Merge pull request #18 from usnistgov/develop

Develop

README.md

@@ -7,24 +7,49 @@ Installation
 -------------------------
 First create a conda environment:
 Install miniconda environment from https://conda.io/miniconda.html
+Based on your system requirements, you'll get a file something like 'Miniconda3-latest-XYZ'.
+
+Now,
+
 ```
 bash Miniconda3-latest-Linux-x86_64.sh (for linux)
 bash Miniconda3-latest-MacOSX-x86_64.sh (for Mac)
+```
 Download 32/64 bit python 3.6 miniconda exe and install (for windows)
 Now, let's make a conda environment, say "version", choose other name as you like::
+```
 conda create --name version python=3.8
 source activate version
 ```
-Now, let's install the package
+
+Now, let's install the package:
 ```
 git clone https://github.com/usnistgov/alignn.git
 cd alignn
 python setup.py develop
 ```
-Example
+Examples
 ---------
+Users can keep their structure files in POSCAR, .cif, or .xyz files in a directory. In the examples below we will use POSCAR format files. In the same directory, there should be id_prop.csv file.
+In this id_prop.csv, the filenames, and correponding target values are kept in comma separated values (csv) format.
+Here is an example of training OptB88vdw bandgaps of 50 materials from JARVIS-DFT. The example is created using the script provided in the script folder.
+Users can modify the script more than 50 data, or make their own dataset in this format. The dataset in split in 80:10:10 as training-validation-test set.
+With the configuration parameters given in config_example_regrssion.json, the model is trained.
+
 ```
 python alignn/scripts/train_folder.py --root_dir "alignn/examples/sample_data" --config "alignn/examples/sample_data/config_example_regrssion.json"
 ```
+While the above example is for regression, the follwoing example shows a classification task for metal/non-metal based on the above bandgap values. We transform the dataset
+into 1 or 0 based on a threshold of 0.01 eV (controlled by the parameter, 'classification_threshold') and train a similar classification model.
+```
+python alignn/scripts/train_folder.py --root_dir "alignn/examples/sample_data" --config "alignn/examples/sample_data/config_example_classification.json"
+```
+While the above example regression was for single-output values, we can train multi-output regression models as well.
+An example is given below for training formation energy per atom, bandgap and total energy per atom simulataneously. The script to generate the example data is provided in the script folder of the sample_data_multi_prop. 
+Another example of training electron and phonon density of states is provided also.
+```
+python alignn/scripts/train_folder_multi_prop.py --root_dir "alignn/examples/sample_data_multi_prop" --config "alignn/examples/sample_data/config_example_regrssion.json"
+```
 
-You can also try multiple example scripts to run multiple dataset training. Look into the 'scripts' folder.
+You can also try multiple example scripts to run multiple dataset training. Look into the 'scripts' folder. 
+These scripts automatically download datasets from jarvis.db.fighshare module in JARVIS-Tools and train several models. Make sure you specify your specific queuing system details in the scripts. 

alignn/__init__.py

@@ -1,2 +1,2 @@
 """Version number."""
-__version__ = "2021.06.20"
+__version__ = "2021.07.05"

alignn/config.py

@@ -72,6 +72,11 @@
     "shear modulus",
     "elastic anisotropy",
     "U0",
+    "HOMO",
+    "LUMO",
+    "R2",
+    "ZPVE",
+    "omega1",
     "mu",
     "alpha",
     "homo",
@@ -86,6 +91,7 @@
     "A",
     "B",
     "C",
+    "all",
     "target",
     "max_efg",
     "avg_elec_mass",
@@ -95,6 +101,9 @@
     "_oqmd_stability",
     "edos_up",
     "pdos_elast",
+    "bandgap",
+    "energy_total",
+    "net_magmom",
 ]
 
 
@@ -112,10 +121,12 @@ class TrainingConfig(BaseSettings):
         "mp_3d_2020",
         "qm9",
         "qm9_dgl",
+        "qm9_std_jctc",
         "user_data",
         "oqmd_3d_no_cfid",
         "edos_up",
         "edos_pdos",
+        "qmof",
     ] = "dft_3d"
     target: TARGET_ENUM = "formation_energy_peratom"
     atom_features: Literal["basic", "atomic_number", "cfid", "cgcnn"] = "cgcnn"

alignn/data.py

@@ -21,7 +21,8 @@
 
 # from sklearn.pipeline import Pipeline
 import pickle as pk
-from sklearn.decomposition import PCA  # ,KernelPCA
+
+# from sklearn.decomposition import PCA  # ,KernelPCA
 from sklearn.preprocessing import StandardScaler
 
 # use pandas progress_apply
@@ -296,6 +297,29 @@ def get_train_val_loaders(
             )
             print("Converting target data into 1 and 0.")
         all_targets = []
+
+        # TODO:make an all key in qm9_dgl
+        if dataset == "qm9_dgl" and target == "all":
+            print("Making all qm9_dgl")
+            tmp = []
+            for ii in d:
+                ii["all"] = [
+                    ii["mu"],
+                    ii["alpha"],
+                    ii["homo"],
+                    ii["lumo"],
+                    ii["gap"],
+                    ii["r2"],
+                    ii["zpve"],
+                    ii["U0"],
+                    ii["U"],
+                    ii["H"],
+                    ii["G"],
+                    ii["Cv"],
+                ]
+                tmp.append(ii)
+            print("Made all qm9_dgl")
+            d = tmp
         for i in d:
             if isinstance(i[target], list):  # multioutput target
                 all_targets.append(torch.tensor(i[target]))
@@ -347,8 +371,20 @@ def get_train_val_loaders(
         if standard_scalar_and_pca:
             y_data = [i[target] for i in dataset_train]
             # pipe = Pipeline([('scale', StandardScaler())])
+            if not isinstance(y_data[0], list):
+                print("Running StandardScalar")
+                y_data = np.array(y_data).reshape(-1, 1)
             sc = StandardScaler()
+
             sc.fit(y_data)
+            print("Mean", sc.mean_)
+            print("Variance", sc.var_)
+            try:
+                print("New max", max(y_data))
+                print("New min", min(y_data))
+            except Exception as exp:
+                print(exp)
+                pass
             # pc = PCA(n_components=output_features)
             # pipe = Pipeline(
             #    [
@@ -357,9 +393,9 @@ def get_train_val_loaders(
             #    ]
             # )
             pk.dump(sc, open("sc.pkl", "wb"))
-            pc = PCA(n_components=40)
-            pc.fit(y_data)
-            pk.dump(pc, open("pca.pkl", "wb"))
+            # pc = PCA(n_components=10)
+            # pc.fit(y_data)
+            # pk.dump(pc, open("pca.pkl", "wb"))
 
         if classification_threshold is None:
             try:

alignn/examples/sample_data_multi_prop/POSCAR-JVASP-10.vasp

@@ -0,0 +1,11 @@
+System
+1.0
+1.6777483798834445 -2.9059452409270157 -1.1e-15
+1.6777483798834438 2.9059452409270126 -7e-16
+-6.5e-15 -8e-16 6.220805465667012
+V Se
+1 2
+direct
+0.0 0.0 0.0 
+0.6666669999999968 0.3333330000000032 0.7479606991085345 
+0.3333330000000032 0.6666669999999968 0.252039300891465 

alignn/examples/sample_data_multi_prop/POSCAR-JVASP-107772.vasp

@@ -0,0 +1,12 @@
+System
+1.0
+4.376835486482439 0.0086562096165887 7.148251977291244
+2.0211490103296166 3.8822313794698684 7.148251977291244
+0.0142338540946976 0.008656214510917 8.38176620441039
+Bi Sb
+3 1
+direct
+0.11687114695010013 0.11687114695010009 0.11687114695134818 
+0.885057350916569 0.8850573509165686 0.8850573509144881 
+0.3806761740317465 0.3806761740317465 0.3806761740305207 
+0.6173953281015849 0.6173953281015848 0.6173953281036431 

alignn/examples/sample_data_multi_prop/POSCAR-JVASP-1372.vasp

@@ -0,0 +1,10 @@
+System
+1.0
+3.5058938597621094 -3.081249e-10 2.0241289627124215
+1.1686312862968669 3.3053879820023613 2.0241289627124215
+-8.715088e-10 -6.162497e-10 4.048256928443838
+Al As
+1 1
+direct
+0.0 0.0 0.0 
+0.24999999999999997 0.25 0.24999999999999997 

alignn/examples/sample_data_multi_prop/POSCAR-JVASP-14014.vasp

@@ -0,0 +1,20 @@
+System
+1.0
+4.157436115454804 -0.0 0.0
+-0.0 7.038494037846648 0.0
+0.0 0.0 7.411178065479046
+Tb Mn Si
+4 4 4
+direct
+0.25 0.5014870121743014 0.1863876024079978 
+0.75 0.49851298782569875 0.8136123975920022 
+0.25 0.0014870121743013 0.3136123975920026 
+0.75 0.9985129878256985 0.6863876024079979 
+0.75 0.8608859093979077 0.06027225572734521 
+0.25 0.1391140906020922 0.939727744272655 
+0.75 0.3608859093979077 0.43972774427265476 
+0.25 0.6391140906020923 0.5602722557273451 
+0.25 0.792634730215007 0.8919146343653592 
+0.75 0.207365269784993 0.10808536563464112 
+0.25 0.2926347302150071 0.6080853656346409 
+0.75 0.707365269784993 0.3919146343653588 

alignn/examples/sample_data_multi_prop/POSCAR-JVASP-14873.vasp

@@ -0,0 +1,15 @@
+System
+1.0
+4.191262576674699 0.0 -0.0
+-0.0 4.191262576674699 0.0
+-0.0 0.0 4.191262576674699
+Sr B
+1 6
+direct
+0.0 0.0 0.0 
+0.2028453684309125 0.5 0.5 
+0.5 0.5 0.7971546315690875 
+0.5 0.5 0.2028453684309125 
+0.5 0.2028453684309125 0.5 
+0.5 0.7971546315690875 0.5 
+0.7971546315690875 0.5 0.5 

alignn/examples/sample_data_multi_prop/POSCAR-JVASP-15345.vasp

@@ -0,0 +1,12 @@
+System
+1.0
+3.5666343258756448 0.0 0.0
+0.0 3.60492483256326 -1.0516480500920402
+0.0 0.0043800721536433 3.7551864245512623
+Y Co C
+1 1 2
+direct
+0.0 0.9970518040927455 0.00294819590736377 
+0.5 0.6150401254054609 0.3849598745944179 
+0.5 0.15192770422861318 0.5440196337270639 
+0.5 0.4559803662731792 0.8480722957711558 

alignn/examples/sample_data_multi_prop/POSCAR-JVASP-1996.vasp

@@ -0,0 +1,10 @@
+System
+1.0
+3.93712543178282 0.0 2.273100275741533
+1.3123751439276066 3.7119571065192623 2.273100275741533
+0.0 0.0 4.546200551483066
+Na I
+1 1
+direct
+0.0 0.0 0.0 
+0.5 0.5 0.5 

alignn/examples/sample_data_multi_prop/POSCAR-JVASP-21210.vasp

@@ -0,0 +1,9 @@
+System
+1.0
+1.6712283e-08 -2.508029669761222 3.5458136263853106
+-2.172017276374766 1.254014874098203 3.545813646368687
+-2.17201803290572 -1.254014795663004 -3.5458136064019246
+Xe
+1
+direct
+0.0 0.0 0.0 

alignn/examples/sample_data_multi_prop/POSCAR-JVASP-22556.vasp

@@ -0,0 +1,13 @@
+System
+1.0
+3.790914410660539 -0.0 0.0
+0.0 3.790914410660539 0.0
+-0.0 -0.0 3.790914410660539
+Sr Fe O
+1 1 3
+direct
+0.4999990000000025 0.4999990000000025 0.4999990000000025 
+0.0 0.0 0.0 
+0.4999990000000025 0.0 0.0 
+0.0 0.4999990000000025 0.0 
+0.0 0.0 0.4999990000000025 

alignn/examples/sample_data_multi_prop/POSCAR-JVASP-28397.vasp

@@ -0,0 +1,16 @@
+System
+1.0
+0.0 -3.9587610833154616 0.0
+-6.655928089533787 0.0 0.0
+0.0 0.0 -23.94045079597872
+Si S
+4 4
+direct
+0.0 0.0 0.5217263648738928 
+0.0 0.5 0.5217263648738928 
+0.5 0.0 0.4661843287970869 
+0.5 0.5 0.4661843287970869 
+0.0524900550457348 0.7500000000000001 0.5792751627500841 
+0.9475099449542649 0.25 0.5792751627500841 
+0.4474867899417421 0.7500000000000001 0.40863414357892813 
+0.5525132100582582 0.25 0.40863414357892813 

alignn/examples/sample_data_multi_prop/POSCAR-JVASP-28565.vasp

@@ -0,0 +1,17 @@
+System
+1.0
+3.3542337275744103 0.0 0.0
+-1.6771168637872051 2.904850045503021 0.0002624763854255
+0.0 0.0027191662067697 29.227006366170723
+Te Mo W Se S
+2 2 1 2 2
+direct
+0.3332869375319439 0.6665748750638851 0.40880470960271253 
+0.3334050587801155 0.6668121175602401 0.2788667649202174 
+0.3335186248083656 0.6670362496167267 0.11458088390572692 
+0.6666801663591748 0.33335933271835083 0.3438741864905265 
+0.3331138284834239 0.6662286569668452 0.5825261819744669 
+0.6664959135073409 0.3329918270146898 0.5252624785410095 
+0.6664002633646019 0.3327995267292049 0.639695963563112 
+0.6669034856871763 0.33380597137434753 0.06276093238425431 
+0.6667957214778616 0.3335914429557147 0.1664678986179713 

alignn/examples/sample_data_multi_prop/POSCAR-JVASP-28634.vasp

@@ -0,0 +1,20 @@
+System
+1.0
+3.2250494729190726 2.216578e-10 2.163e-13
+-1.6125242360333125 2.7929740932668943 -8.4754501616e-06
+6.6135e-12 -9.7386036979e-05 34.14068378844537
+Mo W Se S
+1 3 2 6
+direct
+0.33331504683701274 0.6666300936741254 0.0934536896030271 
+0.33331656497272827 0.6666341299456554 0.47208556491119213 
+0.6666698924990063 0.33333978499784117 0.28247993950367123 
+0.6666991220003533 0.3333972440007849 0.6548932577640771 
+0.33334341592455036 0.6666888318491291 0.3328984548376194 
+0.33332779768550075 0.6666565953711356 0.2320623862553914 
+0.33337390316228616 0.6667488063245635 0.7006769649140224 
+0.6666482437297062 0.33329648745935525 0.0479187308537724 
+0.6666490346540366 0.33329706930798436 0.4262562969343839 
+0.6666480620607286 0.33329512412140816 0.1390473399063092 
+0.6666512247313715 0.33330144946262313 0.5179005541158488 
+0.33335669174272836 0.6667143834853922 0.6090748204006833 

alignn/examples/sample_data_multi_prop/POSCAR-JVASP-28704.vasp

@@ -0,0 +1,20 @@
+System
+1.0
+3.292134155794691 0.0 0.0
+-1.6460670778973454 2.8510786681143565 -3.615752048e-06
+0.0 -0.0001768206441563 34.978246270650075
+Mo W Se S
+3 1 6 2
+direct
+0.3333192249247229 0.6666384498494459 0.0966675137812058 
+0.6666558392026922 0.3333126784053811 0.2791091321598709 
+0.6666926464398288 0.3333852928796507 0.6584266681105794 
+0.3333321581155246 0.666664316231042 0.468712621722382 
+0.3333230348691514 0.6666470697383001 0.3273013286877133 
+0.3333590853838843 0.6667201707677635 0.7065686651425439 
+0.6666558450522622 0.3333116901045159 0.4202334807678751 
+0.6666765519168195 0.3333521038336415 0.5171906710959798 
+0.3333210915930135 0.6666421831860199 0.23092332016136433 
+0.3333560781404684 0.6667131562809414 0.6102231914914819 
+0.6666426939013361 0.33328538780266426 0.0528105185440566 
+0.6666647504603075 0.33332850092061717 0.1405808883349591