Merge pull request #61 from felix5572/devel

support equi/gap/xyz

Author: amcadmus

README.md

@@ -51,6 +51,33 @@ The labels provided in the `OUTCAR`, i.e. energies, forces and virials (if any),
 
 The `System` or `LabeledSystem` can be constructed from the following file formats with the `format key` in the table passed to argument `fmt`:
 
+For `quip/gap xyz` files, single .xyz file may include many different configurations with different atom numbers and atom type.
+
+The Class `dpdata.MultiSystems`  can read data  this kind of file.
+Now Class dpdata.MultiSystems Only support quip/gap xyz  format file.
+
+The following commands relating to `Class dpdata.MultiSystems` may be useful.
+```python
+# load data
+
+xyz_multi_systems = dpdata.MultiSystems.from_file('tests/xyz/xyz_unittest.xyz','quip/gap/xyz')
+
+# print the multi_system infomation
+print(xyz_multi_systems)
+print(xyz_multi_systems.systems) # return a dictionaries
+
+# print the system infomation
+print(xyz_multi_systems.systems['B1C9'].data)
+
+# dump a system's data to ./my_work_dir/B1C9_raw folder
+xyz_multi_systems.systems['B1C9'].to_deepmd_raw('./my_work_dir/B1C9_raw')
+
+# dump all systems
+xyz_multi_systems.to_deepmd_raw('./my_deepmd_data/')
+
+
+```
+
 | Software| format | multi frames | labeled | class	    | format key    |
 | ------- | :---   | :---:        | :---:   | :---          | :---          |
 | vasp	  | poscar | False        | False   | System	    | 'vasp/poscar' | 
@@ -70,6 +97,7 @@ The `System` or `LabeledSystem` can be constructed from the following file forma
 | QE      | log    | False        | True    | LabeledSystem | 'qe/pw/scf'   |
 | QE      | log    | True         | False   | System        | 'qe/cp/traj'  |
 | QE      | log    | True         | True    | LabeledSystem | 'qe/cp/traj'  |
+|quip/gap|xyz|True|True|MultiSystems|'quip/gap/xyz'|
 
 ## Access data
 These properties stored in `System` and `LabeledSystem` can be accessed by operator `[]` with the key of the property supplied, for example

dpdata/system.py

@@ -1,3 +1,4 @@
+#%%
 import os
 import numpy as np
 import dpdata.lammps.lmp
@@ -18,6 +19,7 @@
 from monty.json import MSONable
 from monty.serialization import loadfn,dumpfn
 from dpdata.periodic_table import Element
+from dpdata.xyz.quip_gap_xyz import QuipGapxyzSystems
 
 class System (MSONable) :
     '''
@@ -312,8 +314,8 @@ def append(self, system) :
         for ii in ['atom_numbs', 'atom_names'] :
             assert(system.data[ii] == self.data[ii])
         for ii in ['atom_types','orig'] :
-            eq = (system.data[ii] == self.data[ii])
-            assert(eq.all())
+            eq = [v1==v2 for v1,v2 in zip(system.data[ii], self.data[ii])]
+            assert(all(eq))
         for ii in ['coords', 'cells'] :
             self.data[ii] = np.concatenate((self.data[ii], system[ii]), axis = 0)
         return True
@@ -903,7 +905,7 @@ def sort_atom_types(self):
 class MultiSystems:
     '''A set containing several systems.'''
 
-    def __init__(self, *systems, type_map=None):
+    def __init__(self, *systems,type_map=None):
         """
         Parameters
         ----------
@@ -942,6 +944,22 @@ def __add__(self, others) :
        elif isinstance(others, list):
           return self.__class__(self, *others)
        raise RuntimeError("Unspported data structure")
+    
+    @classmethod
+    def from_file(cls,file_name,fmt):
+        multi_systems = cls()
+        multi_systems.load_systems_from_file(file_name=file_name,fmt=fmt)
+        return multi_systems
+    
+    def load_systems_from_file(self, file_name=None, fmt=None):
+        if file_name is not None:
+            if fmt is None:
+                raise RuntimeError("must specify file format for file {}".format(file_name))
+            elif fmt == 'quip/gap/xyz' or 'xyz':
+                self.from_quip_gap_xyz_file(file_name)
+            else:
+                raise RuntimeError("unknown file format for file {} format {},now supported 'quip/gap/xyz'".format(file_name, fmt))
+
 
     def get_nframes(self) :
         """Returns number of frames in all systems"""
@@ -997,6 +1015,14 @@ def check_atom_names(self, system):
             system.add_atom_names(new_in_self)
         system.sort_atom_names()
 
+    def from_quip_gap_xyz_file(self,filename):
+        # quip_gap_xyz_systems = QuipGapxyzSystems(filename)
+        # print(next(quip_gap_xyz_systems))
+        for info_dict in QuipGapxyzSystems(filename):
+            system=LabeledSystem(data=info_dict)
+            self.append(system)
+
+
     def to_deepmd_raw(self, folder) :
         """
         Dump systems in deepmd raw format to `folder` for each system.
@@ -1053,9 +1079,12 @@ def check_LabeledSystem(data):
 
 def elements_index_map(elements,standard=False,inverse=False):
     if standard:
-       elements.sort(key=lambda x: Element(x).Z)
+        elements.sort(key=lambda x: Element(x).Z)
     if inverse:
-       return dict(zip(range(len(elements)),elements))
+        return dict(zip(range(len(elements)),elements))
     else:
-       return dict(zip(elements,range(len(elements))))
+        return dict(zip(elements,range(len(elements))))
+
+
 
+# %%

dpdata/xyz/quip_gap_xyz.py

@@ -0,0 +1,140 @@
+#!/usr/bin/env python3
+
+#%%
+# with open('./test.xyz', 'r') as xyz_file:
+#     lines = xyz_file.readlines()
+#     print(lines)
+#%% 
+import numpy as np
+from collections import OrderedDict
+import re 
+class QuipGapxyzSystems(object):
+    """
+    deal with QuipGapxyzFile 
+    """
+    def __init__(self, file_name):
+        self.file_object = open(file_name, 'r')
+        self.block_generator = self.get_block_generator()
+    def __iter__(self):
+        return self
+    def __next__(self):
+        return self.handle_single_xyz_frame(next(self.block_generator))
+    def __del__(self):
+        self.file_object.close()
+    
+    def get_block_generator(self):
+        p3 = re.compile(r'^\s*(\d+)\s*')
+        while True:
+            line = self.file_object.readline()
+            if not line:
+                break
+            if p3.match(line):
+                atom_num = int(p3.match(line).group(1))
+                lines = []
+                lines.append(line)
+                for ii in range(atom_num+1):
+                    lines.append(self.file_object.readline())
+                if not lines[-1]:
+                    raise RuntimeError("this xyz file may lack of lines, should be {};lines:{}".format(atom_num+2, lines))
+                yield lines
+    
+    @staticmethod
+    def handle_single_xyz_frame(lines):
+        atom_num = int(lines[0].strip('\n').strip())
+        if len(lines) != atom_num + 2:
+            raise RuntimeError("format error, atom_num=={}, {}!=atom_num+2".format(atom_num, len(lines)))
+        data_format_line = lines[1].strip('\n').strip()+str(' ')
+        p1 = re.compile(r'(?P<key>\S+)=(?P<quote>[\'\"]?)(?P<value>.*?)(?P=quote)\s+')
+        p2 = re.compile(r'(?P<key>\w+?):(?P<datatype>[a-zA-Z]):(?P<value>\d+)')
+        field_list = [kv_dict.groupdict() for kv_dict in p1.finditer(data_format_line)]
+        field_dict = {}
+        for item in field_list:
+            field_dict[item['key']]=item['value']
+        data_format_line = lines[1]
+        data_format_list= [m.groupdict() for m in p1.finditer(data_format_line)]
+        field_dict = {}
+        for item in data_format_list:
+            field_dict[item['key']]=item['value']
+
+        Properties = field_dict['Properties']
+        prop_list = [m.groupdict() for m in p2.finditer(Properties)]
+
+        data_lines = []
+        for line in lines[2:]:
+            data_lines.append(list(filter(bool, line.strip().split())))
+        data_array = np.array(data_lines)
+        used_colomn = 0
+
+        type_array = None
+        coords_array = None
+        Z_array = None
+        force_array = None
+        virials = None 
+        for kv_dict in prop_list:
+            if kv_dict['key'] == 'species':
+                if kv_dict['datatype'] != 'S':
+                    raise RuntimeError("datatype for species must be 'S' instead of {}".format(kv_dict['datatype']))
+                field_length = int(kv_dict['value'])
+                type_array = data_array[:,used_colomn:used_colomn+field_length].flatten()
+                used_colomn += field_length
+                continue
+            elif kv_dict['key'] == 'pos':
+                if kv_dict['datatype'] != 'R':
+                    raise RuntimeError("datatype for pos must be 'R' instead of {}".format(kv_dict['datatype']))
+                field_length = int(kv_dict['value'])
+                coords_array = data_array[:,used_colomn:used_colomn+field_length]
+                used_colomn += field_length
+                continue
+            elif kv_dict['key'] == 'Z':
+                if kv_dict['datatype'] != 'I':
+                    raise RuntimeError("datatype for pos must be 'R' instead of {}".format(kv_dict['datatype']))
+                field_length = int(kv_dict['value'])
+                Z_array = data_array[:,used_colomn:used_colomn+field_length].flatten()
+                used_colomn += field_length
+                continue
+            elif kv_dict['key'] == 'force':
+                if kv_dict['datatype'] != 'R':
+                    raise RuntimeError("datatype for pos must be 'R' instead of {}".format(kv_dict['datatype']))
+                field_length = int(kv_dict['value'])
+                force_array = data_array[:,used_colomn:used_colomn+field_length]
+                used_colomn += field_length
+                continue
+            else:
+                raise RuntimeError("unknown field {}".format(kv_dict['key']))
+
+        type_num_dict = OrderedDict()
+        atom_type_list = []
+        type_map = {}
+        temp_atom_max_index = 0
+        if type_array is None:
+            raise RuntimeError("type_array can't be None type, check .xyz file")
+        for ii in type_array:
+            if ii not in type_map:
+                type_map[ii] = temp_atom_max_index
+                temp_atom_max_index += 1
+                temp_atom_index = type_map[ii]
+                atom_type_list.append(temp_atom_index)
+                type_num_dict[ii] = 1
+            else:
+                temp_atom_index = type_map[ii]
+                atom_type_list.append(temp_atom_index)
+                type_num_dict[ii] += 1
+        type_num_list = []
+        for atom_type,atom_num in type_num_dict.items():
+            type_num_list.append((atom_type,atom_num))
+        type_num_array = np.array(type_num_list)
+        if field_dict.get('virial', None):
+            virials = np.array([np.array(list(filter(bool,field_dict['virial'].split(' ')))).reshape(3,3)]).astype('float32')
+        else:
+            virials = None
+        info_dict = {}
+        info_dict['atom_names'] = list(type_num_array[:,0])
+        info_dict['atom_numbs'] = list(type_num_array[:,1].astype(int))
+        info_dict['atom_types'] = np.array(atom_type_list).astype(int)
+        info_dict['cells'] = np.array([np.array(list(filter(bool,field_dict['Lattice'].split(' ')))).reshape(3,3)]).astype('float32')
+        info_dict['coords'] = np.array([coords_array]).astype('float32')
+        info_dict['energies'] = np.array([field_dict['energy']]).astype('float32')
+        info_dict['forces'] = np.array([force_array]).astype('float32')
+        info_dict['virials'] = virials
+        info_dict['orig'] = [0,0,0]
+        return info_dict

setup.py

@@ -24,7 +24,7 @@
     long_description=readme,
     long_description_content_type="text/markdown",
     url="https://github.com/deepmodeling/dpdata",
-    packages=['dpdata', 'dpdata/vasp', 'dpdata/lammps', 'dpdata/md', 'dpdata/deepmd', 'dpdata/qe', 'dpdata/siesta', 'dpdata/gaussian', 'dpdata/cp2k'],
+    packages=['dpdata', 'dpdata/vasp', 'dpdata/lammps', 'dpdata/md', 'dpdata/deepmd', 'dpdata/qe', 'dpdata/siesta', 'dpdata/gaussian', 'dpdata/cp2k','dpdata/xyz'],
     package_data={'dpdata':['*.json']},
     classifiers=[
         "Programming Language :: Python :: 3.6",

tests/test_quip_gap_xyz.py

@@ -0,0 +1,31 @@
+import os
+import numpy as np
+import unittest
+from context import dpdata
+from comp_sys import CompLabeledSys
+
+class TestQuipGapxyz(unittest.TestCase, CompLabeledSys):
+    def setUp (self) :
+        self.multi_systems = dpdata.MultiSystems.from_file('xyz/xyz_unittest.xyz','quip/gap/xyz')
+        self.system_1 = self.multi_systems.systems['B1C9']
+        self.system_2 = dpdata.LabeledSystem('xyz/B1C9', fmt='deepmd')
+        self.places = 6
+        self.e_places = 6
+        self.f_places = 6
+        self.v_places = 4
+
+class TestQuipGapxyz2(unittest.TestCase, CompLabeledSys):
+    def setUp (self) :
+        self.system_temp0 = dpdata.MultiSystems.from_file(file_name='xyz/xyz_unittest.xyz', fmt='quip/gap/xyz')
+        self.system_1 = self.system_temp0.systems['B5C7']
+        self.system_temp1 = dpdata.LabeledSystem('xyz/B1C9', fmt='deepmd')
+        self.system_temp2 = dpdata.LabeledSystem('xyz/B5C7', fmt='deepmd')
+        self.system_temp3 = dpdata.MultiSystems(self.system_temp2, self.system_temp1)
+        self.system_2 = self.system_temp3.systems['B5C7']
+        self.places = 6
+        self.e_places = 6
+        self.f_places = 6
+        self.v_places = 4
+
+if __name__ == '__main__':
+    unittest.main()

tests/xyz/B1C9/box.raw

@@ -0,0 +1,2 @@
+9.217081069946289062e+00 0.000000000000000000e+00 0.000000000000000000e+00 4.859999762629740871e-07 6.431267261505126953e+00 0.000000000000000000e+00 4.201562881469726562e+00 4.099999983964153216e-08 2.205334901809692383e+00
+5.512073516845703125e+00 0.000000000000000000e+00 0.000000000000000000e+00 4.160000059982849052e-07 5.512073516845703125e+00 0.000000000000000000e+00 2.529999960643181112e-07 2.529999960643181112e-07 3.347159147262573242e+00

tests/xyz/B1C9/coord.raw

@@ -0,0 +1,2 @@
+1.238022994995117188e+01 3.215630054473876953e+00 2.131030082702636719e+00 1.296782970428466797e+01 4.328790187835693359e+00 2.141720056533813477e+00 1.296782970428466797e+01 2.102469921112060547e+00 2.141720056533813477e+00 7.954239845275878906e+00 1.034819960594177246e+00 2.102900028228759766e+00 7.954239845275878906e+00 5.396450042724609375e+00 2.102900028228759766e+00 4.896999895572662354e-01 5.378680229187011719e+00 3.689999878406524658e-02 4.896999895572662354e-01 1.052590012550354004e+00 3.689999878406524658e-02 2.719000101089477539e+00 1.711969971656799316e+00 1.720999926328659058e-01 2.719000101089477539e+00 4.719299793243408203e+00 1.720999926328659058e-01 9.699810028076171875e+00 3.215630054473876953e+00 5.539500117301940918e-01
+4.085810184478759766e+00 4.576230049133300781e+00 2.942379951477050781e+00 1.426270008087158203e+00 9.358400106430053711e-01 2.942379951477050781e+00 9.358400106430053711e-01 4.085810184478759766e+00 2.942379951477050781e+00 4.576230049133300781e+00 1.426270008087158203e+00 2.942379951477050781e+00 2.756040096282958984e+00 2.756040096282958984e+00 3.207600116729736328e-01 2.949359893798828125e+00 1.197110056877136230e+00 3.320899963378906250e+00 2.562710046768188477e+00 4.314960002899169922e+00 3.320899963378906250e+00 4.314960002899169922e+00 2.949359893798828125e+00 3.320899963378906250e+00 1.197110056877136230e+00 2.562710046768188477e+00 3.320899963378906250e+00 2.756040096282958984e+00 2.756040096282958984e+00 1.720729947090148926e+00

tests/xyz/B1C9/energy.raw

@@ -0,0 +1,2 @@
+-5.834249877929687500e+01
+-5.639742660522460938e+01

tests/xyz/B1C9/force.raw

@@ -0,0 +1,2 @@
+-9.355509757995605469e+00 -0.000000000000000000e+00 -2.236460000276565552e-01 7.056536197662353516e+00 1.119171047210693359e+01 1.800871014595031738e+00 7.056534767150878906e+00 -1.119171047210693359e+01 1.800871014595031738e+00 1.710104942321777344e+00 -3.803577899932861328e+00 -1.240210011601448059e-01 1.710106015205383301e+00 3.803577899932861328e+00 -1.240210011601448059e-01 4.354149997234344482e-01 7.533950209617614746e-01 -6.208680272102355957e-01 4.354149997234344482e-01 -7.533950209617614746e-01 -6.208680272102355957e-01 -3.430412054061889648e+00 5.665700137615203857e-02 -4.403569996356964111e-01 -3.430412054061889648e+00 -5.665700137615203857e-02 -4.403569996356964111e-01 -2.187779903411865234e+00 -0.000000000000000000e+00 -1.007603049278259277e+00
+3.989300131797790527e-02 -1.872578978538513184e+00 9.720240235328674316e-01 -3.989300131797790527e-02 1.872578978538513184e+00 9.720240235328674316e-01 1.872578978538513184e+00 3.989300131797790527e-02 9.720240235328674316e-01 -1.872578978538513184e+00 -3.989300131797790527e-02 9.720240235328674316e-01 -0.000000000000000000e+00 -0.000000000000000000e+00 -1.597074985504150391e+00 7.413330078125000000e-01 -2.438414096832275391e+00 -2.333703041076660156e+00 -7.413340210914611816e-01 2.438414096832275391e+00 -2.333703041076660156e+00 2.438414096832275391e+00 7.413340210914611816e-01 -2.333703041076660156e+00 -2.438414096832275391e+00 -7.413340210914611816e-01 -2.333703041076660156e+00 9.999999974752427079e-07 9.999999974752427079e-07 7.043791770935058594e+00

tests/xyz/B1C9/type.raw

@@ -0,0 +1,10 @@
+0
+1
+1
+1
+1
+1
+1
+1
+1
+1