sots-engine/tests/mars_parse/canon.cpp

#include "canon.h"

#include <algorithm>
#include <cstdio>
#include <memory>
#include <utility>
#include <vector>

#include "mars/parse/value.h"

namespace canon {

namespace {

using mars::parse::Entry;
using mars::parse::Node;

// Minimal JSON value: enough to mirror the Python dict/list/scalar shapes.
struct Json {
    enum class Kind { Raw, String, Array, Object } kind = Kind::Raw;
    std::string raw;                                       // Raw: already-rendered literal
    std::string str;                                       // String
    std::vector<Json> items;                               // Array
    std::vector<std::pair<std::string, Json>> members;     // Object (insertion order)

    static Json rawlit(std::string s) { Json j; j.kind = Kind::Raw; j.raw = std::move(s); return j; }
    static Json string(std::string s) { Json j; j.kind = Kind::String; j.str = std::move(s); return j; }
    static Json array() { Json j; j.kind = Kind::Array; return j; }
    static Json object() { Json j; j.kind = Kind::Object; return j; }

    // Python's _add(): first value is scalar, the second turns it into a list.
    void add(const std::string& key, Json value) {
        for (auto& m : members) {
            if (m.first != key) continue;
            if (m.second.kind == Kind::Array && m.second.is_multi) {
                m.second.items.push_back(std::move(value));
            } else {
                Json list = array();
                list.is_multi = true;
                list.items.push_back(std::move(m.second));
                list.items.push_back(std::move(value));
                m.second = std::move(list);
            }
            return;
        }
        members.emplace_back(key, std::move(value));
    }
    bool is_multi = false;  // array created by repetition (vs. a value list)
};

void render(const Json& j, std::string& out) {
    switch (j.kind) {
    case Json::Kind::Raw: out += j.raw; break;
    case Json::Kind::String: out += json_string(j.str); break;
    case Json::Kind::Array:
        out += '[';
        for (std::size_t i = 0; i < j.items.size(); ++i) {
            if (i) out += ',';
            render(j.items[i], out);
        }
        out += ']';
        break;
    case Json::Kind::Object: {
        std::vector<const std::pair<std::string, Json>*> sorted;
        for (const auto& m : j.members) sorted.push_back(&m);
        std::sort(sorted.begin(), sorted.end(),
                  [](auto* a, auto* b) { return a->first < b->first; });
        out += '{';
        for (std::size_t i = 0; i < sorted.size(); ++i) {
            if (i) out += ',';
            out += json_string(sorted[i]->first);
            out += ':';
            render(sorted[i]->second, out);
        }
        out += '}';
        break;
    }
    }
}

// Typed rendering of a bareword, mirroring mars_data.coerce().
Json typed(const std::string& tok) {
    using mars::parse::ScalarKind;
    switch (mars::parse::classify(tok)) {
    case ScalarKind::Int:
        if (auto v = mars::parse::as_int(tok)) return Json::rawlit(std::to_string(*v));
        return Json::string(tok);  // out of 64-bit range: cannot match the oracle, surfaces as a diff
    case ScalarKind::Float: {
        char buf[64];
        std::snprintf(buf, sizeof buf, "%.17g", *mars::parse::as_double(tok));
        return Json::string(std::string("\x01") + buf);
    }
    case ScalarKind::Bool:
        return Json::rawlit(*mars::parse::as_bool(tok) ? "true" : "false");
    case ScalarKind::Text:
        break;
    }
    return Json::string(tok);
}

std::string lower(const std::string& s) {
    std::string out(s);
    for (char& c : out) c = static_cast<char>(mars::parse::ascii_lower(static_cast<unsigned char>(c)));
    return out;
}

Json node_to_json(const Node& n) {
    Json obj = Json::object();
    for (const Entry& e : n.entries) {
        switch (e.kind) {
        case Entry::Kind::Pair:
            obj.add(lower(e.key), e.quoted ? Json::string(e.value) : typed(e.value));
            break;
        case Entry::Kind::Block:
            obj.add(lower(e.key), node_to_json(*e.block));
            break;
        }
    }
    return obj;
}

Json group_to_json(const mars::parse::EffectGroup& g) {
    Json list = Json::array();
    for (const mars::parse::EffectEntry& e : g.entries) {
        Json pair = Json::array();
        pair.items.push_back(Json::string(e.key));
        if (e.is_group()) {
            pair.items.push_back(group_to_json(*e.group));
        } else if (e.values.size() == 1) {
            const auto& v = e.values[0];
            pair.items.push_back(v.quoted ? Json::string(v.text) : typed(v.text));
        } else {
            Json vals = Json::array();
            for (const auto& v : e.values)
                vals.items.push_back(v.quoted ? Json::string(v.text) : typed(v.text));
            pair.items.push_back(std::move(vals));
        }
        list.items.push_back(std::move(pair));
    }
    return list;
}

}  // namespace

std::string json_string(const std::string& s) {
    std::string out;
    out.reserve(s.size() + 2);
    out += '"';
    for (unsigned char c : s) {
        switch (c) {
        case '"': out += "\\\""; break;
        case '\\': out += "\\\\"; break;
        case '\n': out += "\\n"; break;
        case '\r': out += "\\r"; break;
        case '\t': out += "\\t"; break;
        case '\b': out += "\\b"; break;
        case '\f': out += "\\f"; break;
        default:
            if (c < 0x20 || c > 0x7e) {
                char buf[8];
                std::snprintf(buf, sizeof buf, "\\u%04x", c);
                out += buf;
            } else {
                out += static_cast<char>(c);
            }
        }
    }
    out += '"';
    return out;
}

std::string blocks_json(const Node& root) {
    std::string out;
    render(node_to_json(root), out);
    return out;
}

std::string effect_json(const mars::parse::EffectGroup& root) {
    std::string out;
    render(group_to_json(root), out);
    return out;
}

}  // namespace canon