path: root/src/NetUtils.cpp

#include "../include/NetUtils.hpp"
#include "../include/StringUtils.hpp"
#include "../generated/Tlds.hpp"
#include <array>
#include <sstream>
#include <iomanip>
#include <assert.h>
#include <string.h>
#include <unordered_set>

// TODO: Add all html sequences: https://html.spec.whatwg.org/multipage/named-characters.html

namespace QuickMedia {
    struct HtmlEscapeSequence {
        char unescape_char;
        std::string escape_sequence;
    };

    void html_escape_sequences(std::string &str) {
        const std::array<HtmlEscapeSequence, 6> escape_sequences = {
            HtmlEscapeSequence { '&', "&amp;" }, // This should be first, to not accidentally replace a new sequence caused by replacing this
            HtmlEscapeSequence { '"', "&quot;" },
            HtmlEscapeSequence { '\'', "&#39;" },
            HtmlEscapeSequence { '<', "&lt;" },
            HtmlEscapeSequence { '>', "&gt;" },
            HtmlEscapeSequence { '\n', "<br>" }
        };

        for(const HtmlEscapeSequence &escape_sequence : escape_sequences) {
            string_replace_all(str, escape_sequence.unescape_char, escape_sequence.escape_sequence);
        }
    }

    struct HtmlUnescapeSequence {
        std::string escape_sequence;
        std::string unescaped_str;
    };

    static void html_unescape_sequence_numbers(std::string &str) {
        size_t index = 0;
        while(true) {
            index = str.find("&#", index);
            if(index == std::string::npos)
                break;
            
            index += 2;
            size_t end_index = str.find(';', index);
            if(end_index != std::string::npos && end_index - index <= 3) {
                if(str[index] == 'x') {
                    ++index;
                    const size_t num_length = end_index - index;
                    int num;
                    if(to_num_hex(str.c_str() + index, num_length, num)) {
                        const char num_c = (char)num;
                        str.replace(index - 3, 3 + num_length + 1, &num_c, 1);
                        index += (-3 + 1);
                    }
                } else {
                    const size_t num_length = end_index - index;
                    int num;
                    if(to_num(str.c_str() + index, num_length, num)) {
                        const char num_c = (char)num;
                        str.replace(index - 2, 2 + num_length + 1, &num_c, 1);
                        index += (-2 + 1);
                    }
                }
            }
        }
    }

    void html_unescape_sequences(std::string &str) {
        html_unescape_sequence_numbers(str);

        // TODO: Use string find and find & and ; instead of string_replace_all
        const std::array<HtmlUnescapeSequence, 7> unescape_sequences = {
            HtmlUnescapeSequence { "&quot;", "\"" },
            HtmlUnescapeSequence { "&apos;", "'" },
            HtmlUnescapeSequence { "&lt;", "<" },
            HtmlUnescapeSequence { "&gt;", ">" },
            HtmlUnescapeSequence { "&mdash;", "—" },
            HtmlUnescapeSequence { "&nbsp;", " " },
            HtmlUnescapeSequence { "&amp;", "&" } // This should be last, to not accidentally replace a new sequence caused by replacing this
        };

        for(const HtmlUnescapeSequence &unescape_sequence : unescape_sequences) {
            string_replace_all(str, unescape_sequence.escape_sequence, unescape_sequence.unescaped_str);
        }
    }

    static bool is_alpha(char c) {
        return (c >= 'a' && c <= 'z') || (c >= 'A' && c <= 'Z');
    }

    static bool is_digit(char c) {
        return c >= '0' && c <= '9';
    }

    // Returns -1 if its not a hex value
    static int get_hex_value(char c) {
        if(c >= '0' && c <= '9')
            return c - '0';
        else if(c >= 'A' && c <= 'F')
            return 10 + (c - 'A');
        else if(c >= 'a' && c <= 'f')
            return 10 + (c - 'a');
        else
            return -1;
    }

    static bool is_whitespace(char c) {
        return c == ' ' || c == '\t' || c == '\n';
    }

    std::string url_param_encode(const std::string &param) {
        std::ostringstream result;
        result.fill('0');
        result << std::hex;

        for(char c : param) {
            if(is_alpha(c) || is_digit(c) || c == '-' || c == '_' || c == '.' || c == '~') {
                result << c;
            } else {
                result << std::uppercase;
                result << "%" << std::setw(2) << (int)(unsigned char)(c);
            }
        }

        return result.str();
    }

    std::string url_param_decode(const std::string &param) {
        std::string result;
        for(int i = 0; i < (int)param.size();) {
            char c = param[i];
            if(c == '%') {
                if(i < (int)param.size() - 1 && param[i + 1] == '%') {
                    result += c;
                    i += 2;
                } else if(i < (int)param.size() - 2) {
                    int first_c = get_hex_value(param[i + 1]);
                    int second_c = get_hex_value(param[i + 2]);
                    if(first_c != -1 && second_c != -1) {
                        result += (char)((first_c << 4) | second_c);
                        i += 3;
                    } else {
                        i += 1;
                    }
                } else {
                    i += 1;
                }
            } else if(c == '+') {
                result += ' ';
                i += 1;
            } else {
                result += c;
                i += 1;
            }
        }
        return result;
    }

    static bool is_url_start_char(char c) {
        return is_alpha(c) || is_digit(c) || c == '-' || c == '.' || c == '_' || c == '~';
    }

    static bool is_url_domain_char(char c) {
        return is_url_start_char(c);
    }

    // Implementation follows URI standard in general: https://tools.ietf.org/html/rfc3986#section-2.2.
    // Also checks for balanced parentheses to allow text such as: (see: example.com/) that excludes the last parenthesis.
    std::vector<Range> extract_urls(const std::string &str) {
        std::vector<Range> ranges;
        if(str.empty())
            return ranges;

        int parentheses_depth = 0;
        bool is_valid_url = false;
        bool is_domain_part = true;
        bool contains_dot = false;
        size_t url_start = std::string::npos;

        // str.size() is fine, we want to include the NULL character so we can extract url at the end of the string
        for(size_t i = 0; i < (size_t)str.size() + 1;) {
            char c = str[i];

            if(url_start != std::string::npos) {
                if(c == '(')
                    ++parentheses_depth;
                else if(c == ')')
                    --parentheses_depth;
            }

            if(url_start != std::string::npos && c == '.') {
                contains_dot = true;
            }

            if(url_start != std::string::npos && !is_valid_url && contains_dot && (is_whitespace(c) || c == '/' || c == ',' || c == ':' || c == '?' || c == ')' || c == '\0' || (c == '.' && i == str.size()))) {
                size_t tld_end = i - 1;
                char prev_char = str[i - 1];
                // We want to remove the last . or , because the string could contain for example "click on this link: example.com. There you can..."
                // and we want those links to work, I guess?
                if(prev_char == '.' || prev_char == ',' || prev_char == ':' || prev_char == '?')
                    --tld_end;
                else if(prev_char == ')' && parentheses_depth != 0)
                    --tld_end;
                size_t tld_start = tld_end;
                
                while(tld_start > url_start) {
                    if(str[tld_start] == '.')
                        break;
                    --tld_start;
                }

                if(tld_start > url_start && is_tld(str.substr(tld_start + 1, tld_end - tld_start))) {
                    is_valid_url = true;
                    is_domain_part = false;
                }
            }

            if(url_start != std::string::npos && is_domain_part && c == ':') {
                if(i + 2 < (size_t)str.size() + 1 && str[i + 1] == '/' && str[i + 2] == '/') {
                    i += 3;
                    continue;
                } else if(i + 1 < (size_t)str.size() + 1 && is_whitespace(str[i + 1])) {
                    i += 1;
                } else {
                    url_start = std::string::npos;
                    is_valid_url = false;
                    is_domain_part = true;
                    contains_dot = false;
                }
            }

            if(url_start != std::string::npos && is_domain_part && !is_url_domain_char(c)) {
                url_start = std::string::npos;
                is_valid_url = false;
                contains_dot = false;
            }

            if(url_start == std::string::npos && is_url_start_char(c)) {
                url_start = i;
            } else if(url_start != std::string::npos && (is_whitespace(c) || c == '\0')) {
                // Its only an url if there is a dot and the dot is not the last character in the url, for example "example.com" is an url but "example." is not.
                if(is_valid_url) {
                    size_t url_length = i - url_start;
                    char prev_char = str[i - 1];
                    // We want to remove the last . or , because the string could contain for example "click on this link: example.com. There you can..."
                    // and we want those links to work, I guess?
                    if(prev_char == '.' || prev_char == ',' || prev_char == ':' || prev_char == '?')
                        --url_length;
                    else if(prev_char == ')' && parentheses_depth != 0)
                        --url_length;
                    else if(prev_char < 32)
                        --url_length;
                    if(url_length > 0)
                        ranges.push_back({url_start, url_length});
                }

                url_start = std::string::npos;
                is_valid_url = false;
                is_domain_part = true;
                contains_dot = false;
            }

            ++i;
        }

        return ranges;
    }

    std::vector<std::string> ranges_get_strings(const std::string &str, const std::vector<Range> &ranges) {
        std::vector<std::string> strings(ranges.size());
        for(size_t i = 0; i < ranges.size(); ++i) {
            const Range &range = ranges[i];
            strings[i].assign(str.begin() + range.start, str.begin() + range.start + range.length);
        }
        return strings;
    }

    std::string header_extract_value(const std::string &header, const std::string &type) {
        std::string result;
        string_split(header, '\n', [&type, &result](const char *str, size_t size) {
            while(size > 0 && (*str == ' ' || *str == '\t')) { ++str; --size; }
            if(size < type.size() || !strncase_equals(str, type.c_str(), type.size()) || size == type.size())
                return true;

            str += type.size();
            size -= type.size();

            const void *colon_ptr = memchr(str, ':', size);
            if(!colon_ptr)
                return true;

            const size_t colon_offset = (const char*)colon_ptr - str;
            str += (colon_offset + 1);
            size -= (colon_offset + 1);

            // lstrip space
            while(size > 0 && (*str == ' ' || *str == '\t')) { ++str; --size; }
            // rstrip whitespace
            while(size > 0 && (str[size - 1] == ' ' || str[size - 1] == '\t' || str[size - 1] == '\r' || str[size - 1] == '\n')) { --size; }

            result.assign(str, size);
            return false;
        });
        return result;
    }
}