Marivo 0.5.2

Marivo 0.5.2 hardens Session recovery and guided analysis, and makes driver-axis discovery recover missing dimensions from persisted observations when the original replay authority remains available.

Highlights

mv.session.resume(identity) resumes an existing Session by exact stable name or immutable id. If one value matches both a name and a different id, use by="name" or by="id" to choose explicitly; unknown or ambiguous identities fail without creating or changing Session state.
driver_axes can recover missing search axes by replaying both persisted source observations with their original alignment before scoring. Each replay observation and comparison is a separate recoverable Run, and candidate metadata retains original and expanded Artifact refs.
mv.session.abandon_run(session_id=..., run_id=...) provides an atomic repair for a stopped incomplete Run. It records a failed Run and removes only unconsumed Store Artifact registrations, leaving physical files and Evidence markers intact.

Changed

Analysis Help now labels analysis.entry as the route into typed analysis and teaches the Session-bound catalog acquisition path. Catalog readiness errors provide exact typed-entry repair guidance for bare strings.
Session activation preflights recovery and keeps ambiguous incomplete Run reconciliation fail-closed with structured repair instructions.

Fixed

Native percentile attribution now normalizes midnight timestamp coordinates to the canonical daily bucket, avoiding mismatches between replayed and materialized day values.
Driver-axis replay validates current semantic authority, source artifacts, recovered observe parameters, and alignment before admitting the discovery Run.

Documentation

Updated the English and Chinese analysis workflow guidance for Session identity, stopped Run recovery, driver-axis replay, and the progressive analysis/catalog Help routes.

Installation

curl -fsSL https://marivo.io/install.sh | bash

In mainland China, use:

curl -fsSL https://marivo.io/install-marivo-cn.sh | bash

For manual installation:

pip install marivo

Marivo 0.5.2 requires Python 3.10 or newer.