lerim ingest¶

Discover supported trace sessions and extract context records.

Examples¶

lerim ingest
lerim ingest --window 30d
lerim ingest --run-id <run_id> --force
lerim ingest --agent claude,codex

What it does¶

scans connected trace sources
matches sessions to registered projects
queues work
runs selective trace-to-context extraction
writes records into ~/.lerim/context.sqlite3

When --run-id is provided, Lerim targets that single session. If the session is not already in the local session catalog, Lerim asks the selected connected adapter for that run id first. This is the path future completion hooks should call after an agent run finishes.

Flow¶

flowchart TD
    A["Trigger: lerim ingest or daemon"] --> B["Discover and queue changed sessions"]
    B --> C["Extractor receives one session trace"]

    C --> D["Pipeline reads the next trace window"]
    D --> E["DSPy observes typed findings"]
    E --> F{"More trace windows?"}
    F -- "yes" --> D
    F -- "no" --> G["DSPy keeps reusable signal"]
    G --> H["DSPy creates one episode and zero or more durable candidates"]

    H --> I{"Durable records present?"}
    I -- "yes" --> J["Write active durable records"]
    I -- "no" --> K["Write archived episode only"]

    J --> L["SQLite context DB + record_versions"]
    K --> L
    L --> M["Completion summary"]
    M --> N["Ingest artifacts: manifest, graph events, trace"]

Notes¶

--no-extract only indexes and queues work
--dry-run previews the operation
--run-id honors --agent; for example, --agent codex --run-id ... will not process an indexed Claude session with the same id