preps for v1.2 release (#746)

* deleted user manual pdf in src

* fixed issues with user manual tex file

* updated user manual

* bumped ver num

* updated change log

* Update doc/src/sections/hiopbbpy.tex

Co-authored-by: Nai-Yuan Chiang <chiang7@llnl.gov>

---------

Co-authored-by: Nai-Yuan Chiang <chiang7@llnl.gov>

Author: cnpetra

CHANGELOG.md

@@ -1,6 +1,22 @@
 # Change Log
 All notable changes to HiOp are documented in this file.
 
+## Version 1.2.0: Release of HiOpBBpy solver for derivative-free optimization
+* HiOpBBpy Batched Bayesian Optimizer  @thartland in https://github.com/LLNL/hiop/pull/724
+* HiOpBBpy Nonlinear constraints by @nychiang in https://github.com/LLNL/hiop/pull/726
+* HiOpBBpy parallel MPI evaluation manager to utils by @weslleyspereira in https://github.com/LLNL/hiop/pull/733
+* Derivative checker for quasi-Newton and Newton IPMs by @cnpetra in https://github.com/LLNL/hiop/pull/731
+* HiOpBBpy parallel Evaluator Upgrade by @thartland in https://github.com/LLNL/hiop/pull/734
+* Addressing misc compilation warnings by @cnpetra in https://github.com/LLNL/hiop/pull/738
+* Logger for HiOpBBpy by @nychiang in https://github.com/LLNL/hiop/pull/739
+* README update with HiOpBBpy information by @thartland in https://github.com/LLNL/hiop/pull/742
+* Misc fixes for fixed variables, feasibility restoration, Krylov-based IR, and others by @cnpetra in https://github.com/LLNL/hiop/pull/710
+
+## New Contributors
+* @weslleyspereira made their first contribution in https://github.com/LLNL/hiop/pull/733
+
+**Full Changelog**: https://github.com/LLNL/hiop/compare/v1.1.1...v1.2.0
+
 ## Version 1.1.1: Misc updates to quasi-Newton solver and build system, updates to recent RAJA, and preliminary black-box solver
 * Consolidating qn solver by @cnpetra in https://github.com/LLNL/hiop/pull/697 and https://github.com/LLNL/hiop/pull/699
 * Use clang-format in hiop to keep code style by @nychiang in https://github.com/LLNL/hiop/pull/700

CMakeLists.txt

@@ -17,7 +17,7 @@ if (POLICY CMP0074)
   cmake_policy(SET CMP0074 NEW)
 endif ()
 
-project (hiop VERSION "1.1.1")
+project (hiop VERSION "1.2.0")
 
 string(TIMESTAMP HIOP_RELEASE_DATE "%Y-%m-%d")
 

doc/hiop_usermanual.pdf

@@ -19,7 +19,7 @@ \subsubsection{Condensed Linear System}\label{kkt:condensed}
 
 \warningcp{Note:} If equality constraints $c(x)=c_E$ are present, they will be slightly relaxed to inequalities $c_E - C_1\leq c(x)\leq c_E+C_1$, where $C_1$ is a small positive perturbation that will be updated by \Hi internally. Consequently, with the condensed linear algebra, \Hi solves problems with equality constraints as inequality-only problems in the form of~\eqref{spobj_condensed}-\eqref{spbounds_condensed}.\\
 
-Using the notations from \cite{petra_hiop}, the condensed linear system solves the most stable ``xdycyd'' KKT linear system 
+Using the notations from~\cite{petra2019memory}, the condensed linear system solves the most stable ``xdycyd'' KKT linear system 
 \begin{equation} \label{KKT_xdycyd_condensed}
   \begin{bmatrix} 
   H+D_x+\delta_{w}I & 0  & J_d^T\\

doc/src/appendix.tex

@@ -20,8 +20,8 @@ \subsection{HiOpBBpy installation}
 \end{verbatim}
 
 \Hibbpy also supports using nonlinear optimization solver Ipopt \cite{ipopt_impl} as its internal solver. 
-In order to use Ipopt, \Hibbpy optionally depends on {\it{cyipopt}}, which is a python wrapper of Ipopt.
-To use Ipopt via {\it{cyipopt}},, one can install Ipopt from source, and then export the \verb~IPOPT_PATH~ and 
+In order to use Ipopt, \Hibbpy optionally depends on {\it{cyipopt}}, which is a Python wrapper of Ipopt.
+To use Ipopt via {\it{cyipopt}}, one can install Ipopt from source, and then export the \verb~IPOPT_PATH~ and 
 \begin{verbatim}
 > export PKG_CONFIG_PATH=$PKG_CONFIG_PATH:$IPOPT_PATH/lib/pkgconfig
 > export PATH=$PATH:$IPOPT_PATH/bin
@@ -70,7 +70,7 @@ \subsection{The Python interface}
 One example to the \Hibbpy Bayesian optimization algorithm is the following:
 
 
-\begin{lstlisting}
+\begin{lstlisting}[language=Python]
 import numpy as np
 from hiopbbpy.surrogate_modeling import smtKRG
 from hiopbbpy.opt import BOAlgorithm
@@ -96,7 +96,7 @@ \subsection{The Python interface}
 
   acq_type = "EI" # EI or LCB
   options = {
-	'acquisition_type': acq_type,
+	'acquisition\_type': acq_type,
 	'log_level': 'info',
 	'bo_maxiter': 10,
   }
@@ -148,27 +148,23 @@ \subsection{Parallel runs}
 
   acq_type = "EI" # EI or LCB
   options = {
-    'acquisition_type': acq_type,
-    'log_level': 'info',
-    'bo_maxiter': 10,
-    'batch_size': 4,
-    'obj_evaluator': obj_evaluator,
-    'opt_evaluator': opt_evaluator
+     'acquisition_type': acq_type,
+     'log_level': 'info',
+     'bo_maxiter': 10,
+     'batch_size': 4,
+     'obj_evaluator': obj_evaluator,
+     'opt_evaluator': opt_evaluator
   }
   # Instantiate and run Bayesian Optimization
   bo = BOAlgorithm(problem, gp_model, x_train, y_train, options = options)
   bo.optimize()
 \end{lstlisting}
 
-To run this python script \verb~BODriver.py~, in a python environment in which \verb~mpi4py~, and \Hibbpy are installed do the following:
-
-
+To run this Python script \verb~BODriver.py~, in a Python environment in which \verb~mpi4py~, and \Hibbpy, the following command can be used
 \begin{verbatim}
 > mpirun -np 4 python -m mpi4py.futures BODriver.py 
 \end{verbatim}
-
-To run the same script in a slurm based scheduling system do the following:
-
+In a Slurm-based scheduling system, the command should look be issued as 
 \begin{verbatim}
 > srun -n 4 python -m mpi4py.futures BODriver.py 
 \end{verbatim}

doc/src/sections/hiopbbpy.tex

@@ -11,8 +11,10 @@
 \usetikzlibrary{patterns}
 
 \usepackage{adjustbox}
+\usepackage{textcomp}
 \usepackage{listings}
 
+
 \usepackage{color}
 \definecolor{secblue}{HTML}{3B5E7F}
 \definecolor{mycolor1}{RGB}{184,183,255}
@@ -134,7 +136,7 @@
 \vspace{3cm}
 
   {\huge\bfseries \Hi\ -- User Guide} \\[14pt]
-  {\large\bfseries version 1.1.0}
+  {\large\bfseries version 1.2.0}
 
 \vspace{3cm}
 
@@ -156,7 +158,7 @@
 \vspace{4.75cm}
 
  \textcolor{violet}{{\large\bfseries Oct 15, 2017} \\
-{\large\bfseries Updated Sept 22, 2024}}
+{\large\bfseries Updated Sept 10, 2025}}
 
 \vspace{0.75cm}
 
@@ -252,9 +254,9 @@ \section{Introduction}
 The goal of quasi-Newton solver of \Hi is to remove the parallelization limitations of existing state-of-the-art solvers for nonlinear programming (NLP) and match/surpass the parallel scalability of the underlying PDE or DAE solver. Such limitation occurs whenever the dimensionality of the optimization space  is as large as the dimensionality of the discretization of the differential systems of equations governing the optimization. In these cases, the use of existing NLP solvers  results in i. considerable long time spent in optimization, which affects the parallel scalability, and/or ii. memory requirements beyond the memory capacity of the computational node that runs the optimization. \Hi removes these scalability/parallelization bottlenecks (for certain optimization problems described above) by offering interface for a \textit{memory-distributed} specification of the problem and parallelizing the optimization search using specialized parallel linear algebra technique.
 z
 The general computational approach in \Hi is to use existing state-of-the-art NLP algorithms and develop linear algebra kernels tailored to the specific of this
-class of problems. \Hi is based on an interior-point line search filter method~\cite{waecther_05_ipopt2,waecther_05_ipopt} and follows the implementation details from~\cite{waecther_05_ipopt0}, which is the implementation paper for IPOPT open-source NLP solver. The quasi-Newton approach is based on limited-memory secant approximations of the Hessian~\cite{ByrdNocedalSchnabel_94_quasiNewtonRepres}, which is generalized as required by the specific of interior-point methods for constrained optimization problems~\cite{petra_hiop}. The specialized linear algebra decomposition is obtained by using a Schur-complement reduction that
+class of problems. \Hi is based on an interior-point line search filter method~\cite{waecther_05_ipopt2,waecther_05_ipopt} and follows the implementation details from~\cite{waecther_05_ipopt0}, which is the implementation paper for IPOPT open-source NLP solver. The quasi-Newton approach is based on limited-memory secant approximations of the Hessian~\cite{ByrdNocedalSchnabel_94_quasiNewtonRepres}, which is generalized as required by the specific of interior-point methods for constrained optimization problems~\cite{petra2019memory}. The specialized linear algebra decomposition is obtained by using a Schur-complement reduction that
 leverages the fact that the quasi-Newton Hessian matrix has a small number of
-dense blocks that border a low-rank update of a diagonal matrix. The technique is described  in ~\cite{petra_hiop}. The Newton interior-point solver of \Hi uses linear algebra specialized to the particular form of the MDS NLPs supported by this solver, for more details consult Section~\ref{sec:mds}. 
+dense blocks that border a low-rank update of a diagonal matrix. The technique is described  in~\cite{petra2019memory}. The Newton interior-point solver of \Hi uses linear algebra specialized to the particular form of the MDS NLPs supported by this solver, for more details consult Section~\ref{sec:mds}. 
 
 The C++ parallel implementation in \Hi
 is  lightweight and portable since it is expressed and implemented  only in terms of parallel (multi-)vector operations (implemented internally using BLAS level 1 and level 2 operations and MPI for communication) and BLAS level 3 and LAPACK operations for small dense matrices.
@@ -966,7 +968,8 @@ \subsection{Compiling and linking your project with the \Hi library}
 \begin{itemize}
 \item append to the compiler's include path the location of the HiOP's headers: \begin{verbatim} -Ihiop-dir/include \end{verbatim}
 \item specify \texttt{libhiop.a} to the linker, possibly adding the HiOP's library directory to the linker's libraries paths: 
-\begin{verbatim}-Lhiop-dir/lib -lhiop\end{verbatim} \end{itemize}
+\begin{verbatim}-Lhiop-dir/lib -lhiop\end{verbatim} 
+\end{itemize}
 Here, \texttt{hiop-dir} is the \Hi's distribution directory  (created using \Hi's build system, in particular   by using \texttt{make install} command).
 
 In addition, a shared dynamic load library can be also built by using \texttt{HIOP\_BUILD\_SHARED} option with cmake. 

doc/src/techrep_main.pdf

@@ -1,5 +1,5 @@
 '''
-This is the setup file for installing hiopbbpy
+This is the setup file for installing HiOpBBpy
 
 Authors:    Tucker Hartland <hartland1@llnl.gov>
             Nai-Yuan Chiang <chiang7@llnl.gov>
@@ -18,8 +18,8 @@
 
 metadata = dict(
         name="hiopbbpy",
-        version="0.0.8",
-        description="HiOp black box optimization (hiopbbpy)",
+        version="1.2.0",
+        description="HiOp black box optimization (HiOpBBpy)",
         author="Tucker hartland et al.",
         author_email="hartland1@llnl.gov",
         license="BSD-3",