Merge pull request #131 from statisticsnorway/gen_vis

moved repo

Author: Didrikbt

pyproject.toml

@@ -29,6 +29,7 @@ dependencies = [
 dash = [
     "dash>=4.0.0",
     "plotly>=6.5.2",
+    "polars>=1.39.0",
     "ssb-dash-components>=0.11.1",
 ]
 

src/ssb_konjunk/dash/components/gen_vis.py

@@ -0,0 +1,180 @@
+"""Dash component for generic visualization at SSB.
+
+This module provides a dash all-in-one component that are intended to replace
+some functions of myfame for generic visualizations.
+It will only work in SSBs-developer environments.
+
+The module expects the path to JSON-configuration file. 
+
+"dataset_1": {
+        "glob_pattern": "glob/to/files/*.parquet",
+        "index_col": "period", // The columns that contains the periods.
+        "index_pattern": "%Y-%m", // The pattern of the period column. See: https://strftime.org/
+        "groupby_col": "nar" // OPTIONAL: Columns that are supposed to be subset
+        "agg_type": "AVERAGE" // OPTIONAL: For now, only "AVERAGE" is allowed.
+    },
+"dataset_2": ...
+
+Example:
+    ```
+    from ssb_konjunk.dash.gen_vis import GenVis
+    layout = GenVis("./data_setup.json")
+    # Or
+
+    from dash import html
+
+    layout = html.Div(
+        children = [
+            GenVis("./data_setup.json")
+        ]
+    )
+    ```
+
+
+"""
+
+import uuid
+
+from dash import (
+    Input,
+    Output,
+    callback,
+    html,
+)
+
+from ssb_dash_components import Button
+
+from .internal.loading_test import load_datasets
+from .internal.tab_selector import TabSelector
+from .internal.series_selector import SeriesSelector
+from .internal.series_settings_display import SeriesSettingsDisplay
+from .internal.graph_display import GraphDisplay
+from .internal.graph_settings_display import GraphSettingsDisplay
+
+class GenVis(html.Div): 
+
+    class ids:
+        dropdown = lambda aio_id: {
+            "component": "MarkdownWithColorAIO",
+            "subcomponent": "dropdown",
+            "aio_id": aio_id,
+        }
+        markdown = lambda aio_id: {
+            "component": "MarkdownWithColorAIO",
+            "subcomponent": "markdown",
+            "aio_id": aio_id,
+        }
+
+    # Make the ids class a public class
+    ids = ids
+
+    # Define the arguments of the All-in-One component
+    def __init__(self, config_path: str, aio_id=None):
+        """
+        Returns a component that can be used as its own page or as a component in other layouts.
+
+        Args:
+            config_path (str): Path to the config file.
+            aio_id (str | None): An optional id. Will be randomised if not provided.
+        """
+        if aio_id is None:
+            aio_id = str(uuid.uuid4())
+
+        datasets = load_datasets(config_path)
+        select_data = {}
+        for key, value in datasets.items():
+            select_data[key] = value.get_entries()
+
+        super().__init__(
+            [
+                html.Div(
+                    [
+                        html.Div(
+                            children=[
+                                html.Div(
+                                    children=[
+                                        Button(
+                                            "Oppdater datasett",
+                                            id="update-dataset-button",
+                                        ),
+                                    ],
+                                ),
+                                html.Div(
+                                    TabSelector(datasets, aio_id=aio_id, height="200px"),
+                                ),
+                            ],
+                        ),
+                        html.Div(
+                            children=[
+                                html.H3("Velg serier"),
+                                html.Div(
+                                    children=[SeriesSelector(datasets, aio_id=aio_id)],
+                                ),
+                            ],
+                        ),
+                        html.Div(
+                            children=[GraphSettingsDisplay(aio_id)],
+                        ),
+                    ],
+                    style={
+                        "display": "grid",
+                        "gridTemplateColumns": "40% 45% 15%",
+                    },
+                ),
+                html.Div(
+                    children=[
+                        html.Div(
+                            html.Div(
+                                children=[
+                                    SeriesSettingsDisplay(datasets, aio_id),
+                                    GraphDisplay(datasets, aio_id),
+                                ],
+                                id="graph-display",
+                                style={
+                                    "height": "100px",
+                                    "width": "100%",
+                                    "display": "grid",
+                                    "gridTemplateColumns": "30% 70%",
+                                },
+                            )
+                        ),
+                    ],
+                    style={
+                        "gap": "30px",
+                    },
+                ),
+            ],
+            style={"gap": "10px"},
+        )
+
+        @callback(
+            Output(SeriesSelector.ids.store(aio_id), "data"),
+            Input(TabSelector.ids.store(aio_id), "data"),
+        )
+        def update_selected(selected):
+            """Move data from the file selector to the series selector"""
+            return selected
+
+        @callback(
+            Output(SeriesSettingsDisplay.ids.store(aio_id), "data"),
+            Input(SeriesSelector.ids.selected_store(aio_id), "data"),
+        )
+        def update_display(selected):
+            """Move data from the series selcetor to series settings display"""
+            return selected
+
+        @callback(
+            Output(GraphDisplay.ids.series_store(aio_id), "data"),
+            Input(SeriesSettingsDisplay.ids.settings_store(aio_id), "data"),
+        )
+        def update_graph_series(selected):
+            """Moves series settings to the graph display"""
+            return selected
+
+        @callback(
+            Output(GraphDisplay.ids.settings_store(aio_id), "data"),
+            Input(GraphSettingsDisplay.ids.settings_store(aio_id), "data"),
+        )
+        def update_graph_settings(settings):
+            """Move general graph settings to the graph display"""
+            return settings

src/ssb_konjunk/dash/components/internal/__init__.py

@@ -0,0 +1,77 @@
+from typing import Literal
+from datetime import date
+import polars as pl
+
+from .loading_test import AGG_TYPES
+
+class DataSource:
+    def __init__(self, filename: str, index_col: str, date_pattern: str) -> None:
+        self.dt_colname = "dt"
+        self.data = pl.read_parquet(filename).with_columns(
+            dt=pl.col(index_col).cast(pl.String).str.to_date(date_pattern)
+        )
+
+    def filter_dt(self, lower: date, highest: date):
+        return self.data.filter(pl.col(self.dt_colname).is_between(lower, highest))
+
+    def get_unique_dates(self) -> list[date]:
+        return self.data.get_column(self.dt_colname).unique().to_list()
+
+    def get_unique_groupby(self, groupby_col: str):
+        return self.data.get_column(groupby_col).unique().to_list()
+
+    def subset_group(self, groupby_col: str, filter_val):
+        return self.data.filter(pl.col(groupby_col) == filter_val)
+
+    def _set_base_year(
+        self,
+        data: pl.DataFrame,
+        year: str,
+        col: str,
+        method: Literal["discrete", "none"],
+        agg_type: AGG_TYPES,
+    ):
+        # print(data)
+        filtered = data.filter(
+            (pl.col(self.dt_colname) < date(int(year), 12, 31))
+            & (pl.col(self.dt_colname) >= date(int(year), 1, 1))
+        )
+
+        selected_col = filtered.get_column(col).cast(pl.Float64)
+        if method == "discrete":
+            if agg_type == "AVERAGE":
+                index_factor = selected_col.mean()
+            else:
+                index_factor = selected_col.sum()
+        else:
+            index_factor = selected_col.mean()
+
+        altered = data.with_columns(
+            **{col: (pl.col(col).cast(pl.Float64) / index_factor) * 100}
+        )
+        # print(altered)
+        return altered
+
+    def set_base_year(
+        self,
+        data: pl.DataFrame,
+        year: str,
+        col: str,
+        group_col: str | None,
+        method: Literal["discrete", "none"],
+        agg_mapping: AGG_TYPES | dict[str, AGG_TYPES],
+    ) -> pl.DataFrame:
+
+        if isinstance(agg_mapping, dict):
+            agg_type = agg_mapping.get(col, "AVERAGE")
+        elif isinstance(agg_mapping, str):
+            agg_type = agg_mapping
+        else:
+            raise ValueError("")
+
+        if group_col is not None:
+            return data.group_by(group_col).map_groups(
+                lambda x: self._set_base_year(x, year, col, method, agg_type)
+            )
+        else:
+            return self._set_base_year(data, year, col, method, agg_type)