You cannot select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
fluffychat/lib/pangea/analytics_misc/construct_list_model.dart

400 lines
12 KiB
Dart

import 'dart:math';
import 'package:flutter/material.dart';
import 'package:collection/collection.dart';
import 'package:fluffychat/pangea/analytics_misc/analytics_constants.dart';
import 'package:fluffychat/pangea/analytics_misc/construct_type_enum.dart';
import 'package:fluffychat/pangea/analytics_misc/construct_use_model.dart';
import 'package:fluffychat/pangea/analytics_misc/constructs_model.dart';
import 'package:fluffychat/pangea/common/utils/error_handler.dart';
import 'package:fluffychat/pangea/constructs/construct_identifier.dart';
import 'package:fluffychat/pangea/morphs/get_grammar_copy.dart';
/// A wrapper around a list of [OneConstructUse]s, used to simplify
/// the process of filtering / sorting / displaying the events.
class ConstructListModel {
final List<OneConstructUse> _uses = [];
List<OneConstructUse> get uses => _uses;
List<OneConstructUse> get truncatedUses => _uses.take(100).toList();
/// A map of ConstructIdentifiers to ConstructUses, each of which contains a lemma
/// key = lemma + constructType.string, value = ConstructUses
final Map<String, ConstructUses> _constructMap = {};
/// Storing this to avoid re-running the sort operation each time this needs to
/// be accessed. It contains the same information as _constructMap, but sorted.
List<ConstructUses> _constructList = [];
/// A list of unique vocab lemmas
List<String> _vocabLemmasList = [];
/// A list of unique grammar lemmas
List<String> _grammarLemmasList = [];
/// [D] is the "compression factor". It determines how quickly
/// or slowly the level grows relative to XP
final double D = 1500;
List<ConstructIdentifier> unlockedLemmas(
ConstructTypeEnum type, {
int threshold = 0,
}) {
final constructs = constructList(type: type);
final List<ConstructIdentifier> unlocked = [];
final constructsList =
type == ConstructTypeEnum.vocab ? _vocabLemmasList : _grammarLemmasList;
for (final lemma in constructsList) {
final matches = constructs.where((m) => m.lemma == lemma);
final totalPoints = matches.fold<int>(
0,
(total, match) => total + match.points,
);
if (totalPoints > threshold) {
unlocked.add(matches.first.id);
}
}
return unlocked;
}
/// Analytics data consumed by widgets. Updated each time new analytics come in.
int prevXP = 0;
int totalXP = 0;
int level = 0;
ConstructListModel({
required List<OneConstructUse> uses,
int offset = 0,
}) {
updateConstructs(uses, offset);
}
int get totalLemmas => _vocabLemmasList.length + _grammarLemmasList.length;
int get vocabLemmas => _vocabLemmasList.length;
int get grammarLemmas => _grammarLemmasList.length;
List<String> get lemmasList => _vocabLemmasList + _grammarLemmasList;
/// Given a list of new construct uses, update the map of construct
/// IDs to ConstructUses and re-sort the list of ConstructUses
void updateConstructs(List<OneConstructUse> newUses, int offset) {
try {
_updateUsesList(newUses);
_updateConstructMap(newUses);
_updateConstructList();
_updateMetrics(offset);
} catch (err, s) {
ErrorHandler.logError(
e: "Failed to update analytics: $err",
s: s,
data: {
"newUses": newUses.map((e) => e.toJson()),
},
);
}
}
int _sortConstructs(ConstructUses a, ConstructUses b) {
final comp = b.points.compareTo(a.points);
if (comp != 0) return comp;
return a.lemma.compareTo(b.lemma);
}
void _updateUsesList(List<OneConstructUse> newUses) {
newUses.sort((a, b) => b.timeStamp.compareTo(a.timeStamp));
_uses.insertAll(0, newUses);
}
/// A map of lemmas to ConstructUses, each of which contains a lemma
/// key = lemmma + constructType.string, value = ConstructUses
void _updateConstructMap(final List<OneConstructUse> newUses) {
for (final use in newUses) {
final currentUses = _constructMap[use.identifier.string] ??
ConstructUses(
uses: [],
constructType: use.constructType,
lemma: use.lemma,
category: use.category,
);
currentUses.uses.add(use);
currentUses.setLastUsed(use.timeStamp);
_constructMap[use.identifier.string] = currentUses;
}
final broadKeys = _constructMap.keys.where((key) => key.endsWith('other'));
final replacedKeys = [];
for (final broadKey in broadKeys) {
final specificKeyPrefix = broadKey.split("-").first;
final specificKey = _constructMap.keys.firstWhereOrNull(
(key) =>
key != broadKey &&
key.startsWith(specificKeyPrefix) &&
!key.endsWith('other'),
);
if (specificKey == null) continue;
final broadConstructEntry = _constructMap[broadKey];
final specificConstructEntry = _constructMap[specificKey];
specificConstructEntry!.uses.addAll(broadConstructEntry!.uses);
_constructMap[specificKey] = specificConstructEntry;
replacedKeys.add(broadKey);
}
for (final key in replacedKeys) {
_constructMap.remove(key);
}
}
/// A list of ConstructUses, each of which contains a lemma and
/// a list of uses, sorted by the number of uses
void _updateConstructList() {
// TODO check how expensive this is
_constructList = _constructMap.values.toList();
_constructList.sort(_sortConstructs);
}
void _updateMetrics(int offset) {
_vocabLemmasList = constructList(type: ConstructTypeEnum.vocab)
.map((e) => e.lemma)
.toSet()
.toList();
_grammarLemmasList = constructList(type: ConstructTypeEnum.morph)
.map((e) => e.lemma)
.toSet()
.toList();
prevXP = totalXP;
totalXP = (_constructList.fold<int>(
0,
(total, construct) => total + construct.points,
)) +
offset;
if (totalXP < 0) {
totalXP = 0;
}
level = calculateLevelWithXp(totalXP);
}
int calculateLevelWithXp(int totalXP) {
final doubleScore = (1 + sqrt((1 + (8.0 * totalXP / D)) / 2.0));
if (!doubleScore.isNaN && doubleScore.isFinite) {
return doubleScore.floor();
} else {
ErrorHandler.logError(
e: "Calculated level in Nan or Infinity",
data: {
"totalXP": totalXP,
"prevXP": prevXP,
"level": doubleScore,
},
);
return 1;
}
}
int calculateXpWithLevel(int level) {
// If level <= 1, XP should be 0 or negative by this math.
// In practice, you might clamp it to 0:
if (level <= 1) {
return 0;
}
// Convert level to double for the math
final double lc = level.toDouble();
// XP from the inverse formula:
final double xpDouble = (D / 8.0) * (2.0 * pow(lc - 1.0, 2.0) - 1.0);
// Floor or clamp to ensure non-negative.
final int xp = xpDouble.floor();
return (xp < 0) ? 0 : xp;
}
// TODO; make this non-nullable, returning empty if not found
ConstructUses? getConstructUses(ConstructIdentifier identifier) {
final partialKey = "${identifier.lemma}-${identifier.type.string}";
if (_constructMap.containsKey(identifier.string)) {
// try to get construct use entry with full ID key
return _constructMap[identifier.string];
} else if (identifier.category == "other") {
// if the category passed to this function is "other", return the first
// construct use entry that starts with the partial key
return _constructMap.entries
.firstWhereOrNull((entry) => entry.key.startsWith(partialKey))
?.value;
} else {
// if the category passed to this function is not "other", return the first
// construct use entry that starts with the partial key and ends with "other"
return _constructMap.entries
.firstWhereOrNull(
(entry) =>
entry.key.startsWith(partialKey) && entry.key.endsWith("other"),
)
?.value;
}
}
List<ConstructUses> getConstructUsesByLemma(String lemma) {
return _constructList.where((constructUse) {
return constructUse.lemma == lemma;
}).toList();
}
List<ConstructUses> constructList({ConstructTypeEnum? type}) => _constructList
.where(
(constructUse) => type == null || constructUse.constructType == type,
)
.toList();
// uses where points < AnalyticConstants.xpForGreens
List<ConstructUses> get seeds => _constructList
.where(
(use) => use.points < AnalyticsConstants.xpForGreens,
)
.toList();
List<ConstructUses> get greens => _constructList
.where(
(use) =>
use.points >= AnalyticsConstants.xpForGreens &&
use.points < AnalyticsConstants.xpForFlower,
)
.toList();
List<ConstructUses> get flowers => _constructList
.where(
(use) => use.points >= AnalyticsConstants.xpForFlower,
)
.toList();
// Not storing this for now to reduce memory load
// It's only used by downloads, so doesn't need to be accessible on the fly
Map<String, List<ConstructUses>> lemmasToUses({
ConstructTypeEnum? type,
}) {
final Map<String, List<ConstructUses>> lemmasToUses = {};
final constructs = constructList(type: type);
for (final ConstructUses use in constructs) {
final lemma = use.lemma;
lemmasToUses.putIfAbsent(lemma, () => []);
lemmasToUses[lemma]!.add(use);
}
return lemmasToUses;
}
}
class LemmasToUsesWrapper {
final Map<String, List<ConstructUses>> lemmasToUses;
LemmasToUsesWrapper(this.lemmasToUses);
Map<String, List<OneConstructUse>> lemmasToFilteredUses(
bool Function(OneConstructUse) filter,
) {
final Map<String, List<OneConstructUse>> lemmasToOneConstructUses = {};
for (final entry in lemmasToUses.entries) {
final lemma = entry.key;
final uses = entry.value;
lemmasToOneConstructUses[lemma] =
uses.expand((use) => use.uses).toList().where(filter).toList();
}
return lemmasToOneConstructUses;
}
LemmasOverUnderList lemmasByPercent({
required bool Function(OneConstructUse) filter,
required double percent,
required BuildContext context,
}) {
final List<String> correctUseLemmas = [];
final List<String> incorrectUseLemmas = [];
final uses = lemmasToFilteredUses(filter);
for (final entry in uses.entries) {
if (entry.value.isEmpty) continue;
final List<OneConstructUse> correctUses = [];
final List<OneConstructUse> incorrectUses = [];
final lemma = getGrammarCopy(
category: entry.value.first.category,
lemma: entry.key,
context: context,
) ??
entry.key;
final uses = entry.value.toList();
for (final use in uses) {
use.xp > 0 ? correctUses.add(use) : incorrectUses.add(use);
}
final totalUses = correctUses.length + incorrectUses.length;
final percent = totalUses == 0 ? 0 : correctUses.length / totalUses;
percent > 0.8
? correctUseLemmas.add(lemma)
: incorrectUseLemmas.add(lemma);
}
return LemmasOverUnderList(
over: correctUseLemmas,
under: incorrectUseLemmas,
);
}
/// Return an object containing two lists, one of lemmas with
/// any correct uses and one of lemmas no correct uses
LemmasOverUnderList lemmasByCorrectUse({
String Function(ConstructUses)? getCopy,
}) {
final List<String> correctLemmas = [];
final List<String> incorrectLemmas = [];
for (final entry in lemmasToUses.entries) {
final lemma = entry.key;
final constructUses = entry.value;
final copy = getCopy?.call(constructUses.first) ?? lemma;
if (constructUses.any((use) => use.hasCorrectUse)) {
correctLemmas.add(copy);
} else {
incorrectLemmas.add(copy);
}
}
return LemmasOverUnderList(over: correctLemmas, under: incorrectLemmas);
}
int totalXP(String lemma) {
final uses = lemmasToUses[lemma];
if (uses == null) return 0;
if (uses.length == 1) return uses.first.points;
return lemmasToUses[lemma]!.fold<int>(
0,
(total, use) => total + use.points,
);
}
List<String> thresholdedLemmas({
required int start,
int? end,
String Function(ConstructUses)? getCopy,
}) {
final filteredList = lemmasToUses.entries.where((entry) {
final xp = totalXP(entry.key);
return xp >= start && (end == null || xp <= end);
});
return filteredList
.map((entry) => getCopy?.call(entry.value.first) ?? entry.key)
.toList();
}
}
class LemmasOverUnderList {
final List<String> over;
final List<String> under;
LemmasOverUnderList({
required this.over,
required this.under,
});
}