FEATURE: Allow data explorer query result to be sent as recurring PM (#233)

This feature enables admins to create reports automatically based on a recurring schedule.

It introduces a new automation script that includes the new email_group_user field added to discourse-automation, along with a query_id and query_params to pass in parameters to the existing data explorer query.

The output of the report will be sent via pm (as a markdown table) to the recipients entered within the automation script.

The automation (supports individual users, email addresses and groups).

about.json

@@ -0,0 +1,7 @@
+{
+  "tests": {
+    "requiredPlugins": [
+      "https://github.com/discourse/discourse-automation"
+    ]
+  }
+}

config/locales/client.en.yml

@@ -96,3 +96,14 @@ en:
           descriptions:
             discourse_data_explorer:
               run_queries: "Run Data Explorer queries. Restrict the API key to a set of queries by specifying queries IDs."
+    discourse_automation:
+      scriptables:
+        recurring_data_explorer_result_pm:
+          fields:
+            recipients:
+              label: Send to User, Group or Email
+            query_id:
+              label: Data Explorer Query
+            query_params:
+              label: Data Explorer Query parameters
+

config/locales/server.en.yml

@@ -1,3 +1,8 @@
 en:
   site_settings:
     data_explorer_enabled: "Enable the Data Explorer at /admin/plugins/explorer"
+  discourse_automation:
+    scriptables:
+      recurring_data_explorer_result_pm:
+        title: Schedule a PM with Data Explorer results
+        description: Get scheduled reports sent to your messages each month

lib/report_generator.rb

@@ -0,0 +1,76 @@
+# frozen_string_literal: true
+
+module ::DiscourseDataExplorer
+  class ReportGenerator
+    def initialize(creator_user_id)
+      @creator_user_id = creator_user_id
+    end
+
+    def generate(query_id, query_params, recipients)
+      query = DiscourseDataExplorer::Query.find(query_id)
+      return [] unless query
+
+      usernames = filter_recipients_by_query_access(recipients, query)
+      return [] if usernames.empty?
+      params = params_to_hash(query_params)
+
+      result = DataExplorer.run_query(query, params)
+      query.update!(last_run_at: Time.now)
+
+      table = ResultToMarkdown.convert(result[:pg_result])
+
+      build_report_pms(query, table, usernames)
+    end
+
+    def filter_recipients_by_query_access(recipients, query)
+      return [] if recipients.empty?
+      creator = User.find(@creator_user_id)
+      return [] unless Guardian.new(creator).can_send_private_messages?
+
+      recipients.reduce([]) do |names, recipient|
+        if (group = Group.find_by(name: recipient))
+          next names unless query.query_groups.exists?(group_id: group.id)
+          next names.concat group.users.pluck(:username)
+        elsif (user = User.find_by(username: recipient))
+          next names unless Guardian.new(user).user_can_access_query?(query)
+          next names << recipient
+        end
+      end
+    end
+
+    def params_to_hash(query_params)
+      params = JSON.parse(query_params)
+      params_hash = {}
+
+      if !params.blank?
+        param_key, param_value = [], []
+        params.flatten.each.with_index do |data, i|
+          if i % 2 == 0
+            param_key << data
+          else
+            param_value << data
+          end
+        end
+
+        params_hash = Hash[param_key.zip(param_value)]
+      end
+
+      params_hash
+    end
+
+    def build_report_pms(query, table = "", usernames = [])
+      pms = []
+      usernames.flatten.compact.uniq.each do |username|
+        pm = {}
+        pm["title"] = "Scheduled Report for #{query.name}"
+        pm["target_usernames"] = Array(username)
+        pm["raw"] = "Hi #{username}, your data explorer report is ready.\n\n" +
+          "Query Name:\n#{query.name}\n\nHere are the results:\n#{table}\n\n" +
+          "<a href='/admin/plugins/explorer?id=#{query.id}'>View query in Data Explorer</a>\n\n" +
+          "Report created at #{Time.zone.now.strftime("%Y-%m-%d at %H:%M:%S")} (#{Time.zone.name})"
+        pms << pm
+      end
+      pms
+    end
+  end
+end

lib/result_to_markdown.rb

@@ -0,0 +1,49 @@
+# frozen_string_literal: true
+
+include HasSanitizableFields
+
+module ::DiscourseDataExplorer
+  class ResultToMarkdown
+    def self.convert(pg_result)
+      relations, colrender = DataExplorer.add_extra_data(pg_result)
+      result_data = []
+
+      # column names to search in place of id columns (topic_id, user_id etc)
+      cols = %w[name title username]
+
+      # find values from extra data, based on result id
+      pg_result.values.each do |row|
+        row_data = []
+
+        row.each_with_index do |col, col_index|
+          col_name = pg_result.fields[col_index]
+          related = relations.dig(colrender[col_index].to_sym) if col_index < colrender.size
+
+          if related.is_a?(ActiveModel::ArraySerializer)
+            related_row = related.object.find_by(id: col)
+            if col_name.include?("_id")
+              column = cols.find { |c| related_row.try c }
+            else
+              column = related_row.try(col_name)
+            end
+
+            if column.nil?
+              row_data[col_index] = col
+            else
+              row_data[col_index] = related_row[column]
+            end
+          else
+            row_data[col_index] = col
+          end
+        end
+
+        result_data << row_data.map { |c| "| #{sanitize_field(c.to_s)} " }.join + "|\n"
+      end
+
+      table_headers = pg_result.fields.map { |c| " #{c.gsub("_id", "")} |" }.join
+      table_body = pg_result.fields.size.times.map { " :-----: |" }.join
+
+      "|#{table_headers}\n|#{table_body}\n#{result_data.join}"
+    end
+  end
+end

plugin.rb

@@ -70,4 +70,49 @@ after_initialize do
     :discourse_data_explorer,
     { run_queries: { actions: %w[discourse_data_explorer/query#run], params: %i[id] } },
   )
+
+  require_relative "lib/report_generator"
+  require_relative "lib/result_to_markdown"
+  reloadable_patch do
+    if defined?(DiscourseAutomation)
+      DiscourseAutomation::Scriptable::RECURRING_DATA_EXPLORER_RESULT_PM =
+        "recurring_data_explorer_result_pm"
+      add_automation_scriptable(
+        DiscourseAutomation::Scriptable::RECURRING_DATA_EXPLORER_RESULT_PM,
+      ) do
+        queries =
+          DiscourseDataExplorer::Query
+            .where(hidden: false)
+            .map { |q| { id: q.id, translated_name: q.name } }
+        field :recipients, component: :email_group_user, required: true
+        field :query_id, component: :choices, required: true, extra: { content: queries }
+        field :query_params, component: :"key-value", accepts_placeholders: true
+
+        version 1
+        triggerables [:recurring]
+
+        script do |_, fields, automation|
+          recipients = Array(fields.dig("recipients", "value"))
+          query_id = fields.dig("query_id", "value")
+          query_params = fields.dig("query_params", "value")
+
+          unless SiteSetting.data_explorer_enabled
+            Rails.logger.warn "#{DiscourseDataExplorer.plugin_name} - plugin must be enabled to run automation #{automation.id}"
+            next
+          end
+
+          unless recipients.present?
+            Rails.logger.warn "#{DiscourseDataExplorer.plugin_name} - couldn't find any recipients for automation #{automation.id}"
+            next
+          end
+
+          data_explorer_report =
+            DiscourseDataExplorer::ReportGenerator.new(automation.last_updated_by_id)
+          report_pms = data_explorer_report.generate(query_id, query_params, recipients)
+
+          report_pms.each { |pm| utils.send_pm(pm, automation_id: automation.id) }
+        end
+      end
+    end
+  end
 end

spec/automation/recurring_data_explorer_result_pm_spec.rb

@@ -0,0 +1,89 @@
+# frozen_string_literal: true
+
+require "rails_helper"
+
+describe "RecurringDataExplorerResultPm" do
+  fab!(:admin) { Fabricate(:admin) }
+
+  fab!(:user) { Fabricate(:user) }
+  fab!(:another_user) { Fabricate(:user) }
+  fab!(:group_user) { Fabricate(:user) }
+  fab!(:not_allowed_user) { Fabricate(:user) }
+
+  fab!(:group) { Fabricate(:group, users: [user, another_user]) }
+  fab!(:another_group) { Fabricate(:group, users: [group_user]) }
+
+  fab!(:automation) do
+    Fabricate(
+      :automation,
+      script: DiscourseAutomation::Scriptable::RECURRING_DATA_EXPLORER_RESULT_PM,
+      trigger: "recurring",
+    )
+  end
+  fab!(:query) { Fabricate(:query) }
+  fab!(:query_group) { Fabricate(:query_group, query: query, group: group) }
+  fab!(:query_group) { Fabricate(:query_group, query: query, group: another_group) }
+
+  let!(:recipients) do
+    [user.username, not_allowed_user.username, another_user.username, another_group.name]
+  end
+
+  before do
+    SiteSetting.data_explorer_enabled = true
+    SiteSetting.discourse_automation_enabled = true
+
+    automation.upsert_field!("query_id", "choices", { value: query.id })
+    automation.upsert_field!("recipients", "email_group_user", { value: recipients })
+    automation.upsert_field!(
+      "query_params",
+      "key-value",
+      { value: [%w[from_days_ago 0], %w[duration_days 15]] },
+    )
+    automation.upsert_field!(
+      "recurrence",
+      "period",
+      { value: { interval: 1, frequency: "day" } },
+      target: "trigger",
+    )
+    automation.upsert_field!("start_date", "date_time", { value: 2.minutes.ago }, target: "trigger")
+  end
+
+  context "when using recurring trigger" do
+    it "sends the pm at recurring date_date" do
+      freeze_time 1.day.from_now do
+        expect { Jobs::DiscourseAutomationTracker.new.execute }.to change { Topic.count }.by(3)
+
+        title = "Scheduled Report for #{query.name}"
+        expect(Topic.last(3).pluck(:title)).to eq([title, title, title])
+      end
+    end
+
+    it "ensures only allowed users in recipients field receive reports via pm" do
+      expect do
+        automation.update(last_updated_by_id: admin.id)
+        automation.trigger!
+      end.to change { Topic.count }.by(3)
+
+      created_topics = Topic.last(3)
+      expect(created_topics.pluck(:archetype)).to eq(
+        [Archetype.private_message, Archetype.private_message, Archetype.private_message],
+      )
+      expect(created_topics.map { |t| t.allowed_users.pluck(:username) }).to match_array(
+        [
+          [user.username, Discourse.system_user.username],
+          [another_user.username, Discourse.system_user.username],
+          [group_user.username, Discourse.system_user.username],
+        ],
+      )
+    end
+
+    it "has appropriate content from the report generator" do
+      automation.update(last_updated_by_id: admin.id)
+      automation.trigger!
+
+      expect(Post.last.raw).to include(
+        "Hi #{group_user.username}, your data explorer report is ready.\n\nQuery Name:\n#{query.name}",
+      )
+    end
+  end
+end

spec/fabricators/query_fabricator.rb

@@ -1,13 +1,13 @@
 # frozen_string_literal: true
 
-Fabricator(:query, from: "DiscourseDataExplorer::Query") do
+Fabricator(:query, from: DiscourseDataExplorer::Query) do
-  name
+  name { sequence(:name) { |i| "cat#{i}" } }
-  description
+  description { sequence(:desc) { |i| "description #{i}" } }
-  sql
+  sql { sequence(:sql) { |i| "SELECT * FROM users limit #{i}" } }
   user
 end
 
-Fabricator(:query_group, from: "DiscourseDataExplorer::QueryGroup") do
+Fabricator(:query_group, from: DiscourseDataExplorer::QueryGroup) do
   query
   group
 end

spec/report_generator_spec.rb

@@ -0,0 +1,58 @@
+# frozen_string_literal: true
+
+require "rails_helper"
+
+describe DiscourseDataExplorer::ReportGenerator do
+  fab!(:user) { Fabricate(:user) }
+  fab!(:unauthorised_user) { Fabricate(:user) }
+  fab!(:unauthorised_group) { Fabricate(:group) }
+  fab!(:group) { Fabricate(:group, users: [user]) }
+
+  fab!(:query) { DiscourseDataExplorer::Query.find(-1) }
+  fab!(:query_group) { Fabricate(:query_group, query: query, group: group) }
+
+  let(:query_params) { [%w[from_days_ago 0], %w[duration_days 15]] }
+
+  before { SiteSetting.data_explorer_enabled = true }
+
+  describe ".generate" do
+    it "returns [] if the creator cannot send PMs" do
+      result = described_class.new(user.id).generate(query.id, query_params, [user.username])
+
+      expect(result).to eq []
+    end
+
+    it "returns [] if the recipient is not in query group" do
+      result =
+        described_class.new(user.id).generate(
+          query.id,
+          query_params,
+          [unauthorised_user.username, unauthorised_group.name],
+        )
+
+      expect(result).to eq []
+    end
+
+    it "returns a list of pms for authorised users" do
+      SiteSetting.personal_message_enabled_groups = group.id
+      DiscourseDataExplorer::ResultToMarkdown.expects(:convert).returns("le table")
+      freeze_time
+
+      result = described_class.new(user.id).generate(query.id, query_params, [user.username])
+
+      expect(result).to eq(
+        [
+          {
+            "title" => "Scheduled Report for #{query.name}",
+            "target_usernames" => [user.username],
+            "raw" =>
+              "Hi #{user.username}, your data explorer report is ready.\n\n" +
+                "Query Name:\n#{query.name}\n\nHere are the results:\nle table\n\n" +
+                "<a href='/admin/plugins/explorer?id=#{query.id}'>View query in Data Explorer</a>\n\n" +
+                "Report created at #{Time.zone.now.strftime("%Y-%m-%d at %H:%M:%S")} (#{Time.zone.name})",
+          },
+        ],
+      )
+    end
+  end
+end