Validate flow and postman model inputs

[isPANN]

Summary

• Share private constructor checks with deserialization for IntegralFlowBundles, IntegralFlowHomologousArcs, IntegralFlowWithMultipliers, MinimumCostCirculation, MinimumEdgeCostFlow, UndirectedFlowLowerBounds, DirectedTwoCommodityIntegralFlow, UndirectedTwoCommodityIntegralFlow, PathConstrainedNetworkFlow, and MixedChinesePostman; reuse those checks in the five existing create specifications.
• Reject malformed terminals, capacity/cost arrays, bundles, and homologous arc indices; reject negative capacities, negative internal multipliers, and nonpositive bundle and lower-bounded flow requirements consistently.
• Preserve public constructors, JSON fields, signed costs, BruteForce dimensions(), and existing capacity-domain limits.
• Twenty files, based directly on main. Nested DirectedGraph endpoint validation is handled separately in Validate graph invariants during deserialization #1159.

Verification

• Three create/deserialization regressions fail on main; two cost-flow deserialization regressions, two two-commodity flow regressions, and the lower-bounded flow regression fail before their fixes.
• Full workspace all-feature tests: 6,488 passed; 2 existing ignored. Documentation tests on the preceding cost-flow commit: 145 passed; 4 existing ignored.
• Focused graph model coverage run: 1,120 passed. Changed-line coverage: 99.86% across 694 lines; the remaining line is the bundle domain overflow error on narrower usize platforms. Both cost-flow models, both two-commodity models, and the lower-bounded flow model have 100% changed-line coverage.
• Workspace all-target all-feature Clippy with warnings denied, formatting, and diff whitespace checks passed.
• Prescribed-path flow and mixed postman deserialization reuse existing fallible constructors; new rejection tests cover malformed paths, terminals, and weight vectors. Both additions have 100% changed-line coverage.

[codecov]

Codecov Report

❌ Patch coverage is 99.85591% with 1 line in your changes missing coverage. Please review.
✅ Project coverage is 96.05%. Comparing base (7dd5fcd) to head (effeae4).
⚠️ Report is 7 commits behind head on main.

Files with missing lines	Patch %	Lines
src/models/graph/integral_flow_bundles.rs	98.48%	1 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #1162      +/-   ##
==========================================
+ Coverage   95.93%   96.05%   +0.12%     
==========================================
  Files        1074     1074              
  Lines      132106   132935     +829     
==========================================
+ Hits       126730   127685     +955     
+ Misses       5376     5250     -126     

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.
• 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.

[GiggleLiu]

Final review notes

• Pushed one mechanical commit rewriting !(a < b) / !(c > 0) as a >= b / c <= 0 in the three integral-flow models. Older clippy (1.93) denies these under nonminimal_bool; behaviour is identical.
• Same follow-up as Validate graph, set-system, and database deserialization #1161: consider pub fn try_new on the new constructors so reductions can build these targets fallibly, and the test_ prefix on new tests.

[GiggleLiu]

Final review: all new rejections trace to existing asserts; JSON fields preserved; make check passes locally on the merged state after the clippy rewrite. Approving.