Added new "payee" and "capture" directives

The payee directive allows for transformation of payee names during the
parsing of a Ledger file.  This means you can record the payee name in
one form, but always have it reported in another form.  The syntax is
(and will be):

payee PAYEE_NAME  REGEXP

or

payee PAYEE_NAME
  REGEXP1
  REGEXP2
  ...

The account directive sets the account automatically based on the payee
iff the base account name is Unknown.  For example, if you have a bunch
of transaction in Expenses:Unknown, you can assign accounts for several
automatically using:

account ACCOUNT_NAME  PAYEE_REGEXP

or

account ACCOUNT_NAME
  PAYEE_REGEXP1
  PAYEE_REGEXP2
  ...

3 files changed, 62 insertions, 1 deletions

diff --git a/src/journal.h b/src/journal.h
index 2711c3b5..e54814aa 100644
--- a/src/journal.h
+++ b/src/journal.h
@@ -44,6 +44,7 @@
 
 #include "utils.h"
 #include "times.h"
+#include "mask.h"
 
 namespace ledger {
 
@@ -58,6 +59,11 @@ typedef std::list<xact_t *>	   xacts_list;
 typedef std::list<auto_xact_t *>   auto_xacts_list;
 typedef std::list<period_xact_t *> period_xacts_list;
 
+typedef std::pair<mask_t, string>      payee_mapping_t;
+typedef std::list<payee_mapping_t>     payee_mappings_t;
+typedef std::pair<mask_t, account_t *> account_mapping_t;
+typedef std::list<account_mapping_t>   account_mappings_t;
+
 class journal_t : public noncopyable
 {
 public:
@@ -110,6 +116,8 @@ public:
   period_xacts_list	period_xacts;
   std::list<fileinfo_t> sources;
   bool                  was_loaded;
+  payee_mappings_t      payee_mappings;
+  account_mappings_t    account_mappings;
 
   journal_t();
   journal_t(const path& pathname);
diff --git a/src/mask.h b/src/mask.h
index 8c20e554..755a2229 100644
--- a/src/mask.h
+++ b/src/mask.h
@@ -74,6 +74,9 @@ public:
 
   mask_t& operator=(const string& other);
 
+  bool operator<(const mask_t& other) const {
+    return expr < other.expr;
+  }
   bool operator==(const mask_t& other) const {
     return expr == other.expr;
   }
diff --git a/src/textual.cc b/src/textual.cc
index 9ad70c60..764d15e8 100644
--- a/src/textual.cc
+++ b/src/textual.cc
@@ -144,6 +144,8 @@ namespace {
     void end_directive(char * line);
     void alias_directive(char * line);
     void fixed_directive(char * line);
+    void payee_mapping_directive(char * line);
+    void account_mapping_directive(char * line);
     void tag_directive(char * line);
     void define_directive(char * line);
     bool general_directive(char * line);
@@ -735,6 +737,24 @@ void instance_t::fixed_directive(char * line)
   }
 }
 
+void instance_t::payee_mapping_directive(char * line)
+{
+  char * payee = skip_ws(line);
+  char * regex = next_element(payee, true);
+
+  context.journal.payee_mappings.push_back(payee_mapping_t(mask_t(regex), payee));
+}
+
+void instance_t::account_mapping_directive(char * line)
+{
+  char * account_name = skip_ws(line);
+  char * payee_regex  = next_element(account_name, true);
+
+  context.journal.account_mappings.push_back
+    (account_mapping_t(mask_t(payee_regex),
+		       context.top_account()->find_account(account_name)));
+}
+
 void instance_t::tag_directive(char * line)
 {
   string tag(trim_ws(line));
@@ -782,6 +802,13 @@ bool instance_t::general_directive(char * line)
     }
     break;
 
+  case 'c':
+    if (std::strcmp(p, "capture") == 0) {
+      account_mapping_directive(arg);
+      return true;
+    }
+    break;
+
   case 'd':
     if (std::strcmp(p, "def") == 0 || std::strcmp(p, "define") == 0) {
       define_directive(arg);
@@ -810,6 +837,13 @@ bool instance_t::general_directive(char * line)
     }
     break;
 
+  case 'p':
+    if (std::strcmp(p, "payee") == 0) {
+      payee_mapping_directive(arg);
+      return true;
+    }
+    break;
+
   case 't':
     if (std::strcmp(p, "tag") == 0) {
       tag_directive(arg);
@@ -929,6 +963,15 @@ post_t * instance_t::parse_post(char *		line,
     post->account->add_flags(ACCOUNT_KNOWN);
   }
 
+  if (post->account->name == _("Unknown")) {
+    foreach (account_mapping_t& value, context.journal.account_mappings) {
+      if (value.first.match(xact->payee)) {
+	post->account = value.second;
+	break;
+      }
+    }
+  }
+
   // Parse the optional amount
 
   bool saw_amount = false;
@@ -1242,7 +1285,14 @@ xact_t * instance_t::parse_xact(char *		line,
 
   if (next && *next) {
     char * p = next_element(next, true);
-    xact->payee = next;
+    foreach (payee_mapping_t& value, context.journal.payee_mappings) {
+      if (value.first.match(next)) {
+	xact->payee = value.second;
+	break;
+      }
+    }
+    if (xact->payee.empty())
+      xact->payee = next;
     next = p;
   } else {
     xact->payee = _("<Unspecified payee>");