exp show: Add parallel coordinates plot.

dvc/command/experiments/show.py

@@ -3,10 +3,12 @@
 from collections import Counter, OrderedDict, defaultdict
 from datetime import date, datetime
 from fnmatch import fnmatch
+from pathlib import Path
 from typing import TYPE_CHECKING, Dict, Iterable, Optional
 
 from funcy import lmap
 
+from dvc.command import completion
 from dvc.command.base import CmdBase, append_doc_link
 from dvc.command.metrics import DEFAULT_PRECISION
 from dvc.exceptions import DvcException, InvalidArgumentError
@@ -382,6 +384,7 @@ def show_experiments(
     no_timestamp=False,
     csv=False,
     markdown=False,
+    html=False,
     **kwargs,
 ):
     from funcy.seqs import flatten as flatten_list
@@ -429,7 +432,7 @@ def show_experiments(
         kwargs.get("iso"),
     )
 
-    if no_timestamp:
+    if no_timestamp or html:
         td.drop("Created")
 
     for col in ("State", "Executor"):
@@ -466,9 +469,21 @@ def show_experiments(
         }
     )
 
-    if kwargs.get("only_changed", False):
+    if kwargs.get("only_changed", False) or html:
         td.drop_duplicates("cols")
 
+    html_args = {}
+    if html:
+        td.dropna("rows", how="all")
+        td.column("Experiment")[:] = [
+            # remove tree characters
+            str(x).encode("ascii", "ignore").strip().decode()
+            for x in td.column("Experiment")
+        ]
+        out = kwargs.get("out") or "dvc_plots"
+        html_args["output_path"] = (Path.cwd() / out).resolve()
+        html_args["color_by"] = kwargs.get("sort_by") or "Experiment"
+
     td.render(
         pager=pager,
         borders=True,
@@ -477,8 +492,13 @@ def show_experiments(
         row_styles=row_styles,
         csv=csv,
         markdown=markdown,
+        html=html,
+        **html_args,
     )
 
+    if html and kwargs.get("open"):
+        return ui.open_browser(Path(out) / "index.html")
+
 
 def _normalize_headers(names, count):
     return [
@@ -544,6 +564,9 @@ def run(self):
                 csv=self.args.csv,
                 markdown=self.args.markdown,
                 only_changed=self.args.only_changed,
+                html=self.args.html,
+                out=self.args.out,
+                open=self.args.open,
             )
         return 0
 
@@ -693,4 +716,23 @@ def add_parser(experiments_subparsers, parent_parser):
             "across the selected experiments."
         ),
     )
+    experiments_show_parser.add_argument(
+        "--html",
+        action="store_true",
+        default=False,
+        help="Generate a parallel coordinates plot from the tabulated output.",
+    )
+    experiments_show_parser.add_argument(
+        "-o",
+        "--out",
+        default=None,
+        help="Destination folder to save the HTML to",
+        metavar="<path>",
+    ).complete = completion.DIR
+    experiments_show_parser.add_argument(
+        "--open",
+        action="store_true",
+        default=False,
+        help="Open the HTML directly in the browser.",
+    )
     experiments_show_parser.set_defaults(func=CmdExperimentsShow)

dvc/compare.py

@@ -153,6 +153,18 @@ def to_csv(self) -> str:
             writer.writerow(row)
         return buff.getvalue()
 
+    def to_parallel_coordinates(self, output_path, color_by):
+        from dvc.render.html import write
+        from dvc.render.plotly import ParallelCoordinatesRenderer
+
+        index_path = write(
+            output_path,
+            renderers=[
+                ParallelCoordinatesRenderer(self, color_by, self._fill_value)
+            ],
+        )
+        return index_path.as_uri()
+
     def add_column(self, name: str) -> None:
         self._columns[name] = Column([self._fill_value] * len(self))
         self._keys.append(name)
@@ -173,6 +185,14 @@ def render(self, **kwargs: Any):
 
         if kwargs.pop("csv", False):
             ui.write(self.to_csv(), end="")
+
+        elif kwargs.pop("html", False):
+            ui.write(
+                self.to_parallel_coordinates(
+                    kwargs["output_path"], kwargs.get("color_by")
+                )
+            )
+
         else:
             ui.table(self, headers=self.keys(), **kwargs)
 

dvc/render/plotly.py

@@ -0,0 +1,92 @@
+import json
+from collections import defaultdict
+from typing import TYPE_CHECKING, Any, Dict, Optional
+
+from dvc.render.base import Renderer
+
+if TYPE_CHECKING:
+    from dvc.compare import TabularData
+
+
+class ParallelCoordinatesRenderer(Renderer):
+    TYPE = "plotly"
+
+    DIV = """
+    <div id = "{id}">
+        <script type = "text/javascript">
+            var plotly_data = {partial};
+            Plotly.newPlot("{id}", plotly_data.data, plotly_data.layout);
+        </script>
+    </div>
+    """
+
+    SCRIPTS = """
+    <script src="https://cdn.plot.ly/plotly-latest.min.js"></script>
+    """
+
+    # pylint: disable=W0231
+    def __init__(
+        self,
+        tabular_data: "TabularData",
+        color_by: Optional[str] = None,
+        fill_value: str = "",
+    ):
+        self.tabular_data = tabular_data
+        self.color_by = color_by
+        self.filename = "experiments"
+        self.fill_value = fill_value
+
+    def partial_html(self, **kwargs):
+        return self.as_json()
+
+    def as_json(self, **kwargs) -> str:
+        tabular_dict = defaultdict(list)
+        for row in self.tabular_data.as_dict():
+            for col_name, value in row.items():
+                tabular_dict[col_name].append(str(value))
+
+        trace: Dict[str, Any] = {"type": "parcoords", "dimensions": []}
+        for label, values in tabular_dict.items():
+            is_categorical = False
+            try:
+                float_values = [
+                    float(x) if x != self.fill_value else None for x in values
+                ]
+            except ValueError:
+                is_categorical = True
+
+            if is_categorical:
+                non_missing = [x for x in values if x != self.fill_value]
+                unique_values = sorted(set(non_missing))
+                unique_values.append(self.fill_value)
+
+                dummy_values = [unique_values.index(x) for x in values]
+
+                values = [
+                    x if x != self.fill_value else "Missing" for x in values
+                ]
+                trace["dimensions"].append(
+                    {
+                        "label": label,
+                        "values": dummy_values,
+                        "tickvals": dummy_values,
+                        "ticktext": values,
+                    }
+                )
+            else:
+                trace["dimensions"].append(
+                    {"label": label, "values": float_values}
+                )
+
+            if label == self.color_by:
+                trace["line"] = {
+                    "color": dummy_values if is_categorical else float_values,
+                    "showscale": True,
+                    "colorbar": {"title": self.color_by},
+                }
+                if is_categorical:
+                    trace["line"]["colorbar"]["tickmode"] = "array"
+                    trace["line"]["colorbar"]["tickvals"] = dummy_values
+                    trace["line"]["colorbar"]["ticktext"] = values
+
+        return json.dumps({"data": [trace], "layout": {}})

tests/func/experiments/test_show.py

@@ -597,3 +597,78 @@ def test_show_only_changed(tmp_dir, dvc, scm, capsys):
     cap = capsys.readouterr()
 
     assert "bar" not in cap.out
+
+
+def test_show_parallel_coordinates(tmp_dir, dvc, scm, mocker):
+    from dvc.command.experiments import show
+
+    webbroser_open = mocker.patch("webbrowser.open")
+    show_experiments = mocker.spy(show, "show_experiments")
+
+    tmp_dir.gen("copy.py", COPY_SCRIPT)
+    params_file = tmp_dir / "params.yaml"
+    params_data = {
+        "foo": 1,
+        "bar": 1,
+    }
+    (tmp_dir / params_file).dump(params_data)
+
+    dvc.run(
+        cmd="python copy.py params.yaml metrics.yaml",
+        metrics_no_cache=["metrics.yaml"],
+        params=["foo", "bar"],
+        name="copy-file",
+        deps=["copy.py"],
+    )
+    scm.add(
+        [
+            "dvc.yaml",
+            "dvc.lock",
+            "copy.py",
+            "params.yaml",
+            "metrics.yaml",
+            ".gitignore",
+        ]
+    )
+    scm.commit("init")
+
+    dvc.experiments.run(params=["foo=2"])
+
+    assert main(["exp", "show", "--html"]) == 0
+    kwargs = show_experiments.call_args[1]
+
+    html_text = (tmp_dir / "dvc_plots" / "index.html").read_text()
+    assert all(rev in html_text for rev in ["workspace", "master", "[exp-"])
+
+    assert (
+        '{"label": "metrics.yaml:foo", "values": [2.0, 1.0, 2.0]}' in html_text
+    )
+    assert (
+        '{"label": "params.yaml:foo", "values": [2.0, 1.0, 2.0]}' in html_text
+    )
+    assert '"line": {"color": [2, 1, 0]' in html_text
+    assert '"label": "metrics.yaml:bar"' not in html_text
+
+    assert (
+        main(["exp", "show", "--html", "--sort-by", "metrics.yaml:foo"]) == 0
+    )
+    kwargs = show_experiments.call_args[1]
+
+    html_text = (tmp_dir / "dvc_plots" / "index.html").read_text()
+    assert '"line": {"color": [2.0, 1.0, 2.0]' in html_text
+
+    assert main(["exp", "show", "--html", "--out", "experiments"]) == 0
+    kwargs = show_experiments.call_args[1]
+
+    assert kwargs["out"] == "experiments"
+    assert (tmp_dir / "experiments" / "index.html").exists()
+
+    assert main(["exp", "show", "--html", "--open"]) == 0
+
+    webbroser_open.assert_called()
+
+    params_data = {"foo": 1, "bar": 1, "foobar": 2}
+    (tmp_dir / params_file).dump(params_data)
+    assert main(["exp", "show", "--html"]) == 0
+    html_text = (tmp_dir / "dvc_plots" / "index.html").read_text()
+    assert '{"label": "foobar", "values": [2.0, null, null]}' in html_text

tests/unit/command/test_experiments.py

@@ -646,6 +646,7 @@ def test_experiments_init_config(dvc, scm, mocker):
 
     assert isinstance(cmd, CmdExperimentsInit)
     assert cmd.run() == 0
+
     m.assert_called_once_with(
         ANY(Repo),
         name="train",
@@ -758,3 +759,37 @@ def test_experiments_init_extra_args(extra_args, expected_kw, mocker):
 def test_experiments_init_type_invalid_choice():
     with pytest.raises(DvcParserError):
         parse_args(["exp", "init", "--type=invalid", "cmd"])
+
+
+def test_show_experiments_html(tmp_dir, mocker):
+    all_experiments = {
+        "workspace": {
+            "baseline": {
+                "data": {
+                    "timestamp": None,
+                    "params": {"params.yaml": {"data": {"foo": 1}}},
+                    "queued": False,
+                    "running": False,
+                    "executor": None,
+                    "metrics": {
+                        "scores.json": {"data": {"bar": 0.9544670443829399}}
+                    },
+                }
+            }
+        },
+    }
+    experiments_table = mocker.patch(
+        "dvc.command.experiments.show.experiments_table"
+    )
+    td = experiments_table.return_value
+
+    show_experiments(all_experiments, html=True)
+
+    td.dropna.assert_called_with("rows", how="all")
+
+    render_kwargs = td.render.call_args[1]
+
+    for arg in ["html", "output_path", "color_by"]:
+        assert arg in render_kwargs
+    assert render_kwargs["output_path"] == tmp_dir / "dvc_plots"
+    assert render_kwargs["color_by"] == "Experiment"