Implement inspect CLI for measure workflow and update the O/N pesudos

.gitignore

@@ -3,7 +3,6 @@
 *.swp
 ~*
 *~
-*.pdf
 .project
 *.egg*
 .DS_Store
@@ -15,3 +14,7 @@ build/
 dist/
 run_demos
 pip-wheel-metadata/
+
+# plot output for test
+*.pdf
+*_precision_summary.json

aiida_sssp_workflow/calculations/calculate_metric.py

@@ -116,7 +116,7 @@ def metric_analyze(element, configuration, V0, B0, B1, natoms) -> orm.Dict:
 
     results = {
         "birch_murnaghan_results": [V0, B0, B1],
-        "reference_wien2k_V0_B0_B1": [ref_V0, ref_B0, ref_B1],
+        "reference_ae_V0_B0_B1": [ref_V0, ref_B0, ref_B1],
         "V0_B0_B1_units_info": "eV/A^3 for B0",
     }
     # Delta computation

aiida_sssp_workflow/cli/__init__.py

@@ -17,4 +17,5 @@ def cmd_root():
 
 
 from .extract import extract
+from .inspect import inspect
 from .run import launch

aiida_sssp_workflow/cli/inspect.py

@@ -0,0 +1,141 @@
+#!/usr/bin/env python
+# -*- coding: utf-8 -*-
+"""CLI to inspect the results of the workflow"""
+import json
+
+import click
+import matplotlib.pyplot as plt
+import numpy as np
+from aiida import orm
+
+from aiida_sssp_workflow.cli import cmd_root
+
+
+def birch_murnaghan(V, E0, V0, B0, B01):
+    """
+    Return the energy for given volume (V - it can be a vector) according to
+    the Birch Murnaghan function with parameters E0,V0,B0,B01.
+    """
+    r = (V0 / V) ** (2.0 / 3.0)
+    return E0 + 9.0 / 16.0 * B0 * V0 * (
+        (r - 1.0) ** 3 * B01 + (r - 1.0) ** 2 * (6.0 - 4.0 * r)
+    )
+
+
+def eos_plot(
+    ax, ref_ae_data, line_data, energy0, volumes, energies, nu, title="EOS", fontsize=8
+):
+    """plot EOS result on ax"""
+    dense_volume_max = max(volumes)
+    dense_volume_min = min(volumes)
+
+    dense_volumes = np.linspace(dense_volume_min, dense_volume_max, 100)
+
+    ref_V0, ref_B0, ref_B01 = ref_ae_data
+    V0, B0, B01 = line_data
+
+    ae_eos_fit_energy = birch_murnaghan(
+        V=dense_volumes,
+        E0=energy0,  # in future update E0 from referece json, where ACWF has E0 stored.
+        V0=ref_V0,
+        B0=ref_B0,
+        B01=ref_B01,
+    )
+    psp_eos_fit_energy = birch_murnaghan(
+        V=dense_volumes,
+        E0=energy0,
+        V0=V0,
+        B0=B0,
+        B01=B01,
+    )
+
+    # Plot EOS
+    ax.tick_params(axis="y", labelsize=6, rotation=45)
+
+    ax.plot(volumes, energies, "ob", label="RAW equation of state")
+    ax.plot(dense_volumes, ae_eos_fit_energy, "--r", label="AE")
+    ax.axvline(V0, linestyle="--", color="gray")
+
+    ax.plot(dense_volumes, psp_eos_fit_energy, "-b", label=f"Pseudo")
+    ax.fill_between(
+        dense_volumes,
+        ae_eos_fit_energy,
+        psp_eos_fit_energy,
+        alpha=0.5,
+        color="red",
+    )
+
+    center_x = (max(volumes) + min(volumes)) / 2
+    center_y = (max(energies) + min(energies)) / 2
+
+    # write text of nu value in close middle
+    nu = round(nu, 3)
+    ax.text(center_x, center_y, f"$\\nu$={nu}")
+
+    ax.legend(loc="upper center", bbox_to_anchor=(0.5, 1.0), fontsize=fontsize)
+    ax.set_xlabel("Cell volume per formula unit ($\\AA^3$)", fontsize=fontsize)
+    ax.set_ylabel("$E - TS$ (eV)", fontsize=fontsize)
+    ax.set_xticks([])
+    ax.set_yticks([])
+    ax.set_title(title, fontsize=fontsize)
+
+
+@cmd_root.command("inspect")
+@click.argument("node")  # The node to inspect, uuid or pk
+@click.option("--output", "-o", default="output", help="The output file name")
+def inspect(node, output):
+    """Inspect the results of the workflow"""
+    wf_node = orm.load_node(node)
+    if "measure" in wf_node.outputs:
+        if "precision" in wf_node.outputs.measure:
+            precision = wf_node.outputs.measure.precision
+
+            # print summary of the precision to a json file
+            d_str = json.dumps(precision.output_parameters.get_dict(), indent=4)
+            with open(f"{output}_precision_summary.json", "w") as f:
+                f.write(d_str)
+
+            # if there are 5 plots, need 3 rows, since the output_parametres is in the dict len(precision) / 2 is the number of rows
+            rows = len(precision) // 2
+
+            # create a figure with 2 columns and rows rows on a a4 size paper
+            fig, axs = plt.subplots(rows, 2, figsize=(8.27, 11.69), dpi=100)
+
+            # Plot EOS curve and save to a pdf file
+            i = 0
+            for conf, res in precision.items():
+                if conf == "output_parameters":
+                    continue
+
+                # Plot EOS curve
+                try:
+                    ref_data = res.output_parameters.get_dict()["reference_ae_V0_B0_B1"]
+                except KeyError:
+                    # For backward compatibility, where I used a different key name before (It is changed after v4.2.0)
+                    ref_data = res.output_parameters.get_dict()[
+                        "reference_wien2k_V0_B0_B1"
+                    ]
+
+                data = res.output_parameters.get_dict()["birch_murnaghan_results"]
+
+                energy0 = res.eos.output_birch_murnaghan_fit.get_dict()["energy0"]
+                volumes = res.eos.output_volume_energy.get_dict()["volumes"]
+                energies = res.eos.output_volume_energy.get_dict()["energies"]
+
+                nu = res.output_parameters.get_dict()["rel_errors_vec_length"]
+                title = f"EOS of {conf}"
+
+                ax = axs.flat[i]
+                eos_plot(
+                    ax, ref_data, data, energy0, volumes, energies, nu, title=title
+                )
+
+                i += 1
+
+            # fig to pdf
+            fig.tight_layout()
+            fig.savefig(f"{output}_precision.pdf", bbox_inches="tight")
+
+
+if __name__ == "__main__":
+    inspect()