1 /* PSPP - a program for statistical analysis.
2 Copyright (C) 2007, 2009 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/>. */
21 #include <libpspp/message.h>
22 #include <libpspp/misc.h>
27 #define _(msgid) gettext (msgid)
28 #define N_(msgid) (msgid)
34 gnumeric_open_reader (struct gnumeric_read_info *gri, struct dictionary **dict)
36 msg (ME, _("Support for Gnumeric files was not compiled into this installation of PSPP"));
43 #include <data/casereader-provider.h>
45 #include <libpspp/str.h>
46 #include <libpspp/i18n.h>
47 #include <data/dictionary.h>
48 #include <data/variable.h>
52 #include <libxml/xmlreader.h>
56 #include <data/case.h>
57 #include <data/value.h>
59 #include "gnumeric-reader.h"
60 #include <data/identifier.h>
64 static void gnm_file_casereader_destroy (struct casereader *, void *);
66 static struct ccase *gnm_file_casereader_read (struct casereader *, void *);
68 static const struct casereader_class gnm_file_casereader_class =
70 gnm_file_casereader_read,
71 gnm_file_casereader_destroy,
76 /* Convert a string, which is an integer encoded in base26
77 IE, A=0, B=1, ... Z=25 to the integer it represents.
78 ... except that in this scheme, digits with an exponent
79 greater than 1 are implicitly incremented by 1, so
80 AA = 0 + 1*26, AB = 1 + 1*26,
81 ABC = 2 + 2*26 + 1*26^2 ....
84 pseudo_base26 (const char *str)
89 int len = strlen (str);
91 for ( i = len - 1 ; i >= 0; --i)
93 int mantissa = (str[i] - 'A');
95 if ( mantissa < 0 || mantissa > 25 )
101 result += mantissa * multiplier;
111 /* Convert a cell reference in the form "A1:B2", to
112 integers. A1 means column zero, row zero.
113 B1 means column 1 row 0. AA1 means column 26, row 0.
116 convert_cell_ref (const char *ref,
117 int *col0, int *row0,
118 int *coli, int *rowi)
126 int n = sscanf (ref, "%4[a-zA-Z]%d:%4[a-zA-Z]%d",
132 str_uppercase (startcol);
133 *col0 = pseudo_base26 (startcol);
134 str_uppercase (stopcol);
135 *coli = pseudo_base26 (stopcol);
136 *row0 = startrow - 1;
137 *rowi = stoprow - 1 ;
145 STATE_INIT = 0, /* Initial state */
146 STATE_SHEET_START, /* Found the start of a sheet */
147 STATE_SHEET_NAME, /* Found the sheet name */
149 STATE_SHEET_FOUND, /* Found the sheet that we actually want */
150 STATE_CELLS_START, /* Found the start of the cell array */
151 STATE_CELL /* Found a cell */
155 struct gnumeric_reader
157 xmlTextReaderPtr xtr;
159 enum reader_state state;
166 const xmlChar *target_sheet;
167 int target_sheet_index;
174 struct caseproto *proto;
175 struct dictionary *dict;
176 struct ccase *first_case;
177 bool used_first_case;
180 static void process_node (struct gnumeric_reader *r);
182 #define _xml(X) (const xmlChar *)(X)
184 #define _xmlchar_to_int(X) atoi((const char *)X)
187 gnm_file_casereader_destroy (struct casereader *reader UNUSED, void *r_)
189 struct gnumeric_reader *r = r_;
194 xmlFreeTextReader (r->xtr);
196 if ( ! r->used_first_case )
197 case_unref (r->first_case);
199 caseproto_unref (r->proto);
205 process_node (struct gnumeric_reader *r)
207 xmlChar *name = xmlTextReaderName (r->xtr);
209 name = xmlStrdup (_xml ("--"));
212 r->node_type = xmlTextReaderNodeType (r->xtr);
217 if (0 == xmlStrcasecmp (name, _xml("gnm:Sheet")) &&
218 XML_READER_TYPE_ELEMENT == r->node_type)
220 r->state = STATE_SHEET_START;
223 case STATE_SHEET_START:
224 if (0 == xmlStrcasecmp (name, _xml("gnm:Name")) &&
225 XML_READER_TYPE_ELEMENT == r->node_type)
227 r->state = STATE_SHEET_NAME;
229 else if (0 == xmlStrcasecmp (name, _xml("gnm:Name")) &&
230 XML_READER_TYPE_END_ELEMENT == r->node_type)
232 r->state = STATE_INIT;
235 case STATE_SHEET_NAME:
236 if (0 == xmlStrcasecmp (name, _xml("gnm:Name")) &&
237 XML_READER_TYPE_END_ELEMENT == r->node_type)
239 r->state = STATE_SHEET_START;
241 else if (XML_READER_TYPE_TEXT == r->node_type)
244 if ( r->target_sheet != NULL)
246 xmlChar *value = xmlTextReaderValue (r->xtr);
247 if ( 0 == xmlStrcmp (value, r->target_sheet))
248 r->state = STATE_SHEET_FOUND;
251 else if (r->target_sheet_index == r->sheet_index)
253 r->state = STATE_SHEET_FOUND;
257 case STATE_SHEET_FOUND:
258 if (0 == xmlStrcasecmp (name, _xml("gnm:Cells")) &&
259 XML_READER_TYPE_ELEMENT == r->node_type)
261 if (! xmlTextReaderIsEmptyElement (r->xtr))
262 r->state = STATE_CELLS_START;
264 else if (0 == xmlStrcasecmp (name, _xml("gnm:MaxRow")) &&
265 XML_READER_TYPE_ELEMENT == r->node_type)
267 r->state = STATE_MAXROW;
269 else if (0 == xmlStrcasecmp (name, _xml("gnm:Sheet")) &&
270 XML_READER_TYPE_END_ELEMENT == r->node_type)
272 r->state = STATE_INIT;
276 if (0 == xmlStrcasecmp (name, _xml("gnm:MaxRow")) &&
277 XML_READER_TYPE_END_ELEMENT == r->node_type)
279 r->state = STATE_SHEET_FOUND;
281 case STATE_CELLS_START:
282 if (0 == xmlStrcasecmp (name, _xml ("gnm:Cell")) &&
283 XML_READER_TYPE_ELEMENT == r->node_type)
285 xmlChar *attr = NULL;
286 r->state = STATE_CELL;
288 attr = xmlTextReaderGetAttribute (r->xtr, _xml ("Col"));
289 r->col = _xmlchar_to_int (attr);
292 attr = xmlTextReaderGetAttribute (r->xtr, _xml ("Row"));
293 r->row = _xmlchar_to_int (attr);
296 else if (0 == xmlStrcasecmp (name, _xml("gnm:Cells")) &&
297 XML_READER_TYPE_END_ELEMENT == r->node_type)
298 r->state = STATE_SHEET_NAME;
302 if (0 == xmlStrcasecmp (name, _xml("gnm:Cell")) &&
303 XML_READER_TYPE_END_ELEMENT == r->node_type)
304 r->state = STATE_CELLS_START;
315 Sets the VAR of case C, to the value corresponding to the xml string XV
318 convert_xml_string_to_value (struct ccase *c, const struct variable *var,
322 union value *v = case_data_rw (c, var);
324 const char *text = (const char *) xv;
327 n_bytes = MIN (var_get_width (var), strlen (text));
329 if ( var_is_alpha (var))
331 memcpy (value_str_rw (v, var_get_width (var)), text, n_bytes);
337 v->f = strtod (text, &endptr);
338 if ( errno != 0 || endptr == text)
347 xmlChar *first_value;
351 gnumeric_open_reader (struct gnumeric_read_info *gri, struct dictionary **dict)
353 unsigned long int vstart = 0;
355 casenumber n_cases = CASENUMBER_MAX;
357 struct var_spec *var_spec = NULL;
360 struct gnumeric_reader *r = NULL;
362 gzFile gz = gzopen (gri->file_name, "r");
366 msg (ME, _("Error opening \"%s\" for reading as a gnumeric file: %s."),
367 gri->file_name, strerror (errno));
372 r = xzalloc (sizeof *r);
374 r->xtr = xmlReaderForIO ((xmlInputReadCallback) gzread, gzclose, gz,
380 if ( gri->cell_range )
382 if ( ! convert_cell_ref (gri->cell_range,
383 &r->start_col, &r->start_row,
384 &r->stop_col, &r->stop_row))
386 msg (SE, _("Invalid cell range \"%s\""),
399 r->state = STATE_INIT;
400 r->target_sheet = BAD_CAST gri->sheet_name;
401 r->target_sheet_index = gri->sheet_index;
402 r->row = r->col = -1;
405 /* Advance to the start of the cells for the target sheet */
406 while ( (r->state != STATE_CELL || r->row < r->start_row )
407 && 1 == (ret = xmlTextReaderRead (r->xtr)))
411 value = xmlTextReaderValue (r->xtr);
413 if ( r->state == STATE_MAXROW && r->node_type == XML_READER_TYPE_TEXT)
415 n_cases = 1 + _xmlchar_to_int (value) ;
421 /* If a range has been given, then use that to calculate the number
423 if ( gri->cell_range)
425 n_cases = MIN (n_cases, r->stop_row - r->start_row + 1);
428 if ( gri->read_names )
434 /* Read in the first row of cells,
435 including the headers if read_names was set */
437 (( r->state == STATE_CELLS_START && r->row <= r->start_row) || r->state == STATE_CELL )
438 && (ret = xmlTextReaderRead (r->xtr))
444 if ( r->row > r->start_row ) break;
446 if ( r->col < r->start_col ||
447 (r->stop_col != -1 && r->col > r->stop_col))
450 idx = r->col - r->start_col;
452 if ( idx >= n_var_specs )
454 n_var_specs = idx + 1 ;
455 var_spec = xrealloc (var_spec, sizeof (*var_spec) * n_var_specs);
456 var_spec [idx].name = NULL;
457 var_spec [idx].width = -1;
458 var_spec [idx].first_value = NULL;
461 if ( r->node_type == XML_READER_TYPE_TEXT )
463 xmlChar *value = xmlTextReaderValue (r->xtr);
464 const char *text = (const char *) value;
466 if ( r->row < r->start_row)
468 if ( gri->read_names )
470 var_spec [idx].name = xstrdup (text);
475 var_spec [idx].first_value = xmlStrdup (value);
477 if (-1 == var_spec [idx].width )
478 var_spec [idx].width = (gri->asw == -1) ?
479 ROUND_UP (strlen(text), MAX_SHORT_STRING) : gri->asw;
484 else if ( r->node_type == XML_READER_TYPE_ELEMENT
485 && r->state == STATE_CELL)
487 if ( r->row == r->start_row )
490 xmlTextReaderGetAttribute (r->xtr, _xml ("ValueType"));
492 if ( NULL == attr || 60 != _xmlchar_to_int (attr))
493 var_spec [idx].width = 0;
501 /* Create the dictionary and populate it */
502 *dict = r->dict = dict_create ();
504 dict_set_encoding (r->dict, (const char *) xmlTextReaderConstEncoding (r->xtr));
506 for (i = 0 ; i < n_var_specs ; ++i )
508 char name[VAR_NAME_LEN + 1];
510 /* Probably no data exists for this variable, so allocate a
512 if ( var_spec[i].width == -1 )
513 var_spec[i].width = MAX_SHORT_STRING;
515 if ( ! dict_make_unique_var_name (r->dict, var_spec[i].name,
518 msg (ME, _("Cannot create variable name from %s"), var_spec[i].name);
522 dict_create_var (r->dict, name, var_spec[i].width);
525 /* Create the first case, and cache it */
526 r->used_first_case = false;
528 if ( n_var_specs == 0 )
530 msg (MW, _("Selected sheet or range of spreadsheet \"%s\" is empty."),
535 r->proto = caseproto_ref (dict_get_proto (r->dict));
536 r->first_case = case_create (r->proto);
537 case_set_missing (r->first_case);
539 for ( i = 0 ; i < n_var_specs ; ++i )
541 const struct variable *var = dict_get_var (r->dict, i);
543 convert_xml_string_to_value (r->first_case, var,
544 var_spec[i].first_value);
547 for ( i = 0 ; i < n_var_specs ; ++i )
549 free (var_spec[i].first_value);
550 free (var_spec[i].name);
555 return casereader_create_sequential
559 &gnm_file_casereader_class, r);
563 for ( i = 0 ; i < n_var_specs ; ++i )
565 free (var_spec[i].first_value);
566 free (var_spec[i].name);
570 dict_destroy (*dict);
572 gnm_file_casereader_destroy (NULL, r);
578 /* Reads and returns one case from READER's file. Returns a null
579 pointer on failure. */
580 static struct ccase *
581 gnm_file_casereader_read (struct casereader *reader UNUSED, void *r_)
586 struct gnumeric_reader *r = r_;
587 int current_row = r->row;
589 if ( !r->used_first_case )
591 r->used_first_case = true;
592 return r->first_case;
595 c = case_create (r->proto);
596 case_set_missing (c);
598 while ((r->state == STATE_CELL || r->state == STATE_CELLS_START )
599 && r->row == current_row && (ret = xmlTextReaderRead (r->xtr)))
603 if ( r->col < r->start_col || (r->stop_col != -1 &&
604 r->col > r->stop_col))
607 if ( r->col - r->start_col >= caseproto_get_n_widths (r->proto))
610 if ( r->stop_row != -1 && r->row > r->stop_row)
613 if ( r->node_type == XML_READER_TYPE_TEXT )
615 xmlChar *value = xmlTextReaderValue (r->xtr);
617 const int idx = r->col - r->start_col;
619 const struct variable *var = dict_get_var (r->dict, idx);
621 convert_xml_string_to_value (c, var, value);
638 #endif /* GNM_SUPPORT */