budget

func AssemblePacket ¶

func AssemblePacket(ctx *entity.Context, budget int) *assembledPacket

AssemblePacket builds a context packet respecting the token budget.

Allocation tiers:

• Tier 1 (always): constitution, read order, instruction
• Tier 2 (40%): active tasks
• Tier 3 (20%): conventions
• Tier 4+5 (remaining): decisions and learnings, scored by relevance

Parameters:

• ctx: Loaded context containing the files
• budget: Token budget to respect

Returns:

• *AssembledPacket: Assembled packet within budget

func EstimateSliceTokens ¶

func EstimateSliceTokens(items []string) int

EstimateSliceTokens sums token estimates for a string slice.

Parameters:

• items: Strings to estimate

Returns:

• int: Total estimated tokens

func ExtractAllConventions ¶

func ExtractAllConventions(ctx *entity.Context) []string

ExtractAllConventions extracts all bullet items from CONVENTIONS.md (not limited to 5 like the old implementation).

Parameters:

• ctx: Loaded context containing the files

Returns:

• []string: All convention bullet items; nil if the file is not found

func FillSection ¶

func FillSection(entries []score.Entry, budget int) ([]string, []string)

FillSection selects scored entries to fill a budget, with graceful degradation.

Includes full entries by score order until ~80% of the budget is consumed. Remaining entries get title-only summaries.

Parameters:

• entries: Scored entries sorted by score descending
• budget: Token budget for this section

Returns:

• []string: Full entry bodies that fit in the budget
• []string: Title-only summaries for entries that didn't fit

func FitItems ¶

func FitItems(items []string, budget int) []string

FitItems returns items that fit within a token budget.

Items are included in order until the budget would be exceeded.

Parameters:

• items: String items to include
• budget: Maximum token budget

Returns:

• []string: Items that fit within the budget

func OutputAgentJSON ¶

func OutputAgentJSON(
	cmd *cobra.Command, ctx *entity.Context, budget int,
) error

OutputAgentJSON writes the context packet as pretty-printed JSON.

Uses budget-aware assembly to score entries and respect the token budget.

Parameters:

• cmd: Cobra command for output stream
• ctx: Loaded context containing the files
• budget: Token budget for content selection

Returns:

• error: Non-nil if JSON encoding fails

func OutputAgentMarkdown ¶

func OutputAgentMarkdown(
	cmd *cobra.Command, ctx *entity.Context, budget int,
) error

OutputAgentMarkdown writes the context packet as formatted Markdown.

Uses budget-aware assembly to score entries and respect the token budget. Output includes sections for constitution, tasks, conventions, decisions (full body), learnings (full body), and title-only summaries.

Parameters:

• cmd: Cobra command for output stream
• ctx: Loaded context containing the files
• budget: Token budget for content selection

Returns:

• error: Always nil (included for interface consistency)

func ParseEntryBlocks ¶

func ParseEntryBlocks(ctx *entity.Context, fileName string) []index.EntryBlock

ParseEntryBlocks parses a context file into entry blocks.

Parameters:

• ctx: Loaded context
• fileName: Name of the file to parse (e.g., config.Decision)

Returns:

• []index.EntryBlock: Parsed entry blocks; nil if the file is not found

func RenderMarkdownPacket ¶

func RenderMarkdownPacket(pkt *assembledPacket) string

RenderMarkdownPacket renders an assembled packet as Markdown.

Parameters:

• pkt: Assembled packet to render

Returns:

• string: Formatted Markdown output

func Split ¶

func Split(total int, a, b []score.Entry) (int, int)

Split divides a token budget between two scored sections.

Each section gets at least 30% of the budget (if content exists). The remaining 40% is allocated proportionally to content size.

Parameters:

• total: Total tokens to split
• a: First section's scored entries
• b: Second section's scored entries

Returns:

• int: Budget for section a
• int: Budget for section b

func TotalEntryTokens ¶

func TotalEntryTokens(entries []score.Entry) int

TotalEntryTokens sums pre-computed token counts for scored entries.

Parameters:

• entries: Scored entries with token estimates

Returns:

• int: Total tokens