Class Gemma4eArgsParser

src/common/AutoModel/modeling_gemma4e.cpp:33–356

--- Gemma4e tool-args parser ------------------------------------------------- The model emits relaxed JSON for tool arguments: - keys are bare identifiers (e.g. `name:`) — sometimes also "..." quoted - string values are delimited by <|"|>...<|"|>, but the model frequently omits the opener and/or replaces the closer with a plain `"` - values can also be objects {..}, arrays [..], booleans, null, o

//
// We rewrite the input into well-formed JSON via recursive-descent.
struct Gemma4eArgsParser {
    const std::string& s;
    size_t i = 0;
    static constexpr size_t marker_len = 5;
    static constexpr const char* quote_marker_lit = "<|\"|>";

    explicit Gemma4eArgsParser(const std::string& src) : s(src) {}

    void skip_ws() {
        while (i < s.size() &&
               std::isspace(static_cast<unsigned char>(s[i]))) ++i;
    }
    void skip_ws_at(size_t& pos) const {
        while (pos < s.size() &&
               std::isspace(static_cast<unsigned char>(s[pos]))) ++pos;
    }
    bool match_marker(size_t pos) const {
        return pos + marker_len <= s.size() &&
               s.compare(pos, marker_len, quote_marker_lit) == 0;
    }

    static void json_escape_char(std::string& out, char c) {
        switch (c) {
            case '"':  out.append("\\\""); break;
            case '\\': out.append("\\\\"); break;
            case '\n': out.append("\\n");  break;
            case '\r': out.append("\\r");  break;
            case '\t': out.append("\\t");  break;
            case '\b': out.append("\\b");  break;
            case '\f': out.append("\\f");  break;
            default:
                if (static_cast<unsigned char>(c) < 0x20) {
                    char buf[8];
                    std::snprintf(buf, sizeof(buf), "\\u%04x",
                                  static_cast<unsigned char>(c));
                    out.append(buf);
                } else {
                    out.push_back(c);
                }
                break;
        }
    }

    bool match_word(const char* w, size_t len) const {
        if (i + len > s.size()) return false;
        if (s.compare(i, len, w) != 0) return false;
        if (i + len == s.size()) return true;
        unsigned char nc = static_cast<unsigned char>(s[i + len]);
        return !(std::isalnum(nc) || nc == '_');
    }

    // Parse one JSON value. `terminators` is the set of chars that end a
    // bare (undelimited) string at depth 0 (e.g. ",}" inside an object,
    // ",]" inside an array, "" at the very top).
    std::string parse_value(const std::string& terminators) {
        skip_ws();
        if (i >= s.size()) return "null";
        char c = s[i];