Skip to content

Commit

Permalink
* src/import-export/import-backend.c:
Browse files Browse the repository at this point in the history 
	* src/import-export/import-match-map.c:
	* src/import-export/import-match-map.h:
	  Chris Morgan's Baysian Matching code, to match transactions
	  based on Bayesian filtering of previously matched transactions.


git-svn-id: svn+ssh://svn.gnucash.org/repo/gnucash/trunk@8044 57a11ea4-9604-0410-9ed3-97b8803252fd
  • Loading branch information
@derekatkins
derekatkins committed Mar 8, 2003
1 parent 6e14383 commit b2ccbf6
Show file tree
Hide file tree
Showing 4 changed files with 552 additions and 24 deletions.
8 changes: 8 additions & 0 deletions ChangeLog
View file
@@ -1,3 +1,11 @@
2003-03-08 Derek Atkins <derek@ihtfp.com>

* src/import-export/import-backend.c:
* src/import-export/import-match-map.c:
* src/import-export/import-match-map.h:
Chris Morgan's Baysian Matching code, to match transactions
based on Bayesian filtering of previously matched transactions.

2003-03-06 Christian Stimming <stimming@tuhh.de>

* src/import-export/hbci/dialog-hbcitrans.c: Include a latest
Expand Down
169 changes: 150 additions & 19 deletions src/import-export/import-backend.c
View file
Expand Up @@ -44,6 +44,9 @@

#include "gnc-ui-util.h"

#define IMPORT_PAGE "Online Banking & Importing" /* from app-utils/prefs.scm */
#define BAYES_OPTION "Use Bayesian Matching?"

/********************************************************************\
* Constants *
\********************************************************************/
Expand Down Expand Up @@ -90,6 +93,9 @@ struct _transactioninfo
GNCImportAction action;
GNCImportAction previous_action;

/* A list of tokenized strings to use for bayesian matching purposes */
GList * match_tokens;

/* In case of a single destination account it is stored here. */
Account *dest_acc;
gboolean dest_acc_selected_manually;
Expand Down Expand Up @@ -241,6 +247,15 @@ void gnc_import_TransInfo_delete (GNCImportTransInfo *info)
xaccTransDestroy(info->trans);
xaccTransCommitEdit(info->trans);
}
if (info->match_tokens)
{
GList *node;

for (node = info->match_tokens; node; node = node->next)
g_free (node->data);

g_list_free (info->match_tokens);
}
g_free(info);
}
}
Expand Down Expand Up @@ -343,28 +358,128 @@ GdkPixmap* gen_probability_pixmap(gint score_original, GNCImportSettings *settin
* MatchMap- related functions (storing and retrieving)
*/

/* searches using the GNCImportTransInfo through all existing transactions */
/* if there is an exact match of the description and memo */
/* Tokenize a string and append to an existing GList(or an empty GList)
* the tokens
*/
static GList*
tokenize_string(GList* existing_tokens, const char *string)
{
char **tokenized_strings; /* array of strings returned by g_strsplit() */
char **stringpos;

tokenized_strings = g_strsplit(string, " ", 0);
stringpos = tokenized_strings;

/* add each token to the token GList */
while(stringpos && *stringpos)
{
/* prepend the char* to the token GList */
existing_tokens = g_list_prepend(existing_tokens, g_strdup(*stringpos));

/* then move to the next string */
stringpos++;
}

/* free up the strings that g_strsplit() created */
g_strfreev(tokenized_strings);

return existing_tokens;
}

/* create and return a list of tokens for a given transaction info. */
static GList*
TransactionGetTokens(GNCImportTransInfo *info)
{
Transaction* transaction;
GList* tokens;
const char* text;
time_t transtime;
struct tm *tm_struct;
char local_day_of_week[16];
Split* split;
int split_index;

g_return_val_if_fail (info, NULL);
if (info->match_tokens) return info->match_tokens;

transaction = gnc_import_TransInfo_get_trans(info);
g_assert(transaction);

tokens = 0; /* start off with an empty list */

/* make tokens from the transaction description */
text = xaccTransGetDescription(transaction);
tokens = tokenize_string(tokens, text);

/* the day of week the transaction occured is a good indicator of
* what account this transaction belongs in get the date and covert
* it to day of week as a token
*/
transtime = xaccTransGetDate(transaction);
tm_struct = gmtime(&transtime);
if(!strftime(local_day_of_week, sizeof(local_day_of_week), "%A", tm_struct))
{
PERR("TransactionGetTokens: error, strftime failed\n");
}

/* we cannot add a locally allocated string to this array, dup it so
* it frees the same way the rest do
*/
tokens = g_list_prepend(tokens, g_strdup(local_day_of_week));

/* make tokens from the memo of each split of this transaction */
split_index = 0;
while((split = xaccTransGetSplit(transaction, split_index)))
{
text = xaccSplitGetMemo(split);
tokens = tokenize_string(tokens, text);
split_index++; /* next split */
}

/* remember the list of tokens for later.. */
info->match_tokens = tokens;

/* return the pointer to the GList */
return tokens;
}

/* searches using the GNCImportTransInfo through all existing transactions
* if there is an exact match of the description and memo
*/
static Account *
matchmap_find_destination (GncImportMatchMap *matchmap,
GNCImportTransInfo *info)
matchmap_find_destination (GncImportMatchMap *matchmap, GNCImportTransInfo *info)
{
GncImportMatchMap *tmp_map;
Account *result;
GList* tokens;
gboolean useBayes;

g_assert (info);

tmp_map = ((matchmap != NULL) ? matchmap :
gnc_imap_create_from_account
(xaccSplitGetAccount
(gnc_import_TransInfo_get_fsplit (info))));

result = gnc_imap_find_account
(tmp_map, GNCIMPORT_DESC,
xaccTransGetDescription (gnc_import_TransInfo_get_trans (info)));
useBayes = gnc_lookup_boolean_option(IMPORT_PAGE, BAYES_OPTION, TRUE);
if(useBayes)
{
/* get the tokens for this transaction* */
tokens = TransactionGetTokens(info);

/* try to find the destination account for this transaction from its tokens */
result = gnc_imap_find_account_bayes(tmp_map, tokens);

} else {
/* old system of transaction to account matching */
result = gnc_imap_find_account
(tmp_map, GNCIMPORT_DESC,
xaccTransGetDescription (gnc_import_TransInfo_get_trans (info)));
}

/* Disable matching by memo, until bayesian filtering is implemented.
It's currently unlikely to help, and has adverse effects, causing false positives,
since very often the type of the transaction is stored there.
* It's currently unlikely to help, and has adverse effects,
* causing false positives, since very often the type of the
* transaction is stored there.
if (result == NULL)
result = gnc_imap_find_account
Expand All @@ -390,6 +505,9 @@ matchmap_store_destination (GncImportMatchMap *matchmap,
GncImportMatchMap *tmp_matchmap = NULL;
Account *dest;
const char *descr, *memo;
GList *tokens;
gboolean useBayes;

g_assert (trans_info);

/* This will store the destination account of the selected match if
Expand All @@ -410,20 +528,33 @@ matchmap_store_destination (GncImportMatchMap *matchmap,
(xaccSplitGetAccount
(gnc_import_TransInfo_get_fsplit (trans_info))));

descr = xaccTransGetDescription
(gnc_import_TransInfo_get_trans (trans_info));
if (descr && (strlen (descr) > 0))
gnc_imap_add_account (tmp_matchmap,
/* see what matching system we are currently using */
useBayes = gnc_lookup_boolean_option(IMPORT_PAGE, BAYES_OPTION, TRUE);
if(useBayes)
{
/* tokenize this transaction */
tokens = TransactionGetTokens(trans_info);

/* add the tokens to the imap with the given destination account */
gnc_imap_add_account_bayes(tmp_matchmap, tokens, dest);

} else {
/* old matching system */
descr = xaccTransGetDescription
(gnc_import_TransInfo_get_trans (trans_info));
if (descr && (strlen (descr) > 0))
gnc_imap_add_account (tmp_matchmap,
GNCIMPORT_DESC,
descr,
dest);
memo = xaccSplitGetMemo
(gnc_import_TransInfo_get_fsplit (trans_info));
if (memo && (strlen (memo) > 0))
gnc_imap_add_account (tmp_matchmap,
memo = xaccSplitGetMemo
(gnc_import_TransInfo_get_fsplit (trans_info));
if (memo && (strlen (memo) > 0))
gnc_imap_add_account (tmp_matchmap,
GNCIMPORT_MEMO,
memo,
dest);
} /* if(useBayes) */

if (matchmap == NULL)
gnc_imap_destroy (tmp_matchmap);
Expand Down Expand Up @@ -935,7 +1066,7 @@ gnc_import_TransInfo_refresh_destacc (GNCImportTransInfo *transaction_info,
/* if we haven't manually selected a destination account for this transaction */
if(gnc_import_TransInfo_get_destacc_selected_manually(transaction_info) == FALSE)
{
/* Try to find a previous selected destination account string match for the ADD action */
/* Try to find the destination account for this transaction based on prior ones */
new_destacc = matchmap_find_destination(matchmap, transaction_info);
gnc_import_TransInfo_set_destacc(transaction_info, new_destacc, FALSE);
} else
Expand Down

0 comments on commit b2ccbf6

Please sign in to comment.