← All changes | includes/Admin/Importer/Parsers/CSV_Parser.php +63 -10 4.5.0 → 4.9.3 View file →

		@@ -57,21 +57,32 @@
57	57	}
58	58
59	59	$fileContent = str_replace( [ "\r\n", "\r" ], "\n", $fileContent );
60	60
	61	+ // Parse through a stream so fgetcsv() correctly assembles records whose
	62	+ // quoted fields span multiple lines (e.g. multi-line doc content). A plain
	63	+ // explode( "\n" ) + str_getcsv() per line splits such records apart, producing
	64	+ // rows whose column count no longer matches the headers — which makes the
	65	+ // array_combine() calls below fatal (500 error on Sample Docs import).
	66	+ // phpcs:ignore WordPress.WP.AlternativeFunctions.file_system_operations_fopen -- in-memory stream, no filesystem access; required for multi-line CSV records.
61	67	$handle = fopen( 'php://temp', 'r+' );
62		- fwrite( $handle, $fileContent );
63		- rewind( $handle );
64		-
65	68	if ( $handle !== false ) {
66		- while ( ( $row = fgetcsv( $handle, 1000, ',' ) ) !== false ) {
	69	+ fwrite( $handle, $fileContent );
	70	+ rewind( $handle );
	71	+ while ( ( $row = fgetcsv( $handle, 0, ',' ) ) !== false ) {
67	72	$csv_data[] = $row;
68	73	}
	74	+ // phpcs:ignore WordPress.WP.AlternativeFunctions.file_system_operations_fclose -- closing in-memory stream; WP_Filesystem does not apply.
69	75	fclose( $handle );
70	76	}
71	77
72	78	$headers = array_shift( $csv_data );
73	79
	80	+ // Bail out cleanly on an empty/malformed file instead of fataling below.
	81	+ if ( ! is_array( $headers ) || empty( $headers ) ) {
	82	+ return $data;
	83	+ }
	84	+
74	85	// Process specific headers for 'Docs Title'
75	86	if ( $headers[0] == 'Docs Title' ) {
76	87	$replacementMap = [ 'Docs Slug' => 'post_name' ];
77	88
		@@ -93,8 +104,32 @@
93	104	}
94	105
95	106	usort( $csv_data, [ $this, 'csvSort' ] );
96	107
	108	+ // The combined CSV packs three blocks side by side: Docs (from index 1),
	109	+ // then Author, then Term. CSVs exported after the WPML language columns
	110	+ // were added widen the Docs block by 2 columns (Docs language code +
	111	+ // translation source slug), shifting the Author/Term offsets. Locate
	112	+ // each block by its header name rather than a fixed position, so the
	113	+ // layout is detected wherever the columns land (and adding a column to
	114	+ // one block can't silently corrupt the others).
	115	+ $has_wpml_columns = in_array( 'Docs language code', $headers, true );
	116	+
	117	+ $author_offset = array_search( 'Author id', $headers, true );
	118	+ $term_offset = array_search( 'Taxonomy', $headers, true );
	119	+
	120	+ // Fall back to the historical fixed offsets if a block header is missing
	121	+ // (malformed file) so such files still parse exactly as they did before.
	122	+ if ( $author_offset === false ) {
	123	+ $author_offset = $has_wpml_columns ? 24 : 22;
	124	+ }
	125	+ if ( $term_offset === false ) {
	126	+ $term_offset = $has_wpml_columns ? 30 : 28;
	127	+ }
	128	+
	129	+ // The Docs block runs from index 1 up to the start of the Author block.
	130	+ $post_block_len = $author_offset - 1;
	131	+
97	132	foreach ( $csv_data as $row ) {
98	133	$type = $row[0];
99	134
100	135	if ( $type === 'glossaries' ) {
		@@ -122,10 +157,10 @@
122	157	];
123	158
124	159	$data['terms'][] = $term_args;
125	160	} elseif ( $type === 'Term' ) {
126		- $term_headers = array_slice( $headers, 28, 11 );
127		- $term_row = array_slice( $row, 28, 11 );
	161	+ $term_headers = array_slice( $headers, $term_offset, 11 );
	162	+ $term_row = array_slice( $row, $term_offset, 11 );
128	163	$term_row = array_pad( $term_row, count( $term_headers ), '' );
129	164
130	165	$term_data = array_combine( $term_headers, $term_row );
131	166
		@@ -171,10 +206,10 @@
171	206	}
172	207
173	208	$data['terms'][] = $term_args;
174	209	} elseif ( $type === 'Author' ) {
175		- $author_headers = array_slice( $headers, 22, 6 );
176		- $author_row = array_slice( $row, 22, 6 );
	210	+ $author_headers = array_slice( $headers, $author_offset, 6 );
	211	+ $author_row = array_slice( $row, $author_offset, 6 );
177	212	$author_row = array_pad( $author_row, count( $author_headers ), '' );
178	213
179	214	$author_data = array_combine( $author_headers, $author_row );
180	215
		@@ -186,10 +221,13 @@
186	221	'author_first_name' => sanitize_text_field( $author_data['Author first name'] ),
187	222	'author_last_name' => sanitize_text_field( $author_data['Author last name'] )
188	223	];
189	224	} else if ( $type === 'Docs' || $type === 'FAQ' ) {
190		- $post_headers = array_slice( $headers, 1, 21 );
191		- $post_row = array_slice( $row, 1, 21 );
	225	+ // Keep FAQ import (HEAD) and use the dynamic post-block length
	226	+ // from the WPML branch so the variable WPML language columns are
	227	+ // handled instead of a hardcoded count.
	228	+ $post_headers = array_slice( $headers, 1, $post_block_len );
	229	+ $post_row = array_slice( $row, 1, $post_block_len );
192	230	$post_row = array_pad( $post_row, count( $post_headers ), '' );
193	231
194	232	$post_data = array_combine( $post_headers, $post_row );
195	233
		@@ -218,8 +256,23 @@
218	256	$this->searchTermsByIds( $data['terms'], sanitize_text_field( $post_data['Doc Categories'] ) ),
219	257	$this->searchTermsByIds( $data['terms'], sanitize_text_field( $post_data['Doc Tags'] ) ),
220	258	$this->searchTermsByIds( $data['terms'], sanitize_text_field( $post_data['Knowledge Bases'] ) )
221	259	);
	260	+ }
	261	+
	262	+ if ( $has_wpml_columns ) {
	263	+ if ( ! empty( $post_data['Docs language code'] ) ) {
	264	+ $post_args['postmeta'][] = [
	265	+ 'key' => '_betterdocs_wpml_lang',
	266	+ 'value' => sanitize_text_field( $post_data['Docs language code'] ),
	267	+ ];
	268	+ }
	269	+ if ( ! empty( $post_data['Docs translation source slug'] ) ) {
	270	+ $post_args['postmeta'][] = [
	271	+ 'key' => '_betterdocs_wpml_source_slug',
	272	+ 'value' => sanitize_text_field( $post_data['Docs translation source slug'] ),
	273	+ ];
	274	+ }
222	275	}
223	276
224	277	$data['posts'][] = $post_args;
225	278