TNView v1.1.0

TNView v1.1.0

TNView is a terminal-first telemetry and diagnostics tool for tensor-network
runs. It records append-only JSONL from DMRG, TEBD, tensor-network optimizer,
or custom long-running jobs, then lets you watch, diagnose, replay, and compare
runs from a terminal.

Highlights

• tnview watch is now the main live monitoring command.
• Run-log views include a live status line, pressure meters, sparklines,
  deterministic diagnostics, and a compact event ticker.
• Replay topology now uses pressure-aware bond glyphs and marks the selected
  bond.
• Public docs now explain the stable API surface, non-goals, and adapter
  boundaries.
• New integration guide covers plain RunLogger, quimb MPS snapshots, quimb
  TNOptimizer, and TeNPy DMRG observers.

Stable Surface

• RunLogger
• tnview watch
• tnview tail
• tnview diagnose
• tnview compare
• tnview validate
• tnview schema
• tnview init
• stable JSON output for diagnose, compare, validate, and schema

The package version is 1.1.0. The run-log telemetry schema is currently
0.1.

Adapter Status

The quimb and TeNPy adapters are dependency-optional and duck-typed. They are
intended as practical integration helpers, not complete wrappers around those
libraries.

Known Non-Goals

• tensor serialization
• browser dashboards
• full quantum-object inspection
• replacing quimb or TeNPy engines

First Run

python -m pip install -e .
tnview init emit_tnview.py
python emit_tnview.py
tnview watch runs/example.jsonl

For a checked-in demo:

make runlog-demo