mirror of
https://github.com/hanzoai/java-sdk.git
synced 2026-10-30 15:07:36 +00:00
zeekayandHanzo Dev
f5b25946c7
2452 operations over 1737 paths, 263 tags, 1798 schemas — 2333 Java sources under hanzo-java-cloud/src/main/java/ai/hanzo/cloud, none of them written by hand. Published as ai.hanzo:hanzo-java-cloud, versioned to the document's own info.version so a release is: change the spec, regenerate, tag. scripts/generate.sh is a CALL SITE, not a generator invocation. The invocation lives once in hanzoai/openapi generate.py and every knob is data in sdks.yaml beside it, so there is nothing here that can drift on its own — and `--check` is a CI gate, which makes that a fact instead of a convention. The client is okhttp-gson rather than the `native` java.net.http library the java row originally asked for: native's oneof template emits `List<T>.class`, which is not legal Java, and this document has three oneOfs with an inline array variant. That is recorded where the decision lives, in sdks.yaml. The six canonical flows are in examples/<flow>/, one per flows.yaml, and the build compiles them, so they cannot rot. HANZO_API_KEY, HANZO_BASE_URL and HANZO_ORG_ID resolve in exactly one place, examples/Hanzo.java — verified on the wire to send both `Authorization: Bearer` and `X-Org-Id`. The repo root owns the build, README, LICENSE, examples, scripts and CI; generate.py owns one directory and replaces it wholesale. Co-authored-by: Hanzo Dev <[email protected]>
7 lines
220 B
Groovy
7 lines
220 B
Groovy
rootProject.name = 'java-sdk'
|
|
|
|
// hanzo-java-cloud is the published artifact and is generated in full from
|
|
// hanzoai/openapi. examples is the gate that proves it is usable.
|
|
include 'hanzo-java-cloud'
|
|
include 'examples'
|