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/minmax.h"
28 #define _(msgid) gettext (msgid)
29 #define N_(msgid) (msgid)
31 #include "ods-reader.h"
32 #include "spreadsheet-reader.h"
37 ods_open_reader (const struct spreadsheet_read_options *opts,
38 struct dictionary **dict)
40 msg (ME, _("Support for %s files was not compiled into this installation of PSPP"), "OpenDocument");
47 #include "libpspp/zip-reader.h"
53 #include <libxml/xmlreader.h>
56 #include "data/format.h"
57 #include "data/case.h"
58 #include "data/casereader-provider.h"
59 #include "data/dictionary.h"
60 #include "data/identifier.h"
61 #include "data/value.h"
62 #include "data/variable.h"
63 #include "libpspp/i18n.h"
64 #include "libpspp/str.h"
66 #include "gl/xalloc.h"
68 static void ods_file_casereader_destroy (struct casereader *, void *);
70 static struct ccase *ods_file_casereader_read (struct casereader *, void *);
72 static const struct casereader_class ods_file_casereader_class =
74 ods_file_casereader_read,
75 ods_file_casereader_destroy,
82 /* The name of the sheet (utf8 encoding) */
94 STATE_INIT = 0, /* Initial state */
95 STATE_SPREADSHEET, /* Found the start of the spreadsheet doc */
96 STATE_TABLE, /* Found the sheet that we actually want */
97 STATE_ROW, /* Found the start of the cell array */
98 STATE_CELL, /* Found a cell */
99 STATE_CELL_CONTENT /* Found a the text within a cell */
104 struct spreadsheet spreadsheet;
105 struct zip_reader *zreader;
106 xmlTextReaderPtr xtr;
108 enum reader_state state;
113 xmlChar *current_sheet_name;
115 const xmlChar *target_sheet_name;
116 int target_sheet_index;
119 int wanted_row_start;
120 int wanted_col_start;
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;
145 reading_target_sheet (const struct ods_reader *r)
147 if (r->target_sheet_name != NULL)
149 if ( 0 == xmlStrcmp (r->target_sheet_name, r->current_sheet_name))
153 if (r->target_sheet_index == r->current_sheet + 1)
160 static void process_node (struct ods_reader *r);
164 ods_get_sheet_name (struct spreadsheet *s, int n)
167 struct ods_reader *or = (struct ods_reader *) s;
169 assert (n < s->n_sheets);
172 (or->n_allocated_sheets <= n)
174 (1 == (ret = xmlTextReaderRead (or->xtr)))
180 return or->sheets[n].name;
184 ods_get_sheet_range (struct spreadsheet *s, int n)
187 struct ods_reader *or = (struct ods_reader *) s;
189 assert (n < s->n_sheets);
193 (or->n_allocated_sheets <= n)
194 || (or->sheets[n].stop_row == -1) )
196 (1 == (ret = xmlTextReaderRead (or->xtr)))
203 return create_cell_ref (
204 or->sheets[n].start_col,
205 or->sheets[n].start_row,
206 or->sheets[n].stop_col,
207 or->sheets[n].stop_row);
212 ods_file_casereader_destroy (struct casereader *reader UNUSED, void *r_)
214 struct ods_reader *r = r_;
219 xmlFreeTextReader (r->xtr);
222 if ( ! ds_is_empty (&r->ods_errs))
223 msg (ME, "%s", ds_cstr (&r->ods_errs));
225 ds_destroy (&r->ods_errs);
227 if ( ! r->used_first_case )
228 case_unref (r->first_case);
230 caseproto_unref (r->proto);
236 process_node (struct ods_reader *r)
238 xmlChar *name = xmlTextReaderName (r->xtr);
240 name = xmlStrdup (_xml ("--"));
243 r->node_type = xmlTextReaderNodeType (r->xtr);
248 if (0 == xmlStrcasecmp (name, _xml("office:spreadsheet")) &&
249 XML_READER_TYPE_ELEMENT == r->node_type)
251 r->state = STATE_SPREADSHEET;
252 r->current_sheet = -1;
253 r->current_sheet_name = NULL;
256 case STATE_SPREADSHEET:
257 if (0 == xmlStrcasecmp (name, _xml("table:table"))
259 (XML_READER_TYPE_ELEMENT == r->node_type))
261 xmlFree (r->current_sheet_name);
262 r->current_sheet_name = xmlTextReaderGetAttribute (r->xtr, _xml ("table:name"));
266 if (r->current_sheet >= r->n_allocated_sheets)
268 r->sheets = xrealloc (r->sheets, sizeof (*r->sheets) * ++r->n_allocated_sheets);
269 r->sheets[r->n_allocated_sheets - 1].start_col = -1;
270 r->sheets[r->n_allocated_sheets - 1].stop_col = -1;
271 r->sheets[r->n_allocated_sheets - 1].start_row = -1;
272 r->sheets[r->n_allocated_sheets - 1].stop_row = -1;
273 r->sheets[r->n_allocated_sheets - 1].name = xmlStrdup (r->current_sheet_name);
279 r->state = STATE_TABLE;
281 else if (0 == xmlStrcasecmp (name, _xml("office:spreadsheet")) &&
282 XML_READER_TYPE_ELEMENT == r->node_type)
284 r->state = STATE_INIT;
288 if (0 == xmlStrcasecmp (name, _xml("table:table-row")) &&
289 (XML_READER_TYPE_ELEMENT == r->node_type))
292 xmlTextReaderGetAttribute (r->xtr,
293 _xml ("table:number-rows-repeated"));
295 int row_span = value ? _xmlchar_to_int (value) : 1;
300 if (! xmlTextReaderIsEmptyElement (r->xtr))
301 r->state = STATE_ROW;
303 else if (0 == xmlStrcasecmp (name, _xml("table:table")) &&
304 (XML_READER_TYPE_END_ELEMENT == r->node_type))
306 r->state = STATE_SPREADSHEET;
310 // printf ("%s:%d Name is %s\n", __FILE__, __LINE__, name);
311 if ( (0 == xmlStrcasecmp (name, _xml ("table:table-cell")))
313 (XML_READER_TYPE_ELEMENT == r->node_type))
316 xmlTextReaderGetAttribute (r->xtr,
317 _xml ("table:number-columns-repeated"));
319 r->col_span = value ? _xmlchar_to_int (value) : 1;
320 r->col += r->col_span;
322 // printf ("%s:%d %s\n", __FILE__, __LINE__, value);
324 if (! xmlTextReaderIsEmptyElement (r->xtr))
325 r->state = STATE_CELL;
327 else if ( (0 == xmlStrcasecmp (name, _xml ("table:table-row")))
329 (XML_READER_TYPE_END_ELEMENT == r->node_type))
331 r->state = STATE_TABLE;
335 if ( (0 == xmlStrcasecmp (name, _xml("text:p")))
337 ( XML_READER_TYPE_ELEMENT == r->node_type))
339 if (! xmlTextReaderIsEmptyElement (r->xtr))
340 r->state = STATE_CELL_CONTENT;
343 ( (0 == xmlStrcasecmp (name, _xml("table:table-cell")))
345 (XML_READER_TYPE_END_ELEMENT == r->node_type)
348 r->state = STATE_ROW;
351 case STATE_CELL_CONTENT:
352 assert (r->current_sheet >= 0);
353 assert (r->current_sheet < r->n_allocated_sheets);
355 if (r->sheets[r->current_sheet].start_row == -1)
356 r->sheets[r->current_sheet].start_row = r->row - 1;
359 (r->sheets[r->current_sheet].start_col == -1)
361 (r->sheets[r->current_sheet].start_col >= r->col - 1)
363 r->sheets[r->current_sheet].start_col = r->col - 1;
365 r->sheets[r->current_sheet].stop_row = r->row - 1;
367 if ( r->sheets[r->current_sheet].stop_col < r->col - 1)
368 r->sheets[r->current_sheet].stop_col = r->col - 1;
370 if (XML_READER_TYPE_END_ELEMENT == r->node_type)
371 r->state = STATE_CELL;
382 A struct containing the parameters of a cell's value
395 struct xml_value firstval;
399 /* Determine the width that a xmv should probably have */
401 xmv_to_width (const struct xml_value *xmv, int fallback)
403 int width = SPREADSHEET_DEFAULT_WIDTH;
405 /* Non-strings always have zero width */
406 if (xmv->type != NULL && 0 != xmlStrcmp (xmv->type, _xml("string")))
413 width = ROUND_UP (xmlStrlen (xmv->value),
414 SPREADSHEET_DEFAULT_WIDTH);
416 width = ROUND_UP (xmlStrlen (xmv->text),
417 SPREADSHEET_DEFAULT_WIDTH);
423 Sets the VAR of case C, to the value corresponding to the xml data
426 convert_xml_to_value (struct ccase *c, const struct variable *var,
427 const struct xml_value *xmv)
429 union value *v = case_data_rw (c, var);
431 if (xmv->value == NULL && xmv->text == NULL)
432 value_set_missing (v, var_get_width (var));
433 else if ( var_is_alpha (var))
434 /* Use the text field, because it seems that there is no
435 value field for strings */
436 value_copy_str_rpad (v, var_get_width (var), xmv->text, ' ');
440 const struct fmt_spec *fmt = var_get_write_format (var);
441 enum fmt_category fc = fmt_get_category (fmt->type);
443 assert ( fc != FMT_CAT_STRING);
446 xmv->value ? CHAR_CAST (const char *, xmv->value) : CHAR_CAST (const char *, xmv->text);
448 free (data_in (ss_cstr (text), "UTF-8",
457 /* Try to find out how many sheets there are in the "workbook" */
459 get_sheet_count (struct zip_reader *zreader)
461 xmlTextReaderPtr mxtr;
462 struct zip_member *meta = NULL;
463 meta = zip_member_open (zreader, "meta.xml");
468 mxtr = xmlReaderForIO ((xmlInputReadCallback) zip_member_read,
469 (xmlInputCloseCallback) zip_member_finish,
470 meta, NULL, NULL, 0);
472 while (1 == xmlTextReaderRead (mxtr))
474 xmlChar *name = xmlTextReaderName (mxtr);
475 if ( 0 == xmlStrcmp (name, _xml("meta:document-statistic")))
477 xmlChar *attr = xmlTextReaderGetAttribute (mxtr, _xml ("meta:table-count"));
481 int s = _xmlchar_to_int (attr);
490 ods_error_handler (void *ctx, const char *mesg,
491 UNUSED xmlParserSeverities sev, xmlTextReaderLocatorPtr loc)
493 struct ods_reader *r = ctx;
495 msg (MW, _("There was a problem whilst reading the %s file `%s' (near line %d): `%s'"),
497 r->spreadsheet.file_name,
498 xmlTextReaderLocatorLineNumber (loc),
504 init_reader (struct ods_reader *r, bool report_errors)
506 struct zip_member *content = zip_member_open (r->zreader, "content.xml");
507 xmlTextReaderPtr xtr;
509 if ( content == NULL)
512 zip_member_ref (content);
515 xtr = xmlReaderForIO ((xmlInputReadCallback) zip_member_read,
516 (xmlInputCloseCallback) zip_member_finish,
518 report_errors ? 0 : (XML_PARSE_NOERROR | XML_PARSE_NOWARNING) );
524 r->spreadsheet.type = SPREADSHEET_ODS;
527 xmlTextReaderSetErrorHandler (xtr, ods_error_handler, r);
534 ods_probe (const char *filename, bool report_errors)
536 struct ods_reader *r;
537 struct string errs = DS_EMPTY_INITIALIZER;
539 struct zip_reader *zr = zip_reader_create (filename, &errs);
545 msg (ME, _("Cannot open %s as a OpenDocument file: %s"),
546 filename, ds_cstr (&errs));
551 sheet_count = get_sheet_count (zr);
553 r = xzalloc (sizeof *r);
556 if (! init_reader (r, report_errors))
561 r->spreadsheet.n_sheets = sheet_count;
562 r->n_allocated_sheets = 0;
567 r->spreadsheet.file_name = filename;
568 return &r->spreadsheet;
571 zip_reader_destroy (r->zreader);
578 ods_make_reader (struct spreadsheet *spreadsheet,
579 const struct spreadsheet_read_options *opts)
582 xmlChar *type = NULL;
583 unsigned long int vstart = 0;
584 casenumber n_cases = CASENUMBER_MAX;
586 struct var_spec *var_spec = NULL;
589 struct ods_reader *r = (struct ods_reader *) spreadsheet;
590 xmlChar *val_string = NULL;
593 r->read_names = opts->read_names;
594 ds_init_empty (&r->ods_errs);
597 if ( !init_reader (r, true))
601 if ( opts->cell_range )
603 if ( ! convert_cell_ref (opts->cell_range,
604 &r->start_col, &r->start_row,
605 &r->stop_col, &r->stop_row))
607 msg (SE, _("Invalid cell range `%s'"),
621 r->state = STATE_INIT;
622 r->target_sheet_name = BAD_CAST opts->sheet_name;
623 r->target_sheet_index = opts->sheet_index;
627 /* If CELLRANGE was given, then we know how many variables should be read */
628 if ( r->stop_col != -1 )
630 assert (var_spec == NULL);
631 n_var_specs = r->stop_col - r->start_col + 1;
632 var_spec = xrealloc (var_spec, sizeof (*var_spec) * n_var_specs);
633 memset (var_spec, '\0', sizeof (*var_spec) * n_var_specs);
637 /* Advance to the start of the cells for the target sheet */
638 while ( ! reading_target_sheet (r) || r->state != STATE_ROW )
640 if (1 != (ret = xmlTextReaderRead (r->xtr)))
649 msg (MW, _("Selected sheet or range of spreadsheet `%s' is empty."),
650 spreadsheet->file_name);
654 if ( opts->read_names)
656 while (1 == (ret = xmlTextReaderRead (r->xtr)))
662 /* If the row is finished then stop for now */
663 if (r->state == STATE_TABLE && r->row > r->wanted_row_start)
666 idx = r->col - r->wanted_col_start - 1;
668 if (r->state == STATE_CELL_CONTENT
670 XML_READER_TYPE_TEXT == r->node_type)
672 xmlChar *value = xmlTextReaderValue (r->xtr);
674 if ( idx >= n_var_specs)
676 var_spec = xrealloc (var_spec, sizeof (*var_spec) * (idx + 1));
678 /* xrealloc (unlike realloc) doesn't initialise its memory to 0 */
679 memset (var_spec + n_var_specs,
681 (n_var_specs - idx + 1) * sizeof (*var_spec));
682 n_var_specs = idx + 1;
684 var_spec[idx].firstval.text = 0;
685 var_spec[idx].firstval.value = 0;
686 var_spec[idx].firstval.type = 0;
688 var_spec [idx].name = strdup (CHAR_CAST (const char *, value));
694 /* Read in the first row of data */
695 while (1 == xmlTextReaderRead (r->xtr))
700 /* If the row is finished then stop for now */
701 if (r->state == STATE_TABLE && r->row > r->wanted_row_start + (opts->read_names ? 1 : 0))
704 idx = r->col - r->wanted_col_start - 1;
706 if ( r->state == STATE_CELL &&
707 XML_READER_TYPE_ELEMENT == r->node_type)
709 type = xmlTextReaderGetAttribute (r->xtr, _xml ("office:value-type"));
710 val_string = xmlTextReaderGetAttribute (r->xtr, _xml ("office:value"));
713 if ( r->state == STATE_CELL_CONTENT &&
714 XML_READER_TYPE_TEXT == r->node_type)
716 if ( idx >= n_var_specs)
718 var_spec = xrealloc (var_spec, sizeof (*var_spec) * (idx + 1));
719 var_spec [idx].name = NULL;
720 n_var_specs = idx + 1;
723 var_spec [idx].firstval.type = type;
724 var_spec [idx].firstval.text = xmlTextReaderValue (r->xtr);
725 var_spec [idx].firstval.value = val_string;
733 /* Create the dictionary and populate it */
734 r->spreadsheet.dict = r->dict = dict_create (
735 CHAR_CAST (const char *, xmlTextReaderConstEncoding (r->xtr)));
737 for (i = 0; i < n_var_specs ; ++i )
740 struct variable *var = NULL;
741 char *name = dict_make_unique_var_name (r->dict, var_spec[i].name, &vstart);
742 int width = xmv_to_width (&var_spec[i].firstval, opts->asw);
743 dict_create_var (r->dict, name, width);
746 var = dict_get_var (r->dict, i);
748 if ( 0 == xmlStrcmp (var_spec[i].firstval.type, _xml("date")))
755 fmt = fmt_default_for_width (width);
757 var_set_both_formats (var, &fmt);
760 /* Create the first case, and cache it */
761 r->used_first_case = false;
763 if ( n_var_specs == 0 )
765 msg (MW, _("Selected sheet or range of spreadsheet `%s' is empty."),
766 spreadsheet->file_name);
770 r->proto = caseproto_ref (dict_get_proto (r->dict));
771 r->first_case = case_create (r->proto);
772 case_set_missing (r->first_case);
774 for (i = 0 ; i < n_var_specs; ++i)
776 const struct variable *var = dict_get_var (r->dict, i);
778 convert_xml_to_value (r->first_case, var, &var_spec[i].firstval);
781 /* Read in the first row of data */
782 while (1 == xmlTextReaderRead (r->xtr))
786 if (r->state == STATE_ROW)
790 // zip_reader_destroy (zreader);
793 for ( i = 0 ; i < n_var_specs ; ++i )
795 free (var_spec[i].firstval.type);
796 free (var_spec[i].firstval.value);
797 free (var_spec[i].firstval.text);
798 free (var_spec[i].name);
804 return casereader_create_sequential
808 &ods_file_casereader_class, r);
812 // zip_reader_destroy (zreader);
814 for ( i = 0 ; i < n_var_specs ; ++i )
816 free (var_spec[i].firstval.type);
817 free (var_spec[i].firstval.value);
818 free (var_spec[i].firstval.text);
819 free (var_spec[i].name);
824 dict_destroy (r->spreadsheet.dict);
825 r->spreadsheet.dict = NULL;
826 ods_file_casereader_destroy (NULL, r);
833 /* Reads and returns one case from READER's file. Returns a null
834 pointer on failure. */
835 static struct ccase *
836 ods_file_casereader_read (struct casereader *reader UNUSED, void *r_)
838 struct ccase *c = NULL;
839 xmlChar *val_string = NULL;
840 struct ods_reader *r = r_;
841 int current_row = r->row;
843 if (!r->used_first_case)
845 r->used_first_case = true;
846 return r->first_case;
850 /* Advance to the start of a row. (If there is one) */
851 while (r->state != STATE_ROW && 1 == xmlTextReaderRead (r->xtr))
857 if ( ! reading_target_sheet (r) || r->state < STATE_TABLE)
862 c = case_create (r->proto);
863 case_set_missing (c);
865 while (1 == xmlTextReaderRead (r->xtr))
869 if (r->row > current_row && r->state == STATE_ROW)
872 // printf ("%s:%d\n", __FILE__, __LINE__);
873 if (r->state == STATE_CELL &&
874 r->node_type == XML_READER_TYPE_ELEMENT)
876 val_string = xmlTextReaderGetAttribute (r->xtr, _xml ("office:value"));
879 if (r->state == STATE_CELL_CONTENT &&
880 r->node_type == XML_READER_TYPE_TEXT)
883 struct xml_value *xmv = xzalloc (sizeof *xmv);
884 xmv->text = xmlTextReaderValue (r->xtr);
885 xmv->value = val_string;
888 for (col = 0; col < r->col_span; ++col)
890 const int idx = r->col + col - r->wanted_col_start - 1;
891 const struct variable *var = dict_get_var (r->dict, idx);
892 convert_xml_to_value (c, var, xmv);
899 if ( r->state <= STATE_TABLE)