File size: 5,505 Bytes
2eae90c |
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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 |
"""This script uses Plotly to visualize benchmark results.
To use this script run
```shell
.venv/bin/python ./scripts/plot_results.py
```
"""
import os
import plotly.express as px
import polars as pl
from common_utils import DEFAULT_PLOTS_DIR, INCLUDE_IO, TIMINGS_FILE, WRITE_PLOT
# colors for each bar
COLORS = {
"polars": "#f7c5a0",
"dask": "#87f7cf",
"pandas": "#72ccff",
# "modin": "#d4a4eb",
}
# default base template for plot's theme
DEFAULT_THEME = "plotly_dark"
# other configuration
BAR_TYPE = "group"
LABEL_UPDATES = {
"x": "query",
"y": "seconds",
"color": "Solution",
"pattern_shape": "Solution",
}
def add_annotations(fig, limit: int, df: pl.DataFrame):
# order of solutions in the file
# e.g. ['polar', 'pandas', 'dask']
bar_order = (
df["solution"].unique(maintain_order=True).to_frame().with_row_count("index")
)
# every bar in the plot has a different offset for the text
start_offset = 10
offsets = [start_offset + 12 * i for i in range(0, bar_order.height)]
# we look for the solutions that surpassed the limit
# and create a text label for them
df = (
df.filter(pl.col("duration[s]") > limit)
.with_columns(
pl.when(pl.col("success"))
.then(
pl.format(
"{} took {} s", "solution", pl.col("duration[s]").cast(pl.Int32)
).alias("labels")
)
.otherwise(pl.format("{} had an internal error", "solution"))
)
.join(bar_order, on="solution")
.groupby("query_no")
.agg([pl.col("labels"), pl.col("index").min()])
.with_columns(pl.col("labels").list.join(",\n"))
)
# then we create a dictionary similar to something like this:
# anno_data = {
# "q1": (offset, "label"),
# "q3": (offset, "label"),
# }
if df.height > 0:
anno_data = {
v[0]: (offsets[int(v[1])], v[2])
for v in df.select(["query_no", "index", "labels"])
.transpose()
.to_dict(False)
.values()
}
else:
# a dummy with no text
anno_data = {"q1": (0, "")}
for q_name, (x_shift, anno_text) in anno_data.items():
fig.add_annotation(
align="right",
x=q_name,
y=LIMIT,
xshift=x_shift,
yshift=30,
font=dict(color="white"),
showarrow=False,
text=anno_text,
)
def write_plot_image(fig):
if not os.path.exists(DEFAULT_PLOTS_DIR):
os.mkdir(DEFAULT_PLOTS_DIR)
file_name = f"plot_with_io.html" if INCLUDE_IO else "plot_without_io.html"
fig.write_html(os.path.join(DEFAULT_PLOTS_DIR, file_name))
def plot(
df: pl.DataFrame,
x: str = "query_no",
y: str = "duration[s]",
group: str = "solution",
limit: int = 120,
):
"""Generate a Plotly Figure of a grouped bar chart diplaying
benchmark results from a DataFrame.
Args:
df (pl.DataFrame): DataFrame containing `x`, `y`, and `group`.
x (str, optional): Column for X Axis. Defaults to "query_no".
y (str, optional): Column for Y Axis. Defaults to "duration[s]".
group (str, optional): Column for group. Defaults to "solution".
limit: height limit in seconds
Returns:
px.Figure: Plotly Figure (histogram)
"""
# build plotly figure object
fig = px.histogram(
x=df[x],
y=df[y],
color=df[group],
barmode=BAR_TYPE,
template=DEFAULT_THEME,
color_discrete_map=COLORS,
pattern_shape=df[group],
labels=LABEL_UPDATES,
)
fig.update_layout(
bargroupgap=0.1,
paper_bgcolor="rgba(41,52,65,1)",
yaxis_range=[0, limit],
plot_bgcolor="rgba(41,52,65,1)",
margin=dict(t=100),
legend=dict(orientation="h", xanchor="left", yanchor="top", x=0.37, y=-0.1),
)
add_annotations(fig, limit, df)
if WRITE_PLOT:
write_plot_image(fig)
# display the object using available environment context
fig.show()
if __name__ == "__main__":
print("write plot:", WRITE_PLOT)
e = pl.lit(True)
max_query = 8
if INCLUDE_IO:
LIMIT = 15
e = e & pl.col("include_io") & ~(pl.col("solution") == "vaex_feather")
else:
LIMIT = 15
e = e & ~pl.col("include_io")
df = (
pl.scan_csv(TIMINGS_FILE)
.filter(e)
# filter the max query to plot
.filter((pl.col("query_no").str.extract("q(\d+)", 1).cast(int) <= max_query))
# create a version no
.with_columns(
[
pl.when(pl.col("success")).then(pl.col("duration[s]")).otherwise(0),
pl.format("{}-{}", "solution", "version").alias("solution-version"),
]
)
# ensure we get the latest version
.sort(["solution", "version"])
.groupby(["solution", "query_no"], maintain_order=True)
.last()
.collect()
)
order = pl.DataFrame(
{
"solution": [
"polars",
"duckdb",
"pandas",
"fireducks",
"dask",
"spark",
"vaex_parquet",
"modin",
]
}
)
df = order.join(df, on="solution", how="left")
plot(df, limit=LIMIT, group="solution-version")
|