Changelog

Replace the overlay-divider approach for weekly goal progress bars with

truly independent pill segments. Each segment now has its own track and

fill bar, eliminating visual artifacts where fill would bleed through

divider gaps. The new GoalSegmentViewModel drives per-pill rendering

with independent fill widths, corner radii, and funding state colors.

Add two new GoalTargetType enum values: BiweeklySpending (anchored to a

user-selected date, repeating every 14 days, producing 2-3 segments per

month) and SemiMonthlySpending (two user-picked days of month, always 2

segments). GoalService gains CountBiweeklyOccurrences and a generalized

CalculateSegmentProgress method that handles rollover for all segmented

goal types.

Category management now shows an anchor date picker for biweekly goals

and two day-of-month ComboBoxes (1-30 + Last of Month) for semi-monthly

goals. The Category record gains GoalAnchorDate, GoalSemiMonthlyDay1,

and GoalSemiMonthlyDay2 fields. Seed data includes demo biweekly and

semi-monthly categories. Indexable content labels updated for new types.

Version bump 3.13.0 -> 3.14.0 (csproj + PluginMetadata).

Weekly spending goals now let users pick which day of the week their

recurring expense falls on (e.g., Saturday for groceries). The monthly

budget is calculated as weeklyTarget × actual occurrences of that day

in the month (4 or 5), replacing the approximate weeks-remaining math.

The progress bar dynamically renders 3 or 4 dividers via an

ItemsRepeater bound to computed divider positions, instead of

hardcoding 3 dividers for a fixed 4-segment layout.

Changes:

• Category model: added GoalWeekday (DayOfWeek?) field
• GoalService: added CountWeekdayOccurrences() and

CalculateWeeklySegmentProgress(); rewrote CalculateWeeklyForMonth()

to use exact weekday counting; updated CalculateGoalProgress() to

compare allocated against the true monthly total for weekly goals

• CategoryViewModel: added WeeklySegmentCount, WeeklyDividers

(ObservableCollection<WeeklyDividerViewModel>) with computed

Canvas.Left positions; updated GoalProgressDisplay to show monthly

total for weekly goals

• BudgetDashboardView: replaced 3 hardcoded Border dividers with

dynamic ItemsControl bound to WeeklyDividers

• CategoryManagementViewModel: added WeekdayOption record,

WeekdayOptions array, SelectedWeekdayOption property, HasWeekday

computed property; persist GoalWeekday in SaveCategoryAsync

• CategoryManagementView: added day-of-week ComboBox visible only

when goal type is Weekly Spending

• FinanceSeedService: Groceries defaults to Saturday; allocation

uses CountWeekdayOccurrences for accurate monthly totals

• Version bump: 3.12.4 → 3.13.0

Added a SPENT column between BUDGETED and AVAILABLE to make the data

flow clear: BUDGETED -> SPENT -> AVAILABLE. Activity (spending) data

is fetched via new BudgetService.GetCategoryActivityAsync and displayed

as an absolute dollar amount per category.

All column headers now include a (?) indicator with tooltip explanations:

• CATEGORY: grouped spending categories
• BUDGETED: allocated amount, click to edit
• SPENT: total spent this month
• AVAILABLE: remaining after spending, color-coded
• GOAL: funding target progress

Goal bar width reduced from 300px to 260px to accommodate the new

column. Version bumped to 3.12.0.

Local models were running every message as a standalone prompt with zero

conversational context — follow-up questions like "but my income is 4700"

had no reference to the prior exchange. This made multi-turn chat useless.

FormatPrompt now accepts ConversationHistory and builds proper multi-turn

chat templates for all three model families (Llama 3.x, Mistral, Phi-3),

including correct turn delimiters and system prompt positioning. History

is capped at 6 turns to fit within local context windows.

The chat view model now passes BuildConversationHistory() to local model

requests, giving Duncan the same conversational continuity as cloud models.

Bump version to 1.60.0.

Three fixes for Duncan giving hallucinated, context-unaware responses:

1. System prompt now instructs Duncan to use provided data context and

never fabricate numbers — applies to both local and cloud prompts.

2. Local RAG context budget increased from 200 to 400 chars per chunk

(6 chunks max, up from 5) so budget data isn't truncated to nothing.

Mistral 7B context window increased from 4096 to 8192 tokens to

accommodate the richer context.

3. System RAM detection via GC.GetGCMemoryInfo().TotalAvailableMemoryBytes

drives a model recommendation in AI settings: 16+ GB gets Mistral 7B,

8+ GB gets Phi-3 Mini, smaller systems get Llama 3.2 1B. The recommended

model is labeled in the dropdown and an explanation shown below it.

Bump version to 1.59.0.