1 /* PSPP - a program for statistical analysis.
2 Copyright (C) 2011, 2012, 2013 Free Software Foundation, Inc.
4 This program is free software: you can redistribute it and/or modify
5 it under the terms of the GNU General Public License as published by
6 the Free Software Foundation, either version 3 of the License, or
7 (at your option) any later version.
9 This program is distributed in the hope that it will be useful,
10 but WITHOUT ANY WARRANTY; without even the implied warranty of
11 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
12 GNU General Public License for more details.
14 You should have received a copy of the GNU General Public License
15 along with this program. If not, see <http://www.gnu.org/licenses/>. */
19 #include "libpspp/message.h"
20 #include "libpspp/misc.h"
21 #include "libpspp/assertion.h"
23 #include "data/data-in.h"
25 #include "gl/c-strtod.h"
26 #include "gl/minmax.h"
29 #define _(msgid) gettext (msgid)
30 #define N_(msgid) (msgid)
32 #include "ods-reader.h"
33 #include "spreadsheet-reader.h"
38 ods_open_reader (const struct spreadsheet_read_options *opts,
39 struct dictionary **dict)
41 msg (ME, _("Support for %s files was not compiled into this installation of PSPP"), "OpenDocument");
48 #include "libpspp/zip-reader.h"
54 #include <libxml/xmlreader.h>
57 #include "data/format.h"
58 #include "data/case.h"
59 #include "data/casereader-provider.h"
60 #include "data/dictionary.h"
61 #include "data/identifier.h"
62 #include "data/value.h"
63 #include "data/variable.h"
64 #include "libpspp/i18n.h"
65 #include "libpspp/str.h"
67 #include "gl/xalloc.h"
69 static void ods_file_casereader_destroy (struct casereader *, void *);
70 static struct ccase *ods_file_casereader_read (struct casereader *, void *);
73 static const struct casereader_class ods_file_casereader_class =
75 ods_file_casereader_read,
76 ods_file_casereader_destroy,
83 /* The name of the sheet (utf8 encoding) */
95 STATE_INIT = 0, /* Initial state */
96 STATE_SPREADSHEET, /* Found the start of the spreadsheet doc */
97 STATE_TABLE, /* Found the sheet that we actually want */
98 STATE_ROW, /* Found the start of the cell array */
99 STATE_CELL, /* Found a cell */
100 STATE_CELL_CONTENT /* Found a the text within a cell */
105 struct spreadsheet spreadsheet;
106 struct zip_reader *zreader;
107 xmlTextReaderPtr xtr;
110 enum reader_state state;
115 xmlChar *current_sheet_name;
117 xmlChar *target_sheet_name;
118 int target_sheet_index;
128 struct sheet_detail *sheets;
129 int n_allocated_sheets;
131 struct caseproto *proto;
132 struct dictionary *dict;
133 struct ccase *first_case;
134 bool used_first_case;
137 struct string ods_errs;
141 ods_destroy (struct spreadsheet *s)
143 struct ods_reader *r = (struct ods_reader *) s;
145 if (--r->ref_cnt == 0)
149 for (i = 0; i < r->n_allocated_sheets; ++i)
151 xmlFree (r->sheets[i].name);
154 zip_reader_destroy (r->zreader);
163 reading_target_sheet (const struct ods_reader *r)
165 if (r->target_sheet_name != NULL)
167 if ( 0 == xmlStrcmp (r->target_sheet_name, r->current_sheet_name))
171 if (r->target_sheet_index == r->current_sheet + 1)
178 static void process_node (struct ods_reader *r);
182 ods_get_sheet_name (struct spreadsheet *s, int n)
184 struct ods_reader *or = (struct ods_reader *) s;
186 assert (n < s->n_sheets);
189 (or->n_allocated_sheets <= n)
190 || or->state != STATE_SPREADSHEET
193 int ret = xmlTextReaderRead (or->xtr);
200 return or->sheets[n].name;
204 ods_get_sheet_range (struct spreadsheet *s, int n)
206 struct ods_reader *or = (struct ods_reader *) s;
208 assert (n < s->n_sheets);
211 (or->n_allocated_sheets <= n)
212 || (or->sheets[n].stop_row == -1)
213 || or->state != STATE_SPREADSHEET
216 int ret = xmlTextReaderRead (or->xtr);
223 return create_cell_ref (
224 or->sheets[n].start_col,
225 or->sheets[n].start_row,
226 or->sheets[n].stop_col,
227 or->sheets[n].stop_row);
232 ods_file_casereader_destroy (struct casereader *reader UNUSED, void *r_)
234 struct ods_reader *r = r_;
239 xmlFreeTextReader (r->xtr);
242 if ( ! ds_is_empty (&r->ods_errs))
243 msg (ME, "%s", ds_cstr (&r->ods_errs));
245 ds_destroy (&r->ods_errs);
247 if ( ! r->used_first_case )
248 case_unref (r->first_case);
250 caseproto_unref (r->proto);
252 xmlFree (r->current_sheet_name);
253 xmlFree (r->target_sheet_name);
255 ods_destroy (&r->spreadsheet);
263 process_node (struct ods_reader *r)
265 xmlChar *name = xmlTextReaderName (r->xtr);
267 name = xmlStrdup (_xml ("--"));
270 r->node_type = xmlTextReaderNodeType (r->xtr);
275 if (0 == xmlStrcasecmp (name, _xml("office:spreadsheet")) &&
276 XML_READER_TYPE_ELEMENT == r->node_type)
278 r->state = STATE_SPREADSHEET;
279 r->current_sheet = -1;
280 r->current_sheet_name = NULL;
283 case STATE_SPREADSHEET:
284 if (0 == xmlStrcasecmp (name, _xml("table:table"))
286 (XML_READER_TYPE_ELEMENT == r->node_type))
288 xmlFree (r->current_sheet_name);
289 r->current_sheet_name = xmlTextReaderGetAttribute (r->xtr, _xml ("table:name"));
293 if (r->current_sheet >= r->n_allocated_sheets)
295 assert (r->current_sheet == r->n_allocated_sheets);
296 r->sheets = xrealloc (r->sheets, sizeof (*r->sheets) * ++r->n_allocated_sheets);
297 r->sheets[r->n_allocated_sheets - 1].start_col = -1;
298 r->sheets[r->n_allocated_sheets - 1].stop_col = -1;
299 r->sheets[r->n_allocated_sheets - 1].start_row = -1;
300 r->sheets[r->n_allocated_sheets - 1].stop_row = -1;
301 r->sheets[r->n_allocated_sheets - 1].name = CHAR_CAST (char *, xmlStrdup (r->current_sheet_name));
307 r->state = STATE_TABLE;
309 else if (0 == xmlStrcasecmp (name, _xml("office:spreadsheet")) &&
310 XML_READER_TYPE_ELEMENT == r->node_type)
312 r->state = STATE_INIT;
316 if (0 == xmlStrcasecmp (name, _xml("table:table-row")) &&
317 (XML_READER_TYPE_ELEMENT == r->node_type))
320 xmlTextReaderGetAttribute (r->xtr,
321 _xml ("table:number-rows-repeated"));
323 int row_span = value ? _xmlchar_to_int (value) : 1;
328 if (! xmlTextReaderIsEmptyElement (r->xtr))
329 r->state = STATE_ROW;
333 else if (0 == xmlStrcasecmp (name, _xml("table:table")) &&
334 (XML_READER_TYPE_END_ELEMENT == r->node_type))
336 r->state = STATE_SPREADSHEET;
340 if ( (0 == xmlStrcasecmp (name, _xml ("table:table-cell")))
342 (XML_READER_TYPE_ELEMENT == r->node_type))
345 xmlTextReaderGetAttribute (r->xtr,
346 _xml ("table:number-columns-repeated"));
348 r->col_span = value ? _xmlchar_to_int (value) : 1;
349 r->col += r->col_span;
351 if (! xmlTextReaderIsEmptyElement (r->xtr))
352 r->state = STATE_CELL;
356 else if ( (0 == xmlStrcasecmp (name, _xml ("table:table-row")))
358 (XML_READER_TYPE_END_ELEMENT == r->node_type))
360 r->state = STATE_TABLE;
364 if ( (0 == xmlStrcasecmp (name, _xml("text:p")))
366 ( XML_READER_TYPE_ELEMENT == r->node_type))
368 if (! xmlTextReaderIsEmptyElement (r->xtr))
369 r->state = STATE_CELL_CONTENT;
372 ( (0 == xmlStrcasecmp (name, _xml("table:table-cell")))
374 (XML_READER_TYPE_END_ELEMENT == r->node_type)
377 r->state = STATE_ROW;
380 case STATE_CELL_CONTENT:
381 assert (r->current_sheet >= 0);
382 assert (r->current_sheet < r->n_allocated_sheets);
384 if (r->sheets[r->current_sheet].start_row == -1)
385 r->sheets[r->current_sheet].start_row = r->row - 1;
388 (r->sheets[r->current_sheet].start_col == -1)
390 (r->sheets[r->current_sheet].start_col >= r->col - 1)
392 r->sheets[r->current_sheet].start_col = r->col - 1;
394 r->sheets[r->current_sheet].stop_row = r->row - 1;
396 if ( r->sheets[r->current_sheet].stop_col < r->col - 1)
397 r->sheets[r->current_sheet].stop_col = r->col - 1;
399 if (XML_READER_TYPE_END_ELEMENT == r->node_type)
400 r->state = STATE_CELL;
411 A struct containing the parameters of a cell's value
424 struct xml_value firstval;
428 /* Determine the width that a xmv should probably have */
430 xmv_to_width (const struct xml_value *xmv, int fallback)
432 int width = SPREADSHEET_DEFAULT_WIDTH;
434 /* Non-strings always have zero width */
435 if (xmv->type != NULL && 0 != xmlStrcmp (xmv->type, _xml("string")))
442 width = ROUND_UP (xmlStrlen (xmv->value),
443 SPREADSHEET_DEFAULT_WIDTH);
445 width = ROUND_UP (xmlStrlen (xmv->text),
446 SPREADSHEET_DEFAULT_WIDTH);
452 Sets the VAR of case C, to the value corresponding to the xml data
455 convert_xml_to_value (struct ccase *c, const struct variable *var,
456 const struct xml_value *xmv)
458 union value *v = case_data_rw (c, var);
460 if (xmv->value == NULL && xmv->text == NULL)
461 value_set_missing (v, var_get_width (var));
462 else if ( var_is_alpha (var))
463 /* Use the text field, because it seems that there is no
464 value field for strings */
465 value_copy_str_rpad (v, var_get_width (var), xmv->text, ' ');
468 const struct fmt_spec *fmt = var_get_write_format (var);
469 enum fmt_category fc = fmt_get_category (fmt->type);
471 assert ( fc != FMT_CAT_STRING);
473 if ( 0 == xmlStrcmp (xmv->type, _xml("float")))
475 v->f = c_strtod (CHAR_CAST (const char *, xmv->value), NULL);
479 const char *text = xmv->value ?
480 CHAR_CAST (const char *, xmv->value) : CHAR_CAST (const char *, xmv->text);
483 free (data_in (ss_cstr (text), "UTF-8",
493 /* Try to find out how many sheets there are in the "workbook" */
495 get_sheet_count (struct zip_reader *zreader)
497 xmlTextReaderPtr mxtr;
498 struct zip_member *meta = NULL;
499 meta = zip_member_open (zreader, "meta.xml");
504 mxtr = xmlReaderForIO ((xmlInputReadCallback) zip_member_read,
505 (xmlInputCloseCallback) NULL,
506 meta, NULL, NULL, 0);
508 while (1 == xmlTextReaderRead (mxtr))
510 xmlChar *name = xmlTextReaderName (mxtr);
511 if ( 0 == xmlStrcmp (name, _xml("meta:document-statistic")))
513 xmlChar *attr = xmlTextReaderGetAttribute (mxtr, _xml ("meta:table-count"));
517 int s = _xmlchar_to_int (attr);
518 xmlFreeTextReader (mxtr);
528 xmlFreeTextReader (mxtr);
533 ods_error_handler (void *ctx, const char *mesg,
534 UNUSED xmlParserSeverities sev, xmlTextReaderLocatorPtr loc)
536 struct ods_reader *r = ctx;
538 msg (MW, _("There was a problem whilst reading the %s file `%s' (near line %d): `%s'"),
540 r->spreadsheet.file_name,
541 xmlTextReaderLocatorLineNumber (loc),
547 init_reader (struct ods_reader *r, bool report_errors)
549 struct zip_member *content = zip_member_open (r->zreader, "content.xml");
550 xmlTextReaderPtr xtr;
552 if ( content == NULL)
556 xmlFreeTextReader (r->xtr);
558 zip_member_ref (content);
559 xtr = xmlReaderForIO ((xmlInputReadCallback) zip_member_read,
560 (xmlInputCloseCallback) zip_member_finish,
562 report_errors ? 0 : (XML_PARSE_NOERROR | XML_PARSE_NOWARNING) );
568 r->spreadsheet.type = SPREADSHEET_ODS;
571 r->current_sheet = 0;
572 r->state = STATE_INIT;
575 xmlTextReaderSetErrorHandler (xtr, ods_error_handler, r);
582 ods_probe (const char *filename, bool report_errors)
584 struct ods_reader *r;
585 struct string errs = DS_EMPTY_INITIALIZER;
587 struct zip_reader *zr = zip_reader_create (filename, &errs);
593 msg (ME, _("Cannot open %s as a OpenDocument file: %s"),
594 filename, ds_cstr (&errs));
599 sheet_count = get_sheet_count (zr);
601 r = xzalloc (sizeof *r);
605 if (! init_reader (r, report_errors))
610 r->spreadsheet.n_sheets = sheet_count;
611 r->n_allocated_sheets = 0;
616 r->spreadsheet.file_name = filename;
617 return &r->spreadsheet;
620 zip_reader_destroy (r->zreader);
627 ods_make_reader (struct spreadsheet *spreadsheet,
628 const struct spreadsheet_read_options *opts)
631 xmlChar *type = NULL;
632 unsigned long int vstart = 0;
633 casenumber n_cases = CASENUMBER_MAX;
635 struct var_spec *var_spec = NULL;
638 struct ods_reader *r = (struct ods_reader *) spreadsheet;
639 xmlChar *val_string = NULL;
642 r->read_names = opts->read_names;
643 ds_init_empty (&r->ods_errs);
646 if ( !init_reader (r, true))
649 if (opts->cell_range)
651 if ( ! convert_cell_ref (opts->cell_range,
652 &r->start_col, &r->start_row,
653 &r->stop_col, &r->stop_row))
655 msg (SE, _("Invalid cell range `%s'"),
668 r->state = STATE_INIT;
669 r->target_sheet_name = xmlStrdup (BAD_CAST opts->sheet_name);
670 r->target_sheet_index = opts->sheet_index;
674 /* Advance to the start of the cells for the target sheet */
675 while ( ! reading_target_sheet (r)
676 || r->state != STATE_ROW || r->row <= r->start_row )
678 if (1 != (ret = xmlTextReaderRead (r->xtr)))
686 msg (MW, _("Selected sheet or range of spreadsheet `%s' is empty."),
687 spreadsheet->file_name);
691 if ( opts->read_names)
693 while (1 == (ret = xmlTextReaderRead (r->xtr)))
699 /* If the row is finished then stop for now */
700 if (r->state == STATE_TABLE && r->row > r->start_row)
703 idx = r->col - r->start_col -1 ;
708 if (r->stop_col != -1 && idx > r->stop_col - r->start_col)
711 if (r->state == STATE_CELL_CONTENT
713 XML_READER_TYPE_TEXT == r->node_type)
715 xmlChar *value = xmlTextReaderValue (r->xtr);
717 if ( idx >= n_var_specs)
719 var_spec = xrealloc (var_spec, sizeof (*var_spec) * (idx + 1));
721 /* xrealloc (unlike realloc) doesn't initialise its memory to 0 */
722 memset (var_spec + n_var_specs,
724 (idx - n_var_specs + 1) * sizeof (*var_spec));
725 n_var_specs = idx + 1;
727 var_spec[idx].firstval.text = 0;
728 var_spec[idx].firstval.value = 0;
729 var_spec[idx].firstval.type = 0;
731 var_spec [idx].name = strdup (CHAR_CAST (const char *, value));
738 /* Read in the first row of data */
739 while (1 == xmlTextReaderRead (r->xtr))
744 if ( ! reading_target_sheet (r) )
747 /* If the row is finished then stop for now */
748 if (r->state == STATE_TABLE &&
749 r->row > r->start_row + (opts->read_names ? 1 : 0))
752 idx = r->col - r->start_col - 1;
756 if (r->stop_col != -1 && idx > r->stop_col - r->start_col)
759 if ( r->state == STATE_CELL &&
760 XML_READER_TYPE_ELEMENT == r->node_type)
762 type = xmlTextReaderGetAttribute (r->xtr, _xml ("office:value-type"));
763 val_string = xmlTextReaderGetAttribute (r->xtr, _xml ("office:value"));
766 if ( r->state == STATE_CELL_CONTENT &&
767 XML_READER_TYPE_TEXT == r->node_type)
769 if (idx >= n_var_specs)
771 var_spec = xrealloc (var_spec, sizeof (*var_spec) * (idx + 1));
772 memset (var_spec + n_var_specs,
774 (idx - n_var_specs + 1) * sizeof (*var_spec));
776 var_spec [idx].name = NULL;
777 n_var_specs = idx + 1;
780 var_spec [idx].firstval.type = type;
781 var_spec [idx].firstval.text = xmlTextReaderValue (r->xtr);
782 var_spec [idx].firstval.value = val_string;
790 /* Create the dictionary and populate it */
791 r->spreadsheet.dict = r->dict = dict_create (
792 CHAR_CAST (const char *, xmlTextReaderConstEncoding (r->xtr)));
794 for (i = 0; i < n_var_specs ; ++i )
797 struct variable *var = NULL;
798 char *name = dict_make_unique_var_name (r->dict, var_spec[i].name, &vstart);
799 int width = xmv_to_width (&var_spec[i].firstval, opts->asw);
800 dict_create_var (r->dict, name, width);
803 var = dict_get_var (r->dict, i);
805 if ( 0 == xmlStrcmp (var_spec[i].firstval.type, _xml("date")))
812 fmt = fmt_default_for_width (width);
814 var_set_both_formats (var, &fmt);
817 /* Create the first case, and cache it */
818 r->used_first_case = false;
820 if ( n_var_specs == 0 )
822 msg (MW, _("Selected sheet or range of spreadsheet `%s' is empty."),
823 spreadsheet->file_name);
827 r->proto = caseproto_ref (dict_get_proto (r->dict));
828 r->first_case = case_create (r->proto);
829 case_set_missing (r->first_case);
831 for (i = 0 ; i < n_var_specs; ++i)
833 const struct variable *var = dict_get_var (r->dict, i);
835 convert_xml_to_value (r->first_case, var, &var_spec[i].firstval);
838 /* Read in the first row of data */
839 while (1 == xmlTextReaderRead (r->xtr))
843 if (r->state == STATE_ROW)
848 for ( i = 0 ; i < n_var_specs ; ++i )
850 free (var_spec[i].firstval.type);
851 free (var_spec[i].firstval.value);
852 free (var_spec[i].firstval.text);
853 free (var_spec[i].name);
859 return casereader_create_sequential
863 &ods_file_casereader_class, r);
867 for ( i = 0 ; i < n_var_specs ; ++i )
869 free (var_spec[i].firstval.type);
870 free (var_spec[i].firstval.value);
871 free (var_spec[i].firstval.text);
872 free (var_spec[i].name);
877 dict_destroy (r->spreadsheet.dict);
878 r->spreadsheet.dict = NULL;
879 ods_file_casereader_destroy (NULL, r);
885 /* Reads and returns one case from READER's file. Returns a null
886 pointer on failure. */
887 static struct ccase *
888 ods_file_casereader_read (struct casereader *reader UNUSED, void *r_)
890 struct ccase *c = NULL;
891 xmlChar *val_string = NULL;
892 xmlChar *type = NULL;
893 struct ods_reader *r = r_;
895 if (!r->used_first_case)
897 r->used_first_case = true;
898 return r->first_case;
902 /* Advance to the start of a row. (If there is one) */
903 while (r->state != STATE_ROW
904 && 1 == xmlTextReaderRead (r->xtr)
911 if ( ! reading_target_sheet (r)
912 || r->state < STATE_TABLE
913 || (r->stop_row != -1 && r->row > r->stop_row + 1)
919 c = case_create (r->proto);
920 case_set_missing (c);
922 while (1 == xmlTextReaderRead (r->xtr))
926 if ( r->stop_row != -1 && r->row > r->stop_row + 1)
929 if (r->state == STATE_CELL &&
930 r->node_type == XML_READER_TYPE_ELEMENT)
932 type = xmlTextReaderGetAttribute (r->xtr, _xml ("office:value-type"));
933 val_string = xmlTextReaderGetAttribute (r->xtr, _xml ("office:value"));
936 if (r->state == STATE_CELL_CONTENT &&
937 r->node_type == XML_READER_TYPE_TEXT)
940 struct xml_value *xmv = xzalloc (sizeof *xmv);
941 xmv->text = xmlTextReaderValue (r->xtr);
942 xmv->value = val_string;
946 for (col = 0; col < r->col_span; ++col)
948 const struct variable *var;
949 const int idx = r->col - col - r->start_col - 1;
952 if (r->stop_col != -1 && idx > r->stop_col - r->start_col )
954 if (idx >= dict_get_var_cnt (r->dict))
957 var = dict_get_var (r->dict, idx);
958 convert_xml_to_value (c, var, xmv);
962 xmlFree (xmv->value);
966 if ( r->state <= STATE_TABLE)