Update of /cvsroot/mahogany/M/include
In directory sc8-pr-cvs1.sourceforge.net:/tmp/cvs-serv25163/include
Modified Files:
SpamFilter.h
Log Message:
1. renamed dspam.cpp to DspamFilter.cpp
2. moved built in heuristic headers-based spam filter to HeadersFilter
3. added SpamFilter::Train() and implemented it for dspam
4. renamed SpamFilter::ReclassifyAsSpam() to Reclassify(), added
DECLARE_SPAM_FILTER() macro and changed CheckIfSpam() to return a
string explaining why the message was deemed to be a spam
Index: SpamFilter.h
===================================================================
RCS file: /cvsroot/mahogany/M/include/SpamFilter.h,v
retrieving revision 1.1
retrieving revision 1.2
diff -b -u -2 -r1.1 -r1.2
--- SpamFilter.h 8 Jul 2004 15:22:23 -0000 1.1
+++ SpamFilter.h 10 Jul 2004 20:26:45 -0000 1.2
@@ -39,16 +39,16 @@
Mark the message as spam.
- This method is used for training the spam filter and also for correcting
- its mistakes.
+ This method is used for correcting spam filter mistakes, see Train() for
+ training the spam filter.
@param msg the spam message
*/
static void ClassifyAsSpam(const Message& msg)
- { ReclassifyAsSpam(msg, true); }
+ { Reclassify(msg, true); }
/**
Mark the message as non-spam.
- Unlike ClassifyAsSpam() this is normally only used to correct mistakes,
+ Like ClassifyAsSpam() this is normally only used to correct mistakes,
i.e. for reclassifying a false positive.
@@ -56,15 +56,32 @@
*/
static void ClassifyAsHam(const Message& msg)
- { ReclassifyAsSpam(msg, false); }
+ { Reclassify(msg, false); }
/**
Reclassify the message as either spam or ham.
- Common part of ClassifyAsSpam/Ham()
+ Common part of ClassifyAsSpam/Ham().
+
+ Note that this should only be used if the spam filter had accidentally
+ detected that the message was a spam, not to train it.
+
@param msg the message to reclassify
@param isSpam if true, reclassify as spam, otherwise as ham
*/
- static void ReclassifyAsSpam(const Message& msg, bool isSpam);
+ static void Reclassify(const Message& msg, bool isSpam);
+
+ /**
+ Train the spam filter using the specified spam or ham message.
+
+ Unlike Reclassify() which should only be used for the messages
+ which had been mistakenly misclassified by the filter, this method should
+ be used for the messages which had never been seen by the spam filter
+ before.
+
+ @param msg the message to train the filter with
+ @param isSpam if true, this as a spam, otherwise -- ham
+ */
+ static void Train(const Message& msg, bool isSpam);
/**
@@ -72,9 +89,15 @@
@param msg the message to check
- @param probability if non NULL, return the probability of this message
- being a spam, may be NULL if caller not interested
+ @param params array containing the stringified options for the filters
+ (each element is of the form "name=options" where "name" is
+ the name of the filter and options depend on filter)
+ @param result place holder for the message explaining why the message was
+ deemed to be spam, if the filter can tell it, ignored if
+ the return value is false
@return true if the message is deemed to be a spam, false otherwise
*/
- static bool CheckIfSpam(const Message& msg, float *probability = NULL);
+ static bool CheckIfSpam(const Message& msg,
+ const wxArrayString& params,
+ String *result);
/**
@@ -84,5 +107,5 @@
memory used by the spam filters but can also be called during the program
execution to force reloading of the spam filters during the next call to
- ReclassifyAsSpam() or CheckIfSpam().
+ Reclassify() or CheckIfSpam().
*/
static void UnloadAll();
@@ -97,5 +120,5 @@
Loads all available spam filters.
- This is called as needed by ReclassifyAsSpam()/CheckIfSpam(), no need to
+ This is called as needed by Reclassify()/CheckIfSpam(), no need to
call it explicitly.
*/
@@ -116,9 +139,35 @@
//@{
- /// Reclassify a message as either spam or ham
- virtual void Reclassify(const Message& msg, bool isSpam) = 0;
+ /**
+ Reclassify a message as either spam or ham.
+
+ This is used by the public Reclassify().
+ */
+ virtual void DoReclassify(const Message& msg, bool isSpam) = 0;
- /// Process a message and return whether it is a spam
- virtual bool Process(const Message& msg, float *probability) = 0;
+ /**
+ Train filter using this message.
+
+ This is used by the public Train().
+ */
+ virtual void DoTrain(const Message& msg, bool isSpam) = 0;
+
+ /**
+ Process a message and return whether it is a spam.
+
+ This is used by the public CheckIfSpam(). The parameter passed to it is
+ the RHS of the element in the params array passed to CheckIfSpam() having
+ the same name as this spam filter.
+ */
+ virtual bool DoCheckIfSpam(const Message& msg,
+ const String& param,
+ String *result) = 0;
+
+ /**
+ Return the internal name of this spam filter.
+
+ This method is implemented by DECLARE_SPAM_FILTER() macro below.
+ */
+ virtual const wxChar *GetName() const = 0;
//@}
@@ -157,4 +206,15 @@
/**
+ This macro must be used in SpamFilter-derived class declaration.
+
+ It defines GetName() which returns the parameter of the macro.
+
+ There should be no semicolon after this macro.
+ */
+#define DECLARE_SPAM_FILTER(name) \
+ public: \
+ virtual const wxChar *GetName() const { return _T(name); }
+
+/**
This macro must be used in the .cpp file containing the implementation of a
spam filter.
-------------------------------------------------------
This SF.Net email sponsored by Black Hat Briefings & Training.
Attend Black Hat Briefings & Training, Las Vegas July 24-29 -
digital self defense, top technical experts, no vendor pitches,
unmatched networking opportunities. Visit www.blackhat.com
_______________________________________________
Mahogany-cvsupdates mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/mahogany-cvsupdates
