add auto-tag feature

zalepa · zalepa · commit 1d5a3f5b676e · 2026-03-22T08:42:40.000-04:00
diff --git a/app/controllers/admin/meetings_controller.rb b/app/controllers/admin/meetings_controller.rb
@@ -162,6 +162,18 @@ def publish
       end
     end
 
+    def auto_tag
+      @meeting = Meeting.find(params[:id])
+      version = @meeting.current_version
+      items = version&.agenda_items&.to_a || []
+
+      AutoTaggingService.new(items).call
+
+      tagged_count = items.count { |i| i.tags.reload.any? }
+      redirect_to admin_meeting_path(@meeting),
+        notice: "Auto-tagged #{tagged_count} of #{items.size} items."
+    end
+
     def destroy
       @meeting = Meeting.find(params[:id])
       @meeting.destroy!
diff --git a/app/controllers/admin/tag_rules_controller.rb b/app/controllers/admin/tag_rules_controller.rb
@@ -0,0 +1,37 @@
+module Admin
+  class TagRulesController < BaseController
+    def create
+      @tag = Tag.find(params[:tag_id])
+      @rule = @tag.tag_rules.build(rule_params)
+
+      if @rule.save
+        respond_to do |format|
+          format.turbo_stream
+          format.html { redirect_to admin_tags_path }
+        end
+      else
+        respond_to do |format|
+          format.turbo_stream { render turbo_stream: turbo_stream.replace("new_rule_form_#{@tag.id}", partial: "admin/tag_rules/form", locals: { tag: @tag, rule: @rule }) }
+          format.html { redirect_to admin_tags_path, alert: @rule.errors.full_messages.to_sentence }
+        end
+      end
+    end
+
+    def destroy
+      @rule = TagRule.find(params[:id])
+      @tag = @rule.tag
+      @rule.destroy
+
+      respond_to do |format|
+        format.turbo_stream
+        format.html { redirect_to admin_tags_path }
+      end
+    end
+
+    private
+
+    def rule_params
+      params.require(:tag_rule).permit(:pattern, :match_type)
+    end
+  end
+end
diff --git a/app/controllers/admin/tags_controller.rb b/app/controllers/admin/tags_controller.rb
@@ -7,6 +7,7 @@ class TagsController < BaseController
 
     def index
       @tags = Tag.left_joins(:agenda_item_tags)
+                  .includes(:tag_rules)
                   .select("tags.*, COUNT(agenda_item_tags.id) AS items_count")
                   .group("tags.id")
 
@@ -43,6 +44,11 @@ def destroy
       redirect_to admin_tags_path(q: params[:q]), notice: "Tag \"#{@tag.name}\" deleted."
     end
 
+    def seed_rules
+      AutoTaggingService.seed_default_rules!
+      redirect_to admin_tags_path, notice: "Default auto-tag rules seeded."
+    end
+
     def search
       tags = Tag.search(params[:q].to_s).alphabetical.limit(10)
       render json: tags.map { |t| { id: t.id, name: t.name } }
diff --git a/app/javascript/controllers/rule_form_controller.js b/app/javascript/controllers/rule_form_controller.js
@@ -0,0 +1,25 @@
+import { Controller } from "@hotwired/stimulus"
+
+export default class extends Controller {
+  static targets = ["input", "matchType", "typeButton"]
+
+  keydown(event) {
+    if (event.key === "Enter") {
+      event.preventDefault()
+      if (this.inputTarget.value.trim() !== "") {
+        this.element.requestSubmit()
+      }
+    }
+  }
+
+  toggleType() {
+    const current = this.matchTypeTarget.value
+    if (current === "keyword") {
+      this.matchTypeTarget.value = "phrase"
+      this.typeButtonTarget.textContent = "P"
+    } else {
+      this.matchTypeTarget.value = "keyword"
+      this.typeButtonTarget.textContent = "K"
+    }
+  }
+}
diff --git a/app/models/tag.rb b/app/models/tag.rb
@@ -2,6 +2,7 @@ class Tag < ApplicationRecord
   include Starrable
   has_many :agenda_item_tags, dependent: :destroy
   has_many :agenda_items, through: :agenda_item_tags
+  has_many :tag_rules, dependent: :destroy
 
   validates :name, presence: true, uniqueness: { case_sensitive: false }
 
diff --git a/app/models/tag_rule.rb b/app/models/tag_rule.rb
@@ -0,0 +1,22 @@
+class TagRule < ApplicationRecord
+  belongs_to :tag
+
+  enum :match_type, {
+    keyword: "keyword",
+    phrase: "phrase"
+  }
+
+  validates :pattern, presence: true
+  validates :match_type, presence: true
+
+  normalizes :pattern, with: ->(p) { p.strip }
+
+  def to_regex
+    escaped = Regexp.escape(pattern)
+    if keyword?
+      /\b#{escaped}\b/i
+    else
+      /\b#{escaped}/i
+    end
+  end
+end
diff --git a/app/services/agenda_apply_changes_service.rb b/app/services/agenda_apply_changes_service.rb
@@ -6,6 +6,7 @@ def initialize(meeting, new_data, accepted_changes)
     @new_data = new_data
     @accepted_changes = accepted_changes # array of item_numbers to accept
     @errors = []
+    @changed_items = []
   end
 
   def call
@@ -18,6 +19,8 @@ def call
     ActiveRecord::Base.transaction do
       apply_accepted_changes(version)
       raise ActiveRecord::Rollback if @errors.any?
+
+      AutoTaggingService.new(@changed_items).call if @changed_items.any?
     end
 
     self
@@ -56,6 +59,7 @@ def update_item(old_item, new_data)
       url: new_data[:url],
       item_type: new_data[:item_type]
     )
+    @changed_items << old_item
   rescue ActiveRecord::RecordInvalid => e
     @errors << "Failed to update #{old_item.item_number}: #{e.message}"
   end
@@ -76,7 +80,9 @@ def add_item(version, new_data)
       position: position
     )
 
-    unless item.save
+    if item.save
+      @changed_items << item
+    else
       @errors << "Failed to add #{new_data[:item_number]}: #{item.errors.full_messages.join(', ')}"
     end
   end
diff --git a/app/services/agenda_import_service.rb b/app/services/agenda_import_service.rb
@@ -27,6 +27,8 @@ def call
 
       build_sections_and_items
       raise ActiveRecord::Rollback if @errors.any?
+
+      auto_tag_new_items
     end
 
     if @errors.any?
@@ -46,6 +48,10 @@ def new_version?
 
   private
 
+  def auto_tag_new_items
+    AutoTaggingService.new(@agenda_version.agenda_items.reload.to_a).call
+  end
+
   def validate_structure!
     unless @data.is_a?(Hash) && @data["meeting"].is_a?(Hash) && @data["sections"].is_a?(Array)
       @errors << "Invalid JSON structure: must contain 'meeting' object and 'sections' array"
diff --git a/app/services/auto_tagging_service.rb b/app/services/auto_tagging_service.rb
@@ -0,0 +1,103 @@
+class AutoTaggingService
+  # Default rules used to seed tag_rules when no rules exist in the database.
+  # Format: { tag: "tag name", rules: [{ pattern: "text", match_type: "keyword"|"phrase" }] }
+  DEFAULT_RULES = [
+    { tag: "parking",         rules: [ { pattern: "parking", match_type: "keyword" } ] },
+    { tag: "taxes",           rules: [ { pattern: "tax abate", match_type: "phrase" },
+                                        { pattern: "property tax", match_type: "phrase" },
+                                        { pattern: "payroll tax", match_type: "phrase" },
+                                        { pattern: "real estate tax", match_type: "phrase" } ] },
+    { tag: "payroll",         rules: [ { pattern: "payroll", match_type: "keyword" },
+                                        { pattern: "salary", match_type: "keyword" },
+                                        { pattern: "salaries", match_type: "keyword" } ] },
+    { tag: "fees",            rules: [ { pattern: "fee schedule", match_type: "phrase" } ] },
+    { tag: "street safety",   rules: [ { pattern: "vehicles and traffic", match_type: "phrase" } ] },
+    { tag: "bike lanes",      rules: [ { pattern: "bicycle", match_type: "keyword" },
+                                        { pattern: "bike", match_type: "keyword" },
+                                        { pattern: "greenway", match_type: "keyword" },
+                                        { pattern: "scooter", match_type: "phrase" } ] },
+    { tag: "environmental",   rules: [ { pattern: "soil remed", match_type: "phrase" },
+                                        { pattern: "chromate", match_type: "keyword" },
+                                        { pattern: "environmental protection", match_type: "phrase" },
+                                        { pattern: "contaminat", match_type: "phrase" },
+                                        { pattern: "flood hazard", match_type: "phrase" },
+                                        { pattern: "underground storage tank", match_type: "phrase" } ] },
+    { tag: "building",        rules: [ { pattern: "redevelopment plan", match_type: "phrase" },
+                                        { pattern: "land development", match_type: "phrase" } ] },
+    { tag: "leases",          rules: [ { pattern: "lease", match_type: "keyword" } ] },
+    { tag: "Infrastructure",  rules: [ { pattern: "department of infrastructure", match_type: "phrase" },
+                                        { pattern: "division of engineering", match_type: "phrase" } ] },
+    { tag: "budget",          rules: [ { pattern: "budget", match_type: "keyword" },
+                                        { pattern: "appropriation", match_type: "phrase" } ] },
+    { tag: "cannabis",        rules: [ { pattern: "cannabis", match_type: "keyword" } ] },
+    { tag: "parks",           rules: [ { pattern: "park maintenance", match_type: "phrase" },
+                                        { pattern: "recreation and youth", match_type: "phrase" },
+                                        { pattern: "pershing field", match_type: "phrase" },
+                                        { pattern: "liberty state park", match_type: "phrase" } ] },
+    { tag: "housing",         rules: [ { pattern: "affordable housing", match_type: "phrase" },
+                                        { pattern: "housing trust", match_type: "phrase" } ] },
+    { tag: "appointments",    rules: [ { pattern: "appointing", match_type: "keyword" },
+                                        { pattern: "appointment of", match_type: "phrase" } ] },
+    { tag: "claims",          rules: [ { pattern: "payment of a claim", match_type: "phrase" },
+                                        { pattern: "settlement of the action", match_type: "phrase" } ] },
+    { tag: "grants",          rules: [ { pattern: "community development block grant", match_type: "phrase" },
+                                        { pattern: "grant fund", match_type: "phrase" } ] },
+    { tag: "public safety",   rules: [ { pattern: "department of public safety", match_type: "phrase" },
+                                        { pattern: "division of fire", match_type: "phrase" },
+                                        { pattern: "division of police", match_type: "phrase" } ] },
+    { tag: "senior services", rules: [ { pattern: "senior citizen", match_type: "phrase" },
+                                        { pattern: "senior center", match_type: "phrase" },
+                                        { pattern: "senior meal", match_type: "phrase" },
+                                        { pattern: "senior lunch", match_type: "phrase" },
+                                        { pattern: "senior affair", match_type: "phrase" },
+                                        { pattern: "congregate", match_type: "phrase" } ] },
+    { tag: "water",           rules: [ { pattern: "water agreement", match_type: "phrase" },
+                                        { pattern: "jcmua", match_type: "keyword" } ] }
+  ].freeze
+
+  def initialize(agenda_items)
+    @agenda_items = Array(agenda_items)
+  end
+
+  def call
+    return if @agenda_items.empty?
+
+    load_rules
+    @agenda_items.each { |item| tag_item(item) }
+  end
+
+  # Seeds default rules for tags that have no rules yet.
+  # Safe to call multiple times — only adds rules to tags that have none.
+  def self.seed_default_rules!
+    DEFAULT_RULES.each do |entry|
+      tag = Tag.where("LOWER(name) = ?", entry[:tag].downcase).first_or_create!(name: entry[:tag])
+      next if tag.tag_rules.any?
+
+      entry[:rules].each do |rule|
+        tag.tag_rules.create!(pattern: rule[:pattern], match_type: rule[:match_type])
+      end
+    end
+  end
+
+  private
+
+  def load_rules
+    @rules = Tag.joins(:tag_rules)
+                .includes(:tag_rules)
+                .distinct
+                .map do |tag|
+                  regexes = tag.tag_rules.map(&:to_regex)
+                  { tag: tag, regexes: regexes }
+                end
+  end
+
+  def tag_item(item)
+    title = item.title.to_s
+
+    @rules.each do |rule|
+      if rule[:regexes].any? { |pat| pat.match?(title) }
+        item.agenda_item_tags.find_or_create_by!(tag: rule[:tag])
+      end
+    end
+  end
+end
diff --git a/app/views/admin/meetings/show.html.erb b/app/views/admin/meetings/show.html.erb
@@ -32,6 +32,9 @@
           <%= button_to "Unpublish", publish_admin_meeting_path(@meeting, version_id: @agenda_version.id), method: :post, class: "px-3.5 py-2 text-sm font-medium text-amber-700 bg-amber-50 rounded-lg hover:bg-amber-100 ring-1 ring-amber-600/10 transition-colors cursor-pointer", data: { turbo_confirm: "Are you sure you want to unpublish this version? It will no longer be visible to the public." } %>
         <% end %>
       <% end %>
+      <% if @agenda_version %>
+        <%= button_to "Auto Tag", auto_tag_admin_meeting_path(@meeting), method: :post, class: "px-3.5 py-2 text-sm font-medium text-blue-700 bg-blue-50 rounded-lg hover:bg-blue-100 ring-1 ring-blue-600/10 transition-colors cursor-pointer", data: { turbo_confirm: "Auto-tag all items in this meeting based on keyword rules?" } %>
+      <% end %>
       <% if @meeting.minutes_imported? %>
         <%= button_to "Delete Minutes", delete_minutes_admin_meeting_path(@meeting), method: :delete, class: "px-3.5 py-2 text-sm font-medium text-amber-700 bg-amber-50 rounded-lg hover:bg-amber-100 ring-1 ring-amber-600/10 transition-colors cursor-pointer", data: { turbo_confirm: "Are you sure you want to delete minutes data? This will remove all vote records for this meeting." } %>
       <% end %>
diff --git a/app/views/admin/tag_rules/_form.html.erb b/app/views/admin/tag_rules/_form.html.erb
@@ -0,0 +1,12 @@
+<%= form_with model: rule, url: admin_tag_rules_path, class: "inline-flex items-center gap-1", data: { controller: "rule-form" } do |f| %>
+  <%= f.hidden_field :match_type, value: "keyword", data: { rule_form_target: "matchType" } %>
+  <input type="hidden" name="tag_id" value="<%= tag.id %>" />
+  <input type="text" name="tag_rule[pattern]" value="<%= rule.pattern %>" placeholder="+ add rule"
+         class="w-20 focus:w-32 transition-all px-1.5 py-0.5 text-xs rounded border border-dashed border-gray-300 focus:border-blue-400 focus:ring-0 placeholder-gray-400"
+         data-action="keydown->rule-form#keydown" data-rule-form-target="input" autocomplete="off" />
+  <button type="button" data-action="click->rule-form#toggleType" data-rule-form-target="typeButton"
+          class="text-[10px] text-gray-400 hover:text-gray-600 cursor-pointer" title="Toggle: K=keyword (exact word), P=phrase (starts with)">K</button>
+<% end %>
+<% if rule.errors.any? %>
+  <span class="text-xs text-red-600"><%= rule.errors.full_messages.first %></span>
+<% end %>
diff --git a/app/views/admin/tag_rules/_rule.html.erb b/app/views/admin/tag_rules/_rule.html.erb
@@ -0,0 +1,7 @@
+<span id="<%= dom_id(rule) %>" class="inline-flex items-center gap-1 px-2 py-0.5 rounded text-xs bg-gray-100 text-gray-700 group">
+  <span class="<%= rule.phrase? ? "italic" : "" %>"><%= rule.pattern %></span>
+  <span class="text-gray-400 text-[10px]"><%= rule.match_type[0].upcase %></span>
+  <%= button_to admin_tag_rule_path(rule), method: :delete, class: "text-gray-400 hover:text-red-600 cursor-pointer opacity-0 group-hover:opacity-100 transition-opacity", form: { class: "inline" } do %>
+    <svg class="w-3 h-3" fill="none" stroke="currentColor" viewBox="0 0 24 24"><path stroke-linecap="round" stroke-linejoin="round" stroke-width="2" d="M6 18L18 6M6 6l12 12"/></svg>
+  <% end %>
+</span>
diff --git a/app/views/admin/tag_rules/create.turbo_stream.erb b/app/views/admin/tag_rules/create.turbo_stream.erb
@@ -0,0 +1,8 @@
+<%= turbo_stream.append "tag_rules_#{@tag.id}" do %>
+  <%= render partial: "admin/tag_rules/rule", locals: { rule: @rule } %>
+<% end %>
+<%= turbo_stream.replace "new_rule_form_#{@tag.id}" do %>
+  <div id="new_rule_form_<%= @tag.id %>">
+    <%= render partial: "admin/tag_rules/form", locals: { tag: @tag, rule: @tag.tag_rules.build } %>
+  </div>
+<% end %>
diff --git a/app/views/admin/tag_rules/destroy.turbo_stream.erb b/app/views/admin/tag_rules/destroy.turbo_stream.erb
@@ -0,0 +1 @@
+<%= turbo_stream.remove dom_id(@rule) %>
diff --git a/app/views/admin/tags/_tag.html.erb b/app/views/admin/tags/_tag.html.erb
@@ -18,6 +18,16 @@
       <span class="text-gray-600"><%= tag.items_count %> <%= "item".pluralize(tag.items_count) %></span>
     <% end %>
   </td>
+  <td class="px-4 py-3 sm:px-6 text-sm">
+    <div id="tag_rules_<%= tag.id %>" class="flex flex-wrap gap-1.5">
+      <% tag.tag_rules.each do |rule| %>
+        <%= render partial: "admin/tag_rules/rule", locals: { rule: rule } %>
+      <% end %>
+      <div id="new_rule_form_<%= tag.id %>">
+        <%= render partial: "admin/tag_rules/form", locals: { tag: tag, rule: tag.tag_rules.build } %>
+      </div>
+    </div>
+  </td>
   <td class="px-4 py-3 sm:px-6 text-right text-sm space-x-2">
     <button type="button" class="text-blue-600 hover:text-blue-800 font-medium cursor-pointer" data-action="inline-edit#edit">Edit</button>
     <%= button_to "Delete", admin_tag_path(tag),
diff --git a/app/views/admin/tags/index.html.erb b/app/views/admin/tags/index.html.erb
@@ -1,6 +1,15 @@
 <div class="mb-6 sm:mb-8">
-  <h1 class="text-xl sm:text-2xl font-bold text-gray-900">Tags</h1>
-  <p class="mt-1 text-sm text-gray-500">Manage tags across all agenda items. Tags with 0 items can be safely removed.</p>
+  <div class="flex items-center justify-between">
+    <div>
+      <h1 class="text-xl sm:text-2xl font-bold text-gray-900">Tags</h1>
+      <p class="mt-1 text-sm text-gray-500">Manage tags and auto-tag rules. Tags with 0 items can be safely removed.</p>
+    </div>
+    <% if TagRule.none? %>
+      <%= button_to "Seed Default Rules", seed_rules_admin_tags_path, method: :post,
+            class: "px-4 py-2 text-sm font-medium text-blue-700 bg-blue-50 rounded-lg hover:bg-blue-100 ring-1 ring-blue-600/10 transition-colors cursor-pointer",
+            data: { turbo_confirm: "Create default auto-tag rules for common topics?" } %>
+    <% end %>
+  </div>
 </div>
 
 <%= form_with url: admin_tags_path, method: :get, local: true, class: "mb-6" do %>
@@ -25,6 +34,7 @@
         <th class="px-4 py-3 sm:px-6 text-left text-xs font-medium uppercase">
           <%= sortable_column("items", "Items", path_helper: :admin_tags_path, default_column: "items") %>
         </th>
+        <th class="px-4 py-3 sm:px-6 text-left text-xs font-medium text-gray-500 uppercase">Auto-tag Rules</th>
         <th class="px-4 py-3 sm:px-6 text-right text-xs font-medium text-gray-500 uppercase">Actions</th>
       </tr>
     </thead>
diff --git a/config/routes.rb b/config/routes.rb
@@ -16,11 +16,14 @@
       get :preview_agenda, on: :member
       post :apply_agenda, on: :member
       get :reupload_agenda, on: :member
+      post :auto_tag, on: :member
     end
     resources :council_members
     resources :tags, only: [ :index, :update, :destroy ] do
       get :search, on: :collection
+      post :seed_rules, on: :collection
     end
+    resources :tag_rules, only: [ :create, :destroy ]
     resources :agenda_items, only: [ :update, :destroy ]
     resources :agenda_item_tags, only: [ :create, :destroy ] do
       post :copy, on: :collection
diff --git a/db/migrate/20260322123323_create_tag_rules.rb b/db/migrate/20260322123323_create_tag_rules.rb
diff --git a/db/schema.rb b/db/schema.rb
diff --git a/lib/tasks/auto_tag.rake b/lib/tasks/auto_tag.rake
diff --git a/test/fixtures/tag_rules.yml b/test/fixtures/tag_rules.yml
diff --git a/test/fixtures/tags.yml b/test/fixtures/tags.yml
diff --git a/test/services/auto_tagging_service_test.rb b/test/services/auto_tagging_service_test.rb
