Fall back to per-leg margin when option strategy margin cannot be computed

[jhonabreul]

Description

Algorithms legging into multi-leg option positions with sequential market orders could crash mid-run. The option strategy margin model threw when a position group's legs didn't fit the modeled strategy's shape:

Runtime Error: Sequence contains no matching element in OptionStrategyPositionGroupBuyingPowerModel.cs:line 516

• Trigger: leg into spreads with 1-lot market orders, then hit a margin call. The margin call quantity solver probes degenerate trial groups.
• Fix margin call handling for single-lot option strategy position groups #9612 guarded zero-quantity probes. Matched strategies without margin modeling still threw NotImplementedException, and other degenerate shapes InvalidOperationException.

Fix: OptionStrategyPositionGroupBuyingPowerModel falls back to margining each leg individually instead of throwing. Well-formed groups keep strategy margin.

• GetMaintenanceMargin / GetInitialMarginRequirement catch InvalidOperationException and return per-leg margin sums, each leg margined through its own security's BuyingPowerModel.
• The NotImplementedException tails for strategies without margin modeling (e.g. CallBackspread) use the same fallback.
• The fallback is conservative (ignores leg offsets) and logs an error once per strategy/method.

Also unblocks the recommended atomic-entry path and hints at it:

• OptionStrategy.OptionLegData gets a (quantity, right, strike, expiration, orderPrice = null) constructor, accepting Python date and datetime. Previously only a parameterless constructor existed, so building legs from Python failed.
• OptionStrategy.Create and the OptionStrategy constructor no longer NRE on symbol-less legs.
• QCAlgorithm.MarketOrder warns once per algorithm when an option market order is placed while holding other contracts of the same chain, suggesting ComboMarketOrder/OptionStrategies.

Related Issue

N/A

Motivation and Context

Legging into spreads with plain market orders is legitimate API usage. It should never dead-end an algorithm in an engine crash with no user-code workaround.

Requires Documentation Change

N/A

How Has This Been Tested?

• New regression algorithm LeggedInOptionStrategiesMarginCallRegressionAlgorithm (C# + Python): legs into two spreads, forces a margin call, asserts full liquidation. Reproduces the crash before the fix, passes after, in both languages.
• OptionStrategyPositionGroupBuyingPowerModelTests: new per-leg fallback tests for a degenerate group and an unmargined strategy — red before the fix, green after.
• OptionStrategiesTests: new OptionLegData constructor tests, including construction from Python with date and datetime.
• Full QuantConnect.Tests.Common.Securities suite: 5502 passed, 0 failed. AlgorithmTradingTests: 189 passed. Related option strategy regression algorithms all passed.

Types of changes

• Bug fix (non-breaking change which fixes an issue)
• Refactor (non-breaking change which improves implementation)
• Performance (non-breaking change which improves performance. Please add associated performance test and results)
• New feature (non-breaking change which adds functionality)
• Breaking change (fix or feature that would cause existing functionality to change)
• Non-functional change (xml comments/documentation/etc)

Checklist:

• My code follows the code style of this project.
• I have read the CONTRIBUTING document.
• I have added tests to cover my changes.
• All new and existing tests passed.
• My branch follows the naming convention bug-<issue#>-<description> or feature-<issue#>-<description>