/* PSPP - a program for statistical analysis.
- Copyright (C) 2007, 2009, 2010, 2011, 2012 Free Software Foundation, Inc.
+ Copyright (C) 2007, 2009, 2010, 2011, 2012, 2013 Free Software Foundation, Inc.
This program is free software: you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by
#include "spreadsheet-reader.h"
-#include "c-xvasprintf.h"
-
#if !GNM_SUPPORT
struct casereader *
-gnumeric_open_reader (struct spreadsheet_read_info *gri, struct spreadsheet_read_options *opts, struct dictionary **dict)
+gnumeric_open_reader (const struct spreadsheet_read_options *opts, struct dictionary **dict)
{
msg (ME, _("Support for %s files was not compiled into this installation of PSPP"), "Gnumeric");
static struct ccase *gnm_file_casereader_read (struct casereader *, void *);
+
static const struct casereader_class gnm_file_casereader_class =
{
gnm_file_casereader_read,
STATE_SHEET_START, /* Found the start of a sheet */
STATE_SHEET_NAME, /* Found the sheet name */
STATE_MAXROW,
+ STATE_MAXCOL,
STATE_SHEET_FOUND, /* Found the sheet that we actually want */
STATE_CELLS_START, /* Found the start of the cell array */
STATE_CELL /* Found a cell */
};
+struct sheet_detail
+{
+ /* The name of the sheet (utf8 encoding) */
+ char *name;
+
+ int start_col;
+ int stop_col;
+ int start_row;
+ int stop_row;
+
+ int maxcol;
+ int maxrow;
+};
+
struct gnumeric_reader
{
struct spreadsheet spreadsheet;
+ /* The libxml reader for this instance */
xmlTextReaderPtr xtr;
+ /* An internal state variable */
enum reader_state state;
int row;
int col;
int min_col;
int node_type;
- int sheet_index;
+ int current_sheet;
+
+ int start_col;
+ int stop_col;
+ int start_row;
+ int stop_row;
- xmlChar **sheet_names;
+ struct sheet_detail *sheets;
const xmlChar *target_sheet;
int target_sheet_index;
- int start_row;
- int start_col;
- int stop_row;
- int stop_col;
-
struct caseproto *proto;
struct dictionary *dict;
struct ccase *first_case;
};
-const char *
-gnumeric_get_sheet_name (struct spreadsheet *s, int n)
+void
+gnumeric_destroy (struct spreadsheet *s)
{
- struct gnumeric_reader *gr = (struct gnumeric_reader *) s;
- assert (n < s->sheets);
+ struct gnumeric_reader *r = s;
+ int i;
+
+ for (i = 0; i < s->n_sheets; ++i)
+ {
+ xmlFree (r->sheets[i].name);
+ }
+
+ free (r->sheets);
- return gr->sheet_names[n];
+ free (r);
}
+
const char *
-gnumeric_get_sheet_range (struct spreadsheet *s, int n)
+gnumeric_get_sheet_name (struct spreadsheet *s, int n)
{
struct gnumeric_reader *gr = (struct gnumeric_reader *) s;
- assert (n < s->sheets);
+ assert (n < s->n_sheets);
- return "I havent the fogiest idea";
+ return gr->sheets[n].name;
}
static void process_node (struct gnumeric_reader *r);
+
+char *
+gnumeric_get_sheet_range (struct spreadsheet *s, int n)
+{
+ int ret;
+ struct gnumeric_reader *gr = (struct gnumeric_reader *) s;
+
+ assert (n < s->n_sheets);
+
+ while (
+ (gr->sheets[n].stop_col == -1)
+ &&
+ (1 == (ret = xmlTextReaderRead (gr->xtr)))
+ )
+ {
+ process_node (gr);
+ }
+
+ return create_cell_ref (
+ gr->sheets[n].start_col,
+ gr->sheets[n].start_row,
+ gr->sheets[n].stop_col,
+ gr->sheets[n].stop_row);
+}
+
+
static void
gnm_file_casereader_destroy (struct casereader *reader UNUSED, void *r_)
{
- int i;
struct gnumeric_reader *r = r_;
if ( r == NULL)
return ;
if ( r->xtr)
xmlFreeTextReader (r->xtr);
+ r->xtr = NULL;
if ( ! r->used_first_case )
case_unref (r->first_case);
caseproto_unref (r->proto);
-
- for (i = 0; i < r->spreadsheet.sheets; ++i)
- {
- xmlFree (r->sheet_names[i]);
- }
-
- free (r->sheet_names);
-
- free (r);
}
+
static void
process_node (struct gnumeric_reader *r)
{
if (name == NULL)
name = xmlStrdup (_xml ("--"));
-
r->node_type = xmlTextReaderNodeType (r->xtr);
- switch ( r->state)
+ switch (r->state)
{
case STATE_PRE_INIT:
+ r->current_sheet = -1;
if (0 == xmlStrcasecmp (name, _xml("gnm:SheetNameIndex")) &&
XML_READER_TYPE_ELEMENT == r->node_type)
{
r->state = STATE_SHEET_COUNT;
- r->spreadsheet.sheets = 0;
}
break;
if (0 == xmlStrcasecmp (name, _xml("gnm:SheetName")) &&
XML_READER_TYPE_ELEMENT == r->node_type)
{
- r->spreadsheet.sheets++;
- r->sheet_names = xrealloc (r->sheet_names, r->spreadsheet.sheets * sizeof *r->sheet_names);
+ ++r->current_sheet;
+ if (r->current_sheet + 1 > r->spreadsheet.n_sheets)
+ {
+ struct sheet_detail *sd ;
+ r->sheets = xrealloc (r->sheets, (r->current_sheet + 1) * sizeof *r->sheets);
+ sd = &r->sheets[r->current_sheet];
+ sd->start_col = sd->stop_col = sd->start_row = sd->stop_row = -1;
+ r->spreadsheet.n_sheets = r->current_sheet + 1;
+ }
}
else if (0 == xmlStrcasecmp (name, _xml("gnm:SheetNameIndex")) &&
XML_READER_TYPE_END_ELEMENT == r->node_type)
{
r->state = STATE_INIT;
+ r->current_sheet = -1;
}
else if (XML_READER_TYPE_TEXT == r->node_type)
{
- r->sheet_names [r->spreadsheet.sheets - 1] = xmlTextReaderValue (r->xtr);
+ r->sheets [r->spreadsheet.n_sheets - 1].name = CHAR_CAST (char *, xmlTextReaderValue (r->xtr));
}
break;
if (0 == xmlStrcasecmp (name, _xml("gnm:Sheet")) &&
XML_READER_TYPE_ELEMENT == r->node_type)
{
- ++r->sheet_index;
+ ++r->current_sheet;
r->state = STATE_SHEET_START;
}
break;
{
r->state = STATE_INIT;
}
+ else if (0 == xmlStrcasecmp (name, _xml("gnm:Sheet")) &&
+ XML_READER_TYPE_END_ELEMENT == r->node_type)
+ {
+ r->state = STATE_INIT;
+ }
else if (XML_READER_TYPE_TEXT == r->node_type)
{
- if ( r->target_sheet != NULL)
+ if ( r->target_sheet != NULL)
{
xmlChar *value = xmlTextReaderValue (r->xtr);
if ( 0 == xmlStrcmp (value, r->target_sheet))
r->state = STATE_SHEET_FOUND;
free (value);
}
- else if (r->target_sheet_index == r->sheet_index)
+ else if (r->target_sheet_index == r->current_sheet + 1)
+ {
+ r->state = STATE_SHEET_FOUND;
+ }
+ else if (r->target_sheet_index == -1)
{
r->state = STATE_SHEET_FOUND;
}
{
r->state = STATE_MAXROW;
}
+ else if (0 == xmlStrcasecmp (name, _xml("gnm:MaxCol")) &&
+ XML_READER_TYPE_ELEMENT == r->node_type)
+ {
+ r->state = STATE_MAXCOL;
+ }
else if (0 == xmlStrcasecmp (name, _xml("gnm:Sheet")) &&
XML_READER_TYPE_END_ELEMENT == r->node_type)
{
- r->state = STATE_INIT;
+ r->state = STATE_INIT;
}
break;
case STATE_MAXROW:
{
r->state = STATE_SHEET_FOUND;
}
+ else if (r->node_type == XML_READER_TYPE_TEXT)
+ {
+ xmlChar *value = xmlTextReaderValue (r->xtr);
+ r->sheets[r->current_sheet].maxrow = _xmlchar_to_int (value);
+ xmlFree (value);
+ }
+ break;
+ case STATE_MAXCOL:
+ if (0 == xmlStrcasecmp (name, _xml("gnm:MaxCol")) &&
+ XML_READER_TYPE_END_ELEMENT == r->node_type)
+ {
+ r->state = STATE_SHEET_FOUND;
+ }
+ else if (r->node_type == XML_READER_TYPE_TEXT)
+ {
+ xmlChar *value = xmlTextReaderValue (r->xtr);
+ r->sheets[r->current_sheet].maxcol = _xmlchar_to_int (value);
+ xmlFree (value);
+ }
+ break;
case STATE_CELLS_START:
if (0 == xmlStrcasecmp (name, _xml ("gnm:Cell")) &&
XML_READER_TYPE_ELEMENT == r->node_type)
{
xmlChar *attr = NULL;
- r->state = STATE_CELL;
attr = xmlTextReaderGetAttribute (r->xtr, _xml ("Col"));
r->col = _xmlchar_to_int (attr);
attr = xmlTextReaderGetAttribute (r->xtr, _xml ("Row"));
r->row = _xmlchar_to_int (attr);
free (attr);
- }
- else if (0 == xmlStrcasecmp (name, _xml("gnm:Cells")) &&
- XML_READER_TYPE_END_ELEMENT == r->node_type)
- r->state = STATE_SHEET_NAME;
+ if (r->sheets[r->current_sheet].start_row == -1)
+ {
+ r->sheets[r->current_sheet].start_row = r->row;
+ }
+
+ if (r->sheets[r->current_sheet].start_col == -1)
+ {
+ r->sheets[r->current_sheet].start_col = r->col;
+ }
+ if (! xmlTextReaderIsEmptyElement (r->xtr))
+ r->state = STATE_CELL;
+ }
+ else if ( (0 == xmlStrcasecmp (name, _xml("gnm:Cells"))) && (XML_READER_TYPE_END_ELEMENT == r->node_type) )
+ {
+ r->sheets[r->current_sheet].stop_col = r->col;
+ r->sheets[r->current_sheet].stop_row = r->row;
+ r->state = STATE_SHEET_NAME;
+ }
break;
case STATE_CELL:
- if (0 == xmlStrcasecmp (name, _xml("gnm:Cell")) &&
- XML_READER_TYPE_END_ELEMENT == r->node_type)
+ if (0 == xmlStrcasecmp (name, _xml("gnm:Cell")) && XML_READER_TYPE_END_ELEMENT == r->node_type)
{
r->state = STATE_CELLS_START;
}
};
-void
-gnumeric_destroy (struct spreadsheet *s)
+static void
+gnumeric_error_handler (void *ctx, const char *mesg,
+ UNUSED xmlParserSeverities sev, xmlTextReaderLocatorPtr loc)
{
- struct gnumeric_reader *r = s;
- gnm_file_casereader_destroy (NULL, s);
+ struct gnumeric_reader *r = ctx;
+
+ msg (MW, _("There was a problem whilst reading the %s file `%s' (near line %d): `%s'"),
+ "Gnumeric",
+ r->spreadsheet.file_name,
+ xmlTextReaderLocatorLineNumber (loc),
+ mesg);
}
-struct spreadsheet *
-gnumeric_probe (const char *filename)
-{
+static struct gnumeric_reader *
+gnumeric_reopen (struct gnumeric_reader *r, const char *filename, bool show_errors)
+{
int ret;
- struct gnumeric_reader *r = NULL;
+
xmlTextReaderPtr xtr;
+ gzFile gz;
+
+ assert (r == NULL || filename == NULL);
+
+ if (r && r->xtr)
+ xmlFreeTextReader (r->xtr);
- gzFile gz = gzopen (filename, "r");
+ if (filename)
+ gz = gzopen (filename, "r");
+ else
+ gz = gzopen ( r->spreadsheet.file_name, "r");
if (NULL == gz)
return NULL;
+
xtr = xmlReaderForIO ((xmlInputReadCallback) gzread,
- (xmlInputCloseCallback) gzclose, gz,
- NULL, NULL, 0);
+ (xmlInputCloseCallback) gzclose, gz,
+ NULL, NULL,
+ show_errors ? 0 : (XML_PARSE_NOERROR | XML_PARSE_NOWARNING) );
if (xtr == NULL)
- return NULL;
+ {
+ gzclose (gz);
+ return NULL;
+ }
- r = xzalloc (sizeof *r);
+ if (r == NULL)
+ {
+ r = xzalloc (sizeof *r);
+ r->spreadsheet.n_sheets = -1;
+ r->spreadsheet.file_name = filename;
+ }
- r->xtr = xtr;
- r->spreadsheet.sheets = -1;
- r->state = STATE_PRE_INIT;
+ if (show_errors)
+ xmlTextReaderSetErrorHandler (xtr, gnumeric_error_handler, r);
+ r->target_sheet = NULL;
+ r->target_sheet_index = -1;
+
+ r->row = r->col = -1;
+ r->state = STATE_PRE_INIT;
+ r->xtr = xtr;
/* Advance to the start of the workbook.
This gives us some confidence that we are actually dealing with a gnumeric
process_node (r);
}
- if (ret != 1)
+
+ if ( ret != 1)
{
- /* Not a gnumeric spreadsheet */
+ /* Does not seem to be a gnumeric file */
+ xmlFreeTextReader (r->xtr);
free (r);
- gzclose (gz);
return NULL;
}
-
+
r->spreadsheet.type = SPREADSHEET_GNUMERIC;
- r->spreadsheet.file_name = filename;
-
+
+ if (show_errors)
+ {
+ const xmlChar *enc = xmlTextReaderConstEncoding (r->xtr);
+ xmlCharEncoding xce = xmlParseCharEncoding (CHAR_CAST (const char *, enc));
+
+ if ( XML_CHAR_ENCODING_UTF8 != xce)
+ {
+ /* I have been told that ALL gnumeric files are UTF8 encoded. If that is correct, this
+ can never happen. */
+ msg (MW, _("The gnumeric file `%s' is encoded as %s instead of the usual UTF-8 encoding. "
+ "Any non-ascii characters will be incorrectly imported."),
+ r->spreadsheet.file_name,
+ enc);
+ }
+ }
+
+ return r;
+}
+
+
+struct spreadsheet *
+gnumeric_probe (const char *filename, bool report_errors)
+{
+ struct gnumeric_reader *r = gnumeric_reopen (NULL, filename, report_errors);
+
return &r->spreadsheet;
}
struct casereader *
gnumeric_make_reader (struct spreadsheet *spreadsheet,
- const struct spreadsheet_read_info *gri,
- struct spreadsheet_read_options *opts)
+ const struct spreadsheet_read_options *opts)
{
+ int x = 0;
struct gnumeric_reader *r = NULL;
unsigned long int vstart = 0;
int ret;
r = (struct gnumeric_reader *) (spreadsheet);
+ if (r->row != -1)
+ r = gnumeric_reopen (r, NULL, true);
+
if ( opts->cell_range )
{
if ( ! convert_cell_ref (opts->cell_range,
r->stop_row = -1;
}
-
r->target_sheet = BAD_CAST opts->sheet_name;
r->target_sheet_index = opts->sheet_index;
r->row = r->col = -1;
- r->sheet_index = 0;
+ r->current_sheet = -1;
/* Advance to the start of the cells for the target sheet */
while ( (r->state != STATE_CELL || r->row < r->start_row )
free (value);
}
-
/* If a range has been given, then use that to calculate the number
of cases */
if ( opts->cell_range)
n_cases = MIN (n_cases, r->stop_row - r->start_row + 1);
}
- if ( gri->read_names )
+ if ( opts->read_names )
{
r->start_row++;
n_cases --;
if ( r->row < r->start_row)
{
- if ( gri->read_names )
+ if ( opts->read_names )
{
var_spec [idx].name = xstrdup (text);
}
var_spec [idx].first_value = xmlStrdup (value);
if (-1 == var_spec [idx].width )
- var_spec [idx].width = (gri->asw == -1) ?
- ROUND_UP (strlen(text), SPREADSHEET_DEFAULT_WIDTH) : gri->asw;
+ var_spec [idx].width = (opts->asw == -1) ?
+ ROUND_UP (strlen(text), SPREADSHEET_DEFAULT_WIDTH) : opts->asw;
}
free (value);
r->first_case = case_create (r->proto);
case_set_missing (r->first_case);
- int x = 0;
+
for ( i = 0 ; i < n_var_specs ; ++i )
{
+ const struct variable *var;
+
if ( (var_spec[i].name == NULL) && (var_spec[i].first_value == NULL))
continue;
- const struct variable *var = dict_get_var (r->dict, x++);
+ var = dict_get_var (r->dict, x++);
convert_xml_string_to_value (r->first_case, var,
var_spec[i].first_value);
free (var_spec);
-
- if (opts->cell_range == NULL)
- {
- opts->cell_range = c_xasprintf ("%c%d:%c%ld",
- r->start_col + 'A',
- r->start_row,
- r->stop_col + 'A' + caseproto_get_n_widths (r->proto),
- r->start_row + n_cases);
- }
-
+
return casereader_create_sequential
(NULL,
r->proto,
projects / pspp / blobdiff