Files
scriptorium/docs/internal/sources.md

102 lines
4.8 KiB
Markdown

# Source Internals
## Purpose
This document covers Scriptorium-owned source locations and the restricted HTTP
artifact reader. Prompt, profile, schema, and ordinary artifact semantics are
owned by the tagged
[Promptkit format reference](https://gitea.maximumdirect.net/eric/promptkit/src/tag/v0.11.0/docs/formats.md).
## Application Source Locations
`internal/config` resolves `prompt_dir`, `profile_dir`, and `schema_dir` from
Scriptorium defaults, configuration files, and CLI overrides. It also resolves
the application-owned `backends` mapping into sorted engine settings.
`internal/adapter/cli` passes the directories into `promptkit.Config` and maps
each configured backend to Promptkit's public engine registration when
constructing an engine shared by the command path.
Scriptorium does not search, parse, validate, or overlay framework source files
itself. Promptkit owns prompt selection, profile built-ins and overlays, schema
resolution, ordinary file artifacts, and the related error identities.
The [configuration reference](../config.md) owns Scriptorium's source-location
fields and precedence. Maintained files under `examples/` are application
inputs that use Promptkit's tagged formats.
## Application Configuration Boundary
`internal/config` contains a ConfigKit-native `Document` definition. Its
fresh defaults come from Scriptorium's existing default constants; its resolver
records only actual trimming, path cleaning, and render-format normalization.
The validator reports application retry, size-limit, and format errors without
mutating the document or exposing rejected values. Application path values keep
working-directory-relative meaning. ConfigKit owns strict YAML decoding and
presence semantics, including dynamic integer values in backend parameters.
The `Load` boundary selects an explicit file, the first existing implicit
configuration file, or ConfigKit's defaults-only source. Only absence advances
implicit discovery. One ConfigKit load applies explicit application overrides
before validation and returns a result containing the resolved document and
its frozen inspection records. `Settings` converts that document into the
existing runtime settings with backend IDs sorted; it does not read files or
apply another merge. File-backed loads enable explicit ConfigKit imports and
select one named overlay when requested. ConfigKit uses the selected file's
canonical directory as the source root; the application does not widen it.
Application path values retain process-working-directory-relative meaning.
Every configuration-loading command uses this boundary. The CLI maps only
explicitly visited application flags to ConfigKit overrides and prints
structured load diagnostics before engine construction. The old application
YAML decoder and merge path have been removed.
Application `config` inspection reuses the result's frozen effective,
provenance, and history views without constructing an engine or reloading
sources. It does not resolve API-key environment values; ordinary backend
parameters remain visible by the selected inspection policy.
## Restricted HTTP Artifact Reader
`internal/adapter/http` implements `promptkit.ArtifactReader` for HTTP
requests. The `serve` path injects it with
`promptkit.WithArtifactReader`, replacing Promptkit's ordinary reader for
inbound HTTP inputs.
The reader:
- accepts inline references without an artifact root;
- denies file references when no root is configured;
- resolves relative paths below the configured root;
- accepts absolute paths only when they are lexically within that root;
- rejects lexical traversal outside the root;
- applies the configured file byte limit, with zero meaning unlimited;
- preserves content type, body, size, hash, name, and URI metadata; and
- honors context cancellation.
Containment is lexical and does not resolve symlinks. The operating system
follows symlinks after the check. The [HTTP API](../api.md) owns observable
request outcomes, and [operations](../operations.md) owns safe deployment
permissions and root selection.
Reader errors remain identifiable after Promptkit wraps them as artifact-load
failures, allowing the HTTP adapter to preserve Scriptorium status and error
codes.
## Verification And Change Recipe
Inspect:
- `internal/config/document_test.go`
- `internal/config/load_test.go`
- `internal/adapter/cli/run_test.go`
- `internal/adapter/http/artifact_reader_test.go`
- `internal/adapter/http/handler_test.go`
When changing an application source location or HTTP artifact policy:
1. preserve strict configuration precedence and the Promptkit public boundary;
2. keep containment and size policy in Scriptorium;
3. update focused configuration, reader, and handler tests;
4. update the [configuration](../config.md), [HTTP](../api.md), and
[operations](../operations.md) contracts as applicable; and
5. do not duplicate Promptkit loaders, formats, or ordinary artifact behavior.