1 /* PSPP - a program for statistical analysis.
2 Copyright (C) 2007, 2009, 2010, 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"
22 #include "gl/minmax.h"
23 #include "gl/c-strtod.h"
26 #define _(msgid) gettext (msgid)
27 #define N_(msgid) (msgid)
29 #include "spreadsheet-reader.h"
34 gnumeric_open_reader (const struct spreadsheet_read_options *opts, struct dictionary **dict)
36 msg (ME, _("Support for %s files was not compiled into this installation of PSPP"), "Gnumeric");
43 #include "data/gnumeric-reader.h"
48 #include <libxml/xmlreader.h>
51 #include "data/case.h"
52 #include "data/casereader-provider.h"
53 #include "data/dictionary.h"
54 #include "data/identifier.h"
55 #include "data/value.h"
56 #include "data/variable.h"
57 #include "libpspp/i18n.h"
58 #include "libpspp/str.h"
60 #include "gl/xalloc.h"
62 static void gnm_file_casereader_destroy (struct casereader *, void *);
64 static struct ccase *gnm_file_casereader_read (struct casereader *, void *);
67 static const struct casereader_class gnm_file_casereader_class =
69 gnm_file_casereader_read,
70 gnm_file_casereader_destroy,
77 STATE_PRE_INIT = 0, /* Initial state */
78 STATE_SHEET_COUNT, /* Found the sheet index */
79 STATE_INIT , /* Other Initial state */
80 STATE_SHEET_START, /* Found the start of a sheet */
81 STATE_SHEET_NAME, /* Found the sheet name */
84 STATE_SHEET_FOUND, /* Found the sheet that we actually want */
85 STATE_CELLS_START, /* Found the start of the cell array */
86 STATE_CELL /* Found a cell */
91 /* The name of the sheet (utf8 encoding) */
104 struct gnumeric_reader
106 struct spreadsheet spreadsheet;
109 /* The libxml reader for this instance */
110 xmlTextReaderPtr xtr;
112 /* An internal state variable */
113 enum reader_state state;
126 struct sheet_detail *sheets;
128 const xmlChar *target_sheet;
129 int target_sheet_index;
131 struct caseproto *proto;
132 struct dictionary *dict;
133 struct ccase *first_case;
134 bool used_first_case;
139 gnumeric_destroy (struct spreadsheet *s)
141 struct gnumeric_reader *r = (struct gnumeric_reader *) s;
143 if (0 == --r->ref_cnt)
147 for (i = 0; i < s->n_sheets; ++i)
149 xmlFree (r->sheets[i].name);
160 gnumeric_get_sheet_name (struct spreadsheet *s, int n)
162 struct gnumeric_reader *gr = (struct gnumeric_reader *) s;
163 assert (n < s->n_sheets);
165 return gr->sheets[n].name;
169 static void process_node (struct gnumeric_reader *r);
174 gnumeric_get_sheet_range (struct spreadsheet *s, int n)
177 struct gnumeric_reader *gr = (struct gnumeric_reader *) s;
179 assert (n < s->n_sheets);
182 (gr->sheets[n].stop_col == -1)
184 (1 == (ret = xmlTextReaderRead (gr->xtr)))
190 return create_cell_ref (
191 gr->sheets[n].start_col,
192 gr->sheets[n].start_row,
193 gr->sheets[n].stop_col,
194 gr->sheets[n].stop_row);
199 gnm_file_casereader_destroy (struct casereader *reader UNUSED, void *r_)
201 struct gnumeric_reader *r = r_;
206 xmlFreeTextReader (r->xtr);
209 if ( ! r->used_first_case )
210 case_unref (r->first_case);
212 caseproto_unref (r->proto);
214 gnumeric_destroy (&r->spreadsheet);
219 process_node (struct gnumeric_reader *r)
221 xmlChar *name = xmlTextReaderName (r->xtr);
223 name = xmlStrdup (_xml ("--"));
225 r->node_type = xmlTextReaderNodeType (r->xtr);
230 r->current_sheet = -1;
231 if (0 == xmlStrcasecmp (name, _xml("gnm:SheetNameIndex")) &&
232 XML_READER_TYPE_ELEMENT == r->node_type)
234 r->state = STATE_SHEET_COUNT;
238 case STATE_SHEET_COUNT:
239 if (0 == xmlStrcasecmp (name, _xml("gnm:SheetName")) &&
240 XML_READER_TYPE_ELEMENT == r->node_type)
243 if (r->current_sheet + 1 > r->spreadsheet.n_sheets)
245 struct sheet_detail *sd ;
246 r->sheets = xrealloc (r->sheets, (r->current_sheet + 1) * sizeof *r->sheets);
247 sd = &r->sheets[r->current_sheet];
248 sd->start_col = sd->stop_col = sd->start_row = sd->stop_row = -1;
249 r->spreadsheet.n_sheets = r->current_sheet + 1;
252 else if (0 == xmlStrcasecmp (name, _xml("gnm:SheetNameIndex")) &&
253 XML_READER_TYPE_END_ELEMENT == r->node_type)
255 r->state = STATE_INIT;
256 r->current_sheet = -1;
258 else if (XML_READER_TYPE_TEXT == r->node_type)
260 r->sheets [r->spreadsheet.n_sheets - 1].name = CHAR_CAST (char *, xmlTextReaderValue (r->xtr));
265 if (0 == xmlStrcasecmp (name, _xml("gnm:Sheet")) &&
266 XML_READER_TYPE_ELEMENT == r->node_type)
269 r->state = STATE_SHEET_START;
272 case STATE_SHEET_START:
273 if (0 == xmlStrcasecmp (name, _xml("gnm:Name")) &&
274 XML_READER_TYPE_ELEMENT == r->node_type)
276 r->state = STATE_SHEET_NAME;
279 case STATE_SHEET_NAME:
280 if (0 == xmlStrcasecmp (name, _xml("gnm:Name")) &&
281 XML_READER_TYPE_END_ELEMENT == r->node_type)
283 r->state = STATE_INIT;
285 else if (0 == xmlStrcasecmp (name, _xml("gnm:Sheet")) &&
286 XML_READER_TYPE_END_ELEMENT == r->node_type)
288 r->state = STATE_INIT;
290 else if (XML_READER_TYPE_TEXT == r->node_type)
292 if ( r->target_sheet != NULL)
294 xmlChar *value = xmlTextReaderValue (r->xtr);
295 if ( 0 == xmlStrcmp (value, r->target_sheet))
296 r->state = STATE_SHEET_FOUND;
299 else if (r->target_sheet_index == r->current_sheet + 1)
301 r->state = STATE_SHEET_FOUND;
303 else if (r->target_sheet_index == -1)
305 r->state = STATE_SHEET_FOUND;
309 case STATE_SHEET_FOUND:
310 if (0 == xmlStrcasecmp (name, _xml("gnm:Cells")) &&
311 XML_READER_TYPE_ELEMENT == r->node_type)
313 r->min_col = INT_MAX;
314 if (! xmlTextReaderIsEmptyElement (r->xtr))
315 r->state = STATE_CELLS_START;
317 else if (0 == xmlStrcasecmp (name, _xml("gnm:MaxRow")) &&
318 XML_READER_TYPE_ELEMENT == r->node_type)
320 r->state = STATE_MAXROW;
322 else if (0 == xmlStrcasecmp (name, _xml("gnm:MaxCol")) &&
323 XML_READER_TYPE_ELEMENT == r->node_type)
325 r->state = STATE_MAXCOL;
327 else if (0 == xmlStrcasecmp (name, _xml("gnm:Sheet")) &&
328 XML_READER_TYPE_END_ELEMENT == r->node_type)
330 r->state = STATE_INIT;
334 if (0 == xmlStrcasecmp (name, _xml("gnm:MaxRow")) &&
335 XML_READER_TYPE_END_ELEMENT == r->node_type)
337 r->state = STATE_SHEET_FOUND;
339 else if (r->node_type == XML_READER_TYPE_TEXT)
341 xmlChar *value = xmlTextReaderValue (r->xtr);
342 r->sheets[r->current_sheet].maxrow = _xmlchar_to_int (value);
347 if (0 == xmlStrcasecmp (name, _xml("gnm:MaxCol")) &&
348 XML_READER_TYPE_END_ELEMENT == r->node_type)
350 r->state = STATE_SHEET_FOUND;
352 else if (r->node_type == XML_READER_TYPE_TEXT)
354 xmlChar *value = xmlTextReaderValue (r->xtr);
355 r->sheets[r->current_sheet].maxcol = _xmlchar_to_int (value);
359 case STATE_CELLS_START:
360 if (0 == xmlStrcasecmp (name, _xml ("gnm:Cell")) &&
361 XML_READER_TYPE_ELEMENT == r->node_type)
363 xmlChar *attr = NULL;
365 attr = xmlTextReaderGetAttribute (r->xtr, _xml ("Col"));
366 r->col = _xmlchar_to_int (attr);
369 if (r->col < r->min_col)
372 attr = xmlTextReaderGetAttribute (r->xtr, _xml ("Row"));
373 r->row = _xmlchar_to_int (attr);
376 if (r->sheets[r->current_sheet].start_row == -1)
378 r->sheets[r->current_sheet].start_row = r->row;
381 if (r->sheets[r->current_sheet].start_col == -1)
383 r->sheets[r->current_sheet].start_col = r->col;
385 if (! xmlTextReaderIsEmptyElement (r->xtr))
386 r->state = STATE_CELL;
388 else if ( (0 == xmlStrcasecmp (name, _xml("gnm:Cells"))) && (XML_READER_TYPE_END_ELEMENT == r->node_type) )
390 r->sheets[r->current_sheet].stop_col = r->col;
391 r->sheets[r->current_sheet].stop_row = r->row;
392 r->state = STATE_SHEET_NAME;
396 if (0 == xmlStrcasecmp (name, _xml("gnm:Cell")) && XML_READER_TYPE_END_ELEMENT == r->node_type)
398 r->state = STATE_CELLS_START;
410 Sets the VAR of case C, to the value corresponding to the xml string XV
413 convert_xml_string_to_value (struct ccase *c, const struct variable *var,
416 union value *v = case_data_rw (c, var);
419 value_set_missing (v, var_get_width (var));
420 else if ( var_is_alpha (var))
421 value_copy_str_rpad (v, var_get_width (var), xv, ' ');
424 const char *text = CHAR_CAST (const char *, xv);
428 v->f = c_strtod (text, &endptr);
429 if ( errno != 0 || endptr == text)
438 xmlChar *first_value;
443 gnumeric_error_handler (void *ctx, const char *mesg,
444 UNUSED xmlParserSeverities sev, xmlTextReaderLocatorPtr loc)
446 struct gnumeric_reader *r = ctx;
448 msg (MW, _("There was a problem whilst reading the %s file `%s' (near line %d): `%s'"),
450 r->spreadsheet.file_name,
451 xmlTextReaderLocatorLineNumber (loc),
455 static struct gnumeric_reader *
456 gnumeric_reopen (struct gnumeric_reader *r, const char *filename, bool show_errors)
460 xmlTextReaderPtr xtr;
463 assert (r == NULL || filename == NULL);
466 xmlFreeTextReader (r->xtr);
469 gz = gzopen (filename, "r");
471 gz = gzopen ( r->spreadsheet.file_name, "r");
477 xtr = xmlReaderForIO ((xmlInputReadCallback) gzread,
478 (xmlInputCloseCallback) gzclose, gz,
480 show_errors ? 0 : (XML_PARSE_NOERROR | XML_PARSE_NOWARNING) );
490 r = xzalloc (sizeof *r);
491 r->spreadsheet.n_sheets = -1;
492 r->spreadsheet.file_name = filename;
496 xmlTextReaderSetErrorHandler (xtr, gnumeric_error_handler, r);
498 r->target_sheet = NULL;
499 r->target_sheet_index = -1;
501 r->row = r->col = -1;
502 r->state = STATE_PRE_INIT;
506 /* Advance to the start of the workbook.
507 This gives us some confidence that we are actually dealing with a gnumeric
510 while ( (r->state != STATE_INIT )
511 && 1 == (ret = xmlTextReaderRead (r->xtr)))
519 /* Does not seem to be a gnumeric file */
520 xmlFreeTextReader (r->xtr);
525 r->spreadsheet.type = SPREADSHEET_GNUMERIC;
529 const xmlChar *enc = xmlTextReaderConstEncoding (r->xtr);
530 xmlCharEncoding xce = xmlParseCharEncoding (CHAR_CAST (const char *, enc));
532 if ( XML_CHAR_ENCODING_UTF8 != xce)
534 /* I have been told that ALL gnumeric files are UTF8 encoded. If that is correct, this
536 msg (MW, _("The gnumeric file `%s' is encoded as %s instead of the usual UTF-8 encoding. "
537 "Any non-ascii characters will be incorrectly imported."),
538 r->spreadsheet.file_name,
548 gnumeric_probe (const char *filename, bool report_errors)
550 struct gnumeric_reader *r = gnumeric_reopen (NULL, filename, report_errors);
552 return &r->spreadsheet;
557 gnumeric_make_reader (struct spreadsheet *spreadsheet,
558 const struct spreadsheet_read_options *opts)
561 struct gnumeric_reader *r = NULL;
562 unsigned long int vstart = 0;
564 casenumber n_cases = CASENUMBER_MAX;
566 struct var_spec *var_spec = NULL;
569 r = (struct gnumeric_reader *) (spreadsheet);
572 r = gnumeric_reopen (r, NULL, true);
576 if ( opts->cell_range )
578 if ( ! convert_cell_ref (opts->cell_range,
579 &r->start_col, &r->start_row,
580 &r->stop_col, &r->stop_row))
582 msg (SE, _("Invalid cell range `%s'"),
595 r->target_sheet = BAD_CAST opts->sheet_name;
596 r->target_sheet_index = opts->sheet_index;
597 r->row = r->col = -1;
598 r->current_sheet = -1;
600 /* Advance to the start of the cells for the target sheet */
601 while ( (r->state != STATE_CELL || r->row < r->start_row )
602 && 1 == (ret = xmlTextReaderRead (r->xtr)))
606 value = xmlTextReaderValue (r->xtr);
608 if ( r->state == STATE_MAXROW && r->node_type == XML_READER_TYPE_TEXT)
610 n_cases = 1 + _xmlchar_to_int (value) ;
615 /* If a range has been given, then use that to calculate the number
617 if ( opts->cell_range)
619 n_cases = MIN (n_cases, r->stop_row - r->start_row + 1);
622 if ( opts->read_names )
628 /* Read in the first row of cells,
629 including the headers if read_names was set */
631 (( r->state == STATE_CELLS_START && r->row <= r->start_row) || r->state == STATE_CELL )
632 && (ret = xmlTextReaderRead (r->xtr))
638 if ( r->row > r->start_row ) break;
640 if ( r->col < r->start_col ||
641 (r->stop_col != -1 && r->col > r->stop_col))
644 idx = r->col - r->start_col;
646 if ( idx >= n_var_specs )
649 var_spec = xrealloc (var_spec, sizeof (*var_spec) * (idx + 1));
650 for (i = n_var_specs; i <= idx; ++i)
652 var_spec [i].name = NULL;
653 var_spec [i].width = -1;
654 var_spec [i].first_value = NULL;
656 n_var_specs = idx + 1 ;
659 if ( r->node_type == XML_READER_TYPE_TEXT )
661 xmlChar *value = xmlTextReaderValue (r->xtr);
662 const char *text = CHAR_CAST (const char *, value);
664 if ( r->row < r->start_row)
666 if ( opts->read_names )
668 var_spec [idx].name = xstrdup (text);
673 var_spec [idx].first_value = xmlStrdup (value);
675 if (-1 == var_spec [idx].width )
676 var_spec [idx].width = (opts->asw == -1) ?
677 ROUND_UP (strlen(text), SPREADSHEET_DEFAULT_WIDTH) : opts->asw;
682 else if ( r->node_type == XML_READER_TYPE_ELEMENT
683 && r->state == STATE_CELL)
685 if ( r->row == r->start_row )
688 xmlTextReaderGetAttribute (r->xtr, _xml ("ValueType"));
690 if ( NULL == attr || 60 != _xmlchar_to_int (attr))
691 var_spec [idx].width = 0;
699 const xmlChar *enc = xmlTextReaderConstEncoding (r->xtr);
702 /* Create the dictionary and populate it */
703 spreadsheet->dict = r->dict = dict_create (CHAR_CAST (const char *, enc));
706 for (i = 0 ; i < n_var_specs ; ++i )
710 if ( (var_spec[i].name == NULL) && (var_spec[i].first_value == NULL))
713 /* Probably no data exists for this variable, so allocate a
715 if ( var_spec[i].width == -1 )
716 var_spec[i].width = SPREADSHEET_DEFAULT_WIDTH;
718 name = dict_make_unique_var_name (r->dict, var_spec[i].name, &vstart);
719 dict_create_var (r->dict, name, var_spec[i].width);
723 /* Create the first case, and cache it */
724 r->used_first_case = false;
726 if ( n_var_specs == 0 )
728 msg (MW, _("Selected sheet or range of spreadsheet `%s' is empty."),
729 spreadsheet->file_name);
733 r->proto = caseproto_ref (dict_get_proto (r->dict));
734 r->first_case = case_create (r->proto);
735 case_set_missing (r->first_case);
738 for ( i = 0 ; i < n_var_specs ; ++i )
740 const struct variable *var;
742 if ( (var_spec[i].name == NULL) && (var_spec[i].first_value == NULL))
745 var = dict_get_var (r->dict, x++);
747 convert_xml_string_to_value (r->first_case, var,
748 var_spec[i].first_value);
751 for ( i = 0 ; i < n_var_specs ; ++i )
753 free (var_spec[i].first_value);
754 free (var_spec[i].name);
760 return casereader_create_sequential
764 &gnm_file_casereader_class, r);
768 for ( i = 0 ; i < n_var_specs ; ++i )
770 free (var_spec[i].first_value);
771 free (var_spec[i].name);
775 dict_destroy (spreadsheet->dict);
776 spreadsheet->dict = NULL;
778 gnm_file_casereader_destroy (NULL, r);
784 /* Reads and returns one case from READER's file. Returns a null
785 pointer on failure. */
786 static struct ccase *
787 gnm_file_casereader_read (struct casereader *reader UNUSED, void *r_)
792 struct gnumeric_reader *r = r_;
793 int current_row = r->row;
795 if ( !r->used_first_case )
797 r->used_first_case = true;
798 return r->first_case;
801 c = case_create (r->proto);
802 case_set_missing (c);
804 if (r->start_col == -1)
805 r->start_col = r->min_col;
807 while ((r->state == STATE_CELL || r->state == STATE_CELLS_START )
808 && r->row == current_row && (ret = xmlTextReaderRead (r->xtr)))
812 if ( r->col < r->start_col || (r->stop_col != -1 &&
813 r->col > r->stop_col))
816 if ( r->col - r->start_col >= caseproto_get_n_widths (r->proto))
819 if ( r->stop_row != -1 && r->row > r->stop_row)
822 if ( r->node_type == XML_READER_TYPE_TEXT )
824 xmlChar *value = xmlTextReaderValue (r->xtr);
826 const int idx = r->col - r->start_col;
828 const struct variable *var = dict_get_var (r->dict, idx);
830 convert_xml_string_to_value (c, var, value);
847 #endif /* GNM_SUPPORT */