audita/internal/core/reporting/markdown.go

package reporting

import (
	"fmt"
	"sort"
	"strings"

	"gitea.maximumdirect.net/eric/audita/v3/internal/core/redaction"
)

// CompletionLine is the brief successful CLI summary. It uses the same counts
// as both YAML documents and remains independent of file-output selection.
func CompletionLine(snapshot Snapshot) string {
	return fmt.Sprintf("Completed: %d corrections applied; %d work units with errors; %d recoveries.",
		count(snapshot.Summary.CorrectionsApplied), count(snapshot.Summary.WorkUnitsWithErrors), count(snapshot.Summary.Recoveries))
}

func count(value *int) int {
	if value == nil {
		return 0
	}
	return *value
}

// Markdown renders a fixed-layout human report from one processing snapshot.
func Markdown(snapshot Snapshot, secrets ...string) []byte {
	r := redaction.New(secrets)
	safe := func(value string) string { return escapeMarkdown(r.String(value)) }
	var b strings.Builder
	fmt.Fprintf(&b, "# Audita report: %s\n\n", safe(snapshot.RunID))
	fmt.Fprintf(&b, "Status: %s\n\n", safe(snapshot.Status))
	fmt.Fprintf(&b, "Corrections applied: %d  \n", count(snapshot.Summary.CorrectionsApplied))
	fmt.Fprintf(&b, "Work units: %s total; %s completed; %s interrupted; %s not started  \n",
		showCount(snapshot.Summary.WorkUnitsTotal), showCount(snapshot.Summary.WorkUnitsCompleted),
		showCount(snapshot.Summary.WorkUnitsInterrupted), showCount(snapshot.Summary.WorkUnitsNotStarted))
	fmt.Fprintf(&b, "Errors: %s work units; %s run errors  \n", showCount(snapshot.Summary.WorkUnitsWithErrors), showCount(snapshot.Summary.RunErrors))
	fmt.Fprintf(&b, "Recoveries: %s of %s model operations  \n", showCount(snapshot.Summary.Recoveries), showCount(snapshot.Summary.LLMOperations))
	fmt.Fprintf(&b, "Proposal responses: %s; with filtering: %s\n\n",
		showCount(snapshot.Summary.ProposalResponses), showCount(snapshot.Summary.ProposalResponsesWithFiltering))
	fmt.Fprintf(&b, "Transcript input: %s  \n", safe(snapshot.TranscriptPath))
	fmt.Fprintf(&b, "Glossary input: %s  \n", safe(snapshot.GlossaryPath))
	output := snapshot.OutputPath
	if output == "" {
		output = "stdout"
	}
	fmt.Fprintf(&b, "Transcript output: %s\n\n", safe(output))

	fmt.Fprintln(&b, "## Work-unit errors")
	fmt.Fprintln(&b)
	writeMatrix(&b, snapshot, safe, false)
	fmt.Fprintln(&b)
	fmt.Fprintln(&b, "## Recoveries")
	fmt.Fprintln(&b)
	writeMatrix(&b, snapshot, safe, true)
	fmt.Fprintln(&b)
	fmt.Fprintln(&b, "Legend: `—` = not started; `n*` = interrupted with n recorded errors or recoveries. Completed error cells are 0 or 1.")
	fmt.Fprintln(&b)
	writeReasons(&b, "Error reasons", errorReasons(snapshot.Errors), safe)
	writeReasons(&b, "Recovery reasons", recoveryReasons(snapshot.Recoveries), safe)
	fmt.Fprintln(&b, "A recovery can have multiple reasons; reason counts need not sum to the recovery total.")
	fmt.Fprintln(&b)
	fmt.Fprintln(&b, "Details: `errors.yml` and `recoveries.yml`.")
	return []byte(b.String())
}

func showCount(value *int) string {
	if value == nil {
		return "unknown"
	}
	return fmt.Sprint(*value)
}

func writeMatrix(b *strings.Builder, snapshot Snapshot, safe func(string) string, recovery bool) {
	if len(snapshot.StageNames) == 0 || len(snapshot.WorkUnits) == 0 {
		fmt.Fprintln(b, "No work units.")
		return
	}
	columns := snapshot.StageNames
	fmt.Fprint(b, "| Chunk |")
	for _, name := range columns {
		fmt.Fprintf(b, " %s |", safe(name))
	}
	fmt.Fprint(b, "\n| --- |")
	for range columns {
		fmt.Fprint(b, " ---: |")
	}
	fmt.Fprintln(b)
	byChunk := make(map[int]map[int]WorkUnit)
	chunkMeta := make(map[int]WorkUnit)
	keys := make([]int, 0)
	for _, unit := range snapshot.WorkUnits {
		if byChunk[unit.ChunkIndex] == nil {
			byChunk[unit.ChunkIndex] = make(map[int]WorkUnit)
			keys = append(keys, unit.ChunkIndex)
		}
		byChunk[unit.ChunkIndex][unit.StageIndex] = unit
		chunkMeta[unit.ChunkIndex] = unit
	}
	sort.Ints(keys)
	for _, chunk := range keys {
		meta := chunkMeta[chunk]
		fmt.Fprintf(b, "| Chunk %d (segments %d–%d) |", chunk+1, meta.StartSegmentID, meta.EndSegmentID)
		for stage := range columns {
			unit, exists := byChunk[chunk][stage]
			cell := "—"
			if exists && unit.State != "not_started" {
				n := unit.ErrorCount
				if recovery {
					n = unit.RecoveryCount
				}
				cell = fmt.Sprint(n)
				if unit.State == "interrupted" {
					cell += "*"
				}
			}
			fmt.Fprintf(b, " %s |", cell)
		}
		fmt.Fprintln(b)
	}
}

func writeReasons(b *strings.Builder, title string, counts map[string]int, safe func(string) string) {
	fmt.Fprintf(b, "### %s\n\n", title)
	if len(counts) == 0 {
		fmt.Fprintln(b, "None.")
		fmt.Fprintln(b)
		return
	}
	keys := make([]string, 0, len(counts))
	for key := range counts {
		keys = append(keys, key)
	}
	sort.Strings(keys)
	for _, key := range keys {
		fmt.Fprintf(b, "- `%s`: %d\n", safe(key), counts[key])
	}
	fmt.Fprintln(b)
}

func errorReasons(records []ErrorRecord) map[string]int {
	counts := make(map[string]int)
	for _, record := range records {
		seen := make(map[string]bool)
		if record.Scope == "run" {
			seen[record.Reason] = true
		}
		for _, failure := range record.Failures {
			seen[failure.Reason] = true
		}
		for reason := range seen {
			if reason != "" {
				counts[reason]++
			}
		}
	}
	return counts
}

func recoveryReasons(records []RecoveryRecord) map[string]int {
	counts := make(map[string]int)
	for _, record := range records {
		seen := make(map[string]bool)
		for _, reason := range record.Reasons {
			seen[reason] = true
		}
		for reason := range seen {
			counts[reason]++
		}
	}
	return counts
}

func escapeMarkdown(value string) string {
	value = strings.ReplaceAll(value, "\\", "\\\\")
	value = strings.ReplaceAll(value, "|", "\\|")
	value = strings.ReplaceAll(value, "`", "\\`")
	value = strings.ReplaceAll(value, "\r\n", "\n")
	value = strings.ReplaceAll(value, "\r", "\n")
	return strings.ReplaceAll(value, "\n", "<br>")
}