generate_constraint constructs optimization constraints from a constraint object's problem data. Rebalancing constraints receive the current date, holdings, optimization variables, and portfolio information. section 13.3.2
Each constraint subclass implements the method differently, translating portfolio rules into constraint expressions. Basic examples enforce full investment and nonnegative positions. section 13.3.2
Basic constraint generation
A generated full-investment constraint sums variables['positions'] and equates the result to port_info['investment_value']; a long-only constraint checks every position against zero.
