1 /* PSPP - a program for statistical analysis.
2 Copyright (C) 2011, 2012, 2013, 2016 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 "ods-reader.h"
20 #include "spreadsheet-reader.h"
25 #include <libxml/xmlreader.h>
28 #include "data/case.h"
29 #include "data/casereader-provider.h"
30 #include "data/data-in.h"
31 #include "data/dictionary.h"
32 #include "data/format.h"
33 #include "data/identifier.h"
34 #include "data/value.h"
35 #include "data/variable.h"
36 #include "libpspp/assertion.h"
37 #include "libpspp/i18n.h"
38 #include "libpspp/message.h"
39 #include "libpspp/misc.h"
40 #include "libpspp/str.h"
41 #include "libpspp/zip-reader.h"
43 #include "gl/c-strtod.h"
44 #include "gl/minmax.h"
45 #include "gl/xalloc.h"
48 #define _(msgid) gettext (msgid)
49 #define N_(msgid) (msgid)
51 static void ods_file_casereader_destroy (struct casereader *, void *);
52 static struct ccase *ods_file_casereader_read (struct casereader *, void *);
55 static const struct casereader_class ods_file_casereader_class =
57 ods_file_casereader_read,
58 ods_file_casereader_destroy,
65 /* The name of the sheet (utf8 encoding) */
77 STATE_INIT = 0, /* Initial state */
78 STATE_SPREADSHEET, /* Found the start of the spreadsheet doc */
79 STATE_TABLE, /* Found the sheet that we actually want */
80 STATE_ROW, /* Found the start of the cell array */
81 STATE_CELL, /* Found a cell */
82 STATE_CELL_CONTENT /* Found a the text within a cell */
88 struct zip_member *zm;
90 enum reader_state state;
94 xmlChar *current_sheet_name;
100 state_data_destroy (struct state_data *sd)
102 xmlFree (sd->current_sheet_name);
103 sd->current_sheet_name = NULL;
105 xmlFreeTextReader (sd->xtr);
108 zip_member_finish (sd->zm);
114 struct spreadsheet spreadsheet;
115 struct zip_reader *zreader;
117 int target_sheet_index;
118 xmlChar *target_sheet_name;
120 /* State data for the meta data */
121 struct state_data msd;
123 /* State data for the reader */
124 struct state_data rsd;
131 struct sheet_detail *sheets;
132 int n_allocated_sheets;
134 struct caseproto *proto;
135 struct dictionary *dict;
136 struct ccase *first_case;
137 bool used_first_case;
140 struct string ods_errs;
142 struct string zip_errs;
146 ods_unref (struct spreadsheet *s)
148 struct ods_reader *r = (struct ods_reader *) s;
150 if (--s->ref_cnt == 0)
154 state_data_destroy (&r->msd);
155 for (i = 0; i < r->n_allocated_sheets; ++i)
157 xmlFree (r->sheets[i].name);
160 dict_unref (r->dict);
162 zip_reader_destroy (r->zreader);
172 reading_target_sheet (const struct ods_reader *r, const struct state_data *msd)
174 if (r->target_sheet_name != NULL)
176 if (0 == xmlStrcmp (r->target_sheet_name, msd->current_sheet_name))
180 if (r->target_sheet_index == msd->current_sheet + 1)
187 static void process_node (struct ods_reader *or, struct state_data *r);
191 ods_get_sheet_name (struct spreadsheet *s, int n)
193 struct ods_reader *r = (struct ods_reader *) s;
194 struct state_data *or = &r->msd;
196 assert (n < s->n_sheets);
199 (r->n_allocated_sheets <= n)
200 || or->state != STATE_SPREADSHEET
203 int ret = xmlTextReaderRead (or->xtr);
207 process_node (r, or);
210 return r->sheets[n].name;
214 ods_get_sheet_range (struct spreadsheet *s, int n)
216 struct ods_reader *r = (struct ods_reader *) s;
217 struct state_data *or = &r->msd;
219 assert (n < s->n_sheets);
222 (r->n_allocated_sheets <= n)
223 || (r->sheets[n].stop_row == -1)
224 || or->state != STATE_SPREADSHEET
227 int ret = xmlTextReaderRead (or->xtr);
231 process_node (r, or);
234 return create_cell_range (
235 r->sheets[n].start_col,
236 r->sheets[n].start_row,
237 r->sheets[n].stop_col,
238 r->sheets[n].stop_row);
243 ods_file_casereader_destroy (struct casereader *reader UNUSED, void *r_)
245 struct ods_reader *r = r_;
249 state_data_destroy (&r->rsd);
251 if (! ds_is_empty (&r->ods_errs))
252 msg (ME, "%s", ds_cstr (&r->ods_errs));
254 ds_destroy (&r->ods_errs);
256 if (r->first_case && ! r->used_first_case)
257 case_unref (r->first_case);
260 caseproto_unref (r->proto);
263 xmlFree (r->target_sheet_name);
264 r->target_sheet_name = NULL;
267 ods_unref (&r->spreadsheet);
275 process_node (struct ods_reader *or, struct state_data *r)
277 xmlChar *name = xmlTextReaderName (r->xtr);
279 name = xmlStrdup (_xml ("--"));
282 r->node_type = xmlTextReaderNodeType (r->xtr);
287 if (0 == xmlStrcasecmp (name, _xml("office:spreadsheet")) &&
288 XML_READER_TYPE_ELEMENT == r->node_type)
290 r->state = STATE_SPREADSHEET;
291 r->current_sheet = -1;
292 r->current_sheet_name = NULL;
295 case STATE_SPREADSHEET:
296 if (0 == xmlStrcasecmp (name, _xml("table:table"))
298 (XML_READER_TYPE_ELEMENT == r->node_type))
300 xmlFree (r->current_sheet_name);
301 r->current_sheet_name = xmlTextReaderGetAttribute (r->xtr, _xml ("table:name"));
305 if (r->current_sheet >= or->n_allocated_sheets)
307 assert (r->current_sheet == or->n_allocated_sheets);
308 or->sheets = xrealloc (or->sheets, sizeof (*or->sheets) * ++or->n_allocated_sheets);
309 or->sheets[or->n_allocated_sheets - 1].start_col = -1;
310 or->sheets[or->n_allocated_sheets - 1].stop_col = -1;
311 or->sheets[or->n_allocated_sheets - 1].start_row = -1;
312 or->sheets[or->n_allocated_sheets - 1].stop_row = -1;
313 or->sheets[or->n_allocated_sheets - 1].name = CHAR_CAST (char *, xmlStrdup (r->current_sheet_name));
319 r->state = STATE_TABLE;
321 else if (0 == xmlStrcasecmp (name, _xml("office:spreadsheet")) &&
322 XML_READER_TYPE_ELEMENT == r->node_type)
324 r->state = STATE_INIT;
328 if (0 == xmlStrcasecmp (name, _xml("table:table-row")) &&
329 (XML_READER_TYPE_ELEMENT == r->node_type))
332 xmlTextReaderGetAttribute (r->xtr,
333 _xml ("table:number-rows-repeated"));
335 int row_span = value ? _xmlchar_to_int (value) : 1;
340 if (! xmlTextReaderIsEmptyElement (r->xtr))
341 r->state = STATE_ROW;
345 else if (0 == xmlStrcasecmp (name, _xml("table:table")) &&
346 (XML_READER_TYPE_END_ELEMENT == r->node_type))
348 r->state = STATE_SPREADSHEET;
352 if ((0 == xmlStrcasecmp (name, _xml ("table:table-cell")))
354 (XML_READER_TYPE_ELEMENT == r->node_type))
357 xmlTextReaderGetAttribute (r->xtr,
358 _xml ("table:number-columns-repeated"));
360 r->col_span = value ? _xmlchar_to_int (value) : 1;
361 r->col += r->col_span;
363 if (! xmlTextReaderIsEmptyElement (r->xtr))
364 r->state = STATE_CELL;
368 else if ((0 == xmlStrcasecmp (name, _xml ("table:table-row")))
370 (XML_READER_TYPE_END_ELEMENT == r->node_type))
372 r->state = STATE_TABLE;
376 if ((0 == xmlStrcasecmp (name, _xml("text:p")))
378 (XML_READER_TYPE_ELEMENT == r->node_type))
380 if (! xmlTextReaderIsEmptyElement (r->xtr))
381 r->state = STATE_CELL_CONTENT;
384 ((0 == xmlStrcasecmp (name, _xml("table:table-cell")))
386 (XML_READER_TYPE_END_ELEMENT == r->node_type)
389 r->state = STATE_ROW;
392 case STATE_CELL_CONTENT:
393 assert (r->current_sheet >= 0);
394 assert (r->current_sheet < or->n_allocated_sheets);
396 if (or->sheets[r->current_sheet].start_row == -1)
397 or->sheets[r->current_sheet].start_row = r->row - 1;
400 (or->sheets[r->current_sheet].start_col == -1)
402 (or->sheets[r->current_sheet].start_col >= r->col - 1)
404 or->sheets[r->current_sheet].start_col = r->col - 1;
406 or->sheets[r->current_sheet].stop_row = r->row - 1;
408 if (or->sheets[r->current_sheet].stop_col < r->col - 1)
409 or->sheets[r->current_sheet].stop_col = r->col - 1;
411 if (XML_READER_TYPE_END_ELEMENT == r->node_type)
412 r->state = STATE_CELL;
423 A struct containing the parameters of a cell's value
436 struct xml_value firstval;
440 /* Determine the width that a xmv should probably have */
442 xmv_to_width (const struct xml_value *xmv, int fallback)
444 int width = SPREADSHEET_DEFAULT_WIDTH;
446 /* Non-strings always have zero width */
447 if (xmv->type != NULL && 0 != xmlStrcmp (xmv->type, _xml("string")))
454 width = ROUND_UP (xmlStrlen (xmv->value),
455 SPREADSHEET_DEFAULT_WIDTH);
457 width = ROUND_UP (xmlStrlen (xmv->text),
458 SPREADSHEET_DEFAULT_WIDTH);
464 Sets the VAR of case C, to the value corresponding to the xml data
467 convert_xml_to_value (struct ccase *c, const struct variable *var,
468 const struct xml_value *xmv, int col, int row)
470 union value *v = case_data_rw (c, var);
472 if (xmv->value == NULL && xmv->text == NULL)
473 value_set_missing (v, var_get_width (var));
474 else if (var_is_alpha (var))
475 /* Use the text field, because it seems that there is no
476 value field for strings */
477 value_copy_str_rpad (v, var_get_width (var), xmv->text, ' ');
480 const struct fmt_spec *fmt = var_get_write_format (var);
481 enum fmt_category fc = fmt_get_category (fmt->type);
483 assert (fc != FMT_CAT_STRING);
485 if (0 == xmlStrcmp (xmv->type, _xml("float")))
487 v->f = c_strtod (CHAR_CAST (const char *, xmv->value), NULL);
491 const char *text = xmv->value ?
492 CHAR_CAST (const char *, xmv->value) : CHAR_CAST (const char *, xmv->text);
494 char *m = data_in (ss_cstr (text), "UTF-8",
502 char buf [FMT_STRING_LEN_MAX + 1];
503 char *cell = create_cell_ref (col, row);
505 msg (MW, _("Cannot convert the value in the spreadsheet cell %s to format (%s): %s"),
506 cell, fmt_to_string (fmt, buf), m);
515 xml_reader_for_zip_member (void *zm_, char *buffer, int len)
517 struct zip_member *zm = zm_;
518 return zip_member_read (zm, buffer, len);
521 /* Try to find out how many sheets there are in the "workbook" */
523 get_sheet_count (struct zip_reader *zreader)
525 xmlTextReaderPtr mxtr;
526 struct zip_member *meta = NULL;
527 meta = zip_member_open (zreader, "meta.xml");
532 mxtr = xmlReaderForIO (xml_reader_for_zip_member, NULL, meta, NULL, NULL, 0);
534 while (1 == xmlTextReaderRead (mxtr))
536 xmlChar *name = xmlTextReaderName (mxtr);
537 if (0 == xmlStrcmp (name, _xml("meta:document-statistic")))
539 xmlChar *attr = xmlTextReaderGetAttribute (mxtr, _xml ("meta:table-count"));
543 int s = _xmlchar_to_int (attr);
544 xmlFreeTextReader (mxtr);
545 zip_member_finish (meta);
555 xmlFreeTextReader (mxtr);
556 zip_member_finish (meta);
561 ods_error_handler (void *ctx, const char *mesg,
562 xmlParserSeverities sev UNUSED,
563 xmlTextReaderLocatorPtr loc)
565 struct ods_reader *r = ctx;
567 msg (MW, _("There was a problem whilst reading the %s file `%s' (near line %d): `%s'"),
569 r->spreadsheet.file_name,
570 xmlTextReaderLocatorLineNumber (loc),
576 init_reader (struct ods_reader *r, bool report_errors,
577 struct state_data *state)
579 struct zip_member *content = zip_member_open (r->zreader, "content.xml");
580 xmlTextReaderPtr xtr;
585 xtr = xmlReaderForIO (xml_reader_for_zip_member, NULL, content, NULL, NULL,
586 report_errors ? 0 : (XML_PARSE_NOERROR | XML_PARSE_NOWARNING));
591 *state = (struct state_data) { .xtr = xtr,
593 .state = STATE_INIT };
595 r->spreadsheet.type = SPREADSHEET_ODS;
598 xmlTextReaderSetErrorHandler (xtr, ods_error_handler, r);
606 ods_probe (const char *filename, bool report_errors)
609 struct ods_reader *r = xzalloc (sizeof *r);
610 struct zip_reader *zr;
612 ds_init_empty (&r->zip_errs);
614 zr = zip_reader_create (filename, &r->zip_errs);
620 msg (ME, _("Cannot open %s as a OpenDocument file: %s"),
621 filename, ds_cstr (&r->zip_errs));
623 ds_destroy (&r->zip_errs);
628 sheet_count = get_sheet_count (zr);
631 r->spreadsheet.ref_cnt = 1;
633 if (!init_reader (r, report_errors, &r->msd))
636 r->spreadsheet.n_sheets = sheet_count;
637 r->n_allocated_sheets = 0;
640 r->spreadsheet.file_name = strdup (filename);
641 return &r->spreadsheet;
644 ds_destroy (&r->zip_errs);
645 zip_reader_destroy (r->zreader);
651 ods_make_reader (struct spreadsheet *spreadsheet,
652 const struct spreadsheet_read_options *opts)
655 xmlChar *type = NULL;
656 unsigned long int vstart = 0;
657 casenumber n_cases = CASENUMBER_MAX;
659 struct var_spec *var_spec = NULL;
662 struct ods_reader *r = (struct ods_reader *) spreadsheet;
663 xmlChar *val_string = NULL;
666 r->read_names = opts->read_names;
667 ds_init_empty (&r->ods_errs);
668 ++r->spreadsheet.ref_cnt;
670 if (!init_reader (r, true, &r->rsd))
673 r->used_first_case = false;
674 r->first_case = NULL;
676 if (opts->cell_range)
678 if (! convert_cell_ref (opts->cell_range,
679 &r->start_col, &r->start_row,
680 &r->stop_col, &r->stop_row))
682 msg (SE, _("Invalid cell range `%s'"),
695 r->target_sheet_name = xmlStrdup (BAD_CAST opts->sheet_name);
696 r->target_sheet_index = opts->sheet_index;
698 /* Advance to the start of the cells for the target sheet */
699 while (! reading_target_sheet (r, &r->rsd)
700 || r->rsd.state != STATE_ROW || r->rsd.row <= r->start_row)
702 if (1 != (ret = xmlTextReaderRead (r->rsd.xtr)))
705 process_node (r, &r->rsd);
710 msg (MW, _("Selected sheet or range of spreadsheet `%s' is empty."),
711 spreadsheet->file_name);
715 if (opts->read_names)
717 while (1 == xmlTextReaderRead (r->rsd.xtr))
719 process_node (r, &r->rsd);
721 /* If the row is finished then stop for now */
722 if (r->rsd.state == STATE_TABLE && r->rsd.row > r->start_row)
725 int idx = r->rsd.col - r->start_col - 1;
730 if (r->stop_col != -1 && idx > r->stop_col - r->start_col)
733 if (r->rsd.state == STATE_CELL_CONTENT
735 XML_READER_TYPE_TEXT == r->rsd.node_type)
737 xmlChar *value = xmlTextReaderValue (r->rsd.xtr);
738 if (idx >= n_var_specs)
740 var_spec = xrealloc (var_spec, sizeof (*var_spec) * (idx + 1));
742 /* xrealloc (unlike realloc) doesn't initialise its memory to 0 */
743 memset (var_spec + n_var_specs,
745 (idx - n_var_specs + 1) * sizeof (*var_spec));
746 n_var_specs = idx + 1;
748 for (int i = 0; i < r->rsd.col_span; ++i)
750 var_spec[idx - i].firstval.text = 0;
751 var_spec[idx - i].firstval.value = 0;
752 var_spec[idx - i].firstval.type = 0;
753 var_spec[idx - i].name =
754 strdup (CHAR_CAST (const char *, value));
762 /* Read in the first row of data */
763 while (1 == xmlTextReaderRead (r->rsd.xtr))
766 process_node (r, &r->rsd);
768 if (! reading_target_sheet (r, &r->rsd))
771 /* If the row is finished then stop for now */
772 if (r->rsd.state == STATE_TABLE &&
773 r->rsd.row > r->start_row + (opts->read_names ? 1 : 0))
776 idx = r->rsd.col - r->start_col - 1;
780 if (r->stop_col != -1 && idx > r->stop_col - r->start_col)
783 if (r->rsd.state == STATE_CELL &&
784 XML_READER_TYPE_ELEMENT == r->rsd.node_type)
786 type = xmlTextReaderGetAttribute (r->rsd.xtr, _xml ("office:value-type"));
787 val_string = xmlTextReaderGetAttribute (r->rsd.xtr, _xml ("office:value"));
790 if (r->rsd.state == STATE_CELL_CONTENT &&
791 XML_READER_TYPE_TEXT == r->rsd.node_type)
793 if (idx >= n_var_specs)
795 var_spec = xrealloc (var_spec, sizeof (*var_spec) * (idx + 1));
796 memset (var_spec + n_var_specs,
798 (idx - n_var_specs + 1) * sizeof (*var_spec));
800 var_spec [idx].name = NULL;
801 n_var_specs = idx + 1;
804 var_spec [idx].firstval.type = type;
805 var_spec [idx].firstval.text = xmlTextReaderValue (r->rsd.xtr);
806 var_spec [idx].firstval.value = val_string;
814 /* Create the dictionary and populate it */
815 r->spreadsheet.dict = r->dict = dict_create (
816 CHAR_CAST (const char *, xmlTextReaderConstEncoding (r->rsd.xtr)));
818 for (i = 0; i < n_var_specs ; ++i)
821 struct variable *var = NULL;
822 char *name = dict_make_unique_var_name (r->dict, var_spec[i].name, &vstart);
823 int width = xmv_to_width (&var_spec[i].firstval, opts->asw);
824 dict_create_var (r->dict, name, width);
827 var = dict_get_var (r->dict, i);
829 if (0 == xmlStrcmp (var_spec[i].firstval.type, _xml("date")))
836 fmt = fmt_default_for_width (width);
838 var_set_both_formats (var, &fmt);
841 if (n_var_specs == 0)
843 msg (MW, _("Selected sheet or range of spreadsheet `%s' is empty."),
844 spreadsheet->file_name);
848 /* Create the first case, and cache it */
849 r->proto = caseproto_ref (dict_get_proto (r->dict));
850 r->first_case = case_create (r->proto);
851 case_set_missing (r->first_case);
853 for (i = 0 ; i < n_var_specs; ++i)
855 const struct variable *var = dict_get_var (r->dict, i);
857 convert_xml_to_value (r->first_case, var, &var_spec[i].firstval,
858 r->rsd.col - n_var_specs + i,
862 /* Read in the first row of data */
863 while (1 == xmlTextReaderRead (r->rsd.xtr))
865 process_node (r, &r->rsd);
867 if (r->rsd.state == STATE_ROW)
872 for (i = 0 ; i < n_var_specs ; ++i)
874 free (var_spec[i].firstval.type);
875 free (var_spec[i].firstval.value);
876 free (var_spec[i].firstval.text);
877 free (var_spec[i].name);
883 return casereader_create_sequential
887 &ods_file_casereader_class, r);
891 for (i = 0 ; i < n_var_specs ; ++i)
893 free (var_spec[i].firstval.type);
894 free (var_spec[i].firstval.value);
895 free (var_spec[i].firstval.text);
896 free (var_spec[i].name);
901 ods_file_casereader_destroy (NULL, r);
907 /* Reads and returns one case from READER's file. Returns a null
908 pointer on failure. */
909 static struct ccase *
910 ods_file_casereader_read (struct casereader *reader UNUSED, void *r_)
912 struct ccase *c = NULL;
913 struct ods_reader *r = r_;
915 xmlChar *val_string = NULL;
916 xmlChar *type = NULL;
918 if (!r->used_first_case)
920 r->used_first_case = true;
921 return r->first_case;
925 /* Advance to the start of a row. (If there is one) */
926 while (r->rsd.state != STATE_ROW
927 && 1 == xmlTextReaderRead (r->rsd.xtr)
930 process_node (r, &r->rsd);
934 if (! reading_target_sheet (r, &r->rsd)
935 || r->rsd.state < STATE_TABLE
936 || (r->stop_row != -1 && r->rsd.row > r->stop_row + 1)
942 c = case_create (r->proto);
943 case_set_missing (c);
945 while (1 == xmlTextReaderRead (r->rsd.xtr))
947 process_node (r, &r->rsd);
949 if (r->stop_row != -1 && r->rsd.row > r->stop_row + 1)
952 if (r->rsd.state == STATE_CELL &&
953 r->rsd.node_type == XML_READER_TYPE_ELEMENT)
955 type = xmlTextReaderGetAttribute (r->rsd.xtr, _xml ("office:value-type"));
956 val_string = xmlTextReaderGetAttribute (r->rsd.xtr, _xml ("office:value"));
959 if (r->rsd.state == STATE_CELL_CONTENT &&
960 r->rsd.node_type == XML_READER_TYPE_TEXT)
963 struct xml_value *xmv = xzalloc (sizeof *xmv);
964 xmv->text = xmlTextReaderValue (r->rsd.xtr);
965 xmv->value = val_string;
970 for (col = 0; col < r->rsd.col_span; ++col)
972 const struct variable *var;
973 const int idx = r->rsd.col - col - r->start_col - 1;
976 if (r->stop_col != -1 && idx > r->stop_col - r->start_col)
978 if (idx >= dict_get_var_cnt (r->dict))
981 var = dict_get_var (r->dict, idx);
982 convert_xml_to_value (c, var, xmv, idx + r->start_col, r->rsd.row - 1);
986 xmlFree (xmv->value);
990 if (r->rsd.state <= STATE_TABLE)
995 xmlFree (val_string);