Deploy preview for PR #42

pr-previews/pr-42/404.html

@@ -439,19 +439,66 @@
 
 
 
-    <li class="md-nav__item">
-      <a href="/titiler-cmr/benchmark.html" class="md-nav__link">
+
+
+
+
+
+
+
+
+    <li class="md-nav__item md-nav__item--nested">
+
 
+
+        <input class="md-nav__toggle md-toggle " type="checkbox" id="__nav_5" >
+
+
+          <label class="md-nav__link" for="__nav_5" id="__nav_5_label" tabindex="0">
+
 
   <span class="md-ellipsis">
     Performance Benchmarks
   </span>
 
 
+            <span class="md-nav__icon md-icon"></span>
+          </label>
+
+        <nav class="md-nav" data-md-level="1" aria-labelledby="__nav_5_label" aria-expanded="false">
+          <label class="md-nav__title" for="__nav_5">
+            <span class="md-nav__icon md-icon"></span>
+            Performance Benchmarks
+          </label>
+          <ul class="md-nav__list" data-md-scrollfix>
+
+
+
+
+
+
+
+    <li class="md-nav__item">
+      <a href="/titiler-cmr/time_series_performance_benchmarks/" class="md-nav__link">
+
+
+  <span class="md-ellipsis">
+    Time series performance benchmarks
+  </span>
+
+
       </a>
     </li>
 
 
+
+
+          </ul>
+        </nav>
+
+    </li>
+
+
 
   </ul>
 </nav>

pr-previews/pr-42/benchmark_analysis/benchmark_analysis.py

@@ -0,0 +1,86 @@
+"""Functions for processing and plotting the time series benchmark results"""
+
+import json
+from collections import defaultdict
+from typing import Dict
+
+import pandas as pd
+import plotly.express as px
+
+
+def process_benchmark_data() -> Dict[str, pd.DataFrame]:
+    """Read the benchmark json and convert into a dictionary of dataframes"""
+    with open("../benchmark.json", "r") as f:
+        benchmarks = json.load(f)["benchmarks"]
+
+    records = defaultdict(list)
+    for bench in benchmarks:
+        record = {
+            "mean_time": bench["stats"]["mean"],
+            "stddev": bench["stats"]["stddev"],
+            **bench["extra_info"],
+            **bench["params"],
+        }
+        record.update(bench["params"])
+        records[bench["group"]].append(record)
+
+    dfs = {
+        group: pd.DataFrame(records).sort_values(by="bbox_size")
+        for group, records in records.items()
+    }
+
+    for group, df in dfs.items():
+        bbox_dims = sorted(
+            df["bbox_dims"].unique(), key=lambda x: float(x.split("x")[0]) * -1
+        )
+
+        df["bbox_dims"] = pd.Categorical(
+            df["bbox_dims"], categories=bbox_dims, ordered=True
+        )
+
+        dfs[group] = df
+
+    return dfs
+
+
+def plot_line_with_error_bars(df: pd.DataFrame, **kwargs):
+    """Create line plot with vertical error bars"""
+    fig = px.line(
+        df,
+        x="num_timepoints",
+        y="mean_time",
+        error_y="stddev",
+        labels={
+            "mean_time": "mean response time (seconds)",
+            "num_timepoints": "number of points in time series",
+        },
+        **kwargs,
+    )
+
+    return fig
+
+
+def plot_error_rate_heatmap(
+    df: pd.DataFrame,
+    x: str,
+    y: str,
+    z: str,
+    labels: Dict[str, str],
+    title: str,
+):
+    """Plot error rate as a heatmap"""
+    # Create the pivot table for heatmap
+    data = df[[x, y, z]].pivot(index=y, columns=x, values=z)
+
+    # Create the faceted heatmap using plotly
+    fig = px.imshow(
+        data,
+        labels=labels,
+        title=title,
+    )
+
+    return fig
+
+
+# Load and process the data
+dfs = process_benchmark_data()