mlatom 1.2.3__tar.gz → 2.0.0__tar.gz

{MLatom-1.2.3/src/MLatom.egg-info → MLatom-2.0.0}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: MLatom
-Version: 1.2.3
+Version: 2.0.0
 Summary: A Package for Atomistic Simulations with Machine Learning
 Home-page: http://mlatom.com
 Author: Pavlo O. Dral
@@ -20,7 +20,8 @@ Description: # Brief Introduction
         - Estimating accuracy of ML models.
         - Creating ML model and saving it to a file.
         - Loading existing ML model from a file and performing ML calculations with this model.
-        - ML-cross-section to boost the calculation of ML-NEA
+        - ML-accelerated calculation of absorption spectra within nuclear ensemble approach
+        - Learning curves
         
         ## Data Set Operations
         - Converting XYZ coordinates into an input vector (molecular descriptor) for ML.
@@ -54,14 +55,14 @@ Description: # Brief Introduction
           - [ sorted by norms of its rows ](http://dx.doi.org/10.1021/ct400195d);
           - unsorted;
           - permuted.
-        - [ Normalized inverted internuclear distances ](http://mlatom.com/self-correcting-machine-learning-and-structure-based-sampling/)
+        - [ Normalized inverse internuclear distances (RE descriptor)](http://mlatom.com/self-correcting-machine-learning-and-structure-based-sampling/)
           - sorted for user-defined atoms by the sum of their nuclear repulsions to all other atoms;
           - unsorted;
           - permuted.
         
         
         # The KREG model
-        Now we set [ KREG (Kernel-ridge-regression using RE descriptor and the Gaussian kernel function )](http://dx.doi.org/10.1021/acs.jpclett.8b02469) as default ML method. Default hyperparameter optimization options were updated to speed up the hyperparameter search for KREG.
+        The [ KREG (Kernel-ridge-regression using RE descriptor and the Gaussian kernel function )](http://dx.doi.org/10.1021/acs.jpclett.8b02469) model is the default ML method.
         
         
         # Model Validation
@@ -81,12 +82,15 @@ Description: # Brief Introduction
         
         Note that hyperparameter tuning can be performed together with model validation. This means that for example one can perform outer loop of the cross-validation for model validation and tune hyperparameters via inner loop of the cross-validation.
         
+        Apart from natively implemented logarithmic grid search for hyperparameters, MLatom also provides the interface to the [ hyperopt package ](http://hyperopt.github.io/hyperopt/) implementing hyperparameter optimization using Bayesian methods with Tree-structured Parzen Estimator (TPE).
+        
         
         # First Derivatives
-        MLatom can be also used to estimate first derivatives from an ML model. It can be useful, if an ML model has been created for energies, and one wants to estimate forces. Two scenarios are possible:
+        MLatom can be also used to estimate first derivatives from an ML model. Two scenarios are possible:
         
-        - partial derivatives are calculated for each dimension of given input vectors (analytical derivatives for Gaussian kernel, numerical derivatives for other kernels);
-        - first derivatives are calculated in XYZ coordinates for input files containing molecular XYZ coordinates (only numerical derivatives).
+        - partial derivatives are calculated for each dimension of given input vectors (analytical derivatives for Gaussian and Matern kernels);
+        - first derivatives are calculated in XYZ coordinates for input files containing molecular XYZ coordinates (analytical derivatives for the RE and Coulomb matrix descriptors).
+        - derivatives for interfaced models
         
         
         # Cross Section
@@ -97,12 +101,21 @@ Description: # Brief Introduction
         > Bao-Xin Xue, Mario Barbatti*, Pavlo O. Dral*, [ Machine Learning for Absorption Cross Sections ](https://doi.org/10.1021/acs.jpca.0c05310), J. Phys. Chem. A 2020, 124, 7199–7210. DOI: 10.1021/acs.jpca.0c05310.
         > [ Preprint on ChemRxiv ](https://doi.org/10.26434/chemrxiv.12594191), DOI: 10.26434/chemrxiv.12594191.
         
+        # Interfaces to 3<sup>rd</sup>-party software
+        MLatom also provides interfaces to some third-party software where extra ML model types are natively implemented. It allows users to access other popular ML model types within MLatom's workflow. Currently available third-party model types are:
+        
+        - [ANI](https://doi.org/10.1039/c6sc05720a) (through [TorchANI](https://doi.org/10.1021/acs.jcim.0c00451)) 
+        - [DeepPot-SE](https://papers.nips.cc/paper/2018/hash/e2ad76f2326fbc6b56a45a56c59fafdb-Abstract.html) and [DPMD](https://doi.org/10.1103/PhysRevLett.120.143001) (through [DeePMD-kit](https://doi.org/10.1016/j.cpc.2018.03.016)) 
+        - [GAP](https://doi.org/10.1103/Physrevlett.104.136403)-[SOAP](https://doi.org/10.1103/physrevb.87.184115) (through [GAP](www.libatoms.org) suite and [QUIP](http://github.com/libAtoms/QUIP))  
+        - [PhysNet](https://doi.org/10.1021/acs.jctc.9b00181) (through [PhysNet](github.com/MMunibas/PhysNet)) 
+        - [sGDML](https://doi.org/10.1038/s41467-018-06169-2) (through [sGDML](www.sgdml.org)) 
+        
         # About Program
         MLatom: a Package for Atomistic Simulations with Machine Learning    
-        Version 1.2  
+        Version 2.0  
         http://mlatom.com/                             
                                                                                    
-        Copyright (c) 2013-2020 Pavlo O. Dral                   
+        Copyright (c) 2013-2021 Pavlo O. Dral                   
         http://dr-dral.com/                            
                                                                                    
         All rights reserved. No part of MLatom may be used, published or redistributed without written permission by Pavlo Dral.  
@@ -110,8 +123,10 @@ Description: # Brief Introduction
         The software is provided "as is", without warranty of any kind, express  or implied, including but not limited to the warranties of merchantability, fitness for a particular purpose and noninfringement. In no event shall the authors or copyright holders be liable for any claim, damages or other liability, whether in an action of contract, tort or otherwise, arising from, out of or in connection with the software or the use or other dealings in the software.                                   
                                                                                    
         Cite as: 
+        
         1. Pavlo O. Dral, J. Comput. Chem. 2019, 40, 2339-2347             
-        2. Pavlo O. Dral, Bao-Xin Xue, Fuchun Ge, Yi-Fan Hou, MLatom: A Package for Atomistic Simulations with Machine Learning, version 1.2, Xiamen University, Xiamen, China, 2013-2020.               
+        2. Pavlo O. Dral, Fuchun Ge, Bao-Xin Xue, Yi-Fan Hou, Max Pinheiro Jr, Jianxing Huang, Mario Barbatti, Top. Curr. Chem. 2021, 379, 27              
+        3. Pavlo O. Dral, Bao-Xin Xue, Fuchun Ge, Yi-Fan Hou, Max Pinheiro Jr, MLatom: A Package for Atomistic Simulations with Machine Learning, version 2.0, Xiamen University, Xiamen, China, 2013-2021.               
         
         
         # License

{MLatom-1.2.3 → MLatom-2.0.0}/README.md

@@ -12,7 +12,8 @@ A brief overview of MLatom capabilities. See sections below for more details.
 - Estimating accuracy of ML models.
 - Creating ML model and saving it to a file.
 - Loading existing ML model from a file and performing ML calculations with this model.
-- ML-cross-section to boost the calculation of ML-NEA
+- ML-accelerated calculation of absorption spectra within nuclear ensemble approach
+- Learning curves
 
 ## Data Set Operations
 - Converting XYZ coordinates into an input vector (molecular descriptor) for ML.
@@ -46,14 +47,14 @@ Permutationally invariant kernel and self-correction are also supported.
   - [ sorted by norms of its rows ](http://dx.doi.org/10.1021/ct400195d);
   - unsorted;
   - permuted.
-- [ Normalized inverted internuclear distances ](http://mlatom.com/self-correcting-machine-learning-and-structure-based-sampling/)
+- [ Normalized inverse internuclear distances (RE descriptor)](http://mlatom.com/self-correcting-machine-learning-and-structure-based-sampling/)
   - sorted for user-defined atoms by the sum of their nuclear repulsions to all other atoms;
   - unsorted;
   - permuted.
 
 
 # The KREG model
-Now we set [ KREG (Kernel-ridge-regression using RE descriptor and the Gaussian kernel function )](http://dx.doi.org/10.1021/acs.jpclett.8b02469) as default ML method. Default hyperparameter optimization options were updated to speed up the hyperparameter search for KREG.
+The [ KREG (Kernel-ridge-regression using RE descriptor and the Gaussian kernel function )](http://dx.doi.org/10.1021/acs.jpclett.8b02469) model is the default ML method.
 
 
 # Model Validation
@@ -73,12 +74,15 @@ Gaussian, Laplacian, and Matérn kernels have σ and λ tunable hyperparameters.
 
 Note that hyperparameter tuning can be performed together with model validation. This means that for example one can perform outer loop of the cross-validation for model validation and tune hyperparameters via inner loop of the cross-validation.
 
+Apart from natively implemented logarithmic grid search for hyperparameters, MLatom also provides the interface to the [ hyperopt package ](http://hyperopt.github.io/hyperopt/) implementing hyperparameter optimization using Bayesian methods with Tree-structured Parzen Estimator (TPE).
+
 
 # First Derivatives
-MLatom can be also used to estimate first derivatives from an ML model. It can be useful, if an ML model has been created for energies, and one wants to estimate forces. Two scenarios are possible:
+MLatom can be also used to estimate first derivatives from an ML model. Two scenarios are possible:
 
-- partial derivatives are calculated for each dimension of given input vectors (analytical derivatives for Gaussian kernel, numerical derivatives for other kernels);
-- first derivatives are calculated in XYZ coordinates for input files containing molecular XYZ coordinates (only numerical derivatives).
+- partial derivatives are calculated for each dimension of given input vectors (analytical derivatives for Gaussian and Matern kernels);
+- first derivatives are calculated in XYZ coordinates for input files containing molecular XYZ coordinates (analytical derivatives for the RE and Coulomb matrix descriptors).
+- derivatives for interfaced models
 
 
 # Cross Section
@@ -89,12 +93,21 @@ In brief, this feature uses fewer QC calculation to achieve higher precision and
 > Bao-Xin Xue, Mario Barbatti*, Pavlo O. Dral*, [ Machine Learning for Absorption Cross Sections ](https://doi.org/10.1021/acs.jpca.0c05310), J. Phys. Chem. A 2020, 124, 7199–7210. DOI: 10.1021/acs.jpca.0c05310.
 > [ Preprint on ChemRxiv ](https://doi.org/10.26434/chemrxiv.12594191), DOI: 10.26434/chemrxiv.12594191.
 
+# Interfaces to 3<sup>rd</sup>-party software
+MLatom also provides interfaces to some third-party software where extra ML model types are natively implemented. It allows users to access other popular ML model types within MLatom's workflow. Currently available third-party model types are:
+
+- [ANI](https://doi.org/10.1039/c6sc05720a) (through [TorchANI](https://doi.org/10.1021/acs.jcim.0c00451)) 
+- [DeepPot-SE](https://papers.nips.cc/paper/2018/hash/e2ad76f2326fbc6b56a45a56c59fafdb-Abstract.html) and [DPMD](https://doi.org/10.1103/PhysRevLett.120.143001) (through [DeePMD-kit](https://doi.org/10.1016/j.cpc.2018.03.016)) 
+- [GAP](https://doi.org/10.1103/Physrevlett.104.136403)-[SOAP](https://doi.org/10.1103/physrevb.87.184115) (through [GAP](www.libatoms.org) suite and [QUIP](http://github.com/libAtoms/QUIP))  
+- [PhysNet](https://doi.org/10.1021/acs.jctc.9b00181) (through [PhysNet](github.com/MMunibas/PhysNet)) 
+- [sGDML](https://doi.org/10.1038/s41467-018-06169-2) (through [sGDML](www.sgdml.org)) 
+
 # About Program
 MLatom: a Package for Atomistic Simulations with Machine Learning    
-Version 1.2  
+Version 2.0  
 http://mlatom.com/                             
                                                                            
-Copyright (c) 2013-2020 Pavlo O. Dral                   
+Copyright (c) 2013-2021 Pavlo O. Dral                   
 http://dr-dral.com/                            
                                                                            
 All rights reserved. No part of MLatom may be used, published or redistributed without written permission by Pavlo Dral.  
@@ -102,8 +115,10 @@ The above copyright notice and this permission notice shall be included  in all
 The software is provided "as is", without warranty of any kind, express  or implied, including but not limited to the warranties of merchantability, fitness for a particular purpose and noninfringement. In no event shall the authors or copyright holders be liable for any claim, damages or other liability, whether in an action of contract, tort or otherwise, arising from, out of or in connection with the software or the use or other dealings in the software.                                   
                                                                            
 Cite as: 
+
 1. Pavlo O. Dral, J. Comput. Chem. 2019, 40, 2339-2347             
-2. Pavlo O. Dral, Bao-Xin Xue, Fuchun Ge, Yi-Fan Hou, MLatom: A Package for Atomistic Simulations with Machine Learning, version 1.2, Xiamen University, Xiamen, China, 2013-2020.               
+2. Pavlo O. Dral, Fuchun Ge, Bao-Xin Xue, Yi-Fan Hou, Max Pinheiro Jr, Jianxing Huang, Mario Barbatti, Top. Curr. Chem. 2021, 379, 27              
+3. Pavlo O. Dral, Bao-Xin Xue, Fuchun Ge, Yi-Fan Hou, Max Pinheiro Jr, MLatom: A Package for Atomistic Simulations with Machine Learning, version 2.0, Xiamen University, Xiamen, China, 2013-2021.               
 
 
 # License

{MLatom-1.2.3 → MLatom-2.0.0}/setup.py

@@ -43,7 +43,7 @@ with open(os.path.join(here, 'README.md')) as f:
 
 setup(
     name = pkg_name,
-    version = "1.2.3",
+    version = "2.0.0",
     author = "Pavlo O. Dral",
     author_email = "admin@mlatom.com",
     license = 'CC BY-NC-ND 4.0',  # https://creativecommons.org/licenses/by-nc-nd/4.0/
@@ -55,7 +55,7 @@ setup(
     packages = find_packages('src'),
     package_dir = {'' : 'src'},
     scripts = ['mlatom', 'MLatomF'],
-    package_data = {"": ['MLatomF', '*.json']},
+    package_data = {"": ['MLatomF', '*.json', 'cs.so']},
     classifiers = [
         "Programming Language :: Python :: 3",
         "Operating System :: POSIX :: Linux",
@@ -63,7 +63,7 @@ setup(
         "Topic :: Scientific/Engineering :: Chemistry",
         "Environment :: Console",
     ],
-    # install_requires = [],
+    install_requires = ['numpy', 'matplotlib'],
     # entry_points = {
     #     'console_scripts' : ['mlatom=MLatom.MLatom:MLatomMainCls']
     # }

MLatom-1.2.3/src/MLatom/ML-NEA.py → MLatom-2.0.0/src/MLatom/ML_NEA.py

@@ -1,5 +1,6 @@
 #!/usr/bin/env python3
 
+import random
 import time
 from typing import Callable, Iterable, List, Optional, Tuple, Union, Any
 import inspect
@@ -9,6 +10,7 @@ import math
 from subprocess import getstatusoutput, Popen, PIPE
 import multiprocessing as mp
 from contextlib import redirect_stdout
+from ctypes import c_double, CDLL, c_long 
 from functools import wraps, reduce
 from itertools import product
 import json
@@ -19,7 +21,7 @@ try:
     import matplotlib.pyplot as plt
 except:
     pass
-''' version: 1.3   script author: Bao-Xin Xue   Xiamen University
+''' version: 1.5   script author: Bao-Xin Xue   Xiamen University
 ┌────────────────────────────────────────────────────────────────────────────────────┐
 │                                                                                    │
 │   ┌────────────────────────────────┐          ┌────────────────────────────────┐   │
@@ -68,6 +70,9 @@ except:
 _Str_or_Float = Union[str, float]
 _Int_or_Float = Union[int, float]
 
+# python script path
+py_script_path = ''
+
 class Error(Exception):
     def __init__(self):
         # todo: to define various error
@@ -105,7 +110,8 @@ def logger(log_content: str, logfile: str='out.log') -> Callable[[Any], Any]:
             result: Any = func(*args, **kwargs)
             end_time = time.time()
 
-            log_string = f'{log_content} ( {func.__name__} ) finished at {time.ctime()} {time.tzname[0]} |||| totally spent {end_time - start_time} sec'
+            spend_time = end_time - start_time
+            log_string = f'{log_content} ( {func.__name__} ) finished at {time.ctime()} {time.tzname[0]} |||| total spent {spend_time:.2f} sec'
             log_string += '\n' + '=' * 90 + '\n\n'
             print(log_string)
 
@@ -308,6 +314,13 @@ class Molecule():
         if atom_num != self.atom_num:
             # todo 2: raise xyz atom number error problem
             pass
+    
+    @staticmethod
+    def read_element_label(input_file: str, idx: int=1) -> List[str]:
+        idx -= 1
+        with open(input_file) as f:
+            data = f.readlines()
+        return [line.split()[idx] for line in data]
 
     @classmethod
     def read_multi_xyz(cls, file_path: str, au_unit: bool=False) -> List['Molecule']:
@@ -409,7 +422,7 @@ class Plot():
             self.ax.grid()
         self.fig.savefig(out_img)
     
-    def plot_multiple_file(self, files: List[str], labels: List[str], colors: Optional[List[str]]=None, out_img_name: str='all.png', x_column: int=1, y_column: int=2):
+    def plot_multiple_file(self, files: List[str], labels: List[str], colors: Optional[List[str]]=None, x_column: int=1, y_column: int=2):
         if not EnvCheck.check_matplotlib():
             return
         plt.rcParams['legend.fontsize'] = 'x-large'
@@ -438,10 +451,17 @@ class Plot():
         # self.ax.set(xlabel=self.xlabel, ylabel=self.ylabel, title=self.title)
         self.ax.set_xlabel(self.xlabel, fontsize=18)
         self.ax.set_ylabel(self.ylabel, fontsize=18)
+        # if self.grid:
+        #     self.ax.grid()
+        # self.fig.savefig(out_img_name)
+    
+    def plot_spc_bar(self, x_data: List[float], y_data: List[float], width: float=0.025, color: str='g') -> None:
+        self.ax.bar(x_data, y_data, width=width, color=color)
+        
+    def save_fig(self, fig_name: str):
         if self.grid:
             self.ax.grid()
-        self.fig.savefig(out_img_name)
-        
+        self.fig.savefig(fig_name)
 
     def column_data(self, i: int) -> List[float]:
         result = []
@@ -751,6 +771,7 @@ class GaussCalc(LinuxInterface):
         
 
 class NewtonXCalc(LinuxInterface):
+    au = 0.5291772083
     def __init__(self, work_path: str) -> None:
         self.work_path = work_path
         self.temp_path = os.path.join(work_path, 'TEMP')
@@ -781,7 +802,7 @@ class NewtonXCalc(LinuxInterface):
         # check whether trnasfer success (delete /dev/null and then check the output log)
         self.exec(f'$NX/xyz2nx < {geom_path} &> /dev/null')
         
-    def run_dynamic(self, freq_file: str, atom_num: int, NEA_point: int) -> None:
+    def run_dynamic_old(self, freq_file: str, atom_num: int, NEA_point: int) -> None:
         self.mv(freq_file, os.path.join(self.NX_path, 'freq.out') )
         nx_inp = '&dat\n nact = 2\n numat = %d\n npoints = %d\n file_geom = geom\n iprog = 4\n' % (atom_num, NEA_point) + \
                  ' file_nmodes = freq.out\n anh_f = 1\n temp = 0.0\n ics_flg = n\n chk_e = 0\n' + \
@@ -792,8 +813,49 @@ class NewtonXCalc(LinuxInterface):
         if err_code != 0 or out != '':
             err_print(f'Newton-X calculation failed!!!\nstatus code: {err_code} (0 stands for normal exit, none 0 stands for error)\noutput: {out}')
         # todo 3: remove NX temp folder and other uncessary folder
-        
+    
+    def run_dynamic(self, freq_file: str, atom_num: int, NEA_point: int) -> None:
+        inp_name = 'initqp_input'
+        self.mv(freq_file, os.path.join(self.NX_path, 'freq.out') )
+        nx_inp = '&dat\n nact = 2\n numat = %d\n npoints = %d\n file_geom = "geom"\n iprog = 4\n' % (atom_num, NEA_point) + \
+                 ' file_nmodes = "freq.out"\n anh_f = 1\n temp = 0.0\n ics_flg = "n"\n chk_e = 0\n' + \
+                 ' nis = 1\n nfs = 2\n kvert = 1\n de = 100\n prog = 6.5\n iseed = 1234\n lvprt = 1\n/\n'
+        with open(inp_name, 'w') as f:
+            f.write(nx_inp)
+        # err_code, out = self.exec(f'$NX/initcond.pl > NX-MD.log 2>&1')
+        err_code, out = self.exec('$NX/mk_qvector < initqp_input ; ' \
+                                  '$NX/weight < initqp_input &> weight.log ; ' \
+                                  '$NX/initqp < initqp_input &> initqp.log ')
+        if err_code != 0 or out != '':
+            err_print(f'Newton-X calculation failed!!!\nstatus code: {err_code} (0 stands for normal exit, none 0 stands for error)\noutput: {out}')
+    
     def restore_multi_xyz(self, atom_num: int):
+        with open('ini_qv') as f:
+            data = f.readlines()
+        elem = Molecule.read_element_label('geom')
+        cnt = 0
+        mole = Molecule()
+        for line in data[1:]:
+            tmp = line.split()
+            if len(tmp) == 2:
+                # if cnt:
+                self.multi_xyz.append(mole)
+                mole = Molecule()
+                cnt = 0
+            elif len(tmp) == 6:
+                x, y, z = self._process_ini_qv_line(tmp)
+                atom = Atom(x, y, z, elem[cnt])
+                cnt += 1
+                mole.atoms.append(atom)
+            else:
+                err_print('processing ini_qv failed! exit!')
+                exit()
+    
+    def _process_ini_qv_line(self, splited_line: List[str]) -> List[float]:
+        xyz = map(lambda x: float(x.replace('D', 'e')), splited_line[0: 3])
+        return list(map(lambda x: self.au * x, xyz))
+    
+    def restore_multi_xyz_old(self, atom_num: int):
         # working directory: In self.NX_path
         awk_inp = '''atom_num=%d \n''' % (atom_num) + \
                  r'''sed -n /Geometry/,/Velocity\ in\ NX/p final_output | sed  /Velocity/d | awk '{printf ("%s\t", $1)}''' + \
@@ -987,18 +1049,14 @@ class ML_calc(LinuxInterface):
         with open(orig_file) as f:
             orig = f.read().splitlines()
         orig_num = len(orig)
-        rtn_value: LineDate = LineDate('')
+        mixed_data = []
         for idx, e in enumerate(est):
             if idx > orig_num - 1:
-                # use est value
-                if float(e) < 0:
-                    rtn_value += '0'
-                else:
-                    rtn_value += e
-            else:
-                # use orig value
-                rtn_value += orig[idx]
-        return rtn_value.string
+                tmp = e if float(e) > 0.0 else '0.0' 
+            else:  # use orig value
+                tmp = orig[idx]
+            mixed_data.append(tmp)
+        return '\n'.join(mixed_data)
     
     def mix_multiple_file(self, max_state: int, est_path: str, orig_path: str, out_path: str) -> None:
         for i in range(1, max_state + 1):
@@ -1038,6 +1096,8 @@ class ML_calc(LinuxInterface):
         if not self.exist_path(path):
             self.mkdir(path)
         self.cd(path)
+        with open('ml.inp', 'w') as f:
+            f.write(self.inp.string)
         s = StringIO()
         with redirect_stdout(s):
             prg = os.path.join(self.program_path, 'MLatomF')
@@ -1046,6 +1106,8 @@ class ML_calc(LinuxInterface):
                 print(proc.stdout.read().decode('utf-8'))
                 err = proc.stderr.read().decode('utf-8')
         self.ML_output = s.getvalue()
+        with open('log', 'w') as f:
+            f.write(self.ML_output)
         if err:
             warning_print(f'MLatom.run_ML returns a non-empty error: {err}, PAY ATTENTION to it.')
         self.inp: LineDate = LineDate('')
@@ -1173,13 +1235,13 @@ class CrossSection_calc(LinuxInterface):
         #####   data pre-process     #####
         E_max = -10000
         E_min = 10000
-        whole_num_point = 0
+        # whole_num_point = 0
         for E_data in self.E_data_list:
             tmp_max = max(E_data)
             tmp_min = min(E_data)
             E_max = tmp_max if tmp_max > E_max else E_max
             E_min = tmp_min if tmp_min < E_min else E_min
-        whole_num_point = len(self.E_data_list[0])  # todo: check the E and f is has the same number 
+        # whole_num_point = len(self.E_data_list[0])  # todo: check the E and f is has the same number 
         range_min = E_min - self.kappa * self.delta 
         range_max = E_max + self.kappa * self.delta
         self.x_min = range_min
@@ -1194,13 +1256,48 @@ class CrossSection_calc(LinuxInterface):
         while de < range_max:
             de_list.append(de)
             de += self.eps
-        pool = mp.Pool()
-        result_list = pool.map(self._calc_at_de, de_list)
-        pool.close()
-        pool.join()
-        for each in result_list:
-            de, wavelength, cross_section = each
-            file_handle.write('%.4f   %.4f     %.8f\n' % (de, wavelength, cross_section))
+        
+        whole_num_point = len(self.E_data_list[0])
+        c_E_list = ((c_double * whole_num_point) * self.max_state)()
+        c_f_list = ((c_double * whole_num_point) * self.max_state)()
+        for state_index, E_data in enumerate(self.E_data_list):
+            f_data = self.f_data_list[state_index]
+            for target_index, E in enumerate(E_data):
+                f = f_data[target_index]
+                c_E_list[state_index][target_index] = E
+                c_f_list[state_index][target_index] = f
+        de_num = len(de_list)
+        c_cs = (c_double * de_num)()
+        c_de_list = (c_double* de_num)()
+        wavelength = []
+        for idx, de in enumerate(de_list):
+            c_de_list[idx] = de
+            wavelength.append(self.ev2nm / de)
+        cs = CDLL(os.path.join(py_script_path, 'cs.so'))
+        c_delta = c_double(self.delta)
+        c_coeff = c_double(self.coeff)
+        c_cs_coeff = c_double(self.cs_coeff)
+        c_de_num = c_long(de_num)
+        # print(f'coeff: {self.coeff}, divided: {self.devided_num}')
+        cs.cs_calc(c_E_list, c_f_list, 
+                   self.max_state, whole_num_point,
+                   c_delta, c_coeff, c_de_num, c_cs_coeff,
+                   c_cs, c_de_list)
+        cross_section = []
+        for cs_single in c_cs:
+            cross_section.append(cs_single)
+        for idx, de in enumerate(de_list):
+            w, c = wavelength[idx], cross_section[idx]
+            file_handle.write('%.4f   %.4f     %.8f\n' % (de, w, c))
+        
+        
+        # pool = mp.get_context('spawn').Pool()
+        # result_list = pool.map(self._calc_at_de, de_list)
+        # pool.close()
+        # pool.join()
+        # for each in result_list:
+        #     de, wavelength, cross_section = each
+        #     file_handle.write('%.4f   %.4f     %.8f\n' % (de, wavelength, cross_section))
         file_handle.close()
 
     def _gauss(self, E: float, dE_0n: float) -> float:
@@ -1296,7 +1393,8 @@ class ML_cross_section(CrossSection_calc, ML_calc):
     def ML_train_all(self):
         # todo 3: to to record log
         self._iter_calc_QC(0, self.TD_point)
-        self._train_epoch(self.TD_point)
+        rmse = self._train_epoch(self.TD_point)
+        print(f'\nRMSE_geom value for {self.TD_point} point: {rmse}')
         self._train_final(self.TD_point)
         self._collect_ML_result(self.use_model_pred_path)
         self.mix_multiple_file(self.max_state, self.pred_data_path, self.this.gauss.gauss_data_path, self.mix_data_path)
@@ -1400,7 +1498,7 @@ class ML_cross_section(CrossSection_calc, ML_calc):
             self.set_model_inp(MLmodelIn=f'{sym_state}.unf')
             self.add_other('molDescriptor', 'RE'); self.add_other('molDescrType', 'unsorted')
             self.run_ML(os.getcwd())
-        
+
     def _collect_ML_result(self, final_path: str):
         for sym, state in product(['E', 'f'], range(1, self.max_state + 1)):
             sym_state = sym + str(state)
@@ -2204,7 +2302,7 @@ class RecurPaperProcedure(ArgParser, LinuxInterface):
         print_citation()
         
     def invoke_by_API(self, Nexcitations: int=3, nMaxPoints: int=10000, nNEApoints: int=0, MLpoints: int=500, deltaQCNEA: float=0.05):
-        print_citation()
+        #print_citation()
         self.pre_define()
         ##### args #####
         self.max_state: int = Nexcitations
@@ -2222,6 +2320,8 @@ class RecurPaperProcedure(ArgParser, LinuxInterface):
         ##### file existance #####
         self.work_path = os.getcwd()
         self.py_script_path   = os.path.dirname(os.path.abspath(__file__))  #self.py_script_path   = os.path.abspath(__file__[:__file__.rfind('/')])
+        global py_script_path
+        py_script_path = self.py_script_path
         self.gauss_opt_inp_name: str = 'gaussian_optfreq.com' # self.gauss_opt_inp_name: str = 'opt-freq.com'
         self.gauss_TD_inp_name: str = 'gaussian_ef.com' # self.gauss_TD_inp_name: str = 'TD.com'
         self.gauss_opt_log_name: str = 'gaussian_optfreq.log' # self.gauss_opt_log_name: str = 'opt-freq.log'
@@ -2315,6 +2415,7 @@ class RecurPaperProcedure(ArgParser, LinuxInterface):
         self.cd(self.cs_TD.cs_path)
         self.cs_TD.calc(self.gauss.gauss_data_path, self.max_state, self.delta_QC, self.cs_TD.TD_out_name)  # use default delta=0.05
         opt_E_list, opt_f_list = self.gauss.process_one_gauss_log(os.path.join(self.gauss.all_log_path, '0.log'))
+        self.opt_E_list, self.opt_f_list = opt_E_list, opt_f_list
         # self.cs_TD.calc_SPC(opt_E_list, opt_f_list, self.delta_SPC, self.cs_TD.SPC_out_name)
         self.cs_SPC.read_data_from_para(opt_E_list, opt_f_list)
         SPC_delta = 0.3
@@ -2342,12 +2443,13 @@ class RecurPaperProcedure(ArgParser, LinuxInterface):
             plot.plot_file(self.cs_TD.SPC_out_name, spc_nea_name, x_column=1, y_column=3)
             # plot.plot_file(self.cs_TD.SPC_out_name, 'SPC-cross-section.png', x_column=1, y_column=3)
         else:
-            file_names = []
-            legend_names = []
-            colors = []
+            file_names: List[str] = []
+            legend_names: List[str] = []
+            colors: List[str] = []
+            plot = Plot()
             if self.exist_path(os.path.join(self.work_path, 'cross-section_ref.dat')):
                 file_names.append('cross-section_ref.dat')
-                legend_names.append('ref-NEA')
+                legend_names.append('ref')
                 colors.append('k')
                 self.cp(os.path.join(self.work_path, 'cross-section_ref.dat'), self.cs_ML.cs_path)
             if self.exist_path(os.path.join(self.cs_ML.cs_path, self.cs_TD.TD_out_name)) and plotQCNEA:
@@ -2362,6 +2464,8 @@ class RecurPaperProcedure(ArgParser, LinuxInterface):
                 file_names.append(self.cs_SPC.SPC_out_name)
                 legend_names.append('QC-SPC')
                 colors.append('g')
+                plot.plot_spc_bar(self.opt_E_list, self.opt_f_list)
+                # plot.ax.set_xlim(min(self.opt_E_list), max(self.opt_E_list))
             if False:
                 if self.exist_path(os.path.join(self.work_path, 'cross-section_ref.dat')):
                     # file_names.append('cross-section_ref.dat')
@@ -2377,9 +2481,11 @@ class RecurPaperProcedure(ArgParser, LinuxInterface):
             if not file_names:
                 err_print('not caluclated cross-section file exists. Exit!')
                 exit()
-            plot = Plot()
             plot.set_info(xlabel='Energy, eV', ylabel=r'Cross section, $\mathrm{\AA^2}$ molecule$^{–1}$', grid=False)
-            plot.plot_multiple_file(file_names, legend_names, colors, out_img_name='plot.png', x_column=1, y_column=3)
+            plot.plot_multiple_file(file_names, legend_names, colors, x_column=1, y_column=3)
+            plot.ax.set_ylim(bottom=0.0)
+            plot.ax.margins(x=0.0)
+            plot.save_fig('plot.png')
 
     def detect_E_f_and_calc(self):
         for i in range(1, self.max_state + 1):
@@ -2490,7 +2596,7 @@ def parse_api(args: List[str]):
                 plotQCNEA = True
             elif splited[0].strip() == 'plotQCSPC':
                 plotQCSPC = True
-            elif splited[0].strip() == 'cross-section':
+            elif splited[0].strip() in ['crossSection', 'cross-section', 'cross_section', 'MLNEA', 'ML-NEA', 'ML_NEA']:
                 pass
             else:
                 err_print('unrecognised parameter: %s. exit!' % splited[0])
@@ -2502,8 +2608,10 @@ def parse_api(args: List[str]):
                 nMaxPoints = transfer(splited[1], 'nMaxPoints', int)
             elif splited[0].strip().lower() == nQMpoints_str.lower():
                 nQMpoints = transfer(splited[1], 'nQMpoints', int)
-            elif splited[0].strip().lower() == MLpoints_str.lower():
-                MLpoints = transfer(splited[1], 'MLpoints', int)
+            elif splited[0].strip().lower() in [MLpoints_str.lower(), 'nNEpoints'.lower()]:
+                MLpoints = transfer(splited[1], 'nNEpoints', int)
+            # elif splited[0].strip().lower() == MLpoints_str.lower():
+            #     MLpoints = transfer(splited[1], 'MLpoints', int)
             elif splited[0].strip().lower() == deltaQCNEA_str.lower():
                 deltaQCNEA = transfer(splited[1], 'deltaQCNEA', float)
             else:
@@ -2521,7 +2629,7 @@ def print_help():
     usage: MLatom.py cross-section [optional arguments]
 
   optional arguments:
-    Nexcitations=N      number of excited states to calculate. (default=3)
+    nExcitations=N      number of excited states to calculate. (default=3)
     nQMpoints=N         user-defined number of QM calculations for training ML. 
                         (default=0, number of QM calculations will be determined iteratively)
     plotQCNEA           requests plotting QC-NEA cross section
@@ -2530,7 +2638,7 @@ def print_help():
   
   advanced arguments:
     nMaxPoints=N        maximum number of QC calculations in the iterative procedure. (default=10000)
-    MLpoints=N          number of ML calculations. (default=50000)
+    nNEpoints=N          number of ML ensemble prediction. (default=50000)
   
   Environment setting (no needed)
     $NX                     Newton-X environment
@@ -2582,6 +2690,7 @@ def print_citation():
     print('\n %s:\n\t%s'%('NEWTON-X',cs_algorithm))
     print('*'*80)
 
+
 if __name__ == "__main__":
     ControlPartProxy.invoke_paper_recur_self()
     # ControlPartProxy.invoke_paper_recur_api(nTDpoints=200, plotQCNEA=True, deltaQCNEA=0.08)