Create a text PG catalog for noncvs/in_progress_check.php

crontab/ImportPGCatalog.inc

@@ -123,6 +123,9 @@ class ImportPGCatalog extends BackgroundJob
 
     private function process_catalog()
     {
+        global $dyn_dir;
+        $title_authors = [];
+
         $mime_types_not_in_display_mapping = [];
 
         echo "Scanning files in {$this->local_catalog_dir}...\n";
@@ -254,6 +257,24 @@ class ImportPGCatalog extends BackgroundJob
             DPDatabase::query($sql);
 
             $n_rdf_files_processed += 1;
+
+
+            // Pull out title and author for text-based catalog
+            // Use xpath wildcards instead of full xpath
+            $authors = [];
+            foreach ($root->xpath('//pgterms:name') as $author_value_element) {
+                $authors[] = (string)$author_value_element;
+            }
+            $author = join("; ", $authors);
+
+            $title = "";
+            foreach ($root->xpath('//dcterms:title') as $title_value_element) {
+                $t_as_string = $title_value_element->asXML();
+                $t_as_string = str_replace('
', '', $t_as_string);
+                $t_as_string = str_replace(PHP_EOL, ' / ', $t_as_string);
+                $title .= strip_tags($t_as_string);
+            }
+            $title_authors[$etext_number] = [$title, $author];
         }
 
         echo "Finished processing $n_rdf_files_processed RDF files.\n";
@@ -268,6 +289,37 @@ class ImportPGCatalog extends BackgroundJob
             }
         }
 
+        // Create the text-based catalog used by noncvs/in_progress_check.php
+        echo "Creating text catalog...\n";
+        $pg_catalog_converted = "$dyn_dir/pg/catalog.txt";
+        $text_catalog = fopen($pg_catalog_converted, 'w');
+        if (!$text_catalog) {
+            throw new RuntimeException("Unable to open $pg_catalog_converted for writing");
+        }
+
+        // Include the current time as the first line of the output file so
+        // search results can display it.
+        fwrite(
+            $text_catalog,
+            sprintf(
+                "Catalog retrieved: %s (converted on %s)\n",
+                date("F d, Y"),
+                date("F j, Y, g:i a")
+            )
+        );
+
+        ksort($title_authors); // sort numerically by $etext_number
+        foreach ($title_authors as $etext_number => [$title, $author]) {
+            // Skip entries which are placeholder PG numbers by
+            // checking that $author is not empty.
+            if ($author) {
+                $output = "$etext_number \"$title\" by $author";
+                $output = preg_replace('/\s+/', ' ', $output);
+                fwrite($text_catalog, "$output\n");
+            }
+        }
+        fclose($text_catalog);
+
         $this->stop_message = sprintf("Processed %d etexts", $n_rdf_files_processed);
     }
 }