Rewrite LLM runtime documentation

This commit is contained in:
2026-07-26 13:33:45 +00:00
parent 74c793e6a1
commit fdf8c4afd4

View File

@@ -1,252 +1,149 @@
# LLM Runtime Internals
`internal/framework/llm` implements Notarius's transport boundary for structured
completion. It contains the Scriptorium adapter, concurrency scheduler,
prompt/schema registries, selected-profile recording, and provider-error
redaction.
`internal/framework/llm` is Notariuss provider-independent structured
completion boundary. It adapts framework requests to Scriptorium, bounds
provider calls, assembles registered prompt and schema assets, records selected
profiles, and redacts provider errors. The architectural boundary is defined in
[Architecture](../policy/architecture.md#llm-boundary); profile sources,
credentials, and concurrency settings belong in
[Configuration](../config.md#scriptorium-profiles) and
[Configuration](../config.md#concurrency-output-cache-and-debug).
Provider-neutral ownership rules are defined in
[Architecture](../policy/architecture.md#llm-boundary). Profile sources,
credentials, and concurrency settings are defined in
[Configuration](../config.md).
## Structured Completion Boundary
## Structured Contract
Modules and LLM-backed validators depend only on
`contracts.StructuredLLMClient`. A completion request supplies a prompt ID and
version, optional profile and session IDs, named input material, variables, and
a caller-owned decode target. The successful response returns the validated raw
structured bytes together with non-secret provider, model, profile, and token
metadata.
Modules and LLM-backed validators depend on
`contracts.StructuredLLMClient.CompleteStructured`. A request identifies a
prompt and optional profile/session, supplies named input materials and
variables, and provides a caller-owned decoding target. A successful response
contains the validated raw structured bytes plus non-secret provider, model,
profile, and token metadata.
The caller owns the domain behavior: it chooses the prompt, prepares inputs,
selects the private response schema, and interprets the decoded result. The
adapter does not own source evidence, artifact conversion, normalization, or
durable schemas. Those responsibilities remain with the module and its
[integration contract](../integrations/).
The caller owns prompt selection, response-schema selection, and interpretation
of the decoded result. `LLMInputMaterial` keeps source and reference bytes with
their origin metadata so the adapter can pass named artifacts to Scriptorium
without exposing Scriptorium types through stage contracts.
`ScriptoriumClient` validates the request target and prompt identity, maps each
named material to a Scriptorium inline artifact while preserving its origin URI,
forwards session and profile selection, then prepares and runs the prompt. It
returns Scriptoriums validated raw bytes rather than re-encoding the decoded
target. An empty optional material is represented as one space so its named
input is retained by Scriptorium.
## Production Construction
An empty request profile lets the prompt select its configured default. The CLI
prepares every explicitly selected binding profile before a run begins, so a
missing explicit profile fails before stage execution. Calls record the profile
actually selected by Scriptorium; the recorder deduplicates non-secret profile
identity, provider, and model values for manifest use.
`internal/cli` constructs the production runtime by:
## Shared Provider-Call Limit
1. allocating the asset registry populated by the generic, Seriatim, and D&D
package-family registrars;
2. creating a `ScriptoriumClient` from the effective profile source;
3. attaching an `LLMProfileRecorder`;
4. creating a scheduler from the effective concurrency limit;
5. returning a `ScheduledClient` wrapper;
6. decorating that shared client before preparation when debug recording is
enabled; and
7. injecting that one shared client into complete pipeline preparation before
the source file is read or the runner is invoked.
Production construction creates one Scriptorium client and wraps it in one
scheduled client. The scheduler has a fixed, positive permit limit, serves
queued calls in FIFO order, and removes a queued call when its context is
cancelled. A granted permit is released exactly once on every completion path.
The D&D scene chunker and spell, NPC, combat-turn, item-event, NPC-interaction, and
scene-description extractors retain this
injected client and use it for every structured completion. Operation requests
do not carry an LLM client.
The CLI separately gathers explicit profile IDs from resolved LLM-capable stage
and validator bindings. It prepares a small internal check prompt for each ID so
missing or invalid profiles fail before pipeline execution. The runtime profile
override syntax and scope are defined in the
[CLI reference](../cli.md#run); binding rules are defined in
[Configuration](../config.md#module-bindings).
## Scriptorium Adapter
`ScriptoriumClient` converts a Notarius request into a Scriptorium `RunRequest`.
It validates the decoding target and prompt identity, maps named input materials
to inline artifacts, forwards explicit profile and session context, delegates
rendering/provider execution/structured validation, and unmarshals successful
JSON into the caller target.
Empty optional input material is represented by a single space so Scriptorium
retains the named input. The client returns Scriptorium's validated structured
bytes rather than re-encoding the caller target, allowing modules to preserve
the runtime result exactly.
Selected profile, provider, model, and token metadata are mapped into the
Notarius response. The recorder deduplicates profiles by identity and supplies
manifest-safe profile summaries after actual calls; manifest population does
not guess the selected prompt default in advance.
Generated-output validation failures and provider failures are wrapped with
prompt context. Error strings pass through bearer-token redaction before they
cross the runtime boundary.
## Scheduling
`Scheduler` uses a bounded permit count and a FIFO waiter queue. Immediate
acquisition increments the in-flight count; queued acquisition waits for a
permit or context cancellation. Cancellation removes a queued waiter, while a
cancelled waiter that has already received a permit releases it.
`ScheduledClient` acquires a permit around each structured completion and
defers release on every result path. The effective limit and default are
configuration facts in [Configuration](../config.md#defaults).
This provider-call ceiling is independent of the pipeline's extract worker
limit. Concurrent lanes, retries, and validators all use the same scheduled
client, so increasing framework workers cannot exceed `total_llm`. Pipeline
dispatch and cancellation mechanics are documented in
[Pipeline Internals](pipeline.md#execution-flow).
The scheduled wrapper surrounds every `CompleteStructured` call, so concurrent
lanes, pipeline retries, and LLM-backed validators share the same provider-call
ceiling. This ceiling is independent of pipeline worker concurrency; changing
worker counts cannot exceed the configured LLM limit. The configuration field
and its effective default are owned by
[Configuration](../config.md#concurrency-output-cache-and-debug).
## Prompt And Schema Assets
`AssetRegistry` combines caller-owned prompt filesystems under stable prefixes
and rejects invalid or conflicting registrations. Production module packages
register their own prompt and schema assets; generic framework code contains no
D&D prompt content. `internal/framework/promptfs` provides the domain-neutral
filesystem composition helper used to combine module-owned files with shared
domain prompt fragments.
An `AssetRegistry` collects prompt and schema filesystems from production module
families. It flattens registered roots into the Scriptorium filesystems and
rejects invalid roots, unreadable assets, duplicate paths, and missing prompt
or schema files during preparation. The frameworks `promptfs` helper combines
module-owned prompt files with reusable domain fragments without making the
framework depend on D&D content.
The D&D scene chunker and spell, NPC, combat-turn, item-event, NPC-interaction, and
scene-description extractors each declare an
ordered prompt asset manifest. The manifest lists the package-owned YAML and
Markdown files, then the exact shared fragments rendered by that prompt; the
same ordered list drives both filesystem mounting and the prompt fingerprint.
Unused shared assets are neither mounted nor fingerprinted. Universal
extraction-evidence and output policy lives only in the shared extraction
assets; package-owned prompt files retain artifact-specific rules. The scene
prompt keeps its separate output rule because it does not render the
extraction-evidence asset.
Each LLM-backed module owns its prompt declaration, package-specific assets,
and private response schema. Shared D&D wording is owned by the D&D shared
asset package; the detailed D&D conventions are in
[D&D Module Internals](dnd.md). The mounted prompt assets used by a module also
determine its prompt fingerprint. Schema loaders validate JSON, attach identity
and digest metadata, make defensive copies, and expose diagnostics without raw
schema bytes.
### D&D Extraction Prompt Ordering And Cache Boundaries
Private response schemas validate a model transport envelope. They are not the
durable artifact schema and should not be documented as an external wire
contract. Durable formats and compatibility rules remain in the
[integration contracts](../integrations/).
D&D extraction prompts order messages from the most reusable content to the
most variable content. New extraction lanes use these tiers in order:
## Prompt Maintenance And Backend Caching
1. universal shared content, including the system, extraction-evidence, and
in-world identity messages;
2. stable campaign or run context shared across lanes, including campaign
references;
3. stable subset- and lane-specific context and instructions, including an NPC
registry, catalog, task, or extraction instructions when applicable;
4. the chunk transcript as the final user message.
Prompt message order and shared asset bytes are runtime behavior. Backend cache
reuse depends on the same preceding messages and content, not merely equivalent
meaning. Keep reusable shared assets byte-identical and keep stable material
before the inputs that vary per request wherever a prompts declared sequence
supports caching. Preserve the existing manifest order and cache-control hints
when editing a prompt.
This ordering lets requests reuse the longest identical prefix before the
per-chunk transcript changes. Cache reuse requires the preceding message
sequence and content to be exactly identical; semantic similarity is not
sufficient. Cache boundaries belong at the ends of reusable stable tiers,
subject to the provider's cache-boundary limit. The shared identity and
campaign-reference messages form the first two extraction boundaries. Spell,
combat, and interaction prompts add a boundary at the shared NPC registry. Each extraction
prompt places its final boundary on its lane-specific instructions, immediately
before the transcript. The transcript does not carry cache control because no
reusable content follows it.
D&D extraction manifests place the changing chunk transcript at the end of the
prompt after their reusable context. Scene chunking and NPC normalization use
their own declared message sequences because their inputs and work differ. The
family-specific asset and ordering rules belong in [D&D Module Internals](dnd.md).
Do not add tests that enforce a fixed message-prefix length; prompt-asset tests
should instead verify the meaningful asset sequence, inputs, and cache controls
of the prompt being changed.
Accordingly, the common prefix of the spell, NPC, combat, item-event, and interaction
extraction prompts is system,
extraction evidence, identity, and campaign references. The NPC prompt then
renders task, instructions, and transcript. Spell renders the NPC registry,
catalog, task, instructions, and transcript. Combat renders the NPC registry,
task, instructions, and transcript. Item-event renders task, instructions, and
transcript without a generated-artifact input. NPC interaction renders the names-only NPC
registry, task, instructions, and transcript. The
scene chunker is not an extraction lane: it retains its separate system,
transcript, campaign-reference, task, and instruction order and marks its
transcript and campaign-reference messages ephemeral.
## Validation, Repair, And Retries
The scene-description extractor deliberately omits the citation-oriented
`common-dnd-extraction-evidence.md` asset because Notarius attaches the whole
accepted chunk range itself. Its manifest is system, shared identity, shared
campaign references, lane task, lane instructions, then the transcript. The
identity, campaign-reference, and instruction messages are ephemeral cache
boundaries; the transcript is last and has no cache control. Compatible shared
messages remain canonical shared assets rather than copied package text.
Scriptorium performs prompt rendering, provider execution, and the prompts
structured-output validation. The adapter reports an empty result, validation
failure, empty structured body, or decode failure as
`ErrInvalidStructuredOutput`, while retaining the returned raw bytes and debug
material when they exist. Provider failures remain operational errors rather
than output-validation failures.
### D&D NPC Normalization Prompt Ordering And Cache Boundaries
Prompt-declared repair is executed within Scriptoriums structured-output flow.
The current production D&D prompt manifests set repair attempts to zero. That
setting does not replace pipeline retry behavior: a bindings configured retry
count reruns its stage attempt after an error or rejection, and an exhausted
rejection is a recorded output rather than a provider error. The pipeline owns
attempt lifecycle, validation chains, and retry diagnostics; see
[Pipeline Internals](pipeline.md#validation-retries-and-output) and the
[binding reference](../config.md#module-bindings-and-validators).
NPC normalization has a distinct prompt and response-schema identity from NPC
extraction. Its stable message tiers are the common D&D system asset, followed
by package-owned task and normalization instructions. Cache boundaries follow
the shared system tier and the package instructions. The variable tail contains
the private candidate-name-and-range input and a windowed transcript input
whose cited units provide local context; neither has a cache boundary because
it changes with the document.
## Observability And Redaction
This prompt intentionally omits shared identity guidance,
extraction-evidence, and campaign-reference assets: it reconciles existing
records rather than extracting events or adding evidence. Its package-owned
manifest and schema identity are fingerprinted separately, so a normalization
prompt or schema change cannot reuse a prior normalization checkpoint.
When debug recording is enabled, the pipeline decorates the shared client. The
wrapper records prepared prompt and response material, timing, selected profile
and model, and call identifiers in the runs debug bundle, including material
available from a failed structured completion. For a successful completion, a
debug-write failure is surfaced; when the completion already failed, its call
error remains the result. Debug-bundle location, retention, and handling are
operational concerns documented in [Operations](../operations.md#debug-bundles).
Shared wording belongs in the canonical assets under
`internal/modules/dnd/shared`; extraction packages reference those assets in
their manifests instead of copying similar text into package-local files.
Package-local assets contain only lane-specific content. An extraction lane may
depart from the tier order only when prompt-quality evidence or a provider
constraint makes the exception necessary; document the exception and rationale
here when it becomes implemented behavior.
Run manifests receive selected profile summaries and component identities, not
prompt, schema, source, reference, or response content. Provider error text is
wrapped with prompt context and bearer credentials are redacted before it
crosses the runtime boundary. Known-secret redaction is available to other
runtime collaborators; it does not make prompt or response contents safe for
general logging.
Schema helpers load embedded JSON Schema with identity and digest metadata,
return defensive copies, and expose a diagnostics map that omits schema bytes.
The small framework registry contains only generic test schemas; production
schemas remain package-owned.
## Failure Boundaries
The spell, NPC, combat, item-event, NPC-interaction, and scene-description extractors'
package-owned prompts declare their
structured JSON inputs and private response schemas. Each private response
schema remains separate from its durable artifact codec schema; this work does
not use shared schema fragments or schema generation. Those private schemas own
the transport envelope—required fields, JSON types, nullability, and
unknown-field rejection—while deterministic validators own semantic constraints
such as enum membership, non-empty values and collections, and positive
numbers. The spell extractor's prompt declares a required
`application/json` `spell_catalog` input and an optional `application/json`
`npcs` input. The extractor generates
the catalog input from its prepared
effective catalog as `{"spell_names":[...]}` using sorted canonical names only.
The shared D&D prompt assets include a generic NPC grounding fragment directly
after the campaign reference message for spell, combat, and interaction prompts. When an NPC
registry is bound, the
domain registry boundary strictly decodes and identity-validates one durable
artifact, re-encodes canonical JSON for provenance, and separately generates a
names-only prompt projection. The unbound projection is exactly `{"npcs":[]}`.
Prompt input and component-local checkpoint digests cover the projected bytes;
manifests retain the optional full registry digest/count rather than names,
overlay bytes, registry paths, or source metadata. Combat and interaction prompt,
response-schema, mapping, normalization, identity, and registry-projection
fingerprints remain separate semantic inputs to checkpoint identity.
- Construction fails for missing asset registries, mutually exclusive profile
sources, invalid asset registration, or a non-positive scheduler limit.
- Preparation failures, unavailable explicit profiles, provider failures, and
context cancellation propagate to the calling stage with context.
- Malformed or schema-invalid provider output is classified separately as
invalid structured output so the module or pipeline can apply its own retry
and rejection policy.
- Domain semantic checks, evidence decisions, and deterministic normalization
run outside the provider adapter.
## Debug And Redaction Boundaries
## Focused Verification
The pipeline may wrap the client with a debug recorder that captures prepared
prompt/response material for an explicitly requested debug run. Debug summaries
and manifests receive identities, hashes, usage, and selected profile summaries
rather than prompt, source, reference, schema, or response content.
Read the LLM adapter, scheduler, asset registry, schema loader, and redaction
tests when changing this boundary. Prompt changes also require the owning
modules asset tests, and retry or debug changes require focused pipeline or
CLI coverage. The focused runtime and D&D checks are:
The Scriptorium error wrapper removes bearer credential values from surfaced
provider errors; `RedactSecrets` and `ErrorWithSecretsRedacted` support known
secret values elsewhere in the runtime. Config summaries use a separate
clone-and-redact path in `internal/core/config`. These mechanisms implement the
security invariant in
[Architecture](../policy/architecture.md#state-output-and-safety); operator
handling of debug data is defined in [Operations](../operations.md#debug).
## Failure Behavior
- Invalid targets, missing prompt IDs, malformed structured output, and
Scriptorium failures return contextual errors to the calling module.
- Scheduler construction rejects non-positive limits; acquisition respects
context cancellation.
- Asset registration rejects invalid roots, missing content, and path conflicts.
- Schema loading distinguishes missing assets, invalid JSON, and invalid
metadata.
- Profile validation errors occur during CLI preparation when an explicit
selected ID cannot be prepared.
## Tests To Inspect
- `internal/framework/llm/scriptorium_client_test.go`: adapter mapping and local
HTTP integration.
- `internal/framework/llm/scheduler_test.go` and
`scheduled_client_test.go`: permits, FIFO behavior, cancellation, and wrapper
release.
- `internal/framework/llm/asset_registry_test.go` and
`schema_registry_test.go`: asset composition, validation, and defensive
copies.
- `internal/framework/llm/secrets_test.go`: provider-error redaction.
- `internal/cli/run_contract_test.go`: profile validation, production client
wiring, manifest recording, and debug integration.
- Module-local `scriptorium_assets_test.go` files: prompt inputs and package
asset registration.
~~~sh
go test ./internal/framework/llm/... ./internal/modules/dnd/...
~~~