Merge pull request #613 from amcadmus/master

Merge recent devel into master

Author: amcadmus

README.md

@@ -22,6 +22,7 @@
 
 [![GitHub release](https://img.shields.io/github/release/deepmodeling/dpgen.svg?maxAge=86400)](https://github.com/deepmodeling/dpgen/releases/)
 [![doi:10.1016/j.cpc.2020.107206](https://img.shields.io/badge/DOI-10.1016%2Fj.cpc.2020.107206-blue)](https://doi.org/10.1016/j.cpc.2020.107206)
+![Citations](https://citations.njzjz.win/10.1016/j.cpc.2020.107206)
 [![conda install](https://img.shields.io/conda/dn/conda-forge/dpgen?label=conda%20install)](https://anaconda.org/conda-forge/dpgen)
 [![pip install](https://img.shields.io/pypi/dm/dpgen?label=pip%20install)](https://pypi.org/project/dpgen)
 
@@ -544,20 +545,21 @@ The bold notation of key (such aas **type_map**) means that it's a necessary key
 | training_iter0_model_path  | list of string  |  ["/path/to/model0_ckpt/", ...]  | The model used to init the first iter training. Number of element should be equal to `numb_models` |
 | training_init_model  | bool  |  False  | Iteration > 0, the model parameters will be initilized from the model trained at the previous iteration. Iteration == 0, the model parameters will be initialized from `training_iter0_model_path`.  |
 | **default_training_param** | Dict |  | Training parameters for `deepmd-kit` in `00.train`. <br /> You can find instructions from here: (https://github.com/deepmodeling/deepmd-kit)..<br /> |
+| dp_compress | bool | false | Use `dp compress` to compress the model. Default is false. |
 | *#Exploration*
 | **model_devi_dt** | Float | 0.002 (recommend) | Timestep for MD |
 | **model_devi_skip** | Integer | 0 | Number of structures skipped for fp in each MD
-| **model_devi_f_trust_lo** | Float | 0.05 | Lower bound of forces for the selection.
- | **model_devi_f_trust_hi** | Float | 0.15 | Upper bound of forces for the selection
-| **model_devi_v_trust_lo**  | Float | 1e10                                                         | Lower bound of virial for the selection. Should be used with DeePMD-kit v2.x |
-| **model_devi_v_trust_hi**  | Float | 1e10                                                         | Upper bound of virial for the selection. Should be used with DeePMD-kit v2.x |
+| **model_devi_f_trust_lo** | Float or List of float | 0.05 | Lower bound of forces for the selection. If List, should be set for each index in `sys_configs`, respectively. |
+| **model_devi_f_trust_hi** | Float or List of float | 0.15 | Upper bound of forces for the selection. If List, should be set for each index in `sys_configs`, respectively. |
+| **model_devi_v_trust_lo**  | Float or List of float | 1e10 | Lower bound of virial for the selection. If List, should be set for each index in `sys_configs`, respectively. Should be used with DeePMD-kit v2.x. |
+| **model_devi_v_trust_hi**  | Float or List of float | 1e10 | Upper bound of virial for the selection. If List, should be set for each index in `sys_configs`, respectively. Should be used with DeePMD-kit v2.x. |
 | model_devi_adapt_trust_lo  | Boolean | False | Adaptively determines the lower trust levels of force and virial. This option should be used together with `model_devi_numb_candi_f`,  `model_devi_numb_candi_v` and optionally with `model_devi_perc_candi_f` and `model_devi_perc_candi_v`. `dpgen` will make two sets: 1. From the frames with force model deviation lower than `model_devi_f_trust_hi`, select `max(model_devi_numb_candi_f, model_devi_perc_candi_f*n_frames)` frames with largest force model deviation. 2. From the frames with virial model deviation lower than `model_devi_v_trust_hi`, select `max(model_devi_numb_candi_v, model_devi_perc_candi_v*n_frames)` frames with largest virial model deviation. The union of the two sets is made as candidate dataset|
 | model_devi_numb_candi_f  | Int | 10 | See `model_devi_adapt_trust_lo`.|
 | model_devi_numb_candi_v  | Int | 0  | See `model_devi_adapt_trust_lo`.|
 | model_devi_perc_candi_f  | Float | 0.0 | See `model_devi_adapt_trust_lo`.|
 | model_devi_perc_candi_v  | Float | 0.0 | See `model_devi_adapt_trust_lo`.|
 | model_devi_f_avg_relative | Boolean | False | Normalized the force model deviations by the RMS force magnitude along the trajectory. This key should not be used with `use_relative`. |
-| **model_devi_clean_traj**  | Boolean | true                                                         | Deciding whether to clean traj folders in MD since they are too large. |
+| **model_devi_clean_traj**  | Boolean or Int | true                                                         | If type of model_devi_clean_traj is boolean type then it denote whether to clean traj folders in MD since they are too large. If it is Int type, then the most recent n iterations of traj folders will be retained, others will be removed. |
 | **model_devi_nopbc**  | Boolean | False                                                         | Assume open boundary condition in MD simulations. |
 | model_devi_activation_func | List of list of string | [["tanh","tanh"],["tanh","gelu"],["gelu","tanh"],["gelu","gelu"]]	| Set activation functions for models, length of the List should be the same as `numb_models`, and two elements in the list of string respectively assign activation functions to the embedding and fitting nets within each model. *Backward compatibility*: the orginal "List of String" format is still supported, where embedding and fitting nets of one model use the same activation function, and the length of the List should be the same as `numb_models`|
 | **model_devi_jobs**        | [<br/>{<br/>"sys_idx": [0], <br/>"temps": <br/>[100],<br/>"press":<br/>[1],<br/>"trj_freq":<br/>10,<br/>"nsteps":<br/> 1000,<br/> "ensembles": <br/> "nvt" <br />},<br />...<br />] | List of dict | Settings for exploration in `01.model_devi`. Each dict in the list corresponds to one iteration. The index of `model_devi_jobs` exactly accord with index of iterations |

conda/meta.yaml

@@ -25,6 +25,9 @@ requirements:
     - requests
     - dpdata
     - dpdispatcher
+    - ase
+    - GromacsWrapper
+    - custodian
 
   run:
     - python >=3.6
@@ -34,6 +37,9 @@ requirements:
     - dpdata
     - dpdispatcher
     - pymatgen
+    - ase
+    - GromacsWrapper
+    - custodian
 
 test:
   imports:

dpgen/auto_test/EOS.py

@@ -169,8 +169,8 @@ def _compute_lower(self,
                 # vol = self.vol_start + ii * self.vol_step
                 vol = loadfn(os.path.join(all_tasks[ii], 'eos.json'))['volume']
                 task_result = loadfn(all_res[ii])
-                res_data[vol] = task_result['energies'][-1] / task_result['atom_numbs'][0]
-                ptr_data += '%7.3f  %8.4f \n' % (vol, task_result['energies'][-1] / task_result['atom_numbs'][0])
+                res_data[vol] = task_result['energies'][-1] / task_result['atom_numbs'].sum()
+                ptr_data += '%7.3f  %8.4f \n' % (vol, task_result['energies'][-1] / task_result['atom_numbs'].sum())
                 # res_data[vol] = all_res[ii]['energy'] / len(all_res[ii]['force'])
                 # ptr_data += '%7.3f  %8.4f \n' % (vol, all_res[ii]['energy'] / len(all_res[ii]['force']))
 

dpgen/auto_test/VASP.py

@@ -140,8 +140,8 @@ def make_input_file(self,
                 incar['ENCUT'] = cal_setting['encut']
 
             if 'kspacing' in cal_setting:
-                dlog.info("%s setting KSAPCING to %s" % (self.make_input_file.__name__, cal_setting['kspacing']))
-                incar['KSAPCING'] = cal_setting['kspacing']
+                dlog.info("%s setting KSPACING to %s" % (self.make_input_file.__name__, cal_setting['kspacing']))
+                incar['KSPACING'] = cal_setting['kspacing']
 
             if 'kgamma' in cal_setting:
                 dlog.info("%s setting KGAMMA to %s" % (self.make_input_file.__name__, cal_setting['kgamma']))

dpgen/auto_test/common_equi.py

@@ -2,6 +2,7 @@
 import os
 import warnings
 from monty.serialization import dumpfn
+from multiprocessing import Pool
 
 import dpgen.auto_test.lib.crys as crys
 import dpgen.auto_test.lib.util as util
@@ -12,6 +13,8 @@
 from distutils.version import LooseVersion
 from dpgen.dispatcher.Dispatcher import make_submission
 from dpgen.remote.decide_machine import convert_mdata
+from dpgen.auto_test.lib.utils import create_path
+
 lammps_task_type = ['deepmd', 'meam', 'eam_fs', 'eam_alloy']
 
 
@@ -76,11 +79,11 @@ def make_equi(confs,
         poscar = os.path.abspath(os.path.join(ii, 'POSCAR'))
         if not os.path.exists(poscar):
             raise FileNotFoundError('no configuration for autotest')
+        if os.path.exists(os.path.join(ii, 'relaxation', 'jr.json')):
+            os.remove(os.path.join(ii, 'relaxation', 'jr.json'))
+
         relax_dirs = os.path.abspath(os.path.join(ii, 'relaxation', 'relax_task'))    # to be consistent with property in make dispatcher
-        if os.path.exists(relax_dirs):
-            dlog.warning('%s already exists' % relax_dirs)
-        else:
-            os.makedirs(relax_dirs)
+        create_path(relax_dirs)
         task_dirs.append(relax_dirs)
         os.chdir(relax_dirs)
         # copy POSCARs to mp-xxx/relaxation/relax_task
@@ -111,6 +114,47 @@ def make_equi(confs,
         inter.make_input_file(ii, 'relaxation', relax_param)
 
 
+def worker(work_path,
+           run_task,
+           forward_common_files,
+           forward_files,
+           backward_files,
+           mdata,
+           inter_type):
+    machine, resources, command, group_size = util.get_machine_info(mdata, inter_type)
+    disp = make_dispatcher(machine, resources, work_path, [run_task], group_size)
+    print("%s --> Runing... " % (work_path))
+
+    api_version = mdata.get('api_version', '0.9')
+    if LooseVersion(api_version) < LooseVersion('1.0'):
+        warnings.warn(f"the dpdispatcher will be updated to new version."
+                      f"And the interface may be changed. Please check the documents for more details")
+        disp.run_jobs(resources,
+                      command,
+                      work_path,
+                      [run_task],
+                      group_size,
+                      forward_common_files,
+                      forward_files,
+                      backward_files,
+                      outlog='outlog',
+                      errlog='errlog')
+    elif LooseVersion(api_version) >= LooseVersion('1.0'):
+        submission = make_submission(
+            mdata_machine=machine,
+            mdata_resources=resources,
+            commands=[command],
+            work_path=work_path,
+            run_tasks=run_task,
+            group_size=group_size,
+            forward_common_files=forward_common_files,
+            forward_files=forward_files,
+            backward_files=backward_files,
+            outlog='outlog',
+            errlog='errlog'
+        )
+        submission.run_submission()
+
 def run_equi(confs,
              inter_param,
              mdata):
@@ -120,6 +164,11 @@ def run_equi(confs,
     for conf in confs:
         conf_dirs.extend(glob.glob(conf))
     conf_dirs.sort()
+
+    processes = len(conf_dirs)
+    pool = Pool(processes=processes)
+    print("Submit job via %d processes" % processes)
+
     # generate a list of task names like mp-xxx/relaxation/relax_task
     # ...
     work_path_list = []
@@ -150,45 +199,28 @@ def run_equi(confs,
     if len(run_tasks) == 0:
         return
     else:
-        # if LooseVersion()
         run_tasks = [os.path.basename(ii) for ii in all_task]
         machine, resources, command, group_size = util.get_machine_info(mdata, inter_type)
         print('%d tasks will be submited '%len(run_tasks))
+        multiple_ret = []
         for ii in range(len(work_path_list)):
             work_path = work_path_list[ii]
-            disp = make_dispatcher(machine, resources, work_path, [run_tasks[ii]], group_size)
-            print("%s --> Runing... "%(work_path))
-
-        api_version = mdata.get('api_version', '0.9')
-        if LooseVersion(api_version) < LooseVersion('1.0'):
-            warnings.warn(f"the dpdispatcher will be updated to new version."
-                f"And the interface may be changed. Please check the documents for more details")
-            disp.run_jobs(resources,
-                            command,
-                            work_path,
-                            [run_tasks[ii]],
-                            group_size,
-                            forward_common_files,
-                            forward_files,
-                            backward_files,
-                            outlog='outlog',
-                            errlog='errlog')
-        elif LooseVersion(api_version) >= LooseVersion('1.0'):
-            submission = make_submission(
-                mdata_machine=machine,
-                mdata_resource=resources,
-                commands=[command],
-                work_path=work_path,
-                run_tasks=run_tasks,
-                group_size=group_size,
-                forward_common_files=forward_common_files,
-                forward_files=forward_files,
-                backward_files=backward_files,
-                outlog = 'outlog',
-                errlog = 'errlog'
-            )
-            submission.run_submission()
 
+            ret = pool.apply_async(worker, (work_path,
+                                            run_tasks[ii],
+                                            forward_common_files,
+                                            forward_files,
+                                            backward_files,
+                                            mdata,
+                                            inter_type,
+                                            ))
+            multiple_ret.append(ret)
+        pool.close()
+        pool.join()
+        for ii in range(len(multiple_ret)):
+            if not multiple_ret[ii].successful():
+                raise RuntimeError("Task %d is not successful! work_path: %s " % (ii, work_path_list[ii]))
+        print('finished')
 
 def post_equi(confs, inter_param):
     # find all POSCARs and their name like mp-xxx

dpgen/auto_test/common_prop.py

@@ -15,6 +15,7 @@
 from dpgen.dispatcher.Dispatcher import make_dispatcher
 from dpgen.dispatcher.Dispatcher import make_submission
 from dpgen.remote.decide_machine import convert_mdata
+from dpgen.auto_test.lib.utils import create_path
 lammps_task_type = ['deepmd', 'meam', 'eam_fs', 'eam_alloy']
 
 
@@ -73,10 +74,7 @@ def make_property(confs,
             path_to_equi = os.path.join(ii, 'relaxation', 'relax_task')
             path_to_work = os.path.join(ii, property_type + '_' + suffix)
 
-            if os.path.exists(path_to_work):
-                dlog.warning('%s already exists' % path_to_work)
-            else:
-                os.makedirs(path_to_work)
+            create_path(path_to_work)
 
             prop = make_property_instance(jj)
             task_list = prop.make_confs(path_to_work, path_to_equi, do_refine)
@@ -112,6 +110,7 @@ def run_property(confs,
     conf_dirs.sort()
     task_list = []
     work_path_list = []
+    multiple_ret = []
     for ii in conf_dirs:
         sepline(ch=ii, screen=True)
         for jj in property_list:
@@ -159,7 +158,7 @@ def run_property(confs,
             all_task = tmp_task_list
             run_tasks = util.collect_task(all_task, inter_type)
             if len(run_tasks) == 0:
-                return
+                continue
             else:
                 ret = pool.apply_async(worker, (work_path,
                                                 all_task,
@@ -169,23 +168,13 @@ def run_property(confs,
                                                 mdata,
                                                 inter_type,
                                                 ))
-            # run_tasks = [os.path.basename(ii) for ii in all_task]
-            # machine, resources, command, group_size = util.get_machine_info(mdata, inter_type)
-            # disp = make_dispatcher(machine, resources, work_path, run_tasks, group_size)
-            # disp.run_jobs(resources,
-            #               command,
-            #               work_path,
-            #               run_tasks,
-            #               group_size,
-            #               forward_common_files,
-            #               forward_files,
-            #               backward_files,
-            #               outlog='outlog',
-            #               errlog='errlog')
+                multiple_ret.append(ret)
     pool.close()
     pool.join()
-    if ret.successful():
-        print('finished')
+    for ii in range(len(multiple_ret)):
+        if not multiple_ret[ii].successful():
+            raise RuntimeError("Job %d is not successful!" % ii)
+    print('%d jobs are finished' % len(multiple_ret))
 
 
 def worker(work_path,
@@ -215,7 +204,7 @@ def worker(work_path,
     elif LooseVersion(api_version) >= LooseVersion('1.0'):
         submission = make_submission(
                 mdata_machine=machine,
-                mdata_resource=resources,
+                mdata_resources=resources,
                 commands=[command],
                 work_path=work_path,
                 run_tasks=run_tasks,

dpgen/data/gen.py

@@ -464,7 +464,7 @@ def make_vasp_md(jdata, mdata) :
 
     for ii in sys_ps :
         for jj in scale :
-            for kk in range(pert_numb) :
+            for kk in range(pert_numb+1) :
                 path_work = path_md
                 path_work = os.path.join(path_work, ii)
                 path_work = os.path.join(path_work, "scale-%.3f" % jj)
@@ -534,6 +534,8 @@ def coll_vasp_md(jdata) :
                     #dlog.info("md_nstep", md_nstep)
                     if nforce == md_nstep :
                         valid_outcars.append(outcar)
+                    elif md_nstep == 0 and nforce == 1 :
+                        valid_outcars.append(outcar)
                     else:
                         dlog.info("WARNING : in directory %s nforce in OUTCAR is not equal to settings in INCAR"%(os.getcwd()))
         arg_cvt = " "
@@ -767,13 +769,14 @@ def gen_init_bulk(args) :
             dlog.info("Current stage is 1, relax")
             create_path(out_dir)
             shutil.copy2(args.PARAM, os.path.join(out_dir, 'param.json'))
+            skip_relax = jdata['skip_relax']
             if from_poscar :
                 make_super_cell_poscar(jdata)
             else :
                 make_unit_cell(jdata)
                 make_super_cell(jdata)
                 place_element(jdata)
-            if args.MACHINE is not None:
+            if args.MACHINE is not None and not skip_relax:
                make_vasp_relax(jdata, mdata)
                run_vasp_relax(jdata, mdata)
             else:

dpgen/dispatcher/Dispatcher.py

@@ -346,11 +346,17 @@ def make_dispatcher(mdata, mdata_resource=None, work_path=None, run_tasks=None,
 def make_submission(mdata_machine, mdata_resources, commands, work_path, run_tasks, group_size,
     forward_common_files, forward_files, backward_files, outlog, errlog):
 
-    machine = Machine.load_from_dict(mdata_machine)
-    resources = Resources.load_from_dict(mdata_resources)
-
     if mdata_machine['local_root'] != './':
         raise RuntimeError(f"local_root must be './' in dpgen's machine.json.")
+    
+    abs_local_root = os.path.abspath('./')
+
+    abs_mdata_machine = mdata_machine.copy()
+    abs_mdata_machine['local_root'] = abs_local_root
+
+    machine = Machine.load_from_dict(abs_mdata_machine)
+    resources = Resources.load_from_dict(mdata_resources)
+
 
     command = "&&".join(commands)
 

dpgen/dispatcher/LocalContext.py

@@ -117,7 +117,13 @@ def download(self,
                         pass
                     elif (os.path.exists(rfile)) and (not os.path.exists(lfile)) :
                         # trivial case, download happily
-                        shutil.move(rfile, lfile)
+                        # If the file to be downloaded is a softlink, `cp` should be performed instead of `mv`.
+                        # Otherwise, `lfile` is still a file linked to some original file,
+                        # and when this file's removed, `lfile` will be invalid.
+                        if os.path.islink(rfile):
+                            shutil.copyfile(rfile,lfile)
+                        else:
+                            shutil.move(rfile, lfile)
                     elif (os.path.exists(rfile)) and (os.path.exists(lfile)) :
                         # both exists, replace!
                         dlog.info('find existing %s, replacing by %s' % (lfile, rfile))