- simplify tolerance classes to single Tol.Bilateral (like vlangs implementation)
- add Distribution.from_process_capability_index(C_p, C_pk, k) ??
- Monte-Carlo simulation
- Tolerance Optimization Simple Arithmetic, Fig 9-9, 11-7 in McGraw Hill,
- Dimension.from_distribution() (nice to have for montecarlo)
- Add Triangular Distribution
- Stack combined Yield (not for stackup analysis, but combined yield)
- Fix negative distribution
- Separate Dimension and StatisticalDim/Data. (AKA, merge StatisticalDim and Reviewed)
- Basic Dimension
- Reviewed Dimension (Reviewed just points at dimension and Distribution/Data)
- Distribution
- Requirement (For distribution or distribution data)
- WC, RSS, MRSS takes all Basic Dimensions
- SixSigma, MonteCarlo takes all Reviewed Dimensions
- Spec renamed to Requirement
- Plots work with new dims
- move ruff to dev dependency
- Move WC, RSS, MRSS calculations to separate locations
- is Statistical "@ ± {self.target_process_sigma}σ & k={self.k}" needed?
- fix StackPlot.add_dimension() not returning self
- automatic or user specified plot dist data xbins (currently size=0.1)
- Fix negative statistical data
- Show Abs. Bounds instead of Rel. Bounds
- uv managed project
- DivisionByZero regression fix in SixSigma analysis
- fix abs_upper and abs_lower for negative nominal dimensions and add tests
- rename
process_sigma
totarget_process_sigma
- deprecate Z_min/Z_max (abs_upper/abs_lower)
- Fix plotting regression
- remove self.items in MRSS(), they should just be items
- reorder mean and stdev in C_pk()
- rename upper_rel to rel_upper
- rename "k" to "Skew (k)", "a" to "Sens. (a)", ...
- rename stack.items to stack.dims
- organize analysis calculations with dim.dir() first (in analysis and other locations)
- rename title to name
- give stack a description
- make C_p and C_pk only work on normal distribution with +/-6sigma (or explicitly state they are "6 sigma" values)
- add abs_upper and abs_lower (same as Z_min and Z_max)
- remove distribution from Basic Dimension
- restructure the way distributions are handled
- rename rss_func to rss and make it receive array instead of arbitrary length parameters
- WC: is tol.dir needed on t_wc computation? No
- remove mean, stdev from Statistical Dimension
- make k calculated from nominal and distribution mean
- fix slight error in stack calculations in McGraw Hill Tests
- rename tolerance to tol
- Support to positive bilateral tolerances. Ex. 2 + 0.2 / + 0.1
Initial Release