From 0839ceb2d8868de25e3bfa840c3009991389bf65 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?benjamin=20melan=C3=A7on?= <ben@agaric.coop>
Date: Thu, 16 Nov 2023 10:26:59 -0500
Subject: [PATCH 1/2] Not quite working take 1

---
 migration_helpers.module | 68 ++++++++++++++++++++++++++++++++++------
 1 file changed, 58 insertions(+), 10 deletions(-)

diff --git a/migration_helpers.module b/migration_helpers.module
index b443a0b..b04ca01 100644
--- a/migration_helpers.module
+++ b/migration_helpers.module
@@ -1,5 +1,7 @@
 <?php
 
+use Psr\Log\LogLevel;
+
 /**
  * Entity query helper.
  */
@@ -38,7 +40,6 @@ function migration_helpers_entity_query(
  * Move (and optionally filter/transform) terms between term reference fields.
  */
 function migration_helpers_move_terms(
-  $source_vocabulary,
   $destination_vocabulary,
   $source_field,
   $destination_field,
@@ -51,15 +52,62 @@ function migration_helpers_move_terms(
   // Destination bundle is also required but we default to making it
   // the same as the source bundle.
   $destination_bundle ??= $source_bundle;
-    print("
-    $source_entity_type,
-    $source_bundle,
-    $source_field,
-    $destination_entity_type,
-    $destination_bundle,
-    $destination_field,
-    $source_vocabulary,
-    $destination_vocabulary");
+
+  /** @var \Drupal\Core\Entity\EntityTypeManagerInterface $entity_type_manager */
+  $entity_type_manager = \Drupal::service('entity_type.manager');
+
+  // Get any existing terms in destination vocabulary so we can efficiently
+  // prevent duplication.
+  $terms = migration_helpers_entity_query(
+    entity_type: 'taxonomy_term',
+    bundle: $destination_vocabulary,
+  );
+
+  $term_data = [];
+  foreach ($terms as $term) {
+    $term_data[$term->id()] = $term->label();
+  }
+
+  $nodes = migration_helpers_entity_query($source_entity_type, $source_bundle, $source_field);
+  foreach ($nodes as $node) {
+    $tids = [];
+    /** @var \Drupal\node\NodeInterface $node */
+    $orig_values = $node->get($source_field);
+    foreach ($orig_values as $orig_value) {
+      $tid = $orig_value->target_id;
+      $orig_term = $entity_type_manager->getStorage('taxonomy_term')->load($tid);
+      $orig = $orig_term->name->value;
+      $term_name = $mapping[$orig] ?? $orig;
+      // If our map intentionally blanked out the term name, do not create
+      // a term in the new vocabulary and assign it to the destination field.
+      // Note, the term must be blanked out ('') or assigned to FALSE in the
+      // mapping, setting a term to NULL is the same as not mapping it, and it
+      // would be created in the new vocab etc.
+      if (!$term_name) {
+        continue;
+      }
+      $tid = array_search($term_name, $term_data);
+      if (empty($tid)) {
+        $new_term = $entity_type_manager->getStorage('taxonomy_term')->create([
+          'name' => $term_name,
+          'vid' => $destination_vocabulary,
+        ]);
+        $new_term->save();
+        $tid = $new_term->id();
+        \Drupal::logger('mass')->log(LogLevel::INFO, "Term $term_name created with TID $tid in focus area vocabulary.");
+        $tids[] = $tid;
+      }
+    }
+    // Now we have collected all the term IDs that have been created.
+    // We have to go through a surprising number of hoops to make our
+    $already_there_tids = [];
+    $already_there_values = $node->get($destination_field);
+    foreach ($already_there_values as $already_there_value) {
+      $already_there_tids[] = $already_there_value->target_id;
+    }
+    $node->set($destination_field, array_merge($tids, $already_there_tids));
+    $node->save();
+  }
 }
 
 // The below almost serve more as examples that can be followed

From 0f3864850796e2e7a180704b88e44c3d8bacd17e Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?benjamin=20melan=C3=A7on?= <ben@agaric.coop>
Date: Thu, 16 Nov 2023 11:06:05 -0500
Subject: [PATCH 2/2] Simplify and fix

That last bit of the tids being assigned *outside* the if loop
(so existing, not only created ones, get used) is key.
---
 migration_helpers.module | 19 +++++++------------
 1 file changed, 7 insertions(+), 12 deletions(-)

diff --git a/migration_helpers.module b/migration_helpers.module
index b04ca01..4710a58 100644
--- a/migration_helpers.module
+++ b/migration_helpers.module
@@ -43,15 +43,10 @@ function migration_helpers_move_terms(
   $destination_vocabulary,
   $source_field,
   $destination_field,
-  $source_bundle,
-  $destination_bundle = NULL,
-  $source_entity_type = 'node',
-  $destination_entity_type = 'node',
+  $bundle = NULL,
+  $entity_type = 'node',
   $mapping = [],
 ) {
-  // Destination bundle is also required but we default to making it
-  // the same as the source bundle.
-  $destination_bundle ??= $source_bundle;
 
   /** @var \Drupal\Core\Entity\EntityTypeManagerInterface $entity_type_manager */
   $entity_type_manager = \Drupal::service('entity_type.manager');
@@ -68,14 +63,14 @@ function migration_helpers_move_terms(
     $term_data[$term->id()] = $term->label();
   }
 
-  $nodes = migration_helpers_entity_query($source_entity_type, $source_bundle, $source_field);
+  $nodes = migration_helpers_entity_query($entity_type, $bundle, $source_field);
   foreach ($nodes as $node) {
     $tids = [];
     /** @var \Drupal\node\NodeInterface $node */
     $orig_values = $node->get($source_field);
     foreach ($orig_values as $orig_value) {
-      $tid = $orig_value->target_id;
-      $orig_term = $entity_type_manager->getStorage('taxonomy_term')->load($tid);
+      $orig_tid = $orig_value->target_id;
+      $orig_term = $entity_type_manager->getStorage('taxonomy_term')->load($orig_tid);
       $orig = $orig_term->name->value;
       $term_name = $mapping[$orig] ?? $orig;
       // If our map intentionally blanked out the term name, do not create
@@ -94,9 +89,9 @@ function migration_helpers_move_terms(
         ]);
         $new_term->save();
         $tid = $new_term->id();
-        \Drupal::logger('mass')->log(LogLevel::INFO, "Term $term_name created with TID $tid in focus area vocabulary.");
-        $tids[] = $tid;
+        \Drupal::logger('mass')->log(LogLevel::INFO, "Term $term_name created with TID $tid in vocabulary $destination_vocabulary.");
       }
+      $tids[] = $tid;
     }
     // Now we have collected all the term IDs that have been created.
     // We have to go through a surprising number of hoops to make our