Add PEtab v2 overview table to the GitHub Pages site, slim the README table

src/python/MANIFEST.in

@@ -1,2 +1,3 @@
 recursive-include benchmark_models_petab/problems *
 include benchmark_models_petab/schemas/*.json
+include benchmark_models_petab/_site/templates/*.j2

src/python/benchmark_models_petab/_site/__init__.py

@@ -0,0 +1,3 @@
+"""Generation of the GitHub Pages site."""
+
+from .build import create_overview_page

src/python/benchmark_models_petab/_site/build.py

@@ -0,0 +1,204 @@
+"""Build the overview page of the GitHub Pages site."""
+
+import json
+from pathlib import Path
+
+import pandas as pd
+
+from .. import overview, overview_v2, v1, v2
+from ..C import MODELS
+from ..C import (
+    GITHUB_DEFAULT_BRANCH,
+    GITHUB_REPO,
+    PROBLEMS_DIRNAME,
+    V1_DIRNAME,
+    V2_DIRNAME,
+)
+from ..metadata import get_metadata
+
+REPO_URL = f"https://github.com/{GITHUB_REPO}"
+LOGO_URL = (
+    f"https://raw.githubusercontent.com/{GITHUB_REPO}"
+    f"/{GITHUB_DEFAULT_BRANCH}/gfx/logo.svg"
+)
+CONTRIBUTING_URL = f"{REPO_URL}/blob/{GITHUB_DEFAULT_BRANCH}/CONTRIBUTING.md"
+
+#: Rendering of the columns that are not plain text, by column name.
+#:  Other columns of integer dtype are rendered as numbers.
+column_kinds = {
+    "petab_problem_id": "id",
+    "short_description": "description",
+    "possible_discontinuities": "bool",
+    "noise_distributions": "tags",
+    "objective_prior_distributions": "tags",
+    "prior_distributions": "tags",
+    "reference_uris": "links",
+    "sbml4humans_urls": "links",
+}
+
+#: Column groups, in display order, with their columns.
+#:  Columns without a group are shown first.
+column_groups = {
+    "Links": ["reference_uris", "sbml4humans_urls"],
+    "Size": [
+        "conditions",
+        "experiments",
+        "estimated_parameters",
+        "measurements",
+        "observables",
+        "species",
+    ],
+    "Features": [
+        "events",
+        "preequilibration",
+        "postequilibration",
+        "possible_discontinuities",
+    ],
+    "Distributions": [
+        "noise_distributions",
+        "objective_prior_distributions",
+        "prior_distributions",
+    ],
+}
+
+column_group = {
+    col: group for group, cols in column_groups.items() for col in cols
+}
+
+#: Statistics shown as summary tiles, as (v1 column, v2 column, label).
+summary_columns = [
+    ("conditions", "experiments", "simulation conditions"),
+    ("estimated_parameters", "estimated_parameters", "estimated parameters"),
+    ("measurements", "measurements", "measurements"),
+    ("observables", "observables", "observables"),
+]
+
+
+def get_table(
+    df: pd.DataFrame,
+    labels: dict[str, str],
+    descriptions: dict[str, str],
+    version_dirname: str,
+) -> dict:
+    """Get the data and column specification for one overview table.
+
+    :param df: Overview table as returned by `get_overview_df`.
+    :param labels: Column labels, by column name.
+    :param descriptions: Column descriptions, by column name.
+    :param version_dirname: The encoding's directory (`v1`, `v2`).
+    """
+    df = df.reset_index()
+    df.insert(
+        1,
+        "short_description",
+        [get_metadata(id_).short_description for id_ in df.petab_problem_id],
+    )
+    labels = {"short_description": "Description"} | labels
+    descriptions = {
+        "short_description": (
+            "Model type, biological system, and, where informative, the data."
+        )
+    } | descriptions
+    # ungrouped columns first, in their original order, then the groups
+    ungrouped = [col for col in df.columns if col not in column_group]
+    ordered_columns = ungrouped + [
+        col for cols in column_groups.values() for col in cols if col in df
+    ]
+    columns = [
+        {
+            "field": col,
+            "title": labels.get(col, col),
+            "tooltip": descriptions.get(col),
+            "group": column_group.get(col),
+            "kind": column_kinds.get(
+                col,
+                "int" if pd.api.types.is_integer_dtype(df[col]) else "text",
+            ),
+        }
+        for col in ordered_columns
+    ]
+    return {
+        "columns": columns,
+        "rows": json.loads(df[ordered_columns].to_json(orient="records")),
+        "tree_url": (
+            f"{REPO_URL}/tree/{GITHUB_DEFAULT_BRANCH}/{PROBLEMS_DIRNAME}"
+            f"/{{id}}/{version_dirname}"
+        ),
+    }
+
+
+def get_summary_tiles(v1_df: pd.DataFrame, v2_df: pd.DataFrame) -> list:
+    """Get the summary statistics shown above the tables.
+
+    The ranges are computed from the v1 statistics, and from the v2
+    statistics for problems without a v1 encoding.
+
+    :param v1_df: v1 overview table.
+    :param v2_df: v2 overview table.
+    :return: List of ``(value, label, tooltip)`` tuples.
+    """
+    tiles = [
+        (
+            f"{len(MODELS)}",
+            "problems in total",
+            "Number of benchmark problems, in any PEtab version.",
+        ),
+        (
+            f"{len(v1.MODELS)}",
+            "in PEtab v1",
+            "Number of problems with a PEtab v1 encoding.",
+        ),
+        (
+            f"{len(v2.MODELS)}",
+            "in PEtab v2",
+            "Number of problems with a PEtab v2 encoding.",
+        ),
+    ]
+    v2_only = ~v2_df.index.isin(v1_df.index)
+    for v1_col, v2_col, label in summary_columns:
+        values = pd.concat([v1_df[v1_col], v2_df.loc[v2_only, v2_col]])
+        tooltip = f"Smallest and largest number of {label} per problem."
+        if v1_col != v2_col:
+            tooltip += (
+                f" For problems without a PEtab v1 encoding, the number of "
+                f"{overview_v2.columns[v2_col].lower()}."
+            )
+        tiles.append((f"{values.min():,}–{values.max():,}", label, tooltip))
+    return tiles
+
+
+def create_overview_page(dest: Path) -> None:
+    """Create the HTML page with the v1 and v2 overview tables.
+
+    :param dest: Path to the output HTML file.
+    """
+    from jinja2 import Environment, PackageLoader, select_autoescape
+
+    env = Environment(
+        loader=PackageLoader(__package__),
+        autoescape=select_autoescape(["html", "j2"]),
+    )
+    v1_df = overview.get_overview_df()
+    v2_df = overview_v2.get_overview_df()
+    html = env.get_template("index.html.j2").render(
+        repo_url=REPO_URL,
+        logo_url=LOGO_URL,
+        contributing_url=CONTRIBUTING_URL,
+        summary_tiles=get_summary_tiles(v1_df, v2_df),
+        tables={
+            "v2": get_table(
+                v2_df,
+                overview_v2.columns,
+                overview_v2.column_descriptions,
+                V2_DIRNAME,
+            ),
+            "v1": get_table(
+                v1_df,
+                overview.columns,
+                overview.column_descriptions,
+                V1_DIRNAME,
+            ),
+        },
+    )
+    dest.parent.mkdir(parents=True, exist_ok=True)
+    dest.write_text(html, encoding="utf-8")