Merge pull request #16 from materialsproject/feature/lammps

Feature/lammps

Author: Hui Zheng

src/mpmorph/analysis/diffusion.py

@@ -290,7 +290,7 @@ def plot(self, title=None, annotate=True, el="", **kwargs):
             self.y,
             yerr=self.yerr.T,
             label="Q[{}]: ".format(el) + tx + " K",
-            **kwargs
+            **kwargs,
         )
         plt.ylabel("ln(D cm$^2$/s)", fontsize=15)
         plt.xlabel("1000/T K$^{-1}$", fontsize=15)

src/mpmorph/analysis/melting_points.py

@@ -7,29 +7,29 @@
 import matplotlib.pyplot as plt
 import math
 
-class MeltingPointClusterAnalyzer():
 
+class MeltingPointClusterAnalyzer:
     def _get_clusters(self, points):
         clustering = AgglomerativeClustering(n_clusters=2).fit(points)
         cluster1 = points[np.argwhere(clustering.labels_ == 1).squeeze()].T
         cluster2 = points[np.argwhere(clustering.labels_ == 0).squeeze()].T
         return cluster1, cluster2
 
-    def plot_vol_vs_temp(self, ts, vs, plot_title = None):
+    def plot_vol_vs_temp(self, ts, vs, plot_title=None):
         points = np.array(list(zip(ts, vs)))
         cluster1, cluster2 = self._get_clusters(points)
         plt.scatter(*cluster1)
         plt.scatter(*cluster2)
         plt.xlabel("Temperature (K)")
         plt.ylabel("Volume (A^3)")
         Tm = self.estimate_melting_temp(ts, vs)
-        plt.plot([Tm, Tm], [min(vs), max(vs)], color='r')
-      
+        plt.plot([Tm, Tm], [min(vs), max(vs)], color="r")
+
         if plot_title is None:
             plt.title("Volume vs Temperature by Clustering")
         else:
             plt.title(plot_title)
-    
+
     def estimate_melting_temp(self, temps, vols):
         points = np.array(list(zip(temps, vols)))
         cluster1, cluster2 = self._get_clusters(points)
@@ -42,8 +42,8 @@ def estimate_melting_temp(self, temps, vols):
 
         return np.mean([max(solid_range), min(liquid_range)])
 
-class MeltingPointSlopeAnalyzer():
 
+class MeltingPointSlopeAnalyzer:
     def split_dset(self, pts, split_idx):
         return pts[0:split_idx], pts[split_idx:]
 
@@ -56,7 +56,7 @@ def assess_splits(self, xs, ys):
         for idx in pt_idxs:
             _, _, _, _, total_err = self.get_split_fit(xs, ys, idx)
             errs.append(total_err)
-        
+
         return list(zip(pt_idxs, errs))
 
     def get_linear_ys(self, m, b, xs):
@@ -79,56 +79,61 @@ def plot_split(self, xs, ys, split_idx):
 
         plt.scatter(rightxs, rightys)
         plt.plot(rightxs, right_fit_ys)
-        
+
     def get_best_split(self, xs, ys):
         split_errs = self.assess_splits(xs, ys)
         errs = [pt[1] for pt in split_errs]
         idxs = [pt[0] for pt in split_errs]
         best_split_idx = idxs[np.argmin(errs)]
         return best_split_idx
-    
+
     def plot_vol_vs_temp(self, temps, vols):
         split_idx = self.get_best_split(temps, vols)
         self.plot_split(temps, vols, split_idx)
         Tm = self.estimate_melting_temp(temps, vols)
         print(Tm)
-        plt.plot([Tm, Tm], [min(vols), max(vols)], color='r')        
-
+        plt.plot([Tm, Tm], [min(vols), max(vols)], color="r")
 
     def estimate_melting_temp(self, temps, vols):
         best_split_idx = self.get_best_split(temps, vols)
         return np.mean([temps[best_split_idx], temps[best_split_idx - 1]])
 
-class MeltingPointSlopeRMSEAnalyzer(MeltingPointSlopeAnalyzer):
 
+class MeltingPointSlopeRMSEAnalyzer(MeltingPointSlopeAnalyzer):
     def get_split_fit(self, xs, ys, split_idx):
         leftx, rightx = self.split_dset(xs, split_idx)
         lefty, righty = self.split_dset(ys, split_idx)
-        
+
         lslope, lintercept, r_value, p_value, std_err = linregress(leftx, lefty)
         left_y_pred = lintercept + lslope * np.array(leftx)
         lefterr = mean_squared_error(y_true=lefty, y_pred=left_y_pred, squared=False)
 
         rslope, rintercept, r_value, p_value, std_err = linregress(rightx, righty)
         right_y_pred = rintercept + rslope * np.array(rightx)
         righterr = mean_squared_error(y_true=righty, y_pred=right_y_pred, squared=False)
-        
-        combined_err = math.sqrt(lefterr ** 2 + righterr ** 2)
+
+        combined_err = math.sqrt(lefterr**2 + righterr**2)
         combined_err = lefterr + righterr
         return lslope, lintercept, rslope, rintercept, combined_err
 
-class MeltingPointSlopeStdErrAnalyzer(MeltingPointSlopeAnalyzer):
 
+class MeltingPointSlopeStdErrAnalyzer(MeltingPointSlopeAnalyzer):
     def get_split_fit(self, xs, ys, split_idx):
         leftx, rightx = self.split_dset(xs, split_idx)
         lefty, righty = self.split_dset(ys, split_idx)
-        
+
         leftfit = linregress(leftx, lefty)
         lefterr = leftfit.stderr
-        
+
         rightfit = linregress(rightx, righty)
         righterr = rightfit.stderr
-        
-        combined_err = math.sqrt(lefterr ** 2 + righterr ** 2)
+
+        combined_err = math.sqrt(lefterr**2 + righterr**2)
         combined_err = lefterr + righterr
-        return leftfit.slope, leftfit.intercept, rightfit.slope, rightfit.intercept, combined_err
+        return (
+            leftfit.slope,
+            leftfit.intercept,
+            rightfit.slope,
+            rightfit.intercept,
+            combined_err,
+        )

src/mpmorph/analysis/structural_analysis.py

@@ -47,7 +47,6 @@ def polyhedra_connectivity(structures, pair, cutoff, step_freq=1):
             polyhedra_list.append(set(current_poly))
 
         for polypair in itertools.combinations(polyhedra_list, 2):
-
             polyhedra_pair_type = (len(polypair[0]), len(polypair[1]))
 
             shared_vertices = len(polypair[0].intersection(polypair[1]))
@@ -143,7 +142,6 @@ class BondAngleDistribution(object):
     """
 
     def __init__(self, structures, cutoffs, step_freq=1):
-
         self.bond_angle_distribution = None
         self.structures = structures
         self.step_freq = step_freq
@@ -251,7 +249,6 @@ def get_bond_angle_distribution(self):
 
                 # get all pair combinations of neoghbor sites of i:
                 for p in itertools.combinations(neighbors[i], 2):
-
                     # check if pairs are within the defined cutoffs
                     if self._cutoff_type == "dict":
                         if self._check_skip_triplet(s_index, i, p[0][2], p[1][2]):

src/mpmorph/database.py

@@ -30,7 +30,7 @@ def __init__(
         collection="tasks",
         user=None,
         password=None,
-        **kwargs
+        **kwargs,
     ):
         super(VaspMDCalcDb, self).__init__(
             host, port, database, collection, user, password, **kwargs
@@ -75,7 +75,6 @@ def insert_task(
 
         # insert structures  at each ionic step into GridFS
         if parse_ionic_steps and "calcs_reversed" in task_doc:
-
             # Convert from ionic steps dictionary to pymatgen.core.trajectory.Trajectory object
             ionic_steps_dict = task_doc["calcs_reversed"][0]["output"]["ionic_steps"]
             time_step = task_doc["input"]["incar"]["POTIM"]

src/mpmorph/firetasks/dbtasks.py

@@ -222,7 +222,6 @@ def load_trajectories_from_gfs(runs, mmdb, gfs_keys=None):
 
     trajectory = None
     for i, (fs_id, fs) in enumerate(gfs_keys):
-
         if fs == "trajectories_fs" or fs == "rebuild_trajectories_fs":
             # Load stored Trajectory
             print(fs_id, "is stored in trajectories_fs")

src/mpmorph/fireworks/powerups.py

@@ -46,7 +46,7 @@ def aggregate_trajectory(fw, **kwargs):
 def add_cont_structure(fw):
     prev_struct_task = PreviousStructureTask()
     insert_i = 2
-    for (i, task) in enumerate(fw.tasks):
+    for i, task in enumerate(fw.tasks):
         if task.fw_name == "{{atomate.vasp.firetasks.run_calc.RunVaspCustodian}}":
             insert_i = i
             break
@@ -68,7 +68,7 @@ def add_pass_pv(fw, **kwargs):
 
 def add_pv_volume_rescale(fw):
     insert_i = 2
-    for (i, task) in enumerate(fw.tasks):
+    for i, task in enumerate(fw.tasks):
         if task.fw_name == "{{atomate.vasp.firetasks.run_calc.RunVaspCustodian}}":
             insert_i = i
             break
@@ -80,7 +80,7 @@ def add_pv_volume_rescale(fw):
 def add_rescale_volume(fw, **kwargs):
     rsv_task = RescaleVolumeTask(**kwargs)
     insert_i = 2
-    for (i, task) in enumerate(fw.tasks):
+    for i, task in enumerate(fw.tasks):
         if task.fw_name == "{{atomate.vasp.firetasks.run_calc.RunVaspCustodian}}":
             insert_i = i
             break

src/mpmorph/flows/md_flow.py

@@ -14,6 +14,10 @@
 M3GNET_MD_CONVERGED_VOL_FLOW = "M3GNET_MD_CONVERGED_VOL_FLOW"
 LAMMPS_VOL_FLOW = "LAMMPS_VOL_FLOW"
 
+
+# def get_md_temperature_sweeping(structure, temp, steps, converge_first = True, initial_vol_scale = 1, **input_kwargs):
+
+
 def get_md_flow_m3gnet(structure, temp, steps, converge_first = True, initial_vol_scale = 1, **input_kwargs):
     inputs = M3GNetMDInputs(
         temperature=temp,
@@ -91,3 +95,5 @@ def _get_converge_flow(structure: Structure, pv_md_maker: PVFromCalc, production
     flow = Flow([equil_vol_job, final_md_job], output=final_md_job.output, name=M3GNET_MD_CONVERGED_VOL_FLOW)
 
     return flow
+
+

src/mpmorph/flows/vt_flow.py

@@ -8,6 +8,7 @@
 
 VOLUME_TEMPERATURE_SWEEP = "VOLUME_TEMPERATURE_SWEEP"
 
+
 def get_vt_sweep_flow(
     structure,
     lower_bound=100,
@@ -16,35 +17,34 @@ def get_vt_sweep_flow(
     output_name="vt.out",
     steps=2000,
 ):
-
     vs = []
     volume_jobs = []
     temps = list(range(lower_bound, upper_bound, temp_step))
 
     for temp in temps:
-        job = get_equil_vol_flow(
-            structure=structure,
-            temp=temp,
-            steps=steps
-        )
+        job = get_equil_vol_flow(structure=structure, temp=temp, steps=steps)
         volume_jobs.append(job)
         vs.append(job.output.volume)
 
     collect_job = _collect_vt_results(vs, temps, structure, output_name)
 
-    new_flow = Flow([*volume_jobs, collect_job], output=collect_job.output, name=VOLUME_TEMPERATURE_SWEEP)
+    new_flow = Flow(
+        [*volume_jobs, collect_job],
+        output=collect_job.output,
+        name=VOLUME_TEMPERATURE_SWEEP,
+    )
     return new_flow
 
+
 def get_vt_sweep_flow_lammps(
     structure,
     lower_bound=100,
     upper_bound=1100,
     temp_step=100,
     output_name="vt.out",
     steps=2000,
-    mp_id=None
+    mp_id=None,
 ):
-
     v_outputs = []
     volume_jobs = []
     temps = list(range(lower_bound, upper_bound, temp_step))
@@ -60,9 +60,10 @@ def get_vt_sweep_flow_lammps(
 
     collect_job = _collect_vt_results(v_outputs, temps, structure, output_name, mp_id)
 
-
-    flow_name = f'{structure.composition.reduced_formula}-Melting Point'
-    new_flow = Flow([*volume_jobs, collect_job], output=collect_job.output, name=flow_name)
+    flow_name = f"{structure.composition.reduced_formula}-Melting Point"
+    new_flow = Flow(
+        [*volume_jobs, collect_job], output=collect_job.output, name=flow_name
+    )
     return new_flow
 
 
@@ -75,18 +76,18 @@ def _collect_vt_results(v_outputs, ts, structure, output_fn, mp_id):
         "mp_id": mp_id,
         "reduced_formula": structure.composition.reduced_formula,
         "formula": structure.composition.formula,
-        "uuid": str(uuid.uuid4())
+        "uuid": str(uuid.uuid4()),
     }
 
     with open(output_fn, "+w") as f:
         f.write(json.dumps(result))
     return result
 
-def get_converged_vol(v_output):   
+
+def get_converged_vol(v_output):
     df = pd.DataFrame.from_dict(v_output)
     total_steps = (len(df) - 1) * 10
     avging_window = int(total_steps / 30)
-    vols = df.iloc[-avging_window::]['vol']
+    vols = df.iloc[-avging_window::]["vol"]
     eq_vol = vols.values.mean()
     return float(eq_vol)
-

src/mpmorph/io.py

@@ -17,13 +17,13 @@ def get_string_from_struct(
     ):
         format_str = "{{:.{0}f}}".format(significant_figures)
 
-        for (si, structure) in enumerate(structures):
+        for si, structure in enumerate(structures):
             lines = [system, "1.0", str(structure.lattice)]
             lines.append(" ".join(self.get_site_symbols(structure)))
             lines.append(" ".join([str(x) for x in self.get_natoms(structure)]))
 
             lines.append("Direct configuration=     " + str(si + 1))
-            for (i, site) in enumerate(structure):
+            for i, site in enumerate(structure):
                 coords = site.frac_coords
                 line = " ".join([format_str.format(c) for c in coords])
                 line += " " + site.species_string
@@ -72,9 +72,9 @@ def get_string(self, system="unknown system", significant_figures=6):
         #  positions = np.add(self.trajectory[0].frac_coords, self.trajectory.displacements)
         atoms = [site.specie.symbol for site in self.trajectory[0]]
 
-        for (si, position_array) in enumerate(positions):
+        for si, position_array in enumerate(positions):
             lines.append("Direct configuration=     " + str(si + 1))
-            for (i, coords) in enumerate(position_array):
+            for i, coords in enumerate(position_array):
                 line = " ".join([format_str.format(c) for c in coords])
                 line += " " + atoms[i]
                 lines.append(line)

src/mpmorph/jobs/equilibrate_volume.py

@@ -29,7 +29,6 @@ class EquilibriumVolumeSearchMaker(Maker):
     def make(
         self, original_structure: Structure, md_pv_data_docs: List[MDPVDataDoc] = None
     ):
-
         if md_pv_data_docs is not None and len(md_pv_data_docs) > MAX_MD_JOBS:
             raise RuntimeError(
                 "Maximum number of jobs for equilibrium volume search exceeded"