix.python }}
+      - run: python -m pip install '.[test]'
+      - run: python -m pytest
+      - run: python examples/louds.py

pixie_py-0.1.0/.gitignore

@@ -0,0 +1,6 @@
+.venv/
+build/
+dist/
+__pycache__/
+.pytest_cache/
+*.egg-info/

pixie_py-0.1.0/AGENTS.md

@@ -0,0 +1,37 @@
+# pixie-py
+
+- Keep Python bindings and ownership adapters here; consume only public Pixie headers.
+- Keep the Pixie dependency pinned in CMakeLists.txt.
+- Do not expose raw rank/select or RmM components with borrowed source lifetimes.
+- Python succinct trees are immutable; PointerTree is an append-only mutable builder with stable handles. Views and node handles retain their backing owner.
+- Validate inputs and handle provenance before entering unchecked C++ operations.
+- Keep the GIL unless a change explicitly establishes lifetime and concurrency safety.
+- Test navigation against an independent topology oracle, invalid inputs and owner deletion.
+- Use the published core pin by default. For local development, set the shell variable `PIXIE_SOURCE_DIR` to an absolute checkout path and reinstall. Keep local paths out of repository configuration.
+- Run `pytest` and `python examples/louds.py` after building.
+- RMQ persistence must use the CLI's native format and implementation names.
+  Artifacts omit source values; Python must retain an owned copy of supplied values.
+  Default to full validation, reject trailing bytes, and test CLI-to-Python and
+  Python-to-CLI loading with `PIXIE_CLI=/path/to/pixie pytest`.
+
+## Python docstrings
+
+- Use NumPy-style docstrings for every public Python class, constructor, method,
+  function and property, including APIs defined through nanobind.
+- Start with a short sentence explaining what the operation does. A generated
+  signature is not a description.
+- Include a `Parameters` section for every public argument except `self`/`cls`.
+  State its type, purpose, units, defaults, indexing and valid range where relevant.
+  Document each constructor overload and summarize construction in the class docstring.
+- Use `Returns` to explain returned values, especially ownership, snapshots and
+  possible `None` results. Use `Raises` for meaningful validation failures and
+  `Notes` only for constraints such as lifetime or concurrency.
+- Present bound classes under their public `pixie` module. Generated signatures,
+  type representations and argument errors must not expose `pixie._core`.
+  Short names such as `Node` in docstrings refer to the public `pixie.Node` type.
+- Keep descriptions concise; omit empty sections. Properties need a short value
+  description and type, not a parameter section. Keep type stubs consistent.
+- Put native API docstrings in the nanobind registrations so `help()` and notebook
+  tooltips display them. Check the installed runtime docstrings after rebuilding;
+  merely checking that `__doc__` exists does not establish useful documentation.
+- Reference convention: https://numpydoc.readthedocs.io/en/latest/format.html

pixie_py-0.1.0/CMakeLists.txt

@@ -0,0 +1,38 @@
+cmake_minimum_required(VERSION 3.24)
+project(pixie_py LANGUAGES CXX)
+
+find_package(Python 3.10 REQUIRED COMPONENTS Interpreter Development.Module)
+find_package(nanobind CONFIG REQUIRED)
+include(FetchContent)
+
+set(PIXIE_SOURCE_DIR "$ENV{PIXIE_SOURCE_DIR}" CACHE PATH "Optional local Pixie checkout")
+set(PIXIE_REVISION be9f30dcc1b823fadf17c3fc67ee7cd425e98ee8)
+# Consume public headers without importing the core's native CPU flags or tests.
+FetchContent_Declare(pixie
+  GIT_REPOSITORY https://github.com/Malkovsky/pixie.git
+  GIT_TAG ${PIXIE_REVISION}
+  GIT_SUBMODULES ""
+  SOURCE_SUBDIR pixie-py-headers-only)
+if(PIXIE_SOURCE_DIR)
+  set(FETCHCONTENT_SOURCE_DIR_PIXIE "${PIXIE_SOURCE_DIR}")
+endif()
+FetchContent_MakeAvailable(pixie)
+if(NOT EXISTS "${pixie_SOURCE_DIR}/include/pixie/tree/decoded_tree.h")
+  message(FATAL_ERROR "The selected Pixie checkout is missing include/pixie/tree/decoded_tree.h. Use a checkout with the traversal API.")
+endif()
+
+nanobind_add_module(_core NB_STATIC src/bindings.cpp src/rmq_bindings.cpp)
+target_compile_features(_core PRIVATE cxx_std_20)
+target_include_directories(_core PRIVATE ${pixie_SOURCE_DIR}/include)
+set(PIXIE_BUILD_REVISION "${PIXIE_REVISION}")
+if(PIXIE_SOURCE_DIR)
+  execute_process(COMMAND git rev-parse HEAD WORKING_DIRECTORY "${pixie_SOURCE_DIR}"
+    OUTPUT_VARIABLE PIXIE_BUILD_REVISION OUTPUT_STRIP_TRAILING_WHITESPACE)
+  execute_process(COMMAND git diff --quiet -- include WORKING_DIRECTORY "${pixie_SOURCE_DIR}"
+    RESULT_VARIABLE PIXIE_CORE_DIRTY)
+  if(PIXIE_CORE_DIRTY)
+    string(APPEND PIXIE_BUILD_REVISION "-dirty")
+  endif()
+endif()
+target_compile_definitions(_core PRIVATE PIXIE_PY_CORE_REVISION="${PIXIE_BUILD_REVISION}")
+install(TARGETS _core LIBRARY DESTINATION pixie)

pixie_py-0.1.0/DESIGN.md

@@ -0,0 +1,285 @@
+# Python tree experimentation API
+
+This document records the agreed design, not a claim that the APIs below are
+already implemented. The current baseline exposes owning LOUDS, BP, DFUDS and
+pointer trees, retained complete-tree views and node handles. Core C++ changes are allowed where needed
+for a clean public API; Python bindings should consume those public contracts.
+
+The purpose is interactive exploration and experimentation. Individual Python
+calls to navigation primitives need not approach native-call performance.
+Higher-level operations can execute primarily in C++. Preserve native traversal
+performance where possible; a measured 10–15% overhead can be acceptable when
+it buys a better interface, rather than being a target overhead to spend.
+
+## Complete trees and ownership
+
+`LoudsTree`, `DfudsTree` and `BPTree` are immutable top-level objects sharing a
+navigation interface corresponding to C++ `TreeBase`. Encoding, rank/select
+and RmM dependencies remain encapsulated by these owning objects. Do not expose
+standalone borrowed rank/select or RmM components in Python.
+
+Python node handles, tree views, decoded results and other borrowers retain the
+owners they require. C++ borrowers may continue using const references or
+pointers with explicit lifetime requirements. Python validates handle provenance
+before calling unchecked core operations. A handle includes representation-
+specific information; numeric IDs are not assumed to survive tree conversion.
+
+`TreeView` denotes a complete-tree storage view, distinct from a node view or
+an incomplete decoded tree. The view retains its backing owner in Python.
+
+### Text and visualization
+
+`PointerTree()` is a mutable builder initially containing root zero.
+`add_node(parent_id)` appends a last child and returns its insertion-order ID.
+Stable pointer allocations keep existing handles valid; handles retain the
+mutable backing owner. Mutation is serialized by the GIL. This binding-side
+adapter implements the native ordered-tree source contract; the immutable C++
+reference implementation is unchanged.
+
+Succinct string constructors validate and pack their own native encoding
+without converting through a pointer tree. BP uses balanced parentheses;
+DFUDS uses unary degrees in preorder without a sentinel; LOUDS uses unary
+degrees in BFS order and accepts its optional display prefix. PointerTree text
+remains BP. All constructors also accept complete trees/views. Copies are
+independent; pointer copies start with preorder IDs, and succinct conversions
+assign IDs for their encoding. Conversions snapshot mutable topology. Public
+list constructors are removed. `from_bp` remains an explicit parsing helper
+accepting `()` or binary `10` (one means open), rejecting malformed or multiple
+trees. Rejected strings include the reason, a format hint and a valid example.
+Construction and conversion are iterative.
+
+All complete trees now provide `to_louds()`, `to_bp()` and `to_dfuds()`.
+The succinct encodings share Python navigation, index and context classes;
+dispatch selects the native encoding without converting it for traversal.
+Tree-like objects expose `to_dot()` and `visualize()` methods, in addition to
+the module-level functions. Public Python classes, methods and properties carry
+short human-readable docstrings.
+
+`parentheses()` and `str(tree)` render the stored encoding with `1` as `(`
+and `0` as `)`. `start` and `count` select logical bits directly, excluding
+allocation padding. BP contains `2*n` bits; LOUDS and DFUDS contain `2*n - 1`,
+with no stored sentinel. Both unary formats use one `1` per child followed by
+`0`, in BFS order for LOUDS and preorder for DFUDS. Nonempty LOUDS output
+starting at stored bit zero adds a display-only `(`. Its full display is balanced;
+`start` and `count` exclude this prefix. DFUDS adds no prefix. Rendering takes
+time proportional to output length and does not traverse or convert the tree.
+Use `to_bp().parentheses()` for canonical BP topology. Pointer trees have no
+stored bit encoding and return BP topology. Full string output round-trips
+through the same class constructor; `from_bp` always parses BP.
+`parentheses(verbose=True, max_nodes=25)` returns only a text topology diagram
+with source node IDs. BP uses matching-parenthesis brackets and IDs at opening
+parentheses; DFUDS brackets subtree spans and labels node-record closing
+parentheses. Columns widen with ID length. The first max_nodes preorder nodes
+are annotated; omitted encoding regions become `...`. Boundary detection scans
+the complete encoding. LOUDS and pointer trees retain an indented diagram with
+`...` children for truncated branches. Encoding slices do not restrict the
+diagram, which starts at the root.
+
+`repr(tree)` returns encoding text for complete trees and a short description
+for partial trees. Notebook MIME display renders an SVG through optional Graphviz
+and falls back to text when the package or executable is missing or rendering
+fails. Automatic display, `to_dot()` and `visualize()` default to 50 nodes;
+explicit `max_nodes` arguments control truncation in the latter two methods.
+
+## Incomplete-tree interface
+
+Introduce a common interface for represented portions of a complete tree.
+Every incomplete tree is bound to an original complete tree through a const
+reference/pointer and exposes `tree_base()`. Python retains that original owner.
+
+The common inspection operations are:
+
+- `contains(node)`: a node record is represented in this object.
+- `children_complete(node)`: for a represented node, all immediate children are
+  represented and its child list is complete. This is true for a represented
+  leaf. It does not imply that descendants have complete child lists.
+- `tree_base()`: access the original complete tree explicitly.
+
+Use valid node handles initially. An arbitrary numeric-ID-to-handle API is a
+separate future extension, not the same operation as decoded-cache lookup.
+
+Inspection and navigation through the incomplete-tree interface describe only
+represented data. They must not silently fetch additional data or fall back to
+the complete tree. An incomplete child list must not masquerade as a complete
+empty list. Boundary navigation distinguishes information outside the result
+from a relation that does not exist. For example, a decoded subtree root can
+have a parent in `tree_base()` that is absent from the decoded result. Exact
+sentinel/result types will be selected with the concrete implementation.
+
+Full trees can satisfy the same inspection contract trivially. Initial partial
+representations include the root navigation index and decoded microtrees.
+DFS events remain in scope as scan results; a public navigation adapter for
+arbitrary event blocks is deferred because an event slice is not necessarily a
+navigable partial tree. Do not force every event result into a topology interface
+that it cannot meaningfully satisfy.
+
+## Root navigation index
+
+`RootNavigationIndex` is a separate immutable object associated with one tree.
+Different indexes may use different budgets for the same tree. It is not owned
+by the traversal context. C++ contexts hold an optional const pointer/reference;
+Python contexts retain the index and tree owners. Validate that a supplied index
+belongs to the supplied tree. Without one, traversal uses the tree directly.
+
+Expose its represented topology through the incomplete-tree contract. Keep this
+separate from its existing cursor behavior, which can fall back to the original
+tree and therefore return complete answers beyond the cached region.
+
+The current index follows contiguous child-entry ranges from a root cursor; it
+has no arbitrary node-to-entry directory. Implementing `contains` and
+`children_complete` requires inspection support, but does not by itself commit
+to a general decoded-cache directory or public context `lookup` implementation.
+Choose and document the inspection cost rather than assuming node IDs equal
+entry offsets.
+
+The index has its own budget and serialization. Context accounting covers its
+own retained state and scratch; aggregate reports must identify shared index
+memory separately to avoid double counting. Replace the context constructor
+that takes ownership of a restored index with borrowing an externally owned one.
+
+## Traversal context and scanning
+
+Retain the existing decoding algorithms and configuration machinery where
+possible. Extend `TreeTraversalContext` to own useful reusable mutable decoding
+state and larger traversal buffers instead of leaving all of them in temporary
+walkers. It borrows the complete tree and optional immutable root index.
+
+Implement these operations now; `scan` replaces the earlier name `look_ahead`:
+
+```python
+context.scan(node, max_bits=...)
+context.fetch(node, max_bits=...)
+```
+
+Both use the encoding-specific algorithm and the same scan-limit semantics.
+`scan` returns extracted information without retaining the result in the
+context. It may use and modify reusable scratch, so it is logically read-only
+with respect to retained results, not necessarily a C++ const operation.
+`fetch` returns extracted information and retains it in the context for reuse.
+Neither method implicitly advances a traversal cursor. Retention does not yet
+promise automatic cache hits or deduplication by node.
+
+`max_bits` bounds the total logical encoding bits examined across the regions
+scanned by the operation. Multiple regions are allowed; the algorithm may
+navigate to another region instead of scanning an irrelevant gap. Word/SIMD
+loads can touch surrounding valid storage without extending the logical scan.
+This parameter is not a node count, memory budget or strict CPU-work limit.
+Document separate navigation/backend work and logical scan accounting when
+implementing the algorithms; physical reads are not all charged as scan bits.
+
+Extract whatever useful information the bounded scan supports. It need not
+produce a complete subtree. Budget exhaustion returns a useful partial result
+(or an explicit empty/miss result when nothing useful fits), not an error.
+Invalid handles and invalid parameter values remain errors.
+
+The concrete return types will follow the decoder capabilities: decoded
+microtrees, node runs, DFS events or other useful encoding-specific results.
+Their kind and completeness must be explicit. Separate entry points or typed
+results are acceptable if they avoid unnecessary conversion or materialization.
+Do not impose a universal decoded representation at a substantial performance
+cost merely to make all encodings look identical.
+
+### LOUDS
+
+LOUDS currently decodes word-sized delimiter batches with per-level cursors,
+plus child boundaries and traversal frames. It does not currently retain a
+subtree block equivalent to the BP/DFUDS microtree decoder.
+
+A LOUDS result can contain consecutive nodes in a level and their children,
+potentially spanning several short levels. Child encodings may occupy a separate
+region from their parents; multiple-region scanning is explicitly allowed under
+the same total bit limit. Knowing a parent's degree alone does not establish
+`children_complete(parent)`: the child records must also be represented.
+Boundary nodes can be present with incomplete children.
+
+### Results, views and iterators
+
+Returned Python results remain valid after subsequent context operations and
+retain their original tree. Do not expose a silently overwritten scratch buffer
+as a persistent result. Retained decoded blocks may be immutable while context
+scratch and publication state remain mutable. The exact ownership mechanism is
+an implementation choice; avoid unnecessary copies.
+
+Allow multiple live views/iterators associated with a context and sequentially
+interleaved use. One operation must not invalidate data still used by another
+view or iterator. A node view supports arbitrary navigation from a node; a DFS
+iterator tracks event order and traversal progress. These are distinct roles.
+
+Keep small cursor state, node handles, event positions and navigation
+memoization local to the view/iterator where useful. Larger decoded objects and
+reusable buffers belong in the context. A cache-line-sized local cursor is a
+performance goal, not a strict size contract: current microtrees themselves are
+larger. Overflow traversal state may be context-allocated, but must remain
+separate for each active iterator.
+
+Ordinary `NodeView` computes parent and children lazily on access. This differs
+from inspecting only represented data through an incomplete-tree object.
+`TreeNavigationView` is a C++ concept describing the node-view interface, not an
+additional runtime object or a synonym for traversal context. Defer Python
+exposure of `TreePathView`.
+
+### Concurrency
+
+For this iteration, allow only one active operation per context, without locks.
+Sequential handoff between threads is allowed; no permanent thread affinity is
+required. Separate contexts may operate on a shared immutable tree/index when
+the execution layer permits it. Multiple live views are allowed, but operations
+using the same mutable context must not overlap.
+
+Shared mutable session caching and internally parallel search operations are
+deferred until beam search or another concrete workload needs them. Do not add
+locking overhead now. If synchronization is introduced later, compare unguarded,
+mutex and read/write-lock paths on cached reads, small decodes and sustained
+traversal. Scratch-mutating scans cannot use a shared lock merely because they
+are logically read-only.
+
+The Python baseline retains the GIL. Native thread-safety contracts alone do not
+promise Python-thread parallelism; GIL release requires an explicit review of
+the operations and result lifetimes.
+
+## Persistence
+
+Serializable public objects expose `save` and `load`, with a load mode selecting
+RAM or mmap backing, using Pixie's native formats. Do not introduce a
+Python-specific wrapper format. Mmap mode does not promise that all metadata is
+zero-copy; some backends restore owned metadata. Python objects retain mapped
+resources for as long as any dependent view/result needs them.
+
+Loading an index requires the associated complete tree. Use full validation by
+default at the Python boundary, with explicit validation options where supported.
+Structural validation and tree-size agreement do not establish artifact identity.
+
+Do not promise serialization of mutable scratch, active iterators or in-progress
+traversal sessions. Durable context configuration and externally owned indexes
+can be restored separately. Resumable search checkpoints and identity binding
+are future application-level work. Persistence is part of the overall design;
+it can be delivered separately from the current traversal-interface changes.
+
+## Implementation scope and verification
+
+The current iteration establishes the incomplete-tree contracts, externalizes
+the root index, implements bounded multi-region `scan`/`fetch`, adds the relevant
+Python objects with lazy node views, and provides parenthesis output and Python
+visualization. Adapt the C++ core where needed and keep the bindings on public
+interfaces. The common tree interface targets LOUDS, BP and DFUDS; adding all
+Python encodings and persistence need not be bundled into the first traversal
+change.
+
+Deferred items include public context `lookup(node, ...)`, a general decoded-cache
+lookup/deduplication policy, arbitrary ID-to-handle resolution, event-block
+navigation, `TreePathView` bindings, concurrent context sharing and beam-search
+parallelization. Document future lookup semantics as checking retained data
+without decoding, but do not add nonfunctional callable stubs.
+
+Choose concrete result types and buffer layouts after inspecting and measuring
+the decoding paths. A hard context memory limit must have explicit behavior when
+live results prevent reclamation; do not evict live data or silently exceed the
+limit. Caller-retained result memory and reusable context capacity must have
+clear, separate accounting.
+
+Verify topology and completeness against independent tree oracles, boundary
+navigation, scan-limit accounting, multiple-region LOUDS scans, retained result
+lifetimes, and interleaved views/iterators. Benchmark native decoding and repeated
+result consumption against existing paths; Python scalar-call overhead is a
+separate measurement. Record any meaningful performance tradeoff rather than
+assuming the interface fits the accepted 10–15% tolerance.

pixie_py-0.1.0/LICENSE

@@ -0,0 +1,201 @@
+                                 Apache License
+                           Version 2.0, January 2004
+                        http://www.apache.org/licenses/
+
+   TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION
+
+   1. Definitions.
+
+      "License" shall mean the terms and conditions for use, reproduction,
+      and distribution as defined by Sections 1 through 9 of this document.
+
+      "Licensor" shall mean the copyright owner or entity authorized by
+      the copyright owner that is granting the License.
+
+      "Legal Entity" shall mean the union of the acting entity and all
+      other entities that control, are controlled by, or are under common
+      control with that entity. For the purposes of this definition,
+      "control" means (i) the power, direct or indirect, to cause the
+      direction or management of such entity, whether by contract or
+      otherwise, or (ii) ownership of fifty percent (50%) or more of the
+      outstanding shares, or (iii) beneficial ownership of such entity.
+
+      "You" (or "Your") shall mean an individual or Legal Entity
+      exercising permissions granted by this License.
+
+      "Source" form shall mean the preferred form for making modifications,
+      including but not limited to software source code, documentation
+      source, and configuration files.
+
+      "Object" form shall mean any form resulting from mechanical
+      transformation or translation of a Source form, including but
+      not limited to compiled object code, generated documentation,
+      and conversions to other media types.
+
+      "Work" shall mean the work of authorship, whether in Source or
+      Object form, made available under the License, as indicated by a
+      copyright notice that is included in or attached to the work
+      (an example is provided in the Appendix below).
+
+      "Derivative Works" shall mean any work, whether in Source or Object
+      form, that is based on (or derived from) the Work and for which the
+      editorial revisions, annotations, elaborations, or other modifications
+      represent, as a whole, an original work of authorship. For the purposes
+      of this License, Derivative Works shall not include works that remain
+      separable from, or merely link (or bind by name) to the interfaces of,
+      the Work and Derivative Works thereof.
+
+      "Contribution" shall mean any work of authorship, including
+      the original version of the Work and any modifications or additions
+      to that Work or Derivative Works thereof, that is intentionally
+      submitted to Licensor for inclusion in the Work by the copyright owner
+      or by an individual or Legal Entity authorized to submit on behalf of
+      the copyright owner. For the purposes of this definition, "submitted"
+      means any form of electronic, verbal, or written communication sent
+      to the Licensor or its representatives, including but not limited to
+      communication on electronic mailing lists, source code control systems,
+      and issue tracking systems that are managed by, or on behalf of, the
+      Licensor for the purpose of discussing and improving the Work, but
+      excluding communication that is conspicuously marked or otherwise
+      designated in writing by the copyright owner as "Not a Contribution."
+
+      "Contributor" shall mean Licensor and any individual or Legal Entity
+      on behalf of whom a Contribution has been received by Licensor and
+      subsequently incorporated within the Work.
+
+   2. Grant of Copyright License. Subject to the terms and conditions of
+      this License, each Contributor hereby grants to You a perpetual,
+      worldwide, non-exclusive, no-charge, royalty-free, irrevocable
+      copyright license to reproduce, prepare Derivative Works of,
+      publicly display, publicly perform, sublicense, and distribute the
+      Work and such Derivative Works in Source or Object form.
+
+   3. Grant of Patent License. Subject to the terms and conditions of
+      this License, each Contributor hereby grants to You a perpetual,
+      worldwide, non-exclusive, no-charge, royalty-free, irrevocable
+      (except as stated in this section) patent license to make, have made,
+      use, offer to sell, sell, import, and otherwise transfer the Work,
+      where such license applies only to those patent claims licensable
+      by such Contributor that are necessarily infringed by their
+      Contribution(s) alone or by combination of their Contribution(s)
+      with the Work to which such Contribution(s) was submitted. If You
+      institute patent litigation against any entity (including a
+      cross-claim or counterclaim in a lawsuit) alleging that the Work
+      or a Contribution incorporated within the Work constitutes direct
+      or contributory patent infringement, then any patent licenses
+      granted to You under this License for that Work shall terminate
+      as of the date such litigation is filed.
+
+   4. Redistribution. You may reproduce and distribute copies of the
+      Work or Derivative Works thereof in any medium, with or without
+      modifications, and in Source or Object form, provided that You
+      meet the following conditions:
+
+      (a) You must give any other recipients of the Work or
+          Derivative Works a copy of this License; and
+
+      (b) You must cause any modified files to carry prominent notices
+          stating that You changed the files; and
+
+      (c) You must retain, in the Source form of any Derivative Works
+          that You distribute, all copyright, patent, trademark, and
+          attribution notices from the Source form of the Work,
+          excluding those notices that do not pertain to any part of
+          the Derivative Works; and
+
+      (d) If the Work includes a "NOTICE" text file as part of its
+          distribution, then any Derivative Works that You distribute must
+          include a readable copy of the attribution notices contained
+          within such NOTICE file, excluding those notices that do not
+          pertain to any part of the Derivative Works, in at least one
+          of the following places: within a NOTICE text file distributed
+          as part of the Derivative Works; within the Source form or
+          documentation, if provided along with the Derivative Works; or,
+          within a display generated by the Derivative Works, if and
+          wherever such third-party notices normally appear. The contents
+          of the NOTICE file are for informational purposes only and
+          do not modify the License. You may add Your own attribution
+          notices within Derivative Works that You distribute, alongside
+          or as an addendum to the NOTICE text from the Work, provided
+          that such additional attribution notices cannot be construed
+          as modifying the License.
+
+      You may add Your own copyright statement to Your modifications and
+      may provide additional or different license terms and conditions
+      for use, reproduction, or distribution of Your modifications, or
+      for any such Derivative Works as a whole, provided Your use,
+      reproduction, and distribution of the Work otherwise complies with
+      the conditions stated in this License.
+
+   5. Submission of Contributions. Unless You explicitly state otherwise,
+      any Contribution intentionally submitted for inclusion in the Work
+      by You to the Licensor shall be under the terms and conditions of
+      this License, without any additional terms or conditions.
+      Notwithstanding the above, nothing herein shall supersede or modify
+      the terms of any separate license agreement you may have executed
+      with Licensor regarding such Contributions.
+
+   6. Trademarks. This License does not grant permission to use the trade
+      names, trademarks, service marks, or product names of the Licensor,
+      except as required for reasonable and customary use in describing the
+      origin of the Work and reproducing the content of the NOTICE file.
+
+   7. Disclaimer of Warranty. Unless required by applicable law or
+      agreed to in writing, Licensor provides the Work (and each
+      Contributor provides its Contributions) on an "AS IS" BASIS,
+      WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or
+      implied, including, without limitation, any warranties or conditions
+      of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A
+      PARTICULAR PURPOSE. You are solely responsible for determining the
+      appropriateness of using or redistributing the Work and assume any
+      risks associated with Your exercise of permissions under this License.
+
+   8. Limitation of Liability. In no event and under no legal theory,
+      whether in tort (including negligence), contract, or otherwise,
+      unless required by applicable law (such as deliberate and grossly
+      negligent acts) or agreed to in writing, shall any Contributor be
+      liable to You for damages, including any direct, indirect, special,
+      incidental, or consequential damages of any character arising as a
+      result of this License or out of the use or inability to use the
+      Work (including but not limited to damages for loss of goodwill,
+      work stoppage, computer failure or malfunction, or any and all
+      other commercial damages or losses), even if such Contributor
+      has been advised of the possibility of such damages.
+
+   9. Accepting Warranty or Additional Liability. While redistributing
+      the Work or Derivative Works thereof, You may choose to offer,
+      and charge a fee for, acceptance of support, warranty, indemnity,
+      or other liability obligations and/or rights consistent with this
+      License. However, in accepting such obligations, You may act only
+      on Your own behalf and on Your sole responsibility, not on behalf
+      of any other Contributor, and only if You agree to indemnify,
+      defend, and hold each Contributor harmless for any liability
+      incurred by, or claims asserted against, such Contributor by reason
+      of your accepting any such warranty or additional liability.
+
+   END OF TERMS AND CONDITIONS
+
+   APPENDIX: How to apply the Apache License to your work.
+
+      To apply the Apache License to your work, attach the following
+      boilerplate notice, with the fields enclosed by brackets "[]"
+      replaced with your own identifying information. (Don't include
+      the brackets!)  The text should be enclosed in the appropriate
+      comment syntax for the file format. We also recommend that a
+      file or class name and description of purpose be included on the
+      same "printed page" as the copyright notice for easier
+      identification within third-party archives.
+
+   Copyright 2026 Pixie contributors
+
+   Licensed under the Apache License, Version 2.0 (the "License");
+   you may not use this file except in compliance with the License.
+   You may obtain a copy of the License at
+
+       http://www.apache.org/licenses/LICENSE-2.0
+
+   Unless required by applicable law or agreed to in writing, software
+   distributed under the License is distributed on an "AS IS" BASIS,
+   WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+   See the License for the specific language governing permissions and
+   limitations under the License.