메뉴
How lncrawl (lightnovel-crawler) releases work — version bump, tag-triggered release pipeline, changelog, per-OS builds, PyPI, Docker, web sync. Use when cutting a release, editing workflows, or debugging CI.
Codex 또는 Claude로 설치 이 Prompt를 복사해 Codex, Claude 또는 다른 어시스턴트에 붙여 넣으면 Skill 페이지를 검토하고 설치를 진행할 수 있습니다.
SOC 직업 분류 기준
Add or modify a FastAPI endpoint in lncrawl — router aggregation, ensure_user/ensure_admin security, DTO vs DAO models, pagination, ServerErrors. Use when touching lncrawl/server/api/, security.py, or server/models/.
Add or modify a background job type in lncrawl — JobType enum, JobService factory, handler class, registry, notifications. Use when creating a new job kind, changing a handler, or debugging job scheduling/cancellation.
Create or fix a source crawler — base-class choice (SoupTemplate/LegacyCrawler/shared templates), selectors, cleaner tuning, testing via CLI, index-gen. Use when adding a new source site, fixing a broken crawler, or working in sources/ or lncrawl/templates/.
Schema changes in lncrawl — SQLModel DAO models, Alembic migration workflow (dev migrate CLI), Postgres/SQLite dialect differences, enum-sync migrations. Use when adding/changing a DAO model field or writing/reviewing a migration.
Binder (e-book output formats) and Translator services — architecture and recipes for adding an output format or a translation backend. Use when working in services/binder/ or services/translators/.
| name | releasing |
| description | How lncrawl (lightnovel-crawler) releases work — version bump, tag-triggered release pipeline, changelog, per-OS builds, PyPI, Docker, web sync. Use when cutting a release, editing workflows, or debugging CI. |
There is no version-bump GitHub workflow — bumping is local and the release is triggered
by pushing a v* tag.
make major|minor|patch # scripts/bump.py edits lncrawl/VERSION only
<fill in CHANGELOG.md> # add a "## [x.y.z]" section matching lncrawl/VERSION
make push-tag # tags v$(VERSION) and pushes → fires release.yml
release.yml (on tag v*) then runs, in dependency order:
lint.yml with the live crawl test enabled.web.yml to pull the latest web build.build-docker.yml (multi-arch → ghcr.io, tagged latest, the
SHA, and the VERSION).CHANGELOG.md and creates
a draft GitHub release.lint.yml — CI on push/PR touching lncrawl/**, sources/**, or dependency files:
lint (pyright + ruff), wheel build + install smoke test, sources list, and
dev migrate verify (schema-drift gate). A live crawl test is opt-in via input/dispatch.web.yml ("Sync Web") — on repository_dispatch: web-build-updated from the
lncrawl-web repo (or manual/workflow_call): checks out lncrawl-web's artifacts branch
into lncrawl/server/web/ and auto-commits to dev. That directory is committed build
output — never hand-edit it.index-gen.yml — on dev pushes touching sources/** or the index scripts: runs
scripts/index_gen.py and auto-commits the regenerated sources/_index.json/.zip and
README source tables.CHANGELOG.md sections are curated by hand; the release step extracts the section whose
heading matches lncrawl/VERSION for the release notes. scripts/changelog_gen.py is a
manual maintainer tool that back-fills missing sections from GitHub Releases.README.md regions between <!-- auto generated ... --> markers (source tables, CLI help)
are rewritten by scripts/index_gen.py — never hand-edit them.setup_pyi.py: --onedir on Windows, --onefile elsewhere. It bundles the
whole sources/ tree as data and force-imports every source module so dynamically
loaded crawlers survive freezing — a new top-level data dependency may need adding there.installer/installer.iss): the AppId GUID must never change — it identifies
the app for upgrades/uninstalls. Install is per-user (PrivilegesRequired=lowest). The
doubled {{ in the AppId is Inno's escape for a literal brace — don't "fix" it.