mirroring multiflex year system for event dash

as above

event_dash_lib/data_viewer.py

@@ -64,6 +64,8 @@ def write(self, data, data_key: str=None, st_loc=st, columns: list[str]=None):
     def lineplot(
         self,
         df: pd.DataFrame,
+        month_reindex: list[int] = None,
+        year_reindex: list[int] = None,
         totals: pd.Series = None,
         categories: list[str] = None,
         cumulative: bool = False,
@@ -154,8 +156,17 @@ def lineplot(
         fig = plt.figure(figsize=(fig_width, fig_height))
         ax = plt.gca()
 
-        if df.index.name == 'Month':
-            plt.xticks(xs,[calendar.month_abbr[i] for i in xs])
+        is_empty = True
+        for cols in df.columns:
+            if sum(list(df[cols])) != 0:
+                is_empty = False
+                break
+
+        if not is_empty:
+            if df.index.name == 'Reindexed Month':
+                plt.xticks(xs, [calendar.month_abbr[month_reindex[i-1]] for i in xs])
+            elif df.index.name == 'Reindexed Year':
+                plt.xticks(xs, year_reindex)
         for j, category_j in enumerate(categories):
 
             ys = df[category_j]

event_dash_lib/interface.py

@@ -8,6 +8,7 @@
 import numpy as np
 import pandas as pd
 import streamlit as st
+import calendar
 
 import matplotlib
 import matplotlib.pyplot as plt
@@ -91,23 +92,29 @@ def request_data_axes(
                 index = display_defaults.get(key + '_ind', 0),
             )
 
-            if value == 'Year Spotlight':
-                value2, ind2 = selectbox(
-                    st_loc,
-                    'what year do you want to spotlight?',
-                    options = list(range(min_year, (max_year+1), 1))
-                )
-                value = value + ':' + str(value2)
-            if value == 'Month across all Years':
-                value2, ind2 = selectbox(
-                    st_loc,
-                    'what month do you want to spotlight?',
-                    options= ['January', 'February', 'March', 'April', 'May', 'June', 'July','August','September','October','November','December']
-                )
-                value = value + ':' + value2
+            if value == 'Year(Flexible)':
+                month_dict = {'January(Calendar Year)':1, 'February':2, 'March':3,'April(Reporting Year)':4,'May':5,'June':6,'July':7,'August':8,'September(Fiscal Year)':9,'October':10,'November':11,'December':12}
+                col1, col2 = st_loc.columns(2)
+                with col1:
+                    value_month, ind_month = selectbox(
+                        st_loc, 
+                        'starting month for twelve-month recording period',
+                        options = list(month_dict.keys())
+                    )
+                    value = value + ':' + str(month_dict[value_month])
+                with col2:
+                    if month_dict[value_month] >= 9:
+                        min_year = min_year - 1
+                    start_year, end_year = st_loc.select_slider(
+                            'years to view',
+                            options=list(range(min_year,max_year+1)),
+                            value=(min_year, max_year),
+                    )
+                    value = value + ':' + str(start_year) + ':' + str(end_year)
 
             selected_settings[key] = value
             selected_settings[key + '_ind'] = ind
+
         key = 'y_column'
         if key in ask_for:
             if selected_settings['aggregation_method'] == 'count':
@@ -346,6 +353,8 @@ def request_view_settings(
             'color_palette',
             'category_colors',
             'totals',
+            'month_reindex',
+            'year_reindex',
             'kwargs'
        ]
         if ask_for == 'all':

event_dash_lib/pages/base_page.py

@@ -9,8 +9,9 @@
 import types
 
 import streamlit as st
+import pandas as pd
 
-from .. import dash_builder
+from .. import dash_builder, utils
 
 importlib.reload(dash_builder)
 
@@ -24,6 +25,7 @@ def main(config_fp: str, user_utils: types.ModuleType = None):
             Defaults to those in root_dash_lib.
     '''
 
+    pd.options.mode.copy_on_write = True
     # This must be the first streamlit command
     st.set_page_config(layout='wide')
 
@@ -59,15 +61,9 @@ def main(config_fp: str, user_utils: types.ModuleType = None):
         ),
     )
 
-    # Identify year bounds for any range calculations
-    min_year = data['preprocessed']['Calendar Year'].min()
-    max_year = data['preprocessed']['Calendar Year'].max()
-    years_to_display = list(range(min_year,max_year+1, 1))
-
-    # for fiscal year range calcs
-    min_year_fisc = data['preprocessed']['Fiscal Year'].min()
-    max_year_fisc = data['preprocessed']['Fiscal Year'].max()
-    years_to_display_fisc = list(range(min_year_fisc, max_year_fisc+1, 1))
+    # for future reference, if you want to set artificial bounds for year/timescale, do it here
+    min_year = data['preprocessed']['Date'].dt.year.min()
+    max_year = data['preprocessed']['Date'].dt.year.max()
 
     # Data axes
     # entered search category passed down to filter settings for further specification
@@ -90,20 +86,45 @@ def main(config_fp: str, user_utils: types.ModuleType = None):
         builder.settings.common['filters']['categorical'],
     )
 
-    months_to_num = {'January':1, 'February':2, 'March':3,'April':4,'May':5,'June':6,'July':7,'August':8,'September':9,'October':10,'November':11,'December':12}
-    # filters by year binning method
-    if axes_object['x_column_ind'] == 2:
-        # tosses all entries that do not fall in specified calendar year
-        year = int(axes_object['x_column'].split(':')[1])
-        data['time_adjusted'] = data['selected'][data['selected']['Date'].dt.year == year]
-        builder.settings.common['data']['x_column'] = 'Month'
-    elif axes_object['x_column_ind'] == 3:
-        # tosses all entries that do not fall in specified month across all years
-        month = str(axes_object['x_column'].split(':')[1])
-        data['time_adjusted'] = data['selected'][data['selected']['Date'].dt.month == months_to_num[month]]
-        builder.settings.common['data']['x_column'] = 'Calendar Year'
+
+
+    # filters data by year bounds selected (so that only entries which fall into this year-bound are displayed)
+    reverse_month_dict = {1:'January', 2:'February', 3:'March', 4:'April', 5:'May',6:'June', 7:'July', 8:'August', 9:'September', 10:'October', 11:'November', 12:'December'}
+    if (axes_object['x_column_ind'] == 0):
+
+
+        # extracts time information from axes_object
+        time_object = axes_object['x_column'].split(':')
+        month_start = int(time_object[1])
+        year_start = int(time_object[2])
+        year_end = int(time_object[3])
+        years_to_display = list(range(year_start+1, year_end+1))
+
+        month_redef = [x if x<=12 else x-12 for x in range(month_start, 12+month_start)]
+
+
+        data['selected']['Reindexed Year'] = utils.get_year(
+                data['selected']['Date'], "{} 1".format(reverse_month_dict[month_start])
+            )
+        data['time_adjusted'] = data['selected'][data['selected']['Reindexed Year'] == year_start]
+
+        if len(years_to_display) != 0:
+            for i in years_to_display:
+                temp = data['selected'][data['selected']['Reindexed Year'] == i]
+                data['time_adjusted'] = pd.concat([data['time_adjusted'], temp])
+
+            builder.settings.common['data']['x_column'] = 'Reindexed Year'
+        if len(years_to_display) == 0:
+
+            # For Fiscal Month visualizations
+            def month_fisc_converter(month:int, forward=True):
+                return month_redef.index(month)+1
+
+            data['time_adjusted'].loc.__setitem__((slice(None), 'Reindexed Month'), data['time_adjusted'].__getitem__('Date').dt.month.map(month_fisc_converter))
+            builder.settings.common['data']['x_column'] = 'Reindexed Month'
     else:
         data['time_adjusted'] = data['selected']
+        builder.settings.common['data']['x_column'] = 'Calendar Year'
 
     # Aggregate data
     data['aggregated'] = builder.aggregate(
@@ -125,27 +146,23 @@ def main(config_fp: str, user_utils: types.ModuleType = None):
 
     ### adds all years for which we have data back into aggregated dataframe (even if all zero that time bin);
     # more accurately displays trends across multiple years
+    years_to_display.insert(0, year_start)
 
     # If you are going to change the configs for x_columns, make sure they are reflected below!
     if len(list(data['aggregated'].columns)) != 0:
         data['aggregated'] = data['aggregated'].T
         data['totals'] = data['totals'].T
 
-        if builder.settings.common['data']['x_column'] == 'Month':
-            for month in months_to_num.values():
+        if builder.settings.common['data']['x_column'] == 'Reindexed Month':
+            for month in month_redef:
                 if month not in data['aggregated'].columns:
                     data['aggregated'].insert(month-1, month, [0 for i in range(len(data['aggregated'].index))])
                     data['totals'].insert(month-1, month, [0 for i in range(len(data['totals'].index))])
-        elif builder.settings.common['data']['x_column'] == 'Fiscal Year':
-            for years in years_to_display_fisc:
-                if years not in data['aggregated'].columns:
-                    data['aggregated'].insert(years-min_year_fisc, years, [0 for i in range(len(data['aggregated'].index))])
-                    data['totals'].insert(years-min_year_fisc, years, [0 for i in range(len(data['totals'].index))])
-        else:
+        elif builder.settings.common['data']['x_column'] == 'Reindexed Year':
             for years in years_to_display:
                 if years not in data['aggregated'].columns:
-                    data['aggregated'].insert(years-min_year, years, [0 for i in range(len(data['aggregated'].index))])
-                    data['totals'].insert(years-min_year, years, [0 for i in range(len(data['totals'].index))])
+                    data['aggregated'].insert(years-min(years_to_display), years, [0 for i in range(len(data['aggregated'].index))])
+                    data['totals'].insert(years-min(years_to_display), years, [0 for i in range(len(data['totals'].index))])
 
         data['aggregated'] = data['aggregated'].T
         data['totals'] = data['totals'].T
@@ -158,7 +175,7 @@ def main(config_fp: str, user_utils: types.ModuleType = None):
     # Lineplot
     local_key = 'lineplot'
     st.header(config.get('lineplot_header', 'Lineplot'))
-    st.text("Note: some data entries may correspond to multiple categories, and so may be contribute to dataset of each.\n As such, the all categories combined may exceed the total, which only counts each entry once***")
+    st.text("Note: entries may correspond to multiple data tags/categories, and so may contribute to the trendline of each (double-counting).\n this results in a conflict between aggregated data and total trendline, which only counts each entry once.\n Users and audience must be made clear of this discrepancy; in essence, total single-counts, aggregate double-counts")
     with st.expander('Lineplot settings'):
         local_opt_keys, common_opt_keys, unset_opt_keys = builder.settings.get_local_global_and_unset(
             function=builder.data_viewer.lineplot,
@@ -181,12 +198,16 @@ def main(config_fp: str, user_utils: types.ModuleType = None):
     if toggle:
         builder.data_viewer.lineplot(
             df = data['aggregated'],
+            month_reindex = month_redef if builder.settings.common['data']['x_column_ind'] == 0 else None, 
+            year_reindex = years_to_display,
             totals = data['totals'],
             **builder.settings.get_settings(local_key)
         )
     else:
         builder.data_viewer.lineplot(
             df = data['aggregated'],
+            month_reindex = month_redef if builder.settings.common['data']['x_column_ind'] == 0 else None, 
+            year_reindex=years_to_display,
             **builder.settings.get_settings(local_key)
         )
 

event_dash_lib/user_utils.py

@@ -165,7 +165,6 @@ def preprocess_data(cleaned_df, config):
     )
 
     preprocessed_df['Calendar Year'] = preprocessed_df['Date'].dt.year
-    preprocessed_df['Month'] = preprocessed_df['Date'].dt.month
 
     # Tweaks to the press data
     #if 'Title (optional)' in preprocessed_df.columns:

src/config.yml

@@ -48,10 +48,8 @@ numerical_columns: # Numeric columns that can be summed
 date_columns: # Dates
   - Date
 x_columns: # Data will be binned by these along the x-axis
-  - Fiscal Year # For example, a value of 2019 would be grouped into the 2019-2020 financial year
-  - Calendar Year
-  - Year Spotlight
-  - Month across all Years
+  - Year(Flexible) # For example, a value of 2019 would be grouped into the 2019-2020 financial year
+  - Adm
 categorical_columns: # Categorical columns that can be grouped, e.g. all Press Types=="Science" articles
   - Research Topic
   - Calendar Group