newsworthycharts 1.59.0__py3-none-any.whl → 1.60.1__py3-none-any.whl

newsworthycharts/__init__.py

@@ -1,4 +1,4 @@
-__version__ = "1.59.0"
+__version__ = "1.60.1"
 
 from .chart import Chart
 from .choroplethmap import ChoroplethMap

newsworthycharts/choroplethmap.py

@@ -6,6 +6,7 @@ import geopandas as gpd
 import numpy as np
 import pandas as pd
 import mapclassify
+import matplotlib.patches as mpatches
 
 
 class ChoroplethMap(Map):
@@ -29,15 +30,27 @@ class ChoroplethMap(Map):
 
         df = self._prepare_map_data()
 
-        if self.categorical:
-            # We'll categorize manually further down the line,
-            # to easier implement custom coloring
-            pass
-            # df["data"] = pd.Categorical(
-            #     df["data"],
-            #     ordered=True,
-            # )
-        else:
+        args = {
+            "categorical": True,
+            "legend": True,  # bug in geopandas, fixed in master but not released
+            "legend_kwds": {
+                "loc": "upper left",
+                "bbox_to_anchor": (1.05, 1.0),
+            },
+            "edgecolor": "white",
+            "linewidth": 0.2,
+            "missing_kwds": {
+                "color": "gainsboro",
+            },
+        }
+        # This should be adjusted per basemap
+        label_kwargs = {
+            "bbox_to_anchor": (0.92, 0.95),
+            "loc": "upper left",
+        }
+
+        patches = []  # for legend
+        if not self.categorical:
             # mapclassify doesn't work well with nan values,
             # but we to keep them for plotting, hence
             # this hack with cutting out nan's and re-pasting them below
@@ -58,44 +71,57 @@ class ChoroplethMap(Map):
             _dict = _has_value[["id", "cats"]].set_index("id").to_dict()
             df["data"] = df["id"].map(_dict["cats"])
 
-        args = {
-            "categorical": True,
-            "legend": True,  # bug in geopandas, fixed in master but not released
-            "legend_kwds": {
-                "loc": "upper left",
-                "bbox_to_anchor": (1.05, 1.0),
-            },
-            "edgecolor": "white",
-            "linewidth": 0.2,
-            "missing_kwds": {
-                "color": "gainsboro",
-            },
-        }
-        # This should be adjusted per basemap
-        label_kwargs = {
-            "bbox_to_anchor": (0.92, 0.95),
-            "loc": "upper left",
-        }
-        if not self.categorical:
-            args["cmap"] = self.color_ramp
-            args["column"] = "data"
-        if self.categorical:
-            cat = df[~df["data"].isna()]["data"].unique()
+            # args["column"] = "data"
+            # args["cmap"] = self.color_ramp
+            # We can not provide vmin/vmax to geopandas, so we need to
+            # normalize the data ourselves, otherwise the inset maps will be off
+            import matplotlib.cm as cm
+            import matplotlib as mpl
+            norm = mpl.colors.Normalize(vmin=df["data"].min(), vmax=df["data"].max())
+            mapper = cm.ScalarMappable(norm=norm, cmap=self.color_ramp)
+            df["color"] = df["data"].apply(lambda x: mapper.to_rgba(x) if not np.isnan(x) else "gainsboro")
+            df["color"] = df["color"].fillna("gainsboro")
+            args["color"] = df["color"]
+
+            # Add labels legend (manually, Geopandas is too crude as of now)
+            fmt = self._get_value_axis_formatter()
+            for idx, cat in enumerate(binning.bins):
+                # cat is the upper limit of each category
+                if binning.counts[idx] == 1:
+                    txt_val = fmt(cat)
+                elif idx == 0:
+                    txt_val = f"– {fmt(cat)}"
+                else:
+                    txt_val = f"{fmt(binning.bins[idx - 1])} – {fmt(cat)}"
+                patches.append(mpatches.Patch(color=mapper.to_rgba(cat), label=txt_val))
+
+        elif self.categorical:
+            # We'll categorize manually further down the line,
+            # to easier implement custom coloring
+            # df["data"] = pd.Categorical(
+            #     df["data"],
+            #     ordered=True,
+            # )
+
+            cat = df[~df["data"].isna()]["data"].astype(str).unique()
             args["categories"] = cat
             if self.colors:
                 color_map = self.colors
             else:
                 color_map = {}
+                if len(cat) > len(self._nwc_style["qualitative_colors"]):
+                    raise ValueError(
+                        "Too many categories for the available colors in the current style. " +  # noqa:W504
+                        "Add a custom color map, or use a style with more categorical colors!"
+                    )
                 for idx, cat in enumerate(cat):
                     color_map[cat] = self._nwc_style["qualitative_colors"][idx]
-            df["color"] = df["data"].map(color_map)
+            df["color"] = df["data"].astype(str).map(color_map)
             df["color"] = df["color"].fillna("gainsboro")
             args["color"] = df["color"]
 
             # Geopandas does not handle legend if color keyword is used
             # We need to add it ourselves
-            import matplotlib.patches as mpatches
-            patches = []
             for label, color in color_map.items():
                 # A bit of an hack:
                 # Check if this corresponds to one of our predefined
@@ -104,44 +130,20 @@ class ChoroplethMap(Map):
                     color = self._nwc_style[f"{color}_color"]
                 patch = mpatches.Patch(color=color, label=label)
                 patches.append(patch)
-            self.ax.legend(
-                handles=patches,
-                **label_kwargs
-            )
+        if self.missing_label:
+            patches.append(mpatches.Patch(color="gainsboro", label=self.missing_label))
 
-        fig = df.plot(ax=self.ax, **args)
+        df.plot(ax=self.ax, **args)
         # Add outer edge
-        gpd.GeoSeries(df.unary_union).plot(
+        gpd.GeoSeries(df.union_all()).plot(
             ax=self.ax,
             edgecolor="lightgrey",
             linewidth=0.2,
             facecolor="none",
             color="none",
         )
-        self.ax.axis("off")
-
-        # Format numbers in legend
-        if not self.categorical:
-            leg = fig.get_legend()
-            fmt = self._get_value_axis_formatter()
-            remove_last = False
-            for lbl in leg.get_texts():
-                val = lbl.get_text()
-                if val == "NaN":  # as returned by mapclassify
-                    if self.missing_label is not None:
-                        val = self.missing_label
-                    else:
-                        remove_last = True
-                        val = ""
-                else:
-                    val = float(val)
-                    val = fmt(val)
-                lbl.set_text(val)
-            if remove_last:
-                del leg.legend_handles[-1]
-                texts = [lbl.get_text() for lbl in leg.get_texts()]
-                fig.legend(handles=leg.legend_handles, labels=texts, **label_kwargs)
 
+        self.ax.axis("off")
         for inset in self.insets:
             if "prefix" in inset:
                 _df = df[df["id"].str.startswith(inset["prefix"])].copy()
@@ -150,12 +152,11 @@ class ChoroplethMap(Map):
             if _df["data"].isnull().all():
                 # Skip if no data
                 continue
-            if self.categorical:
-                # We need a series matching the filtered data
-                args["color"] = _df["color"]
+
+            args["color"] = _df["color"]
             args["legend"] = False
             axin = self.ax.inset_axes(inset["axes"])
-            gpd.GeoSeries(_df.unary_union).plot(
+            gpd.GeoSeries(_df.union_all()).plot(
                 ax=axin,
                 edgecolor="lightgrey",
                 linewidth=0.3,
@@ -169,3 +170,8 @@ class ChoroplethMap(Map):
             r, (a, b, c, d) = self.ax.indicate_inset_zoom(axin)
             for _line in [a, b, c, d]:
                 _line.set_visible(False)
+
+        self.ax.legend(
+            handles=patches,
+            **label_kwargs,
+        )

newsworthycharts/map.py

@@ -4,7 +4,6 @@ Base class for maps
 from .chart import Chart
 from .lib.geography import haversine
 from .translations.regions import NW_MUNI_TO_CLDR
-from fiona.errors import DriverError
 import geopandas as gpd
 import pathlib
 
@@ -87,10 +86,10 @@ class Map(Chart):
 
     def __init__(self, *args, **kwargs):
         super(Map, self).__init__(*args, **kwargs)
-        self.bins = kwargs.get("bins", 9)
+        self.bins = kwargs.get("bins", 5)
         self.binning_method = kwargs.get("binning_method", "natural_breaks")
         self.colors = kwargs.get("colors", None)
-        self.color_ramp = kwargs.get("color_ramp", "YlOrRd")
+        self.color_ramp = kwargs.get("color_ramp", "YlGn")  # YlOrRd
         self.categorical = kwargs.get("categorical", False)
         self.base_map = None
         self.missing_label = None
@@ -139,12 +138,7 @@ class Map(Chart):
 
         if self.df is None:
             __dir = pathlib.Path(__file__).parent.resolve()
-            try:
-                self.df = gpd.read_file(f"{__dir}/maps/{base_map}-{subdivisions}.gpkg")
-            except DriverError:
-                raise ValueError(
-                    f"No such basemap: {_bm} (parsed as base: {base_map}, subdivisions: {subdivisions})"
-                )
+            self.df = gpd.read_file(f"{__dir}/maps/{base_map}-{subdivisions}.gpkg", engine="pyogrio")
         return base_map, subdivisions, subset, *opts
 
     def _prepare_map_data(self):

{newsworthycharts-1.59.0.dist-info → newsworthycharts-1.60.1.dist-info}/METADATA

@@ -1,28 +1,29 @@
 Metadata-Version: 2.1
 Name: newsworthycharts
-Version: 1.59.0
+Version: 1.60.1
 Summary: Matplotlib wrapper to create charts and publish them on Amazon S3
 Home-page: https://github.com/jplusplus/newsworthycharts
-Download-URL: https://github.com/jplusplus/newsworthycharts/archive/1.59.0.tar.gz
+Download-URL: https://github.com/jplusplus/newsworthycharts/archive/1.60.1.tar.gz
 Author: Jens Finnäs and Leo Wallentin, J++ Stockholm
 Author-email: stockholm@jplusplus.org
 License: MIT
 Requires-Python: >=3.9
 Description-Content-Type: text/x-rst
 License-File: LICENSE.txt
+Requires-Dist: boto3 >=1.26
+Requires-Dist: matplotlib ==3.9.2
+Requires-Dist: langcodes >=3.3
 Requires-Dist: Babel <3,>=2.14.0
-Requires-Dist: Pillow ==10.3.0
 Requires-Dist: PyYAML >=3
 Requires-Dist: adjustText ==0.7.3
-Requires-Dist: boto3 >=1.26
-Requires-Dist: geopandas ==0.14.4
-Requires-Dist: langcodes >=3.3
-Requires-Dist: mapclassify ==2.6.1
-Requires-Dist: matplotlib-label-lines ==0.5.1
-Requires-Dist: matplotlib ==3.9.0
-Requires-Dist: numpy <2,>=1.21.0
+Requires-Dist: numpy >2
 Requires-Dist: python-dateutil <3,>=2
+Requires-Dist: Pillow ==10.4.0
 Requires-Dist: requests >=2.22
+Requires-Dist: matplotlib-label-lines ==0.5.1
+Requires-Dist: geopandas >1
+Requires-Dist: mapclassify ==2.6.1
+Requires-Dist: fiona >=1.10
 
 This  module contains methods for producing graphs and publishing them on Amazon S3, or in the location of your choice.
 
@@ -136,8 +137,8 @@ These settings are available for all chart types:
 - yline = None  # A horizontal line across the chart (Matplotlib: axhline)
 - labels = []  # Optionally one label for each dataset
 - annotations = []  # Manually added annotations
-- interval = None  # yearly|quarterly|monthly|weekly|daily
-- units = 'number'  # number|percent|degrees
+- interval = None  # yearly, quarterly, monthly, weekly, daily
+- units = 'number'  # number, percent, degrees
 - show_ticks = True  # toggle category names, dates, etc
 - subtitle = None
 - note = None
@@ -236,13 +237,37 @@ To deploy a new version to [PyPi](https://pypi.org/project/newsworthycharts):
 
 Roadmap
 -------
-  - Adding more base maps
-  - Getting rid of custom settings-hack
+
+  - Get rid of custom settings-hack
+  - Remove custom charts (add missing api interfaces to Chart class instead)
+  - Remove DataWrapper class (out-of-scope)
   - Custom month locator with equal-width month bars
+  - Add color-ramp choropleth maps, as an alternative to binning
 
 Changelog
 ---------
 
+- 1.60.1
+
+  - Don't require fiona. Geopandas now support Pyogrio
+
+- 1.60.0
+
+  - [BREAKING] The default number of bins in maps is now 5, not 9
+  - [BREAKING] Default color ramp for choropleth maps is now `YlGn`
+  - GeoPandas 1.x
+  - numpy 2.x
+  - matplotlib==3.9.2
+  - Pillow==10.4.0
+  - Improved nan handling in labelling continuous choropleth maps
+  - Replace deprecated `.unary_union` with `.union_all()`
+  - Choropleth map legends now print spans for binned data
+  - Use Matplotlib's legend for choropleth maps, to have the same style as other charts, and for much improved flexibility
+  - Inset maps now (finally!) work with continuous data
+  - Better error handling in categorical maps
+  - Cast all categorical values to strings in categorical maps
+  - Replace deprecated imghdr with puremagic in tests
+
 - 1.59.0
 
   - Added `.highlight_annotation` to enable turning off the textual annotation on the highlighted point

{newsworthycharts-1.59.0.dist-info → newsworthycharts-1.60.1.dist-info}/RECORD

@@ -1,10 +1,10 @@
-newsworthycharts/__init__.py,sha256=ql3nVttdON3JXhF6EJyawP8Yqp7_5CVACZVg67qQnVk,1160
+newsworthycharts/__init__.py,sha256=4ncjyqowiBpTm3mvedckG9MB2y2Rzc4y2I-5_mVFgeU,1160
 newsworthycharts/bubblemap.py,sha256=nkocWmpiFgfjEuJGAsthjY5X7Q56jXWsZHUGXw4PwgE,2587
 newsworthycharts/categoricalchart.py,sha256=LwOZ3VbNy9vzvoK0s77AkbfMt4CXVDSAhnsnBInUIrE,14764
 newsworthycharts/chart.py,sha256=VXEJe_g_KahyiUXqXnM3vDS7SQoW1Ppktcj13-A2_oc,30631
-newsworthycharts/choroplethmap.py,sha256=2b61MuYed8ccc7uZkzuSdZDbqVR7C3OP1ftb_khuaLw,6069
+newsworthycharts/choroplethmap.py,sha256=dVRShxjkBKa57r_Kp6ysdNCXGwXh3eZK28MvVtLe7BQ,6677
 newsworthycharts/datawrapper.py,sha256=RRkAVTpfP4updKxUIBaSmKuBi2RUVPaBRF8HDQhlGGA,11250
-newsworthycharts/map.py,sha256=ruhFEFmGg8DwySDPnEbx_4n57DgHL5LlLKQEnoNmuIU,6225
+newsworthycharts/map.py,sha256=EMpsbvY4V4Y5_ky0Urg4WgvqSEKiDd2kUHXFReMRbb0,6004
 newsworthycharts/rangeplot.py,sha256=NE1W9TnmlpK6T3RvBJOU3nd73EXqkj17OY9i5zlw_cQ,8366
 newsworthycharts/scatterplot.py,sha256=6iaMoiZx__Gc-2Hcdw-8Ga5dSonrFo3oexKNmSFuir4,4959
 newsworthycharts/seasonalchart.py,sha256=rr55yqJUkaYDR9Ik98jes6574oY1U8t8LwoLE3gClW4,1967
@@ -28,8 +28,8 @@ newsworthycharts/rc/newsworthy,sha256=X0btLNrmk2DRrfOsKj_WCSIgeD6btacEN2tRF_B4m8
 newsworthycharts/translations/datawrapper_regions.csv,sha256=fzZcQRX6RFMlNNP8mpgfYNdR3Y0QAlQxDXk8FXTaWWI,9214
 newsworthycharts/translations/regions.py,sha256=Nv1McQjggD4S3JRu82rDMTG3pqUVR13E5-FBpSYbm98,239
 newsworthycharts/translations/se_municipalities.csv,sha256=br_mm-IvzQtj_W55_ATREhJ97jWnCweBFlDAVY2EBxA,7098
-newsworthycharts-1.59.0.dist-info/LICENSE.txt,sha256=Sq6kGICrehbhC_FolNdXf0djKjTpv3YqjFCIYsxdQN4,1069
-newsworthycharts-1.59.0.dist-info/METADATA,sha256=sCsOn00COBf8iSzh5TP71gP4-9vl1TnToi10UrMM4g8,27250
-newsworthycharts-1.59.0.dist-info/WHEEL,sha256=oiQVh_5PnQM0E3gPdiz09WCNmwiHDMaGer_elqB3coM,92
-newsworthycharts-1.59.0.dist-info/top_level.txt,sha256=dn_kzIj8UgUCMsh1PHdVEQJHVGSsN7Z8YJF-8xXa8n0,17
-newsworthycharts-1.59.0.dist-info/RECORD,,
+newsworthycharts-1.60.1.dist-info/LICENSE.txt,sha256=Sq6kGICrehbhC_FolNdXf0djKjTpv3YqjFCIYsxdQN4,1069
+newsworthycharts-1.60.1.dist-info/METADATA,sha256=JIG6aIDf814dmnAg3XYbfGmlsSnN2_KeJCngrP4K8Ow,28235
+newsworthycharts-1.60.1.dist-info/WHEEL,sha256=cVxcB9AmuTcXqmwrtPhNK88dr7IR_b6qagTj0UvIEbY,91
+newsworthycharts-1.60.1.dist-info/top_level.txt,sha256=dn_kzIj8UgUCMsh1PHdVEQJHVGSsN7Z8YJF-8xXa8n0,17
+newsworthycharts-1.60.1.dist-info/RECORD,,

{newsworthycharts-1.59.0.dist-info → newsworthycharts-1.60.1.dist-info}/WHEEL

@@ -1,5 +1,5 @@
 Wheel-Version: 1.0
-Generator: bdist_wheel (0.42.0)
+Generator: setuptools (74.1.2)
 Root-Is-Purelib: true
 Tag: py3-none-any