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;
115 const xmlChar *target_sheet;
116 int target_sheet_index;
127 struct sheet_detail *sheets;
128 int n_allocated_sheets;
130 struct caseproto *proto;
131 struct dictionary *dict;
132 struct ccase *first_case;
133 bool used_first_case;
136 struct string ods_errs;
140 static void process_node (struct ods_reader *r);
144 ods_get_sheet_name (struct spreadsheet *s, int n)
147 struct ods_reader *or = (struct ods_reader *) s;
149 assert (n < s->n_sheets);
152 (or->n_allocated_sheets <= n)
154 (1 == (ret = xmlTextReaderRead (or->xtr)))
160 return or->sheets[n].name;
164 ods_get_sheet_range (struct spreadsheet *s, int n)
167 struct ods_reader *or = (struct ods_reader *) s;
169 assert (n < s->n_sheets);
173 (or->n_allocated_sheets <= n)
174 || (or->sheets[n].stop_row == -1) )
176 (1 == (ret = xmlTextReaderRead (or->xtr)))
183 return create_cell_ref (
184 or->sheets[n].start_col,
185 or->sheets[n].start_row,
186 or->sheets[n].stop_col,
187 or->sheets[n].stop_row);
192 ods_file_casereader_destroy (struct casereader *reader UNUSED, void *r_)
194 struct ods_reader *r = r_;
199 xmlFreeTextReader (r->xtr);
202 if ( ! ds_is_empty (&r->ods_errs))
203 msg (ME, "%s", ds_cstr (&r->ods_errs));
205 ds_destroy (&r->ods_errs);
207 if ( ! r->used_first_case )
208 case_unref (r->first_case);
210 caseproto_unref (r->proto);
216 process_node (struct ods_reader *r)
218 xmlChar *name = xmlTextReaderName (r->xtr);
220 name = xmlStrdup (_xml ("--"));
223 r->node_type = xmlTextReaderNodeType (r->xtr);
228 if (0 == xmlStrcasecmp (name, _xml("office:spreadsheet")) &&
229 XML_READER_TYPE_ELEMENT == r->node_type)
231 r->state = STATE_SPREADSHEET;
232 r->current_sheet = -1;
235 case STATE_SPREADSHEET:
236 if (0 == xmlStrcasecmp (name, _xml("table:table"))
238 (XML_READER_TYPE_ELEMENT == r->node_type))
240 xmlChar *value = xmlTextReaderGetAttribute (r->xtr, _xml ("table:name"));
244 if (r->current_sheet >= r->n_allocated_sheets)
246 r->sheets = xrealloc (r->sheets, sizeof (*r->sheets) * ++r->n_allocated_sheets);
247 r->sheets[r->n_allocated_sheets - 1].start_col = -1;
248 r->sheets[r->n_allocated_sheets - 1].stop_col = -1;
249 r->sheets[r->n_allocated_sheets - 1].start_row = -1;
250 r->sheets[r->n_allocated_sheets - 1].stop_row = -1;
251 r->sheets[r->n_allocated_sheets - 1].name = value;
257 if ( r->target_sheet != NULL)
259 if ( 0 == xmlStrcmp (value, r->target_sheet))
261 r->sheet_found = true;
264 else if (r->target_sheet_index == r->current_sheet)
266 r->sheet_found = true;
268 r->state = STATE_TABLE;
270 else if (0 == xmlStrcasecmp (name, _xml("office:spreadsheet")) &&
271 XML_READER_TYPE_ELEMENT == r->node_type)
273 r->state = STATE_INIT;
274 printf ("%s:%d End of Workbook %d: Rows %d Cols %d\n", __FILE__, __LINE__,
275 r->current_sheet, r->row, r->col);
279 if (0 == xmlStrcasecmp (name, _xml("table:table-row")) &&
280 (XML_READER_TYPE_ELEMENT == r->node_type))
283 xmlTextReaderGetAttribute (r->xtr,
284 _xml ("table:number-rows-repeated"));
286 int row_span = value ? _xmlchar_to_int (value) : 1;
291 if (! xmlTextReaderIsEmptyElement (r->xtr))
292 r->state = STATE_ROW;
294 else if (0 == xmlStrcasecmp (name, _xml("table:table")) &&
295 (XML_READER_TYPE_END_ELEMENT == r->node_type))
297 r->state = STATE_SPREADSHEET;
301 if ( (0 == xmlStrcasecmp (name, _xml ("table:table-cell")))
303 (XML_READER_TYPE_ELEMENT == r->node_type))
306 xmlTextReaderGetAttribute (r->xtr,
307 _xml ("table:number-columns-repeated"));
309 r->col_span = value ? _xmlchar_to_int (value) : 1;
310 r->col += r->col_span;
312 if (! xmlTextReaderIsEmptyElement (r->xtr))
313 r->state = STATE_CELL;
315 else if ( (0 == xmlStrcasecmp (name, _xml ("table:table-row")))
317 (XML_READER_TYPE_END_ELEMENT == r->node_type))
319 r->state = STATE_TABLE;
323 if ( (0 == xmlStrcasecmp (name, _xml("text:p")))
325 ( XML_READER_TYPE_ELEMENT == r->node_type))
327 if (! xmlTextReaderIsEmptyElement (r->xtr))
328 r->state = STATE_CELL_CONTENT;
331 ( (0 == xmlStrcasecmp (name, _xml("table:table-cell")))
333 (XML_READER_TYPE_END_ELEMENT == r->node_type)
336 r->state = STATE_ROW;
339 case STATE_CELL_CONTENT:
340 assert (r->current_sheet >= 0);
341 assert (r->current_sheet < r->n_allocated_sheets);
343 if (r->sheets[r->current_sheet].start_row == -1)
344 r->sheets[r->current_sheet].start_row = r->row - 1;
347 (r->sheets[r->current_sheet].start_col == -1)
349 (r->sheets[r->current_sheet].start_col >= r->col - 1)
351 r->sheets[r->current_sheet].start_col = r->col - 1;
353 r->sheets[r->current_sheet].stop_row = r->row - 1;
355 if ( r->sheets[r->current_sheet].stop_col < r->col - 1)
356 r->sheets[r->current_sheet].stop_col = r->col - 1;
358 if (XML_READER_TYPE_END_ELEMENT == r->node_type)
359 r->state = STATE_CELL;
370 A struct containing the parameters of a cell's value
383 struct xml_value firstval;
387 /* Determine the width that a xmv should probably have */
389 xmv_to_width (const struct xml_value *xmv, int fallback)
391 int width = SPREADSHEET_DEFAULT_WIDTH;
393 /* Non-strings always have zero width */
394 if (xmv->type != NULL && 0 != xmlStrcmp (xmv->type, _xml("string")))
401 width = ROUND_UP (xmlStrlen (xmv->value),
402 SPREADSHEET_DEFAULT_WIDTH);
404 width = ROUND_UP (xmlStrlen (xmv->text),
405 SPREADSHEET_DEFAULT_WIDTH);
411 Sets the VAR of case C, to the value corresponding to the xml data
414 convert_xml_to_value (struct ccase *c, const struct variable *var,
415 const struct xml_value *xmv)
417 union value *v = case_data_rw (c, var);
419 if (xmv->value == NULL && xmv->text == NULL)
420 value_set_missing (v, var_get_width (var));
421 else if ( var_is_alpha (var))
422 /* Use the text field, because it seems that there is no
423 value field for strings */
424 value_copy_str_rpad (v, var_get_width (var), xmv->text, ' ');
428 const struct fmt_spec *fmt = var_get_write_format (var);
429 enum fmt_category fc = fmt_get_category (fmt->type);
431 assert ( fc != FMT_CAT_STRING);
434 xmv->value ? CHAR_CAST (const char *, xmv->value) : CHAR_CAST (const char *, xmv->text);
436 free (data_in (ss_cstr (text), "UTF-8",
445 /* Try to find out how many sheets there are in the "workbook" */
447 get_sheet_count (struct zip_reader *zreader)
449 xmlTextReaderPtr mxtr;
450 struct zip_member *meta = NULL;
451 meta = zip_member_open (zreader, "meta.xml");
456 mxtr = xmlReaderForIO ((xmlInputReadCallback) zip_member_read,
457 (xmlInputCloseCallback) zip_member_finish,
458 meta, NULL, NULL, 0);
460 while (1 == xmlTextReaderRead (mxtr))
462 xmlChar *name = xmlTextReaderName (mxtr);
463 if ( 0 == xmlStrcmp (name, _xml("meta:document-statistic")))
465 xmlChar *attr = xmlTextReaderGetAttribute (mxtr, _xml ("meta:table-count"));
469 int s = _xmlchar_to_int (attr);
478 ods_error_handler (void *ctx, const char *mesg,
479 UNUSED xmlParserSeverities sev, xmlTextReaderLocatorPtr loc)
481 struct ods_reader *r = ctx;
483 msg (MW, _("There was a problem whilst reading the %s file `%s' (near line %d): `%s'"),
485 r->spreadsheet.file_name,
486 xmlTextReaderLocatorLineNumber (loc),
492 init_reader (struct ods_reader *r, bool report_errors)
494 struct zip_member *content = zip_member_open (r->zreader, "content.xml");
495 xmlTextReaderPtr xtr;
497 if ( content == NULL)
500 zip_member_ref (content);
503 xtr = xmlReaderForIO ((xmlInputReadCallback) zip_member_read,
504 (xmlInputCloseCallback) zip_member_finish,
506 report_errors ? 0 : (XML_PARSE_NOERROR | XML_PARSE_NOWARNING) );
512 r->spreadsheet.type = SPREADSHEET_ODS;
515 xmlTextReaderSetErrorHandler (xtr, ods_error_handler, r);
522 ods_probe (const char *filename, bool report_errors)
524 struct ods_reader *r;
525 struct string errs = DS_EMPTY_INITIALIZER;
527 struct zip_reader *zr = zip_reader_create (filename, &errs);
532 sheet_count = get_sheet_count (zr);
534 r = xzalloc (sizeof *r);
537 if (! init_reader (r, report_errors))
542 r->spreadsheet.n_sheets = sheet_count;
543 r->n_allocated_sheets = 0;
548 r->spreadsheet.file_name = filename;
549 return &r->spreadsheet;
552 zip_reader_destroy (r->zreader);
559 ods_make_reader (struct spreadsheet *spreadsheet,
560 const struct spreadsheet_read_options *opts)
563 xmlChar *type = NULL;
564 unsigned long int vstart = 0;
565 casenumber n_cases = CASENUMBER_MAX;
567 struct var_spec *var_spec = NULL;
570 struct ods_reader *r = (struct ods_reader *) spreadsheet;
571 xmlChar *val_string = NULL;
574 r->read_names = opts->read_names;
575 ds_init_empty (&r->ods_errs);
578 if ( !init_reader (r, true))
582 if ( opts->cell_range )
584 if ( ! convert_cell_ref (opts->cell_range,
585 &r->start_col, &r->start_row,
586 &r->stop_col, &r->stop_row))
588 msg (SE, _("Invalid cell range `%s'"),
602 r->state = STATE_INIT;
603 r->target_sheet = BAD_CAST opts->sheet_name;
604 r->target_sheet_index = opts->sheet_index;
607 /* If CELLRANGE was given, then we know how many variables should be read */
608 if ( r->stop_col != -1 )
610 assert (var_spec == NULL);
611 n_var_specs = r->stop_col - r->start_col + 1;
612 var_spec = xrealloc (var_spec, sizeof (*var_spec) * n_var_specs);
613 memset (var_spec, '\0', sizeof (*var_spec) * n_var_specs);
616 /* Advance to the start of the cells for the target sheet */
617 while ( r->current_sheet < r->target_sheet_index - 1 ||
618 r->state != STATE_TABLE
621 if (1 != (ret = xmlTextReaderRead (r->xtr)))
630 msg (MW, _("Selected sheet or range of spreadsheet `%s' is empty."),
631 spreadsheet->file_name);
635 if ( opts->read_names)
637 while (1 == (ret = xmlTextReaderRead (r->xtr)))
643 /* If the row is finished then stop for now */
644 if (r->state == STATE_TABLE && r->row > r->start_row)
647 idx = r->col - r->start_col - 1;
649 if (r->state == STATE_CELL_CONTENT
651 XML_READER_TYPE_TEXT == r->node_type)
653 xmlChar *value = xmlTextReaderValue (r->xtr);
655 if ( idx >= n_var_specs)
657 var_spec = xrealloc (var_spec, sizeof (*var_spec) * (idx + 1));
659 /* xrealloc (unlike realloc) doesn't initialise its memory to 0 */
660 memset (var_spec + n_var_specs,
662 (n_var_specs - idx + 1) * sizeof (*var_spec));
663 n_var_specs = idx + 1;
665 var_spec[idx].firstval.text = 0;
666 var_spec[idx].firstval.value = 0;
667 var_spec[idx].firstval.type = 0;
669 var_spec [idx].name = strdup (CHAR_CAST (const char *, value));
675 /* Read in the first row of data */
676 while (1 == xmlTextReaderRead (r->xtr))
681 /* If the row is finished then stop for now */
682 if (r->state == STATE_TABLE && r->row > r->start_row + (opts->read_names ? 1 : 0))
685 idx = r->col - r->start_col - 1;
687 if ( r->state == STATE_CELL &&
688 XML_READER_TYPE_ELEMENT == r->node_type)
690 type = xmlTextReaderGetAttribute (r->xtr, _xml ("office:value-type"));
691 val_string = xmlTextReaderGetAttribute (r->xtr, _xml ("office:value"));
694 if ( r->state == STATE_CELL_CONTENT &&
695 XML_READER_TYPE_TEXT == r->node_type)
697 if ( idx >= n_var_specs)
699 var_spec = xrealloc (var_spec, sizeof (*var_spec) * (idx + 1));
700 var_spec [idx].name = NULL;
701 n_var_specs = idx + 1;
704 var_spec [idx].firstval.type = type;
705 var_spec [idx].firstval.text = xmlTextReaderValue (r->xtr);
706 var_spec [idx].firstval.value = val_string;
713 /* Create the dictionary and populate it */
714 r->spreadsheet.dict = r->dict = dict_create (
715 CHAR_CAST (const char *, xmlTextReaderConstEncoding (r->xtr)));
717 for (i = 0; i < n_var_specs ; ++i )
720 struct variable *var = NULL;
721 char *name = dict_make_unique_var_name (r->dict, var_spec[i].name, &vstart);
722 int width = xmv_to_width (&var_spec[i].firstval, opts->asw);
723 dict_create_var (r->dict, name, width);
726 var = dict_get_var (r->dict, i);
728 if ( 0 == xmlStrcmp (var_spec[i].firstval.type, _xml("date")))
735 fmt = fmt_default_for_width (width);
737 var_set_both_formats (var, &fmt);
740 /* Create the first case, and cache it */
741 r->used_first_case = false;
743 if ( n_var_specs == 0 )
745 msg (MW, _("Selected sheet or range of spreadsheet `%s' is empty."),
746 spreadsheet->file_name);
750 r->proto = caseproto_ref (dict_get_proto (r->dict));
751 r->first_case = case_create (r->proto);
752 case_set_missing (r->first_case);
754 for (i = 0 ; i < n_var_specs; ++i)
756 const struct variable *var = dict_get_var (r->dict, i);
758 convert_xml_to_value (r->first_case, var, &var_spec[i].firstval);
761 /* Read in the first row of data */
762 while (1 == xmlTextReaderRead (r->xtr))
766 if (r->state == STATE_ROW)
770 // zip_reader_destroy (zreader);
773 for ( i = 0 ; i < n_var_specs ; ++i )
775 free (var_spec[i].firstval.type);
776 free (var_spec[i].firstval.value);
777 free (var_spec[i].firstval.text);
778 free (var_spec[i].name);
784 return casereader_create_sequential
788 &ods_file_casereader_class, r);
792 // zip_reader_destroy (zreader);
794 for ( i = 0 ; i < n_var_specs ; ++i )
796 free (var_spec[i].firstval.type);
797 free (var_spec[i].firstval.value);
798 free (var_spec[i].firstval.text);
799 free (var_spec[i].name);
804 dict_destroy (r->spreadsheet.dict);
805 r->spreadsheet.dict = NULL;
806 ods_file_casereader_destroy (NULL, r);
813 /* Reads and returns one case from READER's file. Returns a null
814 pointer on failure. */
815 static struct ccase *
816 ods_file_casereader_read (struct casereader *reader UNUSED, void *r_)
818 struct ccase *c = NULL;
819 xmlChar *val_string = NULL;
820 struct ods_reader *r = r_;
821 int current_row = r->row;
823 if ( !r->used_first_case )
825 r->used_first_case = true;
826 return r->first_case;
829 if ( r->state > STATE_INIT)
831 c = case_create (r->proto);
832 case_set_missing (c);
835 if (r->state == STATE_SPREADSHEET
837 r->current_sheet == r->target_sheet_index - 1
843 while (1 == xmlTextReaderRead (r->xtr))
847 if (r->row > current_row && r->state == STATE_ROW)
850 if ( r->state == STATE_CELL &&
851 r->node_type == XML_READER_TYPE_ELEMENT )
853 val_string = xmlTextReaderGetAttribute (r->xtr, _xml ("office:value"));
856 if ( r->state == STATE_CELL_CONTENT &&
857 r->node_type == XML_READER_TYPE_TEXT )
860 struct xml_value *xmv = xzalloc (sizeof *xmv);
861 xmv->text = xmlTextReaderValue (r->xtr);
862 xmv->value = val_string;
865 for (col = 0; col < r->col_span; ++col)
867 const int idx = r->col + col - r->start_col - 1;
868 const struct variable *var = dict_get_var (r->dict, idx);
869 convert_xml_to_value (c, var, xmv);
877 if ( r->state < STATE_TABLE)