feat: forward solver_options, and gate reads on an actual incumbent

[FBumann]

Review finding on #148, verified and valid — with a correction to my own first assessment.

I was wrong that it was unreachable

My initial read was "a time limit isn't settable, so this is latent — defer to #28". That was wrong, and for an embarrassing reason: the time limit isn't settable because we forgot to forward solver options at all.

linopy takes a dict and applies it:

for k, v in self.solver_options.items():
    h.setOptionValue(k, v)

We forwarded nothing. Every kwarg on farkas.solve stops at DuckdbExecutor.__init__ — memory_limit, chunk_rows, threads, workdir — and none of those reach HiGHS. No time limit, no MIP gap, no solver knob whatsoever.

So the finding wasn't blocked by unreachability; it was blocked by a missing feature. Both halves are here because neither is testable without the other.

is_ok was too coarse

is_ok is linopy's rollup, and ok covers time_limit, iteration_limit, terminated_by_limit, suboptimal, imprecise — all "stopped early", where an incumbent may or may not exist. Only optimal guarantees one.

HiGHS answers that separately (Info.primal_solution_status), so SolveStatus carries it and readers gate on:

is_readable = is_ok and has_primal

Deterministic once options work:

time_limit=0  cond=time_limit  is_ok=True  has_primal=False  obj=nan
no options    cond=optimal     is_ok=True  has_primal=True   obj=46337878.0

Reading the primal of the first raises, instead of returning 60 zeros.

⚠️ This deliberately diverges from linopy

Stating it loudly, because we just made "lean on linopy" a written convention.

I checked all four signals — primal_solution_status, SolutionStatus, has_solution, info.valid: zero hits in linopy. safe_get_solution gates on is_ok alone, so linopy has this exact bug. A MIP stopped before finding a feasible point has its zero-filled col_value read as an answer.

Copying that would reintroduce #115 one level up, so I didn't. The vocabulary stays linopy's — is_ok still means precisely what linopy means. We add a second, narrower predicate rather than redefining theirs.

If you'd rather match linopy exactly and accept the gap, say so — it's a one-line change to make is_readable return is_ok. I think that would be wrong, but it's your call and it is a real deviation from the convention.

Scope

solver_options threaded through api.solve → executor.solve → sinks.solve_direct, kept separate from build knobs. SPEC.md documents both the option and why is_ok and has_primal differ.

Three new tests: options reach the solver, time-limit-without-incumbent, time-limit-with-incumbent.

337 passed, 1 xfailed · ruff / pyrefly clean.

🤖 Generated with Claude Code

[coderabbitai]

Warning

Review limit reached

@FBumann, you've reached your PR review limit, so we couldn't start this review.

Next review available in: 5 minutes

Enable usage-based reviews in Billing to review now. Otherwise, wait until the next included review is available.
You're only billed for reviews past your plan's rate limits ($0.25/file).

How can I continue?

After more reviews become available, a review can be triggered using the @coderabbitai review command as a PR comment. Alternatively, push new commits to this PR.

To avoid repeated limits, reduce automatic review volume by pausing incremental auto-reviews earlier, using label-based review opt-in, excluding WIP or generated PR titles, or requesting reviews manually when the PR is ready. If your team needs uninterrupted high-volume reviews, an organization admin can enable usage-based reviews.

How do review limits work?

CodeRabbit enforces per-developer PR review limits for each organization. Most developers receive the normal plan review availability.

For paid Pro and Pro+ PR reviews, CodeRabbit uses adaptive limits for sustained high-volume activity. When a developer's recent PR review activity reaches the 95th percentile or higher among CodeRabbit users, additional reviews become available more gradually as earlier reviews age out of the rolling window.

Please refer docs for additional details.

Review details

⚙️ Run configuration

Configuration used: defaults

Review profile: CHILL

Plan: Pro Plus

Run ID: 7efe2fc4-3522-49e5-a68a-7a210cd0c154

📥 Commits

Reviewing files that changed from the base of the PR and between 17e1ea9 and 6b2c0cf.

📒 Files selected for processing (6)

• SPEC.md
• src/farkas/api.py
• src/farkas/relational/executor.py
• src/farkas/relational/sinks/highs.py
• src/farkas/relational/status.py
• tests/test_solve_status.py

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Commit unit tests in branch feat/solver-options-and-primal-gate

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands.}