This release is a pre-release and may not be stable for production use.
tigrbl-kernel
Kernel orchestration for composing Tigrbl runtime plans, bindings, operation dispatch, and optimized ASGI execution.
What is tigrbl-kernel?
Kernel orchestration for composing Tigrbl runtime plans, bindings, operation dispatch, and optimized ASGI execution.
Why use tigrbl-kernel?
Use it when you need this foundational Tigrbl layer directly as a small, focused dependency.
When should I install tigrbl-kernel?
Install it for extension packages, package-local tests, or internals that need this boundary without the whole facade.
Who is tigrbl-kernel for?
Framework maintainers, extension authors, and advanced users composing Tigrbl from split packages.
Where does tigrbl-kernel fit?
tigrbl-kernel lives at pkgs/45_kernel/tigrbl_kernel and serves a focused layer in the split Tigrbl framework.
How does tigrbl-kernel work?
It owns a narrow layer in the split workspace and is consumed by higher-level packages through explicit dependencies.
Install
uv add tigrbl-kernel
pip install tigrbl-kernel
Surface Coverage
| Surface | Value |
|---|---|
| PyPI package | tigrbl-kernel |
| Repository path | pkgs/45_kernel/tigrbl_kernel |
| Python import root | tigrbl_kernel |
| Console scripts | none declared |
| Entry points | none declared |
| Optional extras | none declared |
| Legal files | LICENSE, NOTICE |
| Supported Python | 3.10, 3.11, 3.12, 3.13, 3.14 |
What It Owns
tigrbl-kernel owns the foundational framework package boundary. It should be installed when you need this package's focused responsibility without assuming every other Tigrbl workspace package is present.
Implementation orientation:
tigrbl_kernel: _build, _compile, atoms, cache, callbacks, contract_classification, core, eventkey, eventkey_hooks, events, helpers, hook_types
Package catalog:
tigrbl_kernel/core.py,tigrbl_kernel/_build.py,tigrbl_kernel/_compile.py,tigrbl_kernel/models.py,tigrbl_kernel/types.py, andtigrbl_kernel/payload.py: kernel objects, operation views, packed plans, build/compile helpers, and payload contracts.tigrbl_kernel/ordering.py,tigrbl_kernel/labels.py,tigrbl_kernel/hook_types.py,tigrbl_kernel/callbacks.py, andtigrbl_kernel/eventkey_hooks.py: hook ordering, diagnostic labels, callback shape, and event-key-aware hook plumbing.tigrbl_kernel/cache.py,tigrbl_kernel/trace.py, andtigrbl_kernel/measure.py: cached plans, execution trace support, and measurement views.tigrbl_kernel/atoms.py,tigrbl_kernel/transport_atoms.py, andtigrbl_kernel/transaction_units.py: atom references and transaction/transport plan units.tigrbl_kernel/protocol_bindings.py,tigrbl_kernel/protocol_phase_tree.py,tigrbl_kernel/protocol_chains/,tigrbl_kernel/protocol_completion.py,tigrbl_kernel/protocol_fusion.py, andtigrbl_kernel/protocol_legality_matrix.py: transport-specific plan compilation, phase tree construction, chain definitions, completion semantics, fusion, and legality checks.tigrbl_kernel/eventkey.py,tigrbl_kernel/events.py,tigrbl_kernel/transport_events.py,tigrbl_kernel/webtransport_events.py,tigrbl_kernel/subevent_taxonomy.py, andtigrbl_kernel/subevent_handlers.py: event-key construction, subevent taxonomy, and protocol subevent handler mapping.tigrbl_kernel/opchannel_capabilities.py,tigrbl_kernel/loop_modes.py,tigrbl_kernel/loop_regions.py,tigrbl_kernel/segment_fusion.py, andtigrbl_kernel/contract_classification.py: channel capability checks, loop planning, segment grouping, and contract classification.
Public API and Import Surface
- Import roots:
tigrbl_kernel. - Public symbols:
BatchOpPlan,Kernel,OpView,PackedKernel,SchemaIn,SchemaOut,build_kernel_plan,build_packed_kernel,build_packed_kernel_measurement_view. - Workspace dependencies:
tigrbl-typing,tigrbl-atoms,tigrbl-core. - External runtime dependencies: none declared.
Kernel Semantics
The kernel turns specs, operation inventories, hooks, atoms, and protocol bindings into executable plans. It does not own app authoring syntax, persistence engines, or concrete ASGI classes. Its job is to compile the behavior that runtime-owned routing will execute.
The basic flow is:
specs + model metadata + bindings + hooks
-> operation views
-> ordered phase plan
-> packed kernel
-> runtime execution and diagnostics
Kernel and PackedKernel are the planning surfaces; OpView is the operation-level view of schemas, handlers, hooks, bindings, and labels. build_kernel_plan(...) and build_packed_kernel(...) are useful for tests and framework packages that need to verify the compiled plan without driving a full ASGI request.
Protocol Planning
Protocol planning keeps binding kind, family, framing, exchange, lane, subevent rows, and atom anchors separate. The current planner handles:
| Binding kind | Runtime family | Default framing | Plan shape |
|---|---|---|---|
http.rest / https.rest |
request | JSON | request received, handler invoke, response emit. |
http.jsonrpc / https.jsonrpc |
request | JSON-RPC | framing decode, dispatch, handler invoke, framing encode. |
http.stream / https.stream |
stream | stream | handler invoke, transport emit, stream close. |
http.sse / https.sse |
stream | SSE | encode event, emit event, close stream. |
ws / wss / websocket |
message | text or negotiated JSON-RPC | accept, decode, dispatch, handler, emit/close. |
webtransport |
session, stream, or datagram | no top-level app framing | lane-specific session, stream, or datagram rows with lane-local framing validation. |
Unsupported or ambiguous bindings raise planning errors. For example, WebSocket bindings do not accept HTTP methods, HTTP JSON-RPC requires an RPC method, WebTransport request/response exchange is unsupported, and WebTransport app framing must be declared on the relevant control stream, stream, or datagram lane.
Hook Ordering and Labels
The kernel orders security dependencies, dependencies, system hooks, user hooks, atoms, and handlers into phase plans. Labels are part of the diagnostics contract:
PRE_HANDLER:secdep:myapp.auth.require_user
PRE_HANDLER:dep:myapp.context.load_tenant
HANDLER:hook:sys:handler:create@HANDLER
EGRESS_SHAPE:atom:wire:dump
Use labels to debug and test ordering. Do not use them as application APIs; use specs, decorators, dependencies, and hooks for configuration.
Diagnostics and Traceability
Kernel plans feed /system/kernelz in concrete applications. Trace and measurement helpers make it possible to assert plan shape, count phase work, and inspect compiled execution without relying on a live server. This is especially important for transport work, where a route can look correct while the protocol chain is missing required event rows or capability masks.
Best Practices
- Compile plans from explicit specs and bindings; do not guess transport behavior from URL strings alone.
- Keep protocol support checks fail-closed and close to planning.
- Preserve separate event rows for request, stream, message, session, and datagram families.
- Add tests at the kernel layer when changing hook order, labels, phase rows, protocol legality, or capability masks.
- Use measurement views for plan review instead of parsing private object internals.
- Keep runtime side effects out of kernel code; the kernel should plan, classify, and label, not perform IO.
Usage Examples
Verify the installed package
python -m pip show tigrbl-kernel
python - <<'PY'
from importlib.metadata import version
print(version("tigrbl-kernel"))
PY
Import the package boundary
import importlib
module = importlib.import_module("tigrbl_kernel")
print(module.__name__)
Import a public symbol
from tigrbl_kernel import BatchOpPlan
print(BatchOpPlan)
Use with the facade when building applications
uv add tigrbl tigrbl-kernel
python - <<'PY'
import tigrbl
print(tigrbl.__name__)
PY
How To Choose This Package
Choose tigrbl-kernel when the quick-answer table matches your use case. Choose tigrbl instead when you want the full public facade. Choose a lower-level package such as tigrbl-core, tigrbl-base, or tigrbl-runtime when you are building framework extensions or testing a specific internal boundary.
Related Packages
Documentation Links
- Workspace docs
- Package catalog
- Package layout
- Current target
- Current state
- Next steps
- Documentation pointers
- SSOT registry
- Release workflow
Support
- Community: Discord.
- Issues: GitHub Issues.
- Repository: pkgs/45_kernel/tigrbl_kernel.
Package-local Boundary
This file is a package-local distribution entry point. This README is the package-local distribution entry point for tigrbl-kernel. It answers install, usage, API, ownership, and certification-orientation questions for this package. Broader architectural decisions, release status, and cross-package proof chains remain in the repository-level docs and SSOT registry.
Certification Status
- Package status: governed package in the
tigrbl/tigrblworkspace. - Governance source: SSOT registry.
- Release evidence: publish workflow validates package builds, tests, GitHub release assets, and PyPI publication for managed packages.
- Local certification guard:
pkgs/97_tests/tigrbl_tests/tests/unit/test_package_badges_and_notices.pyverifies every package README keeps the Discord badge, Apache 2.0 badge, explicit Python-version badge,LICENSE, andNOTICE. - Scope note: this README documents the package boundary. Runtime feature support remains governed by
.ssot/entities and the conformance docs linked below.
License
Licensed under the Apache License, Version 2.0. See LICENSE, NOTICE, and the official Apache 2.0 license text.
Metadata
Release files for tigrbl-kernel 0.4.5.dev6
For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.
Source distribution (sdist)
| File | Size | Uploaded | Provenance | Actions |
|---|---|---|---|---|
| tigrbl_kernel-0.4.5.dev6.tar.gz | 98.6 kB | Without provenance | Details for tigrbl_kernel-0.4.5.dev6.tar.gz |
Built distribution (wheel)
| File | Interpreter | ABI | Platform | Actions Reset |
|---|---|---|---|---|
| tigrbl_kernel-0.4.5.dev6-py3-none-any.whl | Python 3 | none | any | Details for tigrbl_kernel-0.4.5.dev6-py3-none-any.whl |
Total release size: 215.9 kB
Release files / tigrbl_kernel-0.4.5.dev6.tar.gz
| Download URL | tigrbl_kernel-0.4.5.dev6.tar.gz |
|---|---|
| Size | 98.6 kB |
| Tags | Source |
|
SHA-256 checksum How to use checksums |
a20a71a45066aa5b20c86bdecd2f459a4425bcf5971b83c1d76d3d3480fc31a9
|
|
BLAKE2b-256 checksum How to use checksums |
2cc77123b87ca004687bd9115a4249106cf8c8ddee6f6fd9a311ab5ef7517f52
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
Yes |
| Uploaded via |
uv/0.13.7 {"installer":{"name":"uv","version":"0.13.7","subcommand":["publish"]},"python":null,"implementation":{"name":null,"version":null},"distro":{"name":"Ubuntu","version":"26.04","id":"resolute","libc":null},"system":{"name":null,"release":null},"cpu":null,"openssl_version":null,"setuptools_version":null,"rustc_version":null,"ci":true}
|
Release files / tigrbl_kernel-0.4.5.dev6-py3-none-any.whl
| Download URL | tigrbl_kernel-0.4.5.dev6-py3-none-any.whl |
|---|---|
| Size | 117.3 kB |
| Tags | Python 3 |
|
SHA-256 checksum How to use checksums |
b84526340847abd93fdc00f257bbe492f095ee55b1c0de8ff7b5dcd4c89d6612
|
|
BLAKE2b-256 checksum How to use checksums |
f8100cbf9df7a14f7641a6469dce544dc1e50c0cab3889f54829a512d0cdcb47
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
Yes |
| Uploaded via |
uv/0.13.7 {"installer":{"name":"uv","version":"0.13.7","subcommand":["publish"]},"python":null,"implementation":{"name":null,"version":null},"distro":{"name":"Ubuntu","version":"26.04","id":"resolute","libc":null},"system":{"name":null,"release":null},"cpu":null,"openssl_version":null,"setuptools_version":null,"rustc_version":null,"ci":true}
|
Security for tigrbl-kernel 0.4.5.dev6
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
No provenance found
None of the files in this release were uploaded with provenance metadata.
A note on package security
PyPI is a repository for distributing packages, not a service for auditing them. While third-party security researchers and our staff do their best to detect and remove malicious packages, PyPI cannot guarantee the quality, security or suitability of any given package.
To learn more about protecting yourself from malicious actors, visit our package security documentation.