Merge pull request #870 from VisLab/develop

WordCoud allows font_path and EventManager now checks for unsorted onsets.

hed/tools/analysis/event_manager.py

@@ -1,5 +1,7 @@
 """ Manager of events of temporal extent. """
+import pandas as pd
 
+from hed.errors import HedFileError
 from hed.models import HedString
 from hed.models.model_constants import DefTagNames
 from hed.models.df_util import get_assembled
@@ -31,7 +33,10 @@ def __init__(self, input_data, hed_schema, extra_defs=None):
         self.hed_schema = hed_schema
         self.input_data = input_data
         self.def_dict = input_data.get_def_dict(hed_schema, extra_def_dicts=extra_defs)
-        self.onsets = input_data.dataframe['onset'].tolist()
+        onsets = pd.to_numeric(input_data.dataframe['onset'], errors='coerce')
+        if not onsets.is_monotonic_increasing:
+            raise HedFileError("OnsetsNotOrdered", "The onset values must be non-decreasing", "")
+        self.onsets = onsets.tolist()
         self.hed_strings = None  # Remaining HED strings copy.deepcopy(hed_strings)
         self._create_event_list(input_data)
 

hed/tools/remodeling/operations/summarize_hed_tags_op.py

@@ -91,6 +91,12 @@ class SummarizeHedTagsOp(BaseOp):
                     "max_font_size": {
                         "type": "number"
                     },
+                    "set_font": {
+                        "type": "boolean"
+                    },
+                    "font_path": {
+                        "type": "string"
+                    },
                     "scale_adjustment": {
                         "type": "number"
                     },
@@ -148,6 +154,8 @@ def __init__(self, parameters):
                 "prefer_horizontal": wc_params.get("prefer_horizontal", 0.75),
                 "min_font_size": wc_params.get("min_font_size", 8),
                 "max_font_size": wc_params.get("max_font_size", 15),
+                "set_font": wc_params.get("set_font", False),
+                "font_path": wc_params.get("font_path", ""),
                 "scale_adjustment": wc_params.get("scale_adjustment", 7),
                 "contour_width": wc_params.get("contour_width", 3),
                 "contour_color": wc_params.get("contour_color", 'black'),
@@ -305,7 +313,8 @@ def save_visualizations(self, save_dir, file_formats=['.svg'], individual_summar
         tag_wc = create_wordcloud(word_dict, mask_path=wc["mask_path"], width=wc["width"], height=wc["height"],
                                   prefer_horizontal=wc["prefer_horizontal"], background_color=wc["background_color"],
                                   min_font_size=wc["min_font_size"], max_font_size=wc["max_font_size"],
-                                  contour_width=wc["contour_width"], contour_color=wc["contour_color"])
+                                  contour_width=wc["contour_width"], contour_color=wc["contour_color"],
+                                  set_font=wc["set_font"], font_path=wc["font_path"])
         svg_data = word_cloud_to_svg(tag_wc)
         cloud_filename = os.path.realpath(os.path.join(save_dir, self.sum_op.summary_name,
                                                        self.sum_op.summary_name + '_word_cloud.svg'))

hed/tools/visualization/tag_word_cloud.py

@@ -2,11 +2,12 @@
 
 import numpy as np
 from PIL import Image
+from hed.errors import HedFileError
 from hed.tools.visualization.word_cloud_util import default_color_func, WordCloud, generate_contour_svg
 import matplotlib.font_manager as fm
 
 
-def create_wordcloud(word_dict, mask_path=None, background_color=None, width=400, height=300, font_path=None, **kwargs):
+def create_wordcloud(word_dict, mask_path=None, background_color=None, width=400, height=300, set_font=False, **kwargs):
     """ Takes a word dict and returns a generated word cloud object.
 
     Parameters:
@@ -45,10 +46,12 @@ def create_wordcloud(word_dict, mask_path=None, background_color=None, width=400
     kwargs.setdefault('relative_scaling', 1)
     kwargs.setdefault('max_font_size', height / 20)
     kwargs.setdefault('min_font_size', 8)
-    if font_path and not font_path.endswith(".ttf") and not font_path.endswith(".otf"):
-        font_path = fm.findfont(font_path)
+    if not set_font or 'font_path' not in kwargs:
+        kwargs['font_path'] = None
+    elif kwargs['font_path'] and not kwargs['font_path'].endswith((".ttf", ".otf", ".TTF", ".OTF")):
+        raise HedFileError("InvalidFontPath", f"Font {kwargs['font_path']} not valid on this system", "")
 
-    wc = WordCloud(font_path=font_path, background_color=background_color, mask=mask_image,
+    wc = WordCloud(background_color=background_color, mask=mask_image,
                    width=width, height=height, mode="RGBA", **kwargs)
 
     wc.generate_from_frequencies(word_dict)
@@ -111,4 +114,4 @@ def load_and_resize_mask(mask_path, width=None, height=None):
                                      (mask_image_array[:, :, 1] > 127) &
                                      (mask_image_array[:, :, 2] > 127)), 255, 0)
 
-        return mask_image_array.astype(np.uint8)
+        return mask_image_array.astype(np.uint8)

tests/tools/analysis/test_event_manager.py

@@ -1,6 +1,9 @@
 import os
 import unittest
+import numpy as np
+import pandas as pd
 
+from hed.errors import HedFileError
 from hed.models.sidecar import Sidecar, HedString
 from hed.models.tabular_input import TabularInput
 from hed.schema.hed_schema_io import load_schema_version
@@ -77,6 +80,38 @@ def test_get_type_defs(self):
         self.assertIsInstance(def_names, list)
         self.assertEqual(11, len(def_names))
 
+    def test_onset_ordering_mixed(self):
+        df = pd.DataFrame({'onset': [1, 2, '3', 3.24, 5],
+                           'HED': ['(Duration/4.0 s, Black)', '(Duration/2 s, Red)', 'Blue', 'Green', 'Label/1']})
+        manager = EventManager(TabularInput(df), self.schema)
+        self.assertIsInstance(manager, EventManager)
+        hed, base, context = manager.unfold_context()
+
+    def test_onset_ordering_bad(self):
+        df = pd.DataFrame({'onset': [1, 2, '3', 'n/a', 5],
+                           'HED': ['(Duration/4.0 s, Black)', '(Duration/2 s, Red)', 'Blue', 'n/a', 'Label/1']})
+        with self.assertRaises(HedFileError) as ex:
+            EventManager(TabularInput(df), self.schema)
+            self.assertEqual(ex.args(0), "OnsetsNotOrdered")
+        df1 = pd.DataFrame({'onset': [1, 2, 1.4, 6, 5],
+                           'HED': ['(Duration/4.0 s, Black)', '(Duration/2 s, Red)', 'Blue', 'n/a', 'Label/1']})
+        with self.assertRaises(HedFileError) as ex1:
+            EventManager(TabularInput(df1), self.schema)
+            self.assertEqual(ex1.args(0), "OnsetsNotOrdered")
+
+        df2 = pd.DataFrame({'onset': [1, np.nan, 1.4, 6, 5],
+                           'HED': ['(Duration/4.0 s, Black)', '(Duration/2 s, Red)', 'Blue', 'n/a', 'Label/1']})
+        with self.assertRaises(HedFileError) as ex2:
+            EventManager(TabularInput(df2), self.schema)
+            self.assertEqual(ex2.args(0), "OnsetsNotOrdered")
+
+    def test_duration_context(self):
+        df = pd.DataFrame({'onset': [1, 2, 3, 4, 5],
+                           'HED': ['(Duration/4.0 s, Black)', '(Duration/2 s, Red)', 'Blue', 'n/a', 'Label/1']})
+        manager = EventManager(TabularInput(df), self.schema)
+        hed, base, context = manager.unfold_context()
+        pass
+
 
 if __name__ == '__main__':
     unittest.main()

tests/tools/visualization/test_tag_word_cloud.py

@@ -43,7 +43,8 @@ def test_create_wordcloud_font_direct(self):
 
         fonts = fm.findSystemFonts()
         first_font = fonts[0]
-
+        x = '/C/Windows/Fonts/timesi.ttf'
+        #y = 'C:\\Windows\\Fonts\\arialbd.ttf'
         wc = tag_word_cloud.create_wordcloud(word_dict, width=width, height=height, font_path=first_font)
 
         self.assertIsInstance(wc, wordcloud.WordCloud)