Use the plugin to call AI Large Language Models (LLM) to generate Wiki documentation for repositories.
cnbcool/codewiki
git_doc_dir: Required. Specifies the save location for codewiki generated documentationknowledge_enabled: Whether to enable knowledge base ingestion, default false. Set to true to automatically ingest generated Wiki documents into the repository knowledge baseknowledge_embedding_model: Embedding model name, default hunyuanknowledge_chunk_size: Text chunk size (in tokens), default 1500knowledge_chunk_overlap: Overlap token count between adjacent chunks, default 0$:
tag_push:
- stages:
- name: Generate Code Wiki
timeout: 4h
image: cnbcool/codewiki:latest
settings:
git_doc_dir: /${CNB_BUILD_WORKSPACE}/${CNB_REPO_SLUG}/codewiki
| Field Name | Description | Notes |
|---|---|---|
| timeout | Sets the timeout for a single task, default is 1 hour. For larger repositories, codewiki generation time may exceed 1 hour. | Sets a timeout for a single task |
CNB_BUILD_WORKSPACEandCNB_REPO_SLUGare built-in default environment variables in CNB pipeline. For more environment variables, click here to get more details
$:
tag_push:
- services:
- docker
docker:
volumes:
- /data/codewiki/${CNB_REPO_SLUG}:data
stages:
- name: Generate Code Wiki
timeout: 4h
image: cnbcool/codewiki:latest
settings:
git_doc_dir: /data/codewiki/${CNB_REPO_SLUG}
| Field Name | Description | Notes |
|---|---|---|
| volumes | Mount directory for temporary storage of generated codewiki documentation. :data indicates creating a temporary data volume that will be automatically cleaned up when the pipeline ends. | Declare data volumes |
It is recommended to use the tag_push event to avoid frequent generation of wiki documentation.