/* Prevent channeling down into known open space. Author: Josh Cooper Created: Aug. 4 2020 Updated: Dec. 8 2022 */ /* This skeletal logic has not been kept up-to-date since ~v0.5 Enable plugin: -> build groups -> manage designations Unpause event: -> build groups -> manage designations Manage Designation(s): -> for each group in groups: -> does any tile in group have a group above -> Yes: set entire group to marker mode -> No: activate entire group (still checks is_safe_to_dig_down before activating each designation) Job started event: -> validate job type (channel) -> check pathing: -> Can: add job/worker to tracking -> Can: set tile to restricted -> Cannot: remove worker -> Cannot: insta-dig & delete job -> Cannot: set designation to Marker Mode (no insta-digging) OnUpdate: -> check worker location: -> CanFall: check if a fall would be safe: -> Safe: do nothing -> Unsafe: remove worker -> Unsafe: insta-dig & delete job (presumes the job is only accessible from directly on the tile) -> Unsafe: set designation to Marker Mode (no insta-digging) -> check tile occupancy: -> HasUnit: check if a fall would be safe: -> Safe: do nothing, let them fall -> Unsafe: remove worker for 1 tick (test if this "pauses" or cancels the job) -> Unsafe: Add feature to teleport unit? Job completed event: -> validate job type (channel) -> verify completion: -> IsOpenSpace: mark done -> IsOpenSpace: manage tile below -> NotOpenSpace: check for designation -> HasDesignation: do nothing -> NoDesignation: mark done (erases from group) -> NoDesignation: manage tile below */ #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include // Debugging namespace DFHack { DBG_DECLARE(channelsafely, plugin, DebugCategory::LINFO); DBG_DECLARE(channelsafely, monitor, DebugCategory::LERROR); DBG_DECLARE(channelsafely, manager, DebugCategory::LERROR); DBG_DECLARE(channelsafely, groups, DebugCategory::LERROR); DBG_DECLARE(channelsafely, jobs, DebugCategory::LERROR); } DFHACK_PLUGIN("channel-safely"); DFHACK_PLUGIN_IS_ENABLED(enabled); REQUIRE_GLOBAL(world); namespace EM = EventManager; using namespace DFHack; using namespace EM::EventType; int32_t mapx, mapy, mapz; Configuration config; PersistentDataItem psetting; PersistentDataItem pfeature; const std::string FCONFIG_KEY = std::string(plugin_name) + "/feature"; const std::string SCONFIG_KEY = std::string(plugin_name) + "/setting"; enum FeatureConfigData { VISION, MONITOR, RESURRECT, INSTADIG, RISKAVERSE }; enum SettingConfigData { REFRESH_RATE, MONITOR_RATE, IGNORE_THRESH, FALL_THRESH }; // dig-now.cpp df::coord simulate_fall(const df::coord &pos) { if unlikely(!Maps::isValidTilePos(pos)) { ERR(plugin).print("Error: simulate_fall(" COORD ") - invalid coordinate\n", COORDARGS(pos)); return {}; } df::coord resting_pos(pos); while (Maps::ensureTileBlock(resting_pos)) { df::tiletype tt = *Maps::getTileType(resting_pos); if (isWalkable(tt)) break; --resting_pos.z; } return resting_pos; } df::coord simulate_area_fall(const df::coord &pos) { df::coord neighbours[8]{}; get_neighbours(pos, neighbours); df::coord lowest = simulate_fall(pos); for (auto p : neighbours) { if unlikely(!Maps::isValidTilePos(p)) continue; auto nlow = simulate_fall(p); if (nlow.z < lowest.z) { lowest = nlow; } } return lowest; } namespace CSP { std::unordered_map endangered_units; std::unordered_map job_id_map; std::unordered_map active_jobs; std::unordered_map active_workers; std::unordered_map last_safe; std::unordered_set dignow_queue; void ClearData() { ChannelManager::Get().destroy_groups(); dignow_queue.clear(); last_safe.clear(); endangered_units.clear(); active_workers.clear(); active_jobs.clear(); job_id_map.clear(); } void SaveSettings() { if (pfeature.isValid() && psetting.isValid()) { try { pfeature.ival(MONITOR) = config.monitoring; pfeature.ival(VISION) = config.require_vision; pfeature.ival(INSTADIG) = false; //config.insta_dig; pfeature.ival(RESURRECT) = config.resurrect; pfeature.ival(RISKAVERSE) = config.riskaverse; psetting.ival(REFRESH_RATE) = config.refresh_freq; psetting.ival(MONITOR_RATE) = config.monitor_freq; psetting.ival(IGNORE_THRESH) = config.ignore_threshold; psetting.ival(FALL_THRESH) = config.fall_threshold; } catch (std::exception &e) { ERR(plugin).print("%s\n", e.what()); } } } void LoadSettings() { pfeature = World::GetPersistentData(FCONFIG_KEY); psetting = World::GetPersistentData(SCONFIG_KEY); if (!pfeature.isValid() || !psetting.isValid()) { pfeature = World::AddPersistentData(FCONFIG_KEY); psetting = World::AddPersistentData(SCONFIG_KEY); SaveSettings(); } else { try { config.monitoring = pfeature.ival(MONITOR); config.require_vision = pfeature.ival(VISION); config.insta_dig = false; //pfeature.ival(INSTADIG); config.resurrect = pfeature.ival(RESURRECT); config.riskaverse = pfeature.ival(RISKAVERSE); config.ignore_threshold = psetting.ival(IGNORE_THRESH); config.fall_threshold = psetting.ival(FALL_THRESH); config.refresh_freq = psetting.ival(REFRESH_RATE); config.monitor_freq = psetting.ival(MONITOR_RATE); } catch (std::exception &e) { ERR(plugin).print("%s\n", e.what()); } } active_workers.clear(); } void UnpauseEvent(bool full_scan = false){ CoreSuspender suspend; // we need exclusive access to df memory and this call stack doesn't already have a lock DEBUG(plugin).print("UnpauseEvent()\n"); ChannelManager::Get().build_groups(full_scan); ChannelManager::Get().manage_groups(); DEBUG(plugin).print("UnpauseEvent() exits\n"); } void JobStartedEvent(color_ostream &out, void* j) { if (enabled && World::isFortressMode() && Maps::IsValid()) { TRACE(jobs).print("JobStartedEvent()\n"); auto job = (df::job*) j; // validate job type if (ChannelManager::Get().exists(job->pos)) { DEBUG(jobs).print(" valid channel job:\n"); df::unit* worker = Job::getWorker(job); // there is a valid worker (living citizen) on the job? right.. if (worker && Units::isAlive(worker) && Units::isCitizen(worker)) { ChannelManager::Get().jobs.mark_active(job->pos); if (config.riskaverse) { if (ChannelManager::Get().jobs.possible_cavein(job->pos)) { cancel_job(job); ChannelManager::Get().manage_one(job->pos, true, true); } else { ChannelManager::Get().manage_group(job->pos, true, false); } } DEBUG(jobs).print(" valid worker:\n"); // track workers on jobs df::coord &pos = job->pos; TRACE(jobs).print(" -> Starting job at (" COORD ")\n", COORDARGS(pos)); if (config.monitoring || config.resurrect) { job_id_map.emplace(job, job->id); active_jobs.emplace(job->id, job); active_workers[job->id] = worker; if (config.resurrect) { // this is the only place we can be 100% sure of "safety" // (excluding deadly enemies that will have arrived) last_safe[worker->id] = worker->pos; } } // set tile to restricted TRACE(jobs).print(" setting job tile to restricted\n"); Maps::getTileDesignation(job->pos)->bits.traffic = df::tile_traffic::Restricted; } } TRACE(jobs).print(" <- JobStartedEvent() exits normally\n"); } } void JobCompletedEvent(color_ostream &out, void* j) { if (enabled && World::isFortressMode() && Maps::IsValid()) { TRACE(jobs).print("JobCompletedEvent()\n"); auto job = (df::job*) j; // we only care if the job is a channeling one if (ChannelManager::Get().exists(job->pos)) { ChannelManager::Get().manage_group(job->pos, true, false); // check job outcome auto block = Maps::getTileBlock(job->pos); df::coord local(job->pos); local.x = local.x % 16; local.y = local.y % 16; // verify completion if (TileCache::Get().hasChanged(job->pos, block->tiletype[Coord(local)])) { // the job can be considered done df::coord below(job->pos); below.z--; DEBUG(jobs).print(" -> (" COORD ") is marked done, managing group below.\n", COORDARGS(job->pos)); // mark done and manage below (and the rest of the group, if there were cavein candidates) block->designation[Coord(local)].bits.traffic = df::tile_traffic::Normal; ChannelManager::Get().mark_done(job->pos); ChannelManager::Get().manage_group(below); if (config.resurrect) { // this is the only place we can be 100% sure of "safety" // (excluding deadly enemies that may have arrived, and future digging) if (active_workers.count(job->id)) { df::unit* worker = active_workers[job->id]; last_safe[worker->id] = worker->pos; } } } // clean up auto jp = active_jobs[job->id]; job_id_map.erase(jp); active_workers.erase(job->id); active_jobs.erase(job->id); } TRACE(jobs).print("JobCompletedEvent() exits\n"); } } void NewReportEvent(color_ostream &out, void* r) { int32_t tick = df::global::world->frame_counter; auto report_id = (int32_t)(intptr_t(r)); if (df::global::world) { df::report* report = df::report::find(report_id); if (!report) { WARN(plugin).print("Error: NewReportEvent() received an invalid report_id - a report* cannot be found\n"); return; } switch (report->type) { case announcement_type::CANCEL_JOB: if (config.insta_dig) { if (report->text.find("cancels Dig") != std::string::npos || report->text.find("path") != std::string::npos) { dignow_queue.emplace(report->pos); } DEBUG(plugin).print("%d, pos: " COORD ", pos2: " COORD "\n%s\n", report_id, COORDARGS(report->pos), COORDARGS(report->pos2), report->text.c_str()); } break; case announcement_type::CAVE_COLLAPSE: if (config.resurrect) { DEBUG(plugin).print("CAVE IN\n%d, pos: " COORD ", pos2: " COORD "\n%s\n", report_id, COORDARGS(report->pos), COORDARGS(report->pos2), report->text.c_str()); df::coord below = report->pos; below.z -= 1; below = simulate_area_fall(below); df::coord areaMin{report->pos}; df::coord areaMax{areaMin}; areaMin.x -= 15; areaMin.y -= 15; areaMax.x += 15; areaMax.y += 15; areaMin.z = below.z; areaMax.z += 1; std::vector units; Units::getUnitsInBox(units, COORDARGS(areaMin), COORDARGS(areaMax)); for (auto unit: units) { endangered_units[unit] = tick; DEBUG(plugin).print(" [id %d] was near a cave in.\n", unit->id); } for (auto unit : world->units.all) { if (last_safe.count(unit->id)) { endangered_units[unit] = tick; DEBUG(plugin).print(" [id %d] is/was a worker, we'll track them too.\n", unit->id); } } } break; default: break; } } } void OnUpdate(color_ostream &out) { CoreSuspender suspend; if (enabled && World::isFortressMode() && Maps::IsValid() && !World::ReadPauseState()) { static int32_t last_tick = df::global::world->frame_counter; static int32_t last_monitor_tick = df::global::world->frame_counter; static int32_t last_refresh_tick = df::global::world->frame_counter; static int32_t last_resurrect_tick = df::global::world->frame_counter; int32_t tick = df::global::world->frame_counter; // Refreshing the group data with full scanning if (tick - last_refresh_tick >= config.refresh_freq) { last_refresh_tick = tick; TRACE(monitor).print("OnUpdate() refreshing now\n"); if (config.insta_dig) { TRACE(monitor).print(" -> evaluate dignow queue\n"); for (auto iter = dignow_queue.begin(); iter != dignow_queue.end();) { auto map_pos = *iter; dig_now(out, map_pos); // teleports units to the bottom of a simulated fall ChannelManager::Get().mark_done(map_pos); iter = dignow_queue.erase(iter); } } UnpauseEvent(false); TRACE(monitor).print("OnUpdate() refresh done\n"); } // Clean up stale df::job* if ((config.monitoring || config.resurrect) && tick - last_tick >= 1) { last_tick = tick; // make note of valid jobs std::unordered_map valid_jobs; for (df::job_list_link* link = &df::global::world->jobs.list; link != nullptr; link = link->next) { df::job* job = link->item; if (job && active_jobs.count(job->id)) { valid_jobs.emplace(job->id, job); } } // erase the active jobs that aren't valid std::unordered_set erase; map_value_difference(active_jobs, valid_jobs, erase); for (auto j : erase) { auto id = job_id_map[j]; job_id_map.erase(j); active_jobs.erase(id); active_workers.erase(id); } } // Monitoring Active and Resurrecting Dead if (config.monitoring && tick - last_monitor_tick >= config.monitor_freq) { last_monitor_tick = tick; TRACE(monitor).print("OnUpdate() monitoring now\n"); // iterate active jobs for (auto pair: active_jobs) { df::job* job = pair.second; df::unit* unit = active_workers[job->id]; if (!unit) continue; if (!Maps::isValidTilePos(job->pos)) continue; TRACE(monitor).print(" -> check for job in tracking\n"); if (Units::isAlive(unit)) { if (!config.monitoring) continue; TRACE(monitor).print(" -> compare positions of worker and job\n"); // check for fall safety if (unit->pos == job->pos && !is_safe_fall(job->pos)) { // unsafe WARN(monitor).print(" -> unsafe job\n"); Job::removeWorker(job); // decide to insta-dig or marker mode if (config.insta_dig) { // delete the job Job::removeJob(job); // queue digging the job instantly dignow_queue.emplace(job->pos); DEBUG(monitor).print(" -> insta-dig\n"); } else if (config.resurrect) { endangered_units.emplace(unit, tick); } else { // set marker mode Maps::getTileOccupancy(job->pos)->bits.dig_marked = true; // prevent algorithm from re-enabling designation for (auto &be: Maps::getBlock(job->pos)->block_events) { if (auto bsedp = virtual_cast( be)) { df::coord local(job->pos); local.x = local.x % 16; local.y = local.y % 16; bsedp->priority[Coord(local)] = config.ignore_threshold * 1000 + 1; break; } } DEBUG(monitor).print(" -> set marker mode\n"); } } } else if (config.resurrect) { resurrect(out, unit->id); if (last_safe.count(unit->id)) { df::coord lowest = simulate_fall(last_safe[unit->id]); Units::teleport(unit, lowest); } } } TRACE(monitor).print("OnUpdate() monitoring done\n"); } // Resurrect Dead Workers if (config.resurrect && tick - last_resurrect_tick >= 1) { last_resurrect_tick = tick; // clean up any "endangered" workers that have been tracked 100 ticks or more for (auto iter = endangered_units.begin(); iter != endangered_units.end();) { if (tick - iter->second >= 1200) { //keep watch 1 day DEBUG(plugin).print("It has been one day since [id %d]'s last incident.\n", iter->first->id); iter = endangered_units.erase(iter); continue; } ++iter; } // resurrect any dead units for (auto pair : endangered_units) { auto unit = pair.first; if (!Units::isAlive(unit)) { resurrect(out, unit->id); if (last_safe.count(unit->id)) { df::coord lowest = simulate_fall(last_safe[unit->id]); Units::teleport(unit, lowest); } } } } } } } command_result channel_safely(color_ostream &out, std::vector ¶meters); DFhackCExport command_result plugin_init(color_ostream &out, std::vector &commands) { commands.push_back(PluginCommand("channel-safely", "Automatically manage channel designations.", channel_safely, false)); return CR_OK; } DFhackCExport command_result plugin_shutdown(color_ostream &out) { EM::unregisterAll(plugin_self); return CR_OK; } DFhackCExport command_result plugin_load_data (color_ostream &out) { CSP::LoadSettings(); if (enabled) { std::vector params; channel_safely(out, params); } return DFHack::CR_OK; } DFhackCExport command_result plugin_enable(color_ostream &out, bool enable) { if (enable && !enabled) { // register events to check jobs / update tracking EM::EventHandler jobStartHandler(CSP::JobStartedEvent, 0); EM::EventHandler jobCompletionHandler(CSP::JobCompletedEvent, 0); EM::EventHandler reportHandler(CSP::NewReportEvent, 0); EM::registerListener(EventType::REPORT, reportHandler, plugin_self); EM::registerListener(EventType::JOB_STARTED, jobStartHandler, plugin_self); EM::registerListener(EventType::JOB_COMPLETED, jobCompletionHandler, plugin_self); // manage designations to start off (first time building groups [very important]) out.print("channel-safely: enabled!\n"); CSP::UnpauseEvent(true); } else if (!enable) { // don't need the groups if the plugin isn't going to be enabled EM::unregisterAll(plugin_self); out.print("channel-safely: disabled!\n"); } enabled = enable; return CR_OK; } DFhackCExport command_result plugin_onstatechange(color_ostream &out, state_change_event event) { switch (event) { case SC_UNPAUSED: if (enabled && World::isFortressMode() && Maps::IsValid()) { // manage all designations on unpause CSP::UnpauseEvent(true); } break; case SC_MAP_LOADED: // cache the map size Maps::getSize(mapx, mapy, mapz); CSP::ClearData(); ChannelManager::Get().build_groups(true); break; case SC_WORLD_UNLOADED: case SC_MAP_UNLOADED: CSP::ClearData(); break; default: return DFHack::CR_OK; } return DFHack::CR_OK; } DFhackCExport command_result plugin_onupdate(color_ostream &out, state_change_event event) { CSP::OnUpdate(out); return DFHack::CR_OK; } command_result channel_safely(color_ostream &out, std::vector ¶meters) { if (!parameters.empty()) { if (parameters[0] == "runonce") { CSP::UnpauseEvent(true); return DFHack::CR_OK; } else if (parameters[0] == "rebuild") { ChannelManager::Get().destroy_groups(); ChannelManager::Get().build_groups(true); } if (parameters.size() >= 2 && parameters.size() <= 3) { bool state = false; bool set = false; if (parameters[0] == "enable") { state = true; } else if (parameters[0] == "disable") { state = false; } else if (parameters[0] == "set") { set = true; } else { return DFHack::CR_WRONG_USAGE; } try { if(parameters[1] == "monitoring"){ if (state != config.monitoring) { config.monitoring = state; // if this is a fresh start if (state && !config.resurrect) { // we need a fresh start CSP::active_workers.clear(); } } } else if (parameters[1] == "risk-averse") { config.riskaverse = state; } else if (parameters[1] == "require-vision") { config.require_vision = state; } else if (parameters[1] == "insta-dig") { //config.insta_dig = state; config.insta_dig = false; } else if (parameters[1] == "resurrect") { if (state != config.resurrect) { config.resurrect = state; // if this is a fresh start if (state && !config.monitoring) { // we need a fresh start CSP::active_workers.clear(); } } } else if (parameters[1] == "refresh-freq" && set && parameters.size() == 3) { config.refresh_freq = std::abs(std::stol(parameters[2])); } else if (parameters[1] == "monitor-freq" && set && parameters.size() == 3) { config.monitor_freq = std::abs(std::stol(parameters[2])); } else if (parameters[1] == "ignore-threshold" && set && parameters.size() == 3) { config.ignore_threshold = std::abs(std::stol(parameters[2])); } else if (parameters[1] == "fall-threshold" && set && parameters.size() == 3) { uint8_t t = std::abs(std::stol(parameters[2])); if (t > 0) { config.fall_threshold = t; } else { out.printerr("fall-threshold must have a value greater than 0 or the plugin does a lot of nothing.\n"); return DFHack::CR_FAILURE; } } else { return DFHack::CR_WRONG_USAGE; } } catch (const std::exception &e) { out.printerr("%s\n", e.what()); return DFHack::CR_FAILURE; } } } else { out.print("Channel-Safely is %s\n", enabled ? "ENABLED." : "DISABLED."); out.print(" FEATURES:\n"); out.print(" %-20s\t%s\n", "risk-averse: ", config.riskaverse ? "on." : "off."); out.print(" %-20s\t%s\n", "monitoring: ", config.monitoring ? "on." : "off."); out.print(" %-20s\t%s\n", "require-vision: ", config.require_vision ? "on." : "off."); //out.print(" %-20s\t%s\n", "insta-dig: ", config.insta_dig ? "on." : "off."); out.print(" %-20s\t%s\n", "resurrect: ", config.resurrect ? "on." : "off."); out.print(" SETTINGS:\n"); out.print(" %-20s\t%" PRIi32 "\n", "refresh-freq: ", config.refresh_freq); out.print(" %-20s\t%" PRIi32 "\n", "monitor-freq: ", config.monitor_freq); out.print(" %-20s\t%" PRIu8 "\n", "ignore-threshold: ", config.ignore_threshold); out.print(" %-20s\t%" PRIu8 "\n", "fall-threshold: ", config.fall_threshold); } CSP::SaveSettings(); return DFHack::CR_OK; }