* update lemma meaning and phonetic transcription repos * chore: simplify progress bar widget * Remove instructions from chat view, and add profile explanation to course participant page * Translate courseParticipantTooltip * fix: in course chats list, sort activities by activity ID * use different text in chat/course participant tooltips * depress disabled toolbar buttons * fix: load course images on course load * fix: on add course plan to space, set m.space.child power level to 0 * chore: add label to emoji selector in vocab analytics * chore: increase text sizes in activity summary * fix: don't show open sessions if user has selected a role * feat: add button to regenerate latest bot message * chore: update morph meaning repo * chore: increase text size and spacing in language selection page, consume language locale emojis * feat: on first select lemma emoji, show snackbar with explanation * chore: use builder to style pressable buttons based on height * chore: add tooltips to each practice mode * initial work to add shimmer to match activity options * show word card in image toolbar mode * use the same widget for word card and vocab details emoji pickers * add shimmer background to match choices * fix: close previous snackbar before opening new mode disabled snackbar * fix: refresh course details when course ID changes in course details * chore: keep message button depressed * only show emoji selection shimmer if no emoji is selected * don't show reaction picker in emoji mode * lemma emoji picker style updates * update loading indicators in word zoom card * feat: show word card in vocab details page * practice buttons shimmer * fixed height audio player * more practice mode updates * more practice tweaks * add space between rows of tokens in practice mode * conditional top padding for practice tooltips * feat: send message info in lemma info request * chore: Focus on word meanings in reaction choices * fix: restrict width of morph icon in practice token button * chore: Expand word card for meanings * chore: When first grammar question active, shimmer choices * chore: Swap seed for hyphen for not-yet-chosen emojis in analytics * chore: Level attention to emoji and audio icons * fix: fix non-token vertical spacing in practice mode * fix: close message overlay when screen size changes * feat: While audio is playing, allow clicking of word to move audio to that spot * feat: play audio on token click and on construct click in vocab analytics * chore: snackbar close button * feat: Stay in audio mode after end of audio * chore: more word card spacing adjustments * fix: use construct id json in route for analytics details page * feat: custom SSO login/signup dialog * chore: add content to distinguish system edit from manual edit * Make input bar suggestion text vertically centered when shrinking * Add Pangea comments * Add background to make dark mode icon stand out in own message grammar practice * chore: re-style sso popup * fix: progress bar min width * fix: change how screen width metric changes are tracked * simplify * fix: fix carousel scroll issue * fix: set emoji placeholder text colot * fix: when not in column mode, don't add padding to top of practice tooltip * chore: prevent running IGC on empty message * fix: allow translation of bot audio transcripts * feat: analytics database * fix: update analytics profile room IDs on change, set via parameter in analytics room knock request (#4949) * chore: center title in add a course page (#4950) * fix: update spacing of activity participant indicators to make them narrower, make user activity summary highlight row scrollable (#4951) * fix: remove clicked new token from new tokens cache immeadiatley instead of waiting for new token animation to finish (#4952) * What now button takes user to top of course plan page (#4946) * Add scrollController to course details pages * Make what now button refresh details tab if needed, remove scrollController * 4907 construct details changes (#4961) * chore: remove delegation analytics page * feat: vocab construct analytics level bar * chore: analytics mobile navigation * feat: cap construct XP * Add background to regeneration request background (#4960) * chore: reduce padding between lines of message in practice mode (#4962) * chore: don't show message regeneration button if message has already been regenerated (#4963) * fix: prevent request regeneration button from altering message height (#4964) * fix: only animate top portion of activity status bar (#4965) * fix: fix white box error and add opacity variation to construct levels in progress bar (#4966) * fix: don't close word card on click (#4967) * feat: after user exits IT three times, show them a popup with the option to disable automatic language assistance (#4968) * feat: allow token feedback for word card in vocab analytics (#4900) * feat: allow token feedback for word card in vocab analytics * fix: remove duplicate global keys * 4726 word card in arabic goes way to the side (#4970) * fix: initial work for word card positioning on RTL system * fix: fix practice mode animation for RTL languages * chore: close lemma emoji snackbar on parent widget disposed (#4972) * fix: remove user summary testing code (#4974) * feat: On hover of the Nav Bar, expand to show current icon tooltip text (#4976) * feat: On hover of the Nav Bar, expand to show current icon tooltip text * animate menu transition * chore: delete construct navigation (#4984) * chore: Use hyphen instead of seed/sprout/flower in list view (#4985) * chore: update analytics page on construct update (#4987) * fix: fix word card overlay in mobile vocab details page (#4988) * fix: Latest sent message sinks when clicked on Mobile (#4989) * fix: don't highlight new tokens until analytics initialize (#4990) * chore: calculate times closed out of IT based on all message in session (#4991) * chore: add feedback response dialog (#4992) * chore: move request generation button into message bubble (#4993) * fix: show request regen button in overlay message (#4996) * fix: separate block construct and update construct updates in vocab list view (#4998) * feat: Do gold shimmer every 5 seconds on unselected emojis (#4999) * simplify message token renderer (#4994) * simplify message token renderer * token rendering and new word collection for tokens in activity summary / menu * make tokens hoverable * Model key cleanup (#4983) * refactor: Group redundant ModelKey entries * Add python script to find and replace hardcoded ModelKey values * Edited Python script to not automatically use ModelKey for files not already using it * refactor: Ran script and accepted obvious changes * rename 'duration' model key --------- Co-authored-by: ggurdin <ggurdin@gmail.com> * fix: return bot local stt, ensure stt rep exists in request stt translation function (#5003) * chore: set max lines for word card phonetic transcription (#5005) * chore: Don't show shimmer for unavailable modes (#5006) * chore: Delay until screen darkening (#5009) * chore: add focus node to vocab list view search bar (#5011) * chore: collapse navigation rail on navigate (#5013) * When user saves course edits, return to details page (#5012) * fix: don't lowercase construct keys in morph analytics list view (#5014) * 4860 dms all chats (#5015) * feat: initial work for dms => all chats * more navigation updates * change all chats tooltip * fix: set exact reactions length in overlay (#5016) * fix: fix message list rendering (#5017) * chore: disable lemma emoji selection for word card in token feedback dialog (#5026) * fix: don't add XP update if no new construct uses were added (#5027) * chore: hide request regeneration button in practice mode (#5028) * chore: use root navigator for chat details dialogs (#5029) * fix: rebuild word card on new word overlay dismissed (#5030) * Ensure consistency of pressable button height after animation (#5025) * Ensure consistency of pressable button height after animation * Use variable instead of hardcoded value * fix: fix overlay reactions bouncing around (#5031) * fix: add horizontal padding to prevent choice animation cutoff (#5032) * 4919 further optimizing message info (#5033) * remove original sent from message content * don't add null fields to message content JSON * fix: only show disable language assistance popup is user manually closes IT (#5034) * fix: only exclude xp gained analytics events if blocked constructs has entry (#5035) * fix: on analytics DB init, don't clear DB unless stored userID doesn't match client userID (#5036) * don't log missing POS error for POS 'other' (#5039) * don't long missing POS error for POS 'other' * don't long error for missing grammar copy if lemma is 'other' * chore: rebuild input bar hint text on language update (#5042) * fix: clear database on reinitialize (#5045) * chore: default to reactions maxWidth null if not available (#5047) * fix: remove duplicate navigator pop in member actions popup (#5048) * Reduce gap between lines in practice modes (#5041) * fix: prevent word card overflow in vocab details (#5049) * chore: style tokens in transcription like other clickable tokens (#5055) * fix: always align space nav rail children to the left (#5059) * chore: update message analytics feedback popup background color (#5061) * chore: increase padding in span card scroll view to prevent choice animation overflow (#5062) * chore: Don't use dropdown if only one item (#5063) * chore: Disable ability to send video/files (slash anything else that the bot doesn’t know what to do with) in bot chats (#5065) * chore: show more specific error in audio recording dialog (#5068) * chore: stack expanded space navigation menu over screen in one column mode (#5069) * feat: when screen size gets too short, show warning dialog (#5070) * 5053 can get points from lemma with max score (#5078) * make uses a private field for ConstructUses * expose capped list of uses in ConstructUses * filter capped construct uses in getUses * fix: don't show send button if error in recording dialog (#5079) * chore: allow users to highlight main word in word card * fix: in emoji picker, don't set selected emoji based on old stream data * remove duplicate subscription cancel * fix: fix recording dialog import error * fix: disable new token collection for token not in L2 * chore: use activity plan CEFR level in saved activity display * chore: apply border to dialog directly in delete space dialog (#5093) * chore: hide nav rail item tooltips when expanded (#5094) * chore: reduce min height of span card feedback section (#5095) * chore: force span card to always go above input bar (#5096) * fix: always enable small screen warning dialog on web (#5097) * fix: add new blocks to merge table before fetching previous constructs when calculating points added by construct update (#5098) * fix: remove reaction subscription to prevent overlay jumping (#5100) * 4825 vocabulary practice (#4826) * chore: move logic for lastUsedByActivityType into ConstructIdentifier * feat: vocab practice * add vocab activity progress bar * fix: shuffle audio practice choices * update UI of vocab practice Added buttons, increased text size and change position, cards flip over and turn red/green on click and respond to hover input * add xp sparkle, shimmering choice card placeholder * spacing changes fix padding, make choice cards spacing/sizing responsive to screen size, replace shimmer cards with stationary circle indicator * don't include duplicate lemma choices * use constructID and show lemma/emoji on choice cards add method to clear cache in case the results was an error, and add a retry button on error * gain xp immediately and take out continue session also refactor the choice cards to have separate widgets for each type and a parent widget to give each an id for xp sparkle * add practice finished page with analytics * Color tweaks on completed page and time card placeholder * add timer * give XP for bonuses and change timer to use stopwatch * simplify card logic, lock practice when few vocab words * merge analytics changes and fix bugs * reload on language change - derive XP data from new analytics - Don't allow any clicks after correct answer selected * small fixes, added tooltip, added copy to l10 * small tweaks and comments * formatting and import sorting --------- Co-authored-by: avashilling <165050625+avashilling@users.noreply.github.com> * feat: Directing to click messages with shimmer (#5106) * fix: use standard loading dialog on submit delete space dialog (#5107) * chore: don't show practice tooltip if mode is complete (#5108) * chore: don't restrict token length (#5112) * fix: in recording dialog, throw exception on permission denied (#5114) * chore: remove margin from last entry in user activity summary list (#5115) * chore: make emoji choice shimmer background match word card background (#5116) * feat: allow users to update bot's voice settings (#5119) * fix: hide ability to change bot chat settings from non-admins (#5120) * fix: remove extra text from end of download file name (#5121) * fix: remove invalid expanded widget (#5124) * fix: add guard to prevent showing screen size popup when expanding screen after showing popup (#5127) * chore: normalize accents in vocab search (#5128) * chore: base level icon spacing on xp needed to reach level in vocab details (#5131) * chore: add padding to bottom of vocab list view so practice button doesn't block last vocab entries (#5132) * fix: fix practice record construct id assignment for morph activities (#5133) * fix: coerce existing aggregate analytics database entries into correct format before merging to avoid data loss (#5136) * feat: make construct aggregated case-insensitive (#5137) * chore: prevent user from spamming disabled vocab practice button (#5138) * fix: reset voice on langauge update (#5140) * chore: make emoji base shimmer transparent (#5142) * chore: update sort order in space participants list (#5144) * chore: remove padding from last entry in activity list (#5146) * fix: disable emoji setting for non-L2 constructs (#5148) * fix: add reaction notifier to rebuild reaction picker and reaction display on reaction change (#5151) * chore: decrease text sizes in vocab practice complete page in one column mode (#5152) * chore: hide download button in download dialogs if download is complete (#5157) * fix: show morph as unlocked in analytics if ever used (#5158) * chore: reduce span card spacing to reduce unneeded scroll (#5160) * chore: reduce span card spacing to reduce unneeded scroll * remove debugging code * fix: don't double space ID on navigation (#5163) * chore: reduce negative points to 1 (#5162) To eliminate the chance of having negative total, minimum upon completion now is 30XP * fix: remove duplicates from answer choices (#5161) * fix: use canonical activity time in display for completed activity (#5164) * chore: refresh language cache to add voices (#5165) * chore: don't show loading dialog on reaction redaction (#5166) * build: bump version --------- Co-authored-by: Kelrap <kel.raphael3@outlook.com> Co-authored-by: Kelrap <99418823+Kelrap@users.noreply.github.com> Co-authored-by: avashilling <165050625+avashilling@users.noreply.github.com>
272 lines
8.6 KiB
Dart
272 lines
8.6 KiB
Dart
import 'dart:math';
|
|
|
|
import 'package:matrix/matrix.dart';
|
|
|
|
import 'package:fluffychat/pangea/analytics_misc/construct_use_type_enum.dart';
|
|
import 'package:fluffychat/pangea/analytics_misc/constructs_model.dart';
|
|
import 'package:fluffychat/pangea/choreographer/choreo_record_model.dart';
|
|
import 'package:fluffychat/pangea/choreographer/igc/pangea_match_status_enum.dart';
|
|
import 'package:fluffychat/pangea/common/utils/error_handler.dart';
|
|
import 'package:fluffychat/pangea/events/models/pangea_token_model.dart';
|
|
import 'package:fluffychat/pangea/speech_to_text/speech_to_text_response_model.dart';
|
|
import 'package:fluffychat/widgets/matrix.dart';
|
|
|
|
/// this class is contained within a [RepresentationEvent]
|
|
/// this event is the child of a [EventTypes.Message]
|
|
/// the event has two potential children events -
|
|
/// [PangeaTokensEvent] and [PangeaIGCEvent]
|
|
/// these events contain [PangeaMessageTokens] and [ChoreoRecordModel], respectively.
|
|
class PangeaRepresentation {
|
|
/// system-detected language, possibly condensed from a list,
|
|
/// but only with high certainty
|
|
/// cannot be "unk"
|
|
String langCode;
|
|
|
|
/// final sent text
|
|
/// if this was a process, a [PangeaIGCEvent] will contain changes
|
|
String text;
|
|
|
|
bool originalSent;
|
|
bool originalWritten;
|
|
|
|
// a representation can be create via speech to text on the original message
|
|
SpeechToTextResponseModel? speechToText;
|
|
|
|
// how do we know which representation was sent by author?
|
|
// RepresentationEvent.text == PangeaMessageEvent.event.body
|
|
// use: to know whether directUse
|
|
|
|
// how do we know which representation was original L1 message that was translated (if it exists)?
|
|
// (of l2 rep) RepresentationEvent.igc.steps.first.text = RepresentationEvent.text (of L1 rep)
|
|
// use: for base text for future translations
|
|
|
|
// os = true and ow = false
|
|
// rep that went through IGC/IT
|
|
|
|
// os = false and ow = false
|
|
// rep added by other user
|
|
|
|
// os = true and ow = true
|
|
// potentially L1 language use, maybe with limited IGC, and ignored out of target cries
|
|
// potentially perfect L2 use
|
|
|
|
// os = false and ow = true
|
|
// L1 message that then went through significant IGC and/or IT
|
|
// L2 message with errors that went through IGC
|
|
|
|
PangeaRepresentation({
|
|
required this.langCode,
|
|
required this.text,
|
|
required this.originalSent,
|
|
required this.originalWritten,
|
|
this.speechToText,
|
|
});
|
|
|
|
factory PangeaRepresentation.fromJson(Map<String, dynamic> json) {
|
|
return PangeaRepresentation(
|
|
langCode: json[_langCodeKey],
|
|
text: json[_textKey],
|
|
originalSent: json[_originalSentKey] ?? false,
|
|
originalWritten: json[_originalWrittenKey] ?? false,
|
|
speechToText: json[_speechToTextKey] == null
|
|
? null
|
|
: SpeechToTextResponseModel.fromJson(json[_speechToTextKey]),
|
|
);
|
|
}
|
|
|
|
static const _textKey = "txt";
|
|
static const _langCodeKey = "lang";
|
|
static const _originalSentKey = "snt";
|
|
static const _originalWrittenKey = "wrttn";
|
|
static const _speechToTextKey = "stt";
|
|
|
|
Map<String, dynamic> toJson() {
|
|
final data = <String, dynamic>{};
|
|
data[_textKey] = text;
|
|
data[_langCodeKey] = langCode;
|
|
if (originalSent) data[_originalSentKey] = originalSent;
|
|
if (originalWritten) data[_originalWrittenKey] = originalWritten;
|
|
if (speechToText != null) {
|
|
data[_speechToTextKey] = speechToText!.toJson();
|
|
}
|
|
return data;
|
|
}
|
|
|
|
bool get langCodeMatchesL2 =>
|
|
langCode.split("-").first ==
|
|
MatrixState.pangeaController.userController.userL2?.langCodeShort;
|
|
|
|
/// Get construct uses for the message that weren't captured during language assistance.
|
|
/// Takes a list of tokens and a choreo record, which is searched
|
|
/// through for each token for its construct use type.
|
|
/// Also takes either an event (typically when the Representation itself is
|
|
/// available) or construct use metadata (when the event is not available,
|
|
/// i.e. immediately after message send) to create the construct use.
|
|
List<OneConstructUse> vocabAndMorphUses({
|
|
required List<PangeaToken> tokens,
|
|
Event? event,
|
|
ConstructUseMetaData? metadata,
|
|
ChoreoRecordModel? choreo,
|
|
}) {
|
|
final List<OneConstructUse> uses = [];
|
|
|
|
// missing vital info so return
|
|
if (event?.roomId == null && metadata?.roomId == null) {
|
|
// debugger(when: kDebugMode);
|
|
return uses;
|
|
}
|
|
|
|
metadata ??= ConstructUseMetaData(
|
|
roomId: event!.roomId!,
|
|
eventId: event.eventId,
|
|
timeStamp: event.originServerTs,
|
|
);
|
|
|
|
// for each token, record whether selected in ga, ta, or wa
|
|
List<PangeaToken> tokensToSave =
|
|
tokens.where((token) => token.lemma.saveVocab).toList();
|
|
if (choreo != null && choreo.pastedStrings.isNotEmpty) {
|
|
tokensToSave = tokensToSave
|
|
.where(
|
|
(token) => !choreo.pastedStrings.any(
|
|
(pasted) => pasted
|
|
.toLowerCase()
|
|
.contains(token.text.content.toLowerCase()),
|
|
),
|
|
)
|
|
.toList();
|
|
}
|
|
|
|
if (choreo == null || choreo.choreoSteps.isEmpty) {
|
|
for (final token in tokensToSave) {
|
|
uses.addAll(
|
|
token.allUses(
|
|
ConstructUseTypeEnum.wa,
|
|
metadata,
|
|
ConstructUseTypeEnum.wa.pointValue,
|
|
),
|
|
);
|
|
}
|
|
|
|
return uses;
|
|
}
|
|
|
|
for (final token in tokensToSave) {
|
|
ChoreoRecordStepModel? tokenStep;
|
|
for (final step in choreo.choreoSteps) {
|
|
final igcMatch = step.acceptedOrIgnoredMatch;
|
|
final itStep = step.itStep;
|
|
if (itStep == null && igcMatch == null) {
|
|
continue;
|
|
}
|
|
|
|
final choices = step.choices;
|
|
if (choices == null || choices.isEmpty) {
|
|
continue;
|
|
}
|
|
|
|
final stepContainsToken = choices.any(
|
|
(choice) => choice.contains(token.text.content),
|
|
);
|
|
|
|
// if the step contains the token, and the token hasn't been assigned a step
|
|
// (or the assigned step is an IGC step, but an IT step contains the token)
|
|
// then assign the token to the step
|
|
if (stepContainsToken &&
|
|
(tokenStep == null ||
|
|
(tokenStep.itStep == null && step.itStep != null))) {
|
|
tokenStep = step;
|
|
}
|
|
}
|
|
|
|
if (tokenStep == null ||
|
|
tokenStep.acceptedOrIgnoredMatch?.status ==
|
|
PangeaMatchStatusEnum.automatic) {
|
|
// if the token wasn't found in any IT or IGC step, so it was wa
|
|
uses.addAll(
|
|
token.allUses(
|
|
ConstructUseTypeEnum.wa,
|
|
metadata,
|
|
ConstructUseTypeEnum.wa.pointValue,
|
|
),
|
|
);
|
|
continue;
|
|
}
|
|
|
|
if (tokenStep.acceptedOrIgnoredMatch != null &&
|
|
tokenStep.acceptedOrIgnoredMatch?.status !=
|
|
PangeaMatchStatusEnum.accepted) {
|
|
uses.addAll(
|
|
token.allUses(
|
|
ConstructUseTypeEnum.ga,
|
|
metadata,
|
|
0,
|
|
),
|
|
);
|
|
continue;
|
|
}
|
|
|
|
if (tokenStep.itStep != null) {
|
|
final selectedChoices = tokenStep.itStep!.continuances
|
|
.where((choice) => choice.wasClicked)
|
|
.length;
|
|
if (selectedChoices == 0) {
|
|
ErrorHandler.logError(
|
|
e: "No selected choices for IT step",
|
|
data: {
|
|
"token": token.text.content,
|
|
"step": tokenStep.toJson(),
|
|
},
|
|
);
|
|
continue;
|
|
}
|
|
|
|
final corITPoints = ConstructUseTypeEnum.corIt.pointValue;
|
|
final incITPoints = ConstructUseTypeEnum.incIt.pointValue;
|
|
final xp = max(
|
|
0,
|
|
corITPoints + (incITPoints * (selectedChoices - 1)),
|
|
);
|
|
|
|
uses.addAll(
|
|
token.allUses(
|
|
ConstructUseTypeEnum.ta,
|
|
metadata,
|
|
xp,
|
|
),
|
|
);
|
|
} else if (tokenStep.acceptedOrIgnoredMatch!.match.choices != null) {
|
|
final selectedChoices = tokenStep.acceptedOrIgnoredMatch!.match.choices!
|
|
.where((choice) => choice.selected)
|
|
.length;
|
|
if (selectedChoices == 0) {
|
|
ErrorHandler.logError(
|
|
e: "No selected choices for IGC step",
|
|
data: {
|
|
"token": token.text.content,
|
|
"step": tokenStep.toJson(),
|
|
},
|
|
);
|
|
continue;
|
|
}
|
|
|
|
final corIGCPoints = ConstructUseTypeEnum.corIGC.pointValue;
|
|
final incIGCPoints = ConstructUseTypeEnum.incIGC.pointValue;
|
|
final xp = max(
|
|
0,
|
|
corIGCPoints + (incIGCPoints * (selectedChoices - 1)),
|
|
);
|
|
|
|
uses.addAll(
|
|
token.allUses(
|
|
ConstructUseTypeEnum.ga,
|
|
metadata,
|
|
xp,
|
|
),
|
|
);
|
|
}
|
|
}
|
|
|
|
return uses;
|
|
}
|
|
}
|