ec2b3841ee
All checks were successful
standards workspaces build + test / which workspaces this change touches (pull_request) Successful in 12s
standards workspaces build + test / macOS energy backends (pull_request) Has been skipped
repository checks / lockfiles, tests in CI, workflow hygiene, spec and section references, public text (pull_request) Successful in 3m11s
standards workspaces build + test / mesh-rs (mesh/mesh-rs, -p mesh-engine-gguf -p mesh-openai --features mesh-engine-gguf/gguf,mesh-openai/server, mesh-rs, linux-arm64) (pull_request) Successful in 59s
standards workspaces build + test / joulehook-rs (joulehook/joulehook-rs, joulehook-rs, linux) (pull_request) Successful in 1m13s
The grant work (#294, #301, #306) was reachable only through the library. The reference HTTP server passed no grant to every request, had no way to learn which steward a grant must name, and no way to witness a receipt against the journal. Its binaries kept the journal in memory. It also read whatever request body length a client declared. - **Grants over HTTP.** A request's body carries its presentation in `jcp_presentation` (`presentation_in`). The presentation is signed for a request id, so a granted request is served as exactly the `ServeRequest` the body alone determines (`granted_serve_request`, plus a `_responses` variant). Its prompt-token figure stays the body's estimate rather than the server's tokenizer count, which a caller cannot reproduce. The body must name its own `metadata.joule_ceiling_uj`, since a server default would otherwise enter the id unseen. - **Endpoints.** `GET /v1/mesh/steward` returns the steward's did:key, the subject a grant must name. `POST /v1/mesh/witness`, with the `receipt` and `envelope_hex` a response carried, answers 200 when the journal holds the receipt and 409 when it does not. `Mesh::witness_wire` checks a receipt in that wire form: the envelope must verify under this steward and seal exactly that wire receipt. - **Statuses** (`http_status`): 401 when the caller is not the grant's issuer, 403 when the steward refuses the grant, 503 once it has stopped honouring grants, 500 for an unwritable journal, 400 otherwise. - **Journal in the binaries.** With `MESH_JOURNAL` set, the reference server example and serve_openai_falcon record to that file and replay it before serving, reporting any unresolved serve. - **Body cap.** A body declared over 4 MiB is refused (413) before it is read. - **Spec.** Mesh SPEC §2.3 describes the binding (informative). Checked on the build host: - mesh-rs: 81 tests pass. mesh-openai with the server feature passes 15, including an end-to-end socket test: learn the steward, grant to it, present over HTTP, witness the receipt, refuse a tampered one, refuse an oversize body. - The gguf and openai server features build, every Cargo.lock is current, and clippy is clean. - Mutations each fail tests: a server that drops the grant, a granted path that recounts tokens, a wire witness that skips the seal check, and no body cap. Co-Authored-By: Claude Opus 5.5 <[email protected]>
60 lines
2.6 KiB
TOML
60 lines
2.6 KiB
TOML
[package]
|
|
name = "mesh-openai"
|
|
version.workspace = true
|
|
edition.workspace = true
|
|
rust-version.workspace = true
|
|
license.workspace = true
|
|
repository.workspace = true
|
|
description = "OpenAI-compatible API ingest for Mesh — translate a /v1/chat/completions request into a Mesh ServeRequest (joule ceiling + fidelity floor) and a served result back into an OpenAI chat-completion response carrying the sealed energy receipt as a usage extension. Lets any OpenAI client (vLLM/TGI/Ollama/SDKs) drive energy-metered serving without changing transport. Ingest, not compete."
|
|
|
|
[dependencies]
|
|
mesh-schema = { workspace = true }
|
|
mesh-core = { workspace = true }
|
|
serde_json = { workspace = true }
|
|
thiserror = { workspace = true }
|
|
hex = { workspace = true }
|
|
# Only to choose the reference server example's meter (the features below):
|
|
# the library itself never meters, it carries what `mesh-core` sealed.
|
|
mesh-meter = { workspace = true, optional = true }
|
|
# The GPU load for `ProcessMeter`'s liveness probe, with the `metal` feature.
|
|
jouleclaw-backend-apple = { workspace = true, optional = true }
|
|
|
|
[dev-dependencies]
|
|
mesh-engine = { workspace = true }
|
|
mesh-meter = { workspace = true }
|
|
jcp-schema = { workspace = true }
|
|
jcp-grant = { workspace = true }
|
|
mesh-router = { workspace = true }
|
|
mesh-receipt = { workspace = true }
|
|
ed25519-dalek = { workspace = true }
|
|
|
|
[features]
|
|
default = []
|
|
# A minimal, std-only /v1/chat/completions reference server (no async runtime,
|
|
# no extra deps). Routes through mesh-core and returns the sealed receipt.
|
|
server = []
|
|
# The meter the reference server example serves with (its module docs). With
|
|
# none of these it is the load estimator, and every receipt says `Estimator`.
|
|
# `hw-meter`: the family's machine meter (`HwMeter::detect`), a hardware
|
|
# reading only where a live counter covers the whole request (a RAPL package).
|
|
hw-meter = ["dep:mesh-meter", "mesh-meter/hw-meter"]
|
|
hw-meter-rapl = ["hw-meter", "mesh-meter/hw-meter-rapl"]
|
|
hw-meter-nvml = ["hw-meter", "mesh-meter/hw-meter-nvml"]
|
|
hw-meter-apple = ["hw-meter", "mesh-meter/hw-meter-apple"]
|
|
# macOS: this process's own CPU and GPU energy over each serve
|
|
# (`ProcessMeter`), since the reference engine runs in this process.
|
|
metal = ["hw-meter-apple", "dep:jouleclaw-backend-apple"]
|
|
|
|
# The reference server example uses the `server` module, so it only builds with
|
|
# the feature on (keeps the default `--all-targets` build clean).
|
|
[[example]]
|
|
name = "serve"
|
|
required-features = ["server"]
|
|
|
|
# Drives the example's own meter and mesh through one served request.
|
|
[[test]]
|
|
name = "serve_example"
|
|
required-features = ["server"]
|
|
|
|
[lints]
|
|
workspace = true
|