Agent-System-Backend/internal/runmgr/material_sources.go

package runmgr

import (
	"agent-system-backend/internal/runtime"
	core "agent-system-core"
	"crypto/sha256"
	"encoding/hex"
	"encoding/json"
	"fmt"
	"sort"
	"strings"
)

type materialInputItem struct {
	ArtifactID   core.ID
	Parsed       core.ParsedDocument
	OriginalHash string
}
type materialSource struct {
	Ref      core.TaskSourceRef
	Text     string
	Provided string
}
type materialSourceRegistry map[string]materialSource

func materialInputRevision(items []materialInputItem, processingRevision string) string {
	docs := make([]core.ParsedDocument, 0, len(items))
	for _, item := range items {
		d := item.Parsed
		origin := core.DocumentOrigin{}
		if d.Meta.Origin != nil {
			origin = *d.Meta.Origin
		}
		origin.ContentHash = item.OriginalHash
		d.Meta.Origin = &origin
		docs = append(docs, d)
	}
	sum := sha256.Sum256([]byte(materialKey(docs) + "\n" + processingRevision))
	return hex.EncodeToString(sum[:])
}
func materialSources(docs []core.ParsedDocument, perDoc int) materialSourceRegistry {
	r := materialSourceRegistry{}
	left := materialInlineTotal
	if perDoc <= 0 {
		perDoc = materialPerDocDef
	}
	for _, d := range docs {
		cap := perDoc
		if cap > left {
			cap = left
		}
		text := strings.TrimSpace(d.PlainText())
		rr := []rune(text)
		if len(rr) < cap {
			cap = len(rr)
		}
		provided := string(rr[:cap])
		left -= cap
		if d.Meta.Origin == nil || d.Meta.Origin.ArtifactID == "" || d.Meta.Origin.ContentHash == "" {
			continue
		}
		for index, block := range d.Blocks {
			full := strings.TrimSpace((core.ParsedDocument{Blocks: []core.DocBlock{block}}).PlainText())
			if full == "" || !strings.Contains(provided, full) {
				continue
			}
			sum := sha256.Sum256([]byte(fmt.Sprintf("%s:%s:%d", d.Meta.Origin.ArtifactID, d.Meta.Origin.ContentHash, index)))
			id := "src_" + hex.EncodeToString(sum[:12])
			r[id] = materialSource{Ref: core.TaskSourceRef{ID: id, ArtifactID: d.Meta.Origin.ArtifactID, ContentHash: d.Meta.Origin.ContentHash, ParserRevision: d.Meta.Origin.ParserRevision, BlockIndex: index, Label: fmt.Sprintf("%s · 段落%d", d.Meta.Filename, index+1), Availability: "available"}, Text: full, Provided: provided}
		}
	}
	return r
}
func (r materialSourceRegistry) validate(q core.TaskSourceQuote) (core.TaskSourceRef, error) {
	src, ok := r[q.SourceID]
	if !ok || q.BlockIndex != src.Ref.BlockIndex || strings.TrimSpace(q.Quote) == "" || !strings.Contains(src.Text, q.Quote) || !strings.Contains(src.Provided, q.Quote) {
		return core.TaskSourceRef{}, fmt.Errorf("checkpoint.source: quote not in provided block")
	}
	ref := src.Ref
	ref.Quote = q.Quote
	return ref, nil
}
func (r materialSourceRegistry) prompt() string {
	keys := make([]string, 0, len(r))
	for k := range r {
		keys = append(keys, k)
	}
	sort.Strings(keys)
	var b strings.Builder
	if len(keys) > 0 {
		b.WriteString("\n\n=== 可引用的已提供来源（不含未提供的全文） ===\n")
	}
	for _, id := range keys {
		src := r[id]
		fmt.Fprintf(&b, "sourceId=%s blockIndex=%d %s\n", id, src.Ref.BlockIndex, src.Ref.Label)
	}
	return b.String()
}

func (c *runCtx) materialRevision() string {
	perDoc := int(c.eff.GetNumber(materialPerDocKey, materialPerDocDef))
	if perDoc <= 0 {
		perDoc = materialPerDocDef
	}
	recipe := "material-recipe.v3"
	if runner, ok := c.mgr.subRunner.(*runtime.SubLLMRunner); ok && runner.Prompts != nil {
		if body, err := runner.Prompts.AgentPrompt(c.run.Owner, materialSubagentID); err == nil {
			recipe += "\n" + body
		}
		if spec, ok := runner.Prompts.SubagentSpec(c.run.Owner, materialSubagentID); ok {
			raw, _ := json.Marshal(spec)
			recipe += "\n" + string(raw)
		}
	}
	if perDoc == materialPerDocDef && recipe == "material-recipe.v3" {
		return materialKey(c.attachments)
	}
	items := make([]materialInputItem, 0, len(c.attachments))
	for _, d := range c.attachments {
		item := materialInputItem{Parsed: d}
		if d.Meta.Origin != nil {
			item.ArtifactID = d.Meta.Origin.ArtifactID
			item.OriginalHash = d.Meta.Origin.ContentHash
		}
		items = append(items, item)
	}
	return materialInputRevision(items, fmt.Sprintf("%s\nperDoc=%d,total=%d", recipe, perDoc, materialInlineTotal))
}