| Age | Commit message (Collapse) | Author |
|
The access log said who (`user`) but not with what: with short tokens a
request can be made by a shared credential, so "which person" and "which
credential" are now separate fields -- `scheme` ("jwt" | "opaque") and,
for an opaque token, `token` (its registry handle, e.g.
`agent-ravanan-3`). Never the token or its hash, and the path still
carries no query string.
The line also gains `ts` (RFC 3339 UTC), so a redirected stderr file is
self-describing rather than relying on journald's prefix. Hand-rolled
civil-from-days rather than a new crate: one format, and no Cargo.lock
drift.
scripts/gn4-log.py reads such a log and shows it as a table with the
handle joined against tokens.sqlite: the subject as the registry has it
now (flagging a handle whose subject has been repointed), and the token's
status today -- [revoked], [expired], [unknown] -- which is how "an agent
was working at 14:02" is told apart from "an agent was working at 14:02
with a credential revoked since". It reads the registry read-only, passes
non-JSON lines through in table mode (the process log interleaves
backtraces), and offers --json and --summary for jq and for counting.
CallerSlot now carries the whole Caller (scheme, id, subject) rather than
just the subject string.
|
|
A second credential format beside gn-auth's JWT, for agents and scripts:
`gn4ai_` + 56 body characters + 2 checksum characters = 64, from an
alphabet with no look-alikes, so a typo is refused before the registry is
consulted. Opaque by design -- nothing is stored in the token; every fact
lives in one SQLite database (`tokens.sqlite`, GN4_API_TOKENS_DB, mode
0600, gitignored) keyed by the SHA-256 of the secret. Revoking is a row,
takes effect on the next request, and a token in a log line reveals
nothing; lose the registry and every short token stops working.
Administration is local-only: GET/POST /tokens, DELETE /tokens/{id} and
GET /tokens/who are behind a router layer that refuses any caller whose
TCP peer is not 127.x.x.x/::1, and any request claiming to come through a
proxy. A layer rather than a per-handler check, so nothing of a remote
request is parsed before it is refused; there is no configuration switch,
because minting credentials must never be reachable from the network.
A short token reads public data only: every token-only endpoint, and
`#[gnauth(secured)]` endpoints for public rows; a confidential dataset is
refused with a 403 that names the JWT route, since gn-auth cannot validate
an opaque token (the on-the-fly-bridge and registry-grant alternatives
were both rejected for v1).
scripts/gn4-token.py create|list|revoke|who drives the routes, so there is
one implementation of the format and the revocation rules; the token is
printed exactly once.
Tests: the localhost guard over all four routes (remote peer, no peer,
proxied header), and mint -> authenticate -> who -> revoke -> 401.
|
|
Answering "is robw using this successfully, and the others?" from the
access logs took an hour of inference -- user agents, search terms, which
traits were mapped -- because nginx records the address and the subject is
only in the token. Two callers can share an address (an agent and a person
behind one NAT), one caller can move between addresses (an LLM's egress is
a /24), and the answer for a given person was "probably 80.115.1.132".
The API now sets `x-gn4-user: <token sub>` on every **authenticated**
response, and the `gn4` log format records it as `user=`. The subject is
only known inside the handler (the `AuthToken` extractor verifies the
token there), so a middleware opens a slot in the request extensions, the
extractor fills it, and the header is set on the way out; a request with no
valid token gets no header, which is the honest answer and reads as
`user=-`. The caller only ever sees an identity it already presented.
* src/auth.rs: `USER_HEADER`, `CallerSlot`, and the extractor recording
its subject.
* src/lib.rs: the `tag_caller` middleware, outermost so the slot exists
before any extractor runs.
* tests/auth.rs: `an_authenticated_response_names_its_caller` -- the header
equals the token's `sub`, and `/version` (no identity) carries none.
* etc/nginx-vesuvio.pu.ac.ke.conf (and the deployed copy of it): `user=`
in the log format.
* README.md, docs/nginx.md (the format, an `awk` one-liner for per-user
counts, and the caveat that old and new lines differ in field count),
docs/www/operator-checklist.md (`user=-` means no token, not a token
that failed late).
281 tests pass; the same four as before do not.
|
|
Pure formatting sweep (import ordering, line wrapping); the tree
had accumulated rustfmt drift. Separated from any semantic change
so it stays reviewable and revertable on its own.
|
|
Adds a gnauth-macros proc-macro crate and a private-field Authorized
capability type in src/auth.rs, so #[gnauth(secured)] is a real,
compile-time-checked guarantee that a handler performs the gn-auth
resource-authorisation round trip before sharing trait/probeset
values, rather than just a comment. #[gnauth(token_only)] marks
handlers that require a token but skip the round trip, and
#[gnauth(disabled)] marks the two handlers that need no token at all.
Also fixes groups.rs: group_trait_values, group_trait_info and
group_traits previously performed no gn-auth round trip at all for
published-trait data; they now resolve the group's PublishFreeze row
and call the same auth::check used elsewhere, denying by default when
no such row exists.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
|
|
Breaking: every success body is now {"data", "meta": {"state",
"took_ms", "request_id"}, "notes", "next"}; clients read `data`. Only
`/` (plain text) and `/openapi.json` are unwrapped.
* src/response.rs (Envelope, Meta, State, Ctx, RequestInfo): handlers
take Ctx first and build the response with ctx.envelope(data), so
timing and the request id are never filled in by hand. `state` is
`complete` or `partial`; notes are server-authored only; `next` are
concrete relative hrefs (path_segment percent-encodes names).
* src/routes/*.rs, src/lib.rs (version): every handler returns
Envelope<..> with next links; /search is `partial` when the locus
lookup fails or a sorted 1000-hit window may be truncated, and
/publications/{pubmed_id}/traits when datasets were left out for
authorization.
* src/error.rs (ApiError): add `took_ms`, an RFC 7807 extension member.
* src/auth.rs (AuthToken): the 401 now carries request_id and took_ms
(it never had a request_id); Ctx stores RequestInfo in the request
extensions for it.
* tests: common::data / envelope_data parse every success body through
the envelope; tests/envelope.rs and test/python/test_envelope.py check
the shape, errors, and that every `next` link resolves;
tests/envelope_partial.rs checks the `partial` search without a
database; tests/openapi.rs fails if a JSON success response is not an
Envelope_* schema.
* README.md, docs/SKILL.md, docs/gn-auth.md, RELEASE_NOTES.md: document
the envelope and took_ms on errors.
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
|
|
Clippy type_complexity fired three times on the OnceLock<Mutex<
HashMap<...>> statics. One shared alias per cache shape keeps the
signatures readable without changing any behaviour.
|
|
A new, additional endpoint family that addresses both trait categories
the same way -- (dataset, trait) -- matching GeneNetwork2's own generic
wqflask.api.router.trait_sample_data/get_trait_info, rather than the
two category-specific shapes genenetwork4 already has
(/datasets/{dataset}/probesets... and /groups/{group}/traits...), which
are unchanged and keep working exactly as before.
- db::resolve_trait_dataset(name) resolves a dataset name against
ProbeSetFreeze first, then the new db::PublishDataset (a PublishFreeze
row, via the new get_publish_freeze_by_name) -- whichever matches
decides the category. This means only dataset names are ever
"guessed" across categories, not trait names/ids, which is the much
smaller and safer surface: GN2's own equivalent tries ProbeSet.Name
and PublishXRef.Id against the same bare trait string and silently
falls back, which can return the wrong trait on a same-named
collision. get_publish_freeze_by_name accepts a PublishFreeze's own
Name/ShortName/Id or its owning group's Name/InbredSetName/FullName/
Id, mirroring GN2's own flexible matching (e.g. "BXDPublish" or
"BXD").
- GET /traits/{dataset}: list every trait in a dataset, tagged by
category.
- GET /traits/{dataset}/{trait}: one trait's metadata. A non-numeric
trait against a Publish dataset is a 404 (can't match a
PublishXRef.Id), not a 500.
- GET /traits/{dataset}/{trait}/values: that trait's per-strain values.
TraitValue and ProbeSetValue are already identical in shape, so this
unifies with no information loss, unlike the metadata endpoint above.
Metadata responses are a tagged union ({"category": "probeset"|
"publish", ...}) wrapping ProbeSetInfo/TraitInfo unchanged, rather than
a lossy common-fields schema -- the two share little beyond
mean/locus/lrs/additive. utoipa generates this as a oneOf with a
category discriminator, verified against /openapi.json.
Gated the same way as /datasets/{dataset}/probesets..., extended to
Publish data for the first time: /groups/{group}/traits itself is
still not gated, but reaching the same data through /traits/{group}
now requires a token and checks PublishFreeze.public/confidentiality.
No gn-auth-side change was needed -- POST /auth/data/authorisation
already dispatches *Publish-suffixed dataset names to its phenotype
resource resolver (gn_auth/auth/authorisation/data/phenotypes.py),
the exact mechanism this integration already relies on for mRNA data.
auth::VIEW_RESOURCE_PRIVILEGE moves from routes::datasets (private) to
routes::auth (pub) so routes::traits can reuse it instead of
duplicating the string.
tests/traits.rs and its Python mirror cover both categories (including
both ways of addressing a Publish dataset), the 401/403/404 cases, and
the non-numeric-trait-against-Publish-dataset edge case, run against a
real gn-auth instance. The Publish branch's deny path itself is only
exercised via gn-auth's "unlinked dataset" fallback (already proven for
ProbeSet) -- no non-public PublishFreeze exists in the database this
was tested against.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01XDDpXvZBQat9uSgDeEMJ8d
|
|
List endpoints for the two trait domains, each the listing equivalent
of an existing singular endpoint:
- GET /groups/{group}/traits: every published (phenotype) trait in a
group (db::get_group_traits), same fields/join as
/groups/{group}/traits/{trait_id}, no token required -- matching its
singular sibling, which isn't gated either. A large group (BXD has
15000+ phenotype traits) can take several seconds and tens of
megabytes to return in full; no pagination yet, matching
GeneNetwork2's own wqflask.api.router.fetch_traits, which has no
default limit either.
- GET /datasets/{dataset}/probesets: every expression (ProbeSet) trait
in a dataset (db::get_dataset_probesets), same fields/join as
/datasets/{dataset}/probesets/{probeset}. Gated the same way as that
singular endpoint -- a valid token is required unconditionally, and
a non-public dataset is checked against gn-auth once for the whole
dataset via a new auth::authorise_dataset, rather than once per
probeset in the list. Leaving this listing ungated would have let a
caller read out a whole non-public dataset's trait data in one
request, bypassing the gate just added to the singular endpoint.
auth::authorise_dataset wraps authorise_trait with an empty trait
name: gn-auth resolves a trait's resource purely from its dataset
name, so the trait half of a "Dataset::Trait" string never affects
which privileges come back. Verified directly against the live
instance both for an unlinked non-public dataset (403, deny) and
folded into dataset_probesets's own test for the public/embargo-aged
bypass path.
tests/group_traits.rs, tests/dataset_probesets.rs and their Python
mirrors cover both, run against a real gn-auth instance for the gated
one.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01XDDpXvZBQat9uSgDeEMJ8d
|
|
Require a valid gn-auth bearer token on this endpoint unconditionally.
For a dataset that isn't publicly readable, check the
group:resource:view-resource privilege via a new auth::authorise_trait,
the "Dataset::Trait"-keyed equivalent of authorise (POST
/auth/data/authorisation instead of /auth/resource/authorisation) --
gn-auth resolves the resource itself from the dataset name, so no
separate resource-id lookup is needed here. A trait never linked to a
gn-auth resource (gn-auth answers 404) is treated as "no privileges",
not a request failure. Denied requests get a 403; a failure to
reach/parse gn-auth's answer gets a 502, distinct from both that and
this service's own 500s.
is_publicly_readable now also takes the dataset's CreateTime and treats
anything older than 5 years as public regardless of its
public/confidentiality flags -- a confidentiality embargo isn't meant
to last forever. No admin/sysadmin bypass is implemented.
Register a "bearer_token" OpenAPI security scheme (src/lib.rs) so the
new security(...) annotation on this endpoint's #[utoipa::path] has
something to point at.
tests/dataset_probeset_info.rs and its Python mirror gain cases for the
missing-token (401) and denied-but-existing (403) outcomes, run against
a real gn-auth instance set up as described in docs/gn-auth.md; the
403 case uses UTHSC_BXD_Cochlea_0225, a real non-public, non-embargoed
dataset never linked into gn-auth. tests/common gains get_with_token
and auth_server_available/auth_token helpers for this.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01XDDpXvZBQat9uSgDeEMJ8d
|
|
* Cargo.toml (jsonwebtoken, ureq): New dependencies, RS256 JWT
verification and gn-auth's HTTP API respectively. ureq pulls in ring
for RSA verification, needing a C compiler at build time -- already
provided by cargo-build-system's standard inputs, nothing added to
guix.scm for it. Promote serde_json from dev-dependencies to
dependencies; src/auth.rs needs it outside tests now.
(mysql-common-derive, simple_asn1, time): Pin to the newest versions
still compatible with the Guix-packaged rustc 1.85.1 (same trap as
before: ureq's dependency chain pulls darling 0.23/time 0.3.46+, both
needing rustc 1.88).
* rust-package.scm: Regenerate with `guix import crate -f Cargo.lock
genenetwork4`.
* guix.scm (genenetwork4-git)[native-inputs]: Update crate list to
match.
* src/auth.rs (AuthToken): New extractor. Verifies a bearer token's
RS256 signature and standard time claims (exp/nbf, checked only when
present -- gn-auth's own validator treats them as optional too)
against gn-auth's JWKS, cached and refreshed periodically
(OnceLock<Mutex<HashMap<...>>>, same shape as db::pool()) rather than
fetched per request.
(is_publicly_readable): New function. Matches GeneNetwork2's own
threshold (public > 0 AND confidentiality < 1, used throughout
wqflask/api/router.py) for whether a dataset needs an authorization
check at all; InbredSet/group rows carry a public column too, but GN2
never filters on it, so groups/species endpoints need no gating.
(authorise): New function. Wraps gn-auth's POST
/auth/resource/authorisation, caching the per-(user, resource) answer
briefly (AUTHORISATION_TTL) so repeated requests for the same resource
by the same user don't each cost a round trip. Not yet called from any
route -- see RELEASE_NOTES.md.
* src/config.rs (Config)[auth_server_url]: New field, AUTH_SERVER_URL
config key, matching GeneNetwork3's own key of the same name.
* src/error.rs (ApiError): Derive Debug; needed by tests/auth.rs's
.unwrap() on a Result<_, ApiError>, and every error type should
implement it regardless.
* src/lib.rs: Register the new auth module.
* tests/auth.rs: New file. Runs AuthToken verification and authorise()
(including its cache) against a real gn-auth instance; set
GN4_TEST_AUTH_TOKEN/GN4_TEST_RESOURCE_ID (see docs/gn-auth.md), skips
otherwise.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01XDDpXvZBQat9uSgDeEMJ8d
|