/* bzflag
 * Copyright (c) 1993-2025 Tim Riker
 *
 * This package is free software;  you can redistribute it and/or
 * modify it under the terms of the license found in the file
 * named COPYING that should have accompanied this file.
 *
 * THIS PACKAGE IS PROVIDED ``AS IS'' AND WITHOUT ANY EXPRESS OR
 * IMPLIED WARRANTIES, INCLUDING, WITHOUT LIMITATION, THE IMPLIED
 * WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE.
 */

#ifndef __WORDFILTER_H__
#define __WORDFILTER_H__

#include "common.h"

/* system interface headers */
#include <string>
#include <set>
#include <algorithm>
#include <string.h>

#include <fstream>
#include <iostream>

#include <ctype.h>
#include <sys/types.h>

/* common interface headers */
#include "TextUtils.h"


/* words are stored by the index of their first letter of
 * UTF-8 format.  it would be nice to eventually support
 * full indexing of
 */
#define MAX_FILTER_SETS 256


/** WordFilter will load a list of words and phrases from a file or one at
 * a time or manually.
 *
 * Loading from file assumes that words/phrases are listed one per line;
 * comments are possible using the shell-style "#" delimiter.    Words are
 * matched case-insensitive; punctuation and non-newline white-space are
 * always ignored with the default filter.
 *
 * By default (aggressive filtering), the filter will match many
 * additional combinations to avoid the need to list all potential spelling
 * variations and hacks in the filter file.
 *
 * e.g. "dumb ass" should match "dumbass!", "DUMB a s s", "d_u_m_b! a_S_S", etc
 *
 * Strings should also match repetitive identical letter expansions.
 *
 * e.g. "dumb ass" should match "dumb asssss", "dumbaaaass", "dumb as s", etc
 *
 * Strings should also match l33t-speak. (l=1, o=0, e=3, a=@, s=z, i=l, f=ph)
 *
 * e.g. "ass whipe" should match "@sz wh1p3", etc
 *
 * Strings should also match common word suffixes (at least for English)
 *   (dom|ity|memt|sion|tion|ness|ance|ence|er|or|ist) for nouns
 *   (ive|en|ic|al|able|y|ous|ful|less) for adjectives
 *   (en|ize|ate|ify|fy|ed) for verbs
 *   (ly) for adverbs
 *   (a|z|r|ah|io|rs|rz|in|n|ster|meister) for slang
 *   (s|es) for plurality
 *   (ing|let) for imperfect verb, and diminutive
 *
 * e.g. "dumb ass" should match "dumb assness", "dumb asses", "dumb assly", etc
 *
 * Strings should also match common word prefixes (at least for English)
 *   (bz|beze) for bzflag-centric words
 *
 * e.g. "bitch" should also match "bzbitch", "beezzeebitch", etc
 *
 * Since all of the above matchings are done for free with aggressive matching,
 * only root words need to be provided.  For foreign languages, it may be
 * necessary to list all tenses of certain verbs, unless the rules can be
 * strictly and simply quantified.
 *
 * There is also a simple filter mode which is not as resource intensive and
 * performs a literal match with the filter words (so you have to specify
 * absolutely everything you want to filter and all variations).    It is still
 * case-insensitive and ignores punctuation.
 */
class WordFilter
{
public:

    /** structure for a single filter word, and a compiled regular expression
     */
    typedef struct filterStruct
    {
        std::string word;
        regex_t *compiled;
    } filter_t;


private:

    /** used by the simple filter */
    std::string alphabet;

    /** set of characters used to replace filtered content */
    std::string filterChars;

    /** word expressions to be filtered including compiled regexp versions */
    struct expressionCompare
    {
        bool operator() (const filter_t& word1, const filter_t& word2) const
        {
            return (strncasecmp(word1.word.c_str(), word2.word.c_str(), 1024) < 0);
        }
    };

    typedef std::set<filter_t, expressionCompare> ExpCompareSet;

    /** main collection of what to filter.  items are stored into
     * the array indexed by the first character of the filter word.
     * this means a sparse array, but it's a small price for
     * minimal hashing and rather fast lookups.
     */
    // XXX consider making a numeric hash to avoid array overflows
    ExpCompareSet filters[MAX_FILTER_SETS];


    /** used by the aggressive filter */
    ExpCompareSet suffixes;

    /** used by the aggressive filter */
    ExpCompareSet prefixes;


    /** utility method performs an actual replacement of
     * characters in an input character ray within a specified
     * range.
     */
    inline int filterCharacters(char *input, unsigned int start, size_t length, bool filterSpaces) const;

    /** utility method adds a letter to a string if it is not
     * already in the string
     */
    inline void appendUniqueChar(std::string& string, char c) const;

    /** utility method to add words to the prefix set
     */
    inline void addPrefix(const char *word);

    /** utility method to add words to the suffix set
     */
    inline void addSuffix(const char *word);


protected:

    /** This filter does a simple case-insensitive
     * word comparison that compares all filter
     * words with all alphabetic string sets in the
     * input string. If test is a filter word, then
     * input strings "test", "testy", and "test;"
     * will get filtered to "****", "testy", and
     * "****;" respectively.
     */
    bool simpleFilter(char *input) const;


    /** This filter will take a filter word and
     * create a rather complex regular expression
     * that catches a variety of variations on a
     * word.  Variations include automatic internal
     * expansion, optional punctuation and
     * whitespace, suffix matching (including
     * plurality), leet-speak conversion and more.
     * See the header above for more details.  This
     * filter should be the default.
     */
    bool aggressiveFilter(char *input) const;

    /** provides a pointer to a fresh compiled
     * expression for some given expression
     */
    regex_t *getCompiledExpression(const std::string &expression) const;

    /** returns a set of characters that represent the
     * given character in "l33t-speak"
     */
    std::string l33tspeakSetFromCharacter(const char c) const;

    /** returns what alphabetic character a given char
     * corresponds to (e.g. 3 => e, | => il)
     */
    std::string alphabeticSetFromCharacter(const char c) const;

    /** expands a word into an uncompiled regular
     *  expression.
     */
    std::string expressionFromString(const std::string &word) const;


public:

    WordFilter(void);
    WordFilter(const WordFilter& filter);
    ~WordFilter(void);

    /** loads a set of bad words from a specified file */
    unsigned int loadFromFile(const std::string &fileName, bool verbose=false);

    /** adds a new filter to the existing filter list */
    bool addToFilter(const std::string &word, const std::string &expression);

    /** given an input string, filter the input
     * using either the simple or aggressive filter
     */
    bool filter(char *input, const bool simple=false) const;
    bool filter(std::string &input, const bool simple=false) const;

    /** dump a list of words in the filter to stdout */
    void outputWords(void) const;
    /** dump the filter to stdout (including expressions) */
    void outputFilter(void) const;
    /** retuns a count of how many words are in the filter */
    unsigned long int wordCount(void) const;

    /** Clears the Filter */
    void clear(void);
};



/** utility method performs an actual replacement of
 * characters in an input character ray within a specified
 * range.
 */
inline int WordFilter::filterCharacters(char *input, unsigned int start, size_t length, bool filterSpaces=false) const
{
    if (input == NULL)
        return -1;
    if (length <= 0)
        return -1;
    if (strlen(input) < start)
        return 0;

    int randomCharPos, previousCharPos = -1;
    int maxFilterChar = (int)filterChars.size();
    int count=0;
    for (unsigned int j=0; j < (unsigned int)length; j++)
    {
        char c = input[start + j];

        // don't repeat random chars
        do
        {
            randomCharPos = (int)((float)maxFilterChar * (float)bzfrand());
        }
        while (randomCharPos == previousCharPos);
        previousCharPos = randomCharPos;

        /* when filterspaces is true, we filter everything.
          * otherise the ascii character code ranges for a-z, A-Z, and 0-9
          * are filtered.
          */
        if (filterSpaces)
        {
            input[start + j] = filterChars[randomCharPos];
            count++;
        }
        else if (TextUtils::isAlphanumeric(c))
        {
            input[start + j] = filterChars[randomCharPos];
            count++;
        } /* else it is non-letters so we can ignore */

    }
    return count;
}


inline void WordFilter::appendUniqueChar(std::string& string, char c) const
{
#ifdef HAVE_STD__COUNT
// ISO standard std::count
    if (std::count(string.begin(), string.end(), c) == 0)
        string += c;
#else
// old HP-style std::count (SunPRO for instance)
    int n = 0;
    std::count(string.begin(), string.end(), c, n);
    if (n == 0)
        string += c;
#endif
}


inline void WordFilter::addPrefix(const char *word)
{
    filter_t fix;
    std::pair<ExpCompareSet::iterator, bool> result;
    fix.word = std::string(word);
    fix.compiled = getCompiledExpression(expressionFromString(fix.word));
    result = prefixes.insert(fix);
    if (!result.second && fix.compiled)
    {
        regfree(fix.compiled);
        free(fix.compiled);
    }
}

inline void WordFilter::addSuffix(const char *word)
{
    filter_t fix;
    std::pair<ExpCompareSet::iterator, bool> result;
    fix.word = std::string(word);
    fix.compiled = getCompiledExpression(expressionFromString(fix.word));
    result = suffixes.insert(fix);
    if (!result.second && fix.compiled)
    {
        regfree(fix.compiled);
        free(fix.compiled);
    }
}


#else
class WordFilter;
#endif

// Local Variables: ***
// mode: C++ ***
// tab-width: 4 ***
// c-basic-offset: 4 ***
// indent-tabs-mode: nil ***
// End: ***
// ex: shiftwidth=4 tabstop=4