Merge pull request #88 from GENESIS-EFRC/dev

Address several performance issues, add FREED data, support for metastable CPDs

requirements-testing.txt

@@ -4,7 +4,7 @@ pytest-cov==3.0.0
 pycodestyle==2.8.0
 pydocstyle==6.1.1
 flake8==4.0.1
-mypy==0.910
+mypy==0.931
 mypy-extensions==0.4.3
 pylint==2.12.2
-black==21.12b0
+black==22.1.0

scripts/generate_free_data.py

@@ -0,0 +1,57 @@
+""" 
+This script generates the data/freed/compounds.json file by using the s4 package as an interface.
+"""
+
+import numpy as np
+from pymatgen.core.composition import Composition
+from collections import OrderedDict
+import json
+import gzip
+
+from s4.thermo.exp.freed import ExpThermoDatabase
+
+
+if __name__ == "__main__":
+    xp = ExpThermoDatabase()
+    temps = np.arange(300, 2100, 100)
+    data = {}
+
+    # Acquire data
+    for c in xp.compositions:
+        if c.is_element:
+            continue
+
+        formula, factor = c.get_integer_formula_and_factor()
+
+        if not data.get(formula):
+            data[formula] = dict()
+
+        phases = xp.compositions[c]
+
+        for t in temps:
+            energies = []
+            for p in phases:
+                if t > p.tmax:
+                    continue
+                energies.append(p.dgf(t))
+
+            if not energies:
+                continue
+
+            final_energy = min(energies)
+            g = final_energy / factor
+
+            if data[formula].get(t):
+                data[formula][t].append(g)
+            else:
+                data[formula][t] = [g]
+
+    # Clean the data
+    cleaned_data = {}
+    for f, energy_dict in data.items():
+        cleaned_data[f] = {str(t): min(e) for t, e in energy_dict.items()}
+
+    # Write to file
+    cleaned_data = OrderedDict(sorted(cleaned_data.items()))
+    with gzip.open("compounds.json.gz", "wt", encoding="ascii") as zipfile:
+        json.dump(cleaned_data, zipfile)

scripts/generate_janaf_data.py

@@ -4,6 +4,7 @@
 """
 
 import os
+import gzip
 import json
 from collections import OrderedDict
 import pandas
@@ -74,6 +75,7 @@
         formula, factor = Composition(f).get_integer_formula_and_factor()
         all_data[formula] = {temp: g / factor for temp, g in zip(temps, dGf)}
 
-all_data = OrderedDict(sorted(all_data.items()))
-with open("compounds.json", "w") as f:
-    json.dump(all_data, f)
+    all_data = OrderedDict(sorted(all_data.items()))
+
+    with gzip.open("compounds.json.gz", "wt", encoding="ascii") as zipfile:
+        json.dump(all_data, zipfile)

setup.cfg

@@ -24,6 +24,8 @@ ignore = D105,D2,D4
 ignore_missing_imports = True
 
 [pylint.MASTER]
+ignored-modules=scipy
+extension-pkg-whitelist=scipy
 ignore-patterns=test_
 disable=unsubscriptable-object,
         invalid-name,

src/rxn_network/costs/competitiveness.py

@@ -40,6 +40,7 @@ def __init__(
         use_minimize=False,
         basic_enumerator_kwargs: Optional[Dict] = None,
         minimize_enumerator_kwargs: Optional[Dict] = None,
+        quiet=True,
         name: str = "c_score",
     ):
         """
@@ -70,6 +71,10 @@ def __init__(
         if not self.minimize_enumerator_kwargs.get("calculators"):
             self.minimize_enumerator_kwargs["calculators"] = calcs
 
+        if quiet:
+            self.basic_enumerator_kwargs["quiet"] = True
+            self.minimize_enumerator_kwargs["quiet"] = True
+
     def calculate(self, rxn: ComputedReaction) -> float:
         """
         Calculates the competitiveness score for a given reaction by enumerating
@@ -105,6 +110,8 @@ def get_competing_rxns(self, rxn: ComputedReaction) -> List[ComputedReaction]:
             A list of competing reactions
 
         """
+        chemsys = rxn.chemical_system.split("-")
+
         precursors = [r.reduced_formula for r in rxn.reactants]
         open_elem = self.open_elem
 
@@ -116,6 +123,9 @@ def get_competing_rxns(self, rxn: ComputedReaction) -> List[ComputedReaction]:
         if open_phases:
             precursors = list(set(precursors) - set(open_phases))
 
+        entries = self.entries.filter_by_stability(0.0).get_subset_in_chemsys(chemsys)
+        entries.update(rxn.entries)  # add back entries which may have been filtered out
+
         enumerators = []
         if self.use_basic:
             kwargs = self.basic_enumerator_kwargs.copy()
@@ -143,7 +153,7 @@ def get_competing_rxns(self, rxn: ComputedReaction) -> List[ComputedReaction]:
 
         rxns = set()
         for e in enumerators:
-            rxns.update(e.enumerate(self.entries))
+            rxns.update(e.enumerate(entries))
 
         if rxn in rxns:
             rxns.remove(rxn)

src/rxn_network/data/__init__.py

@@ -10,6 +10,7 @@
 cwd = Path(__file__).parent.resolve()
 
 PATH_TO_BARIN = cwd / "barin"
+PATH_TO_FREED = cwd / "freed"
 PATH_TO_NIST = cwd / "nist"
 
 G_ELEMS = loadfn(cwd / "elements.json")