Fix timeline (#174)

* fix timeline

* add error handling and parsing of unspecified dates

* add more error handling

* check for exmpty strings; fix assigning to wrong year if object has negative year

* make coderabbit happy

* keep coderabbit happy

* improve error handling

* fix date parsing (year "0000")

* refactoring code

* prettier

* change sort order, remove period heading

* fix regex

* prettier

* improve naming

---------

Co-authored-by: Moritz Mähr <[email protected]>

_layouts/timeline_edtf.html

@@ -10,16 +10,17 @@
 {% else %}
   {%- assign items = site.data[site.metadata] | where_exp: 'item', 'item.objectid and item.parentid == nil' -%}
 {% endif %}
+{% assign regex = '[-]?[\dXu]{4,}' %}
 {%- assign items = items | where_exp: 'item', 'item[field]' -%}
 {%- assign raw-dates = items | map: field | compact | uniq -%}
-{% assign regex = '[\dX]{4}' %}
 {%- capture clean-years -%}{% for date in raw-dates %}{{date | regex_match: regex | join: ';' }}{% unless forloop.last %};{% endunless %}{%- endfor -%}{%- endcapture -%}
 {%- assign uniqueYears = clean-years | remove: ' ' | replace: ';;', ';' | split: ';' | uniq | sort -%}
+{% assign sorted_years = uniqueYears | sort_edtf %}
 {%- if site.data.theme['year-navigation'] -%}
   {%- assign navYears = site.data.theme['year-navigation'] | split: ';' -%}
 {%- elsif site.data.theme['year-nav-increment'] -%}
   {%- capture navYears -%}
-{%- for i in uniqueYears -%}{%- assign t = i | modulo: site.data.theme.year-nav-increment -%}
+{%- for i in sorted_years -%}{%- assign t = i | modulo: site.data.theme.year-nav-increment -%}
 {%- if t == 0 -%}{{ i }}{% unless forloop.last %};{% endunless %}{% endif %}{% endfor %}{%- endcapture -%}
   {%- assign navYears = navYears | split: ';' -%}
 {%- endif -%}
@@ -37,30 +38,53 @@
     </button>
     <div class="dropdown-menu" aria-labelledby="yearButton">
       {% for y in navYears %}
-        <a class="dropdown-item" href="#y{{ y }}">{{ y }}</a>
-      {%- endfor %}
+        {% assign year_parts = y | split: ':' %}
+        {% if year_parts.size == 2 %}
+          <a class="dropdown-item" href="#y{{ year_parts[1] }}">{{ year_parts[0] }}</a>
+        {% else %}
+          <a class="dropdown-item" href="#y{{ y }}">{{ y }}</a>
+        {% endif %}
+      {% endfor %}
     </div>
   </div>
 {%- endif -%}
 
 {{ content }}
-
-<h2>
-  <a href="#y{{ uniqueYears | first }}">{{ uniqueYears | first }}</a>&#8211;<a href="#y{{ uniqueYears | last }}">
-    {{- uniqueYears | last -}}
-  </a>
-</h2>
+{% comment %}uncomment the next lines to display a heading indicating the period of the data in text form{% endcomment %}
+{% comment %}
+  <h2>
+    {% assign first_year = sorted_years | first | split: ':' %}
+    {% assign last_year = sorted_years | last | split: ':' %}
+    {% if sorted_years.size > 0 %}
+      {% if first_year.size == 3
+        and last_year.size == 3
+        and first_year[0] != ''
+        and first_year[1] != ''
+        and last_year[0] != ''
+        and last_year[1] != ''
+      %}
+        <a href="#y{{ first_year[1] }}">{{ first_year[0] }}</a> &#8211;
+        <a href="#y{{ last_year[1]}}">{{ last_year[0] }}</a>
+      {% else %}
+        <span class="text-muted">Invalid date range</span>
+      {% endif %}
+    {% else %}
+      <span class="text-muted">No items found</span>
+    {% endif %}
+  </h2>
+{% endcomment %}
 
 <table id="timeline" class="table table-striped">
   <tbody>
-    {% for year in uniqueYears %}
-      <tr id="y{{ year }}">
+    {% for date in sorted_years %}
+      {% assign year = date | split: ':' %}
+      <tr id="y{{ year[1] }}">
         <th>
-          <h3>{{ year }}</h3>
+          <h3>{{ year[0] }}</h3>
         </th>
         <td>
           <div class="row">
-            {%- assign inYear = items | where_exp: 'item', 'item[field] contains year' -%}
+            {%- assign inYear = items | filter_items_by_year: year[1], '[-]?[\dXu]{4,}' -%}
             {% for item in inYear %}
               <div class="col-lg-4 col-md-6">
                 <figure class="figure">

_plugins/jekyll_regex_match.rb

@@ -20,6 +20,36 @@ def regex_match_once(input_str, regex_str)
             return nil
           end
         end
+
+        def filter_items_by_year(items, year, regex_pattern = '[-]?[\dXu]{4,}')
+          raise ArgumentError, 'Regex pattern too complex' if regex_pattern.length > 50
+          return [] unless items.is_a?(Array) && !year.nil?
+          begin
+
+          # Clear cache if it grows too large
+          if defined?(@cache_size) && @cache_size > 1000
+            items.each { |item| item.delete('cached_dates') }
+            @cache_size = 0
+          end
+
+          regex = Regexp.new(regex_pattern)  
+          items.select do |item|
+            next false unless item.is_a?(Hash) && item['date'].is_a?(String)
+
+            # Track cache size
+            @cache_size ||= 0
+            @cache_size += 1 unless item['cached_dates']
+
+            dates = item['cached_dates'] ||= item['date'].scan(regex).map(&:to_s) 
+            dates.include?(year)  
+          end
+          rescue RegexpError => e
+            Jekyll.logger.error "RegexMatch:", "Invalid regex pattern: #{e.message}"
+            # Handle invalid regex pattern
+            []
+          end
+        end
+
     end
   end
 

_plugins/sort_edtf.rb

@@ -1,10 +1,50 @@
 module Jekyll
-    module SortEDTF
-        def sort_edtf(array_of_strings)
-            sorted = array_of_strings.map { |str| str.gsub('X', '0') }
-            sorted.sort_by { |str| str[/\d+/].to_i }
-        end
+  module SortEDTF
+    def sort_edtf(array_of_strings)
+      non_empty_dates = array_of_strings.reject { |str| str.strip.empty? }
+
+      processed_dates = non_empty_dates.map do |str|
+      normalized_date_str = replace_x_with_parsable_0(str)
+      date_as_integer = normalize_to_integer(normalized_date_str)
+      display_format = create_display_format(str)
+
+      { numeric: date_as_integer, original: str, display_format: display_format }
+      end
+
+      sorted_dates = processed_dates.sort_by { |date| -date[:numeric] }
+      sorted_dates.map { |date| "#{date[:display_format]}:#{date[:original]}" }
     end
+
+    private
+
+    def replace_x_with_parsable_0(str)
+      str.gsub(/[Xu]/, '0')
+    end
+
+    def normalize_to_integer(year_str)
+      numeric_str = if year_str.start_with?('-')
+                        year_str.sub(/^-0+/, '-')
+                       else
+                        year_str.sub(/^0+(?=\d)/, '')
+                       end
+
+      if numeric_str.match?(/^-?\d+$/)
+        numeric_str.to_i
+      else
+        raise ArgumentError, "Invalid year format: #{year_str}"
+      end
+    end
+
+    def create_display_format(str)
+      if str.start_with?('-')
+        "#{str[1..-1].sub(/^0+/, '')} v. u. Z."
+      else
+        str.sub(/^0+(?=\d)/, '')
+      end
+    end
+
+
+  end
 end
 
 Liquid::Template.register_filter(Jekyll::SortEDTF)