test: add multi-model pure-active and shutdown_only coverage (#1189)

[microsasa]

Closes #1189

Summary

Adds two tests addressing test coverage gaps identified in the issue:

Gap 1 — render_cost_view multi-model pure-active session

test_cost_view_multi_model_pure_active verifies that a pure-active session with two models in model_metrics:

• Shows '—' in Requests and Premium Cost columns for both model rows
• Grand total includes output tokens from both models (500 + 300 = 800)
• Grand total Requests/Premium stay at 0 (no inflation from synthetic zeros)
• No '↳ Since last shutdown' sub-row appears

Gap 2 — _compute_session_totals pure-active with shutdown_only=True

test_shutdown_only_no_effect_for_pure_active_session verifies that when shutdown_only=True is passed for a pure-active session (has_shutdown_metrics=False):

• model_calls returns the full total (8, not 8−5=3)
• user_messages returns the full total (4, not 4−2=2)

The subtraction branch requires has_shutdown_metrics=True, so pure-active sessions should pass through unchanged.

Testing

Both tests follow the exact spec from the issue and are placed in their respective test classes (TestRenderCostViewPureActiveNoShutdownRow and TestComputeSessionTotals).

Warning

⚠️ Firewall blocked 4 domains

The following domains were blocked by the firewall during workflow execution:

• astral.sh
• index.crates.io
• pypi.org
• releaseassets.githubusercontent.com

To allow these domains, add them to the network.allowed list in your workflow frontmatter:

network:
  allowed:
    - defaults
    - "astral.sh"
    - "index.crates.io"
    - "pypi.org"
    - "releaseassets.githubusercontent.com"

See Network Configuration for more information.

Generated by Issue Implementer · ● 13.6M · ◷

[Copilot]

Pull request overview

Adds targeted regression tests in test_report.py to cover two pure-active session edge cases in reporting logic, strengthening confidence around cost rendering and shutdown-only totals in the CLI reporting layer.

Changes:

• Added a render_cost_view test for a multi-model pure-active session to verify per-model — display, correct grand totals, and absence of the shutdown sub-row.
• Added a _compute_session_totals(..., shutdown_only=True) test confirming pure-active sessions do not subtract active counters when has_shutdown_metrics=False.