File size: 4,806 Bytes
afe68b4
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
from pathlib import Path

import numpy as np
import pandas as pd
import streamlit as st

from mlip_arena.models import REGISTRY as MODELS

DATA_DIR = Path(__file__).parents[2] / "benchmarks" / "combustion"

valid_models = [
    model
    for model, metadata in MODELS.items()
    if Path(__file__).stem in metadata.get("gpu-tasks", [])
]

@st.cache_data
def get_data(models):
    dfs = [
        pd.read_json(DATA_DIR / MODELS[model]["family"].lower() / f"{model}_H256O128.json") for model in models
    ]
    df = pd.concat(dfs, ignore_index=True)
    df.drop_duplicates(inplace=True, subset=["formula", "method"])
    return df


df = get_data(valid_models)


@st.cache_data
def get_com_drifts(df):
    df_exploded = df.explode(["timestep", "energies", "com_drifts"]).reset_index(
        drop=True
    )

    # Convert the 'com_drifts' column (which are arrays) into separate columns for x, y, and z components
    df_exploded[["com_drift_x", "com_drift_y", "com_drift_z"]] = pd.DataFrame(
        df_exploded["com_drifts"].tolist(), index=df_exploded.index
    )

    # Drop the original 'com_drifts' column
    df_flat = df_exploded.drop(columns=["com_drifts"])

    df_flat["total_com_drift"] = np.sqrt(
        df_flat["com_drift_x"] ** 2
        + df_flat["com_drift_y"] ** 2
        + df_flat["com_drift_z"] ** 2
    )

    df_flat = df_flat.drop(columns=["com_drift_x", "com_drift_y", "com_drift_z"])

    return df_flat


df_exploded = get_com_drifts(df)

exp_ref = -68.3078  # kcal/mol

for method, row in df_exploded.groupby("method"):
    #     # row = df[df["method"] == method].iloc[0]
    energies = np.array(row["energies"])
    df_exploded.loc[df_exploded["method"] == method, "reaction_enthlapy_diff"] = (
        (energies[-1] - energies[0]) / 128 * 23.0
    ) - exp_ref
    df_exploded.loc[df_exploded["method"] == method, "final_com_drift"] = np.array(
        row["total_com_drift"]
    )[-1]


df_exploded.drop(
    columns=[
        "temperatures",
        "pressures",
        "total_steps",
        "energies",
        "kinetic_energies",
        "timestep",
        "nproducts",
        "total_com_drift",
        "target_steps",
        "reaction",
        "formula",
        "natoms",
        "seconds_per_step",
        "seconds_per_step_per_atom",
        "final_step",
        "total_time_seconds",
    ],
    axis=1,
    inplace=True,
)

df_exploded.drop_duplicates(inplace=True, subset=["method"])

print(df_exploded.columns)

df_exploded.set_index("method", inplace=True)

df_exploded.rename(columns={"method": "Model"}, inplace=True)


table = pd.DataFrame()

for index, row in df_exploded.iterrows():
    new_row = {
        "Model": index,
        "Reaction enthalpy error [kcal/mol]": row["reaction_enthlapy_diff"],
        "Final COM drift [Å]": row["final_com_drift"],
        "Steps per second": row["steps_per_second"],
        "Yield [%]": row["yield"] * 100,
    }

    table = pd.concat([table, pd.DataFrame([new_row])], ignore_index=True)

table.set_index("Model", inplace=True)

table.sort_values("Reaction enthalpy error [kcal/mol]", ascending=True, inplace=True)
table["Rank"] = np.argsort(
    np.abs(table["Reaction enthalpy error [kcal/mol]"].to_numpy())
)

table.sort_values("Final COM drift [Å]", ascending=True, inplace=True)
table["Rank"] += np.argsort(table["Final COM drift [Å]"].to_numpy())

table.sort_values("Steps per second", ascending=False, inplace=True)
table["Rank"] += np.argsort(-table["Steps per second"].to_numpy())

table.sort_values("Yield [%]", ascending=False, inplace=True)
table["Rank"] += np.argsort(-table["Yield [%]"].to_numpy())

table["Rank"] += 1

table.sort_values(["Rank"], ascending=True, inplace=True)

table["Rank aggr."] = table["Rank"]
table["Rank"] = table["Rank aggr."].rank(method="min").astype(int)


table = table.reindex(
    columns=[
        "Rank",
        "Rank aggr.",
        "Reaction enthalpy error [kcal/mol]",
        "Final COM drift [Å]",
        "Steps per second",
        "Yield [%]",
    ]
)


@st.cache_data
def get_table():
    return table


def render():
    s = (
        get_table()
        .style.background_gradient(
            cmap="Oranges",
            subset=["Reaction enthalpy error [kcal/mol]"],
        )
        .background_gradient(
            cmap="Oranges",
            subset=["Final COM drift [Å]"],
            gmap=np.log10(table["Final COM drift [Å]"].to_numpy() + 1e-10),
        )
        .background_gradient(cmap="Oranges_r", subset=["Steps per second", "Yield [%]"])
        .background_gradient(
            cmap="Blues",
            subset=["Rank", "Rank aggr."],
        )
        .format(
            "{:.3e}",
            subset=["Final COM drift [Å]"],
        )
    )

    st.dataframe(
        s,
        use_container_width=True,
    )