qiweimanager-master/helper/after_sales_file.go

package main

import (
	"fmt"
	"os"
	"path/filepath"
	"strings"
)

const (
	afterSalesFileContentLimit = 6000
	afterSalesFilePromptLimit  = 1800

	afterSalesFileStatusReady         = "parsed"
	afterSalesFileStatusSaved         = "saved"
	afterSalesFileStatusUnsupported   = "unsupported"
	afterSalesFileStatusMissing       = "missing"
	afterSalesFileStatusDownloadFail  = "download_failed"
	afterSalesFileStatusExtractFailed = "extract_failed"
)

func extractAfterSalesFileFromMessage(msg autoReplyMessage, raw map[string]interface{}, sourceMessageID string) AfterSalesFileAttachment {
	if !looksLikeAfterSalesFileMessage(msg, raw) {
		return AfterSalesFileAttachment{}
	}
	attachment := AfterSalesFileAttachment{
		Name:            strings.TrimSpace(msg.MediaFileName),
		Ref:             firstNonEmpty(strings.TrimSpace(msg.MediaURL), strings.TrimSpace(msg.MediaFileID)),
		SourceMessageID: strings.TrimSpace(sourceMessageID),
	}
	for _, value := range []string{msg.MediaLocalPath, firstLocalMediaPathFromValue(raw)} {
		if path := normalizedExistingFilePath(value); path != "" {
			attachment.Path = path
			break
		}
	}
	if attachment.Path == "" {
		if path, err := ensureAutoReplyMediaLocalPath(msg); err == nil {
			attachment.Path = normalizedExistingFilePath(path)
		} else {
			attachment.ExtractStatus = afterSalesFileStatusDownloadFail
		}
	}
	if attachment.Name == "" {
		attachment.Name = firstNonEmpty(filepath.Base(attachment.Path), filepath.Base(strings.TrimSpace(msg.MediaURL)), strings.TrimSpace(msg.MediaFileID), "客户附件")
	}
	if attachment.Path == "" {
		if attachment.ExtractStatus == "" {
			attachment.ExtractStatus = afterSalesFileStatusDownloadFail
		}
		return attachment
	}
	content, status := extractAfterSalesFileContent(attachment.Path)
	attachment.Content = content
	attachment.ExtractStatus = status
	return normalizeAfterSalesFileAttachment(attachment)
}

func looksLikeAfterSalesFileMessage(msg autoReplyMessage, raw map[string]interface{}) bool {
	if strings.TrimSpace(msg.MediaKind) == "file" || msg.RawType == 11045 {
		return true
	}
	if raw != nil {
		if typeVal, ok := raw["type"]; ok && intFromAny(typeVal) == 11045 {
			return true
		}
		if event := strings.TrimSpace(stringFromAny(raw["event"])); event == "20005" {
			return true
		}
	}
	return false
}

func normalizedExistingFilePath(value string) string {
	value = strings.Trim(strings.TrimSpace(value), "\"'")
	if value == "" {
		return ""
	}
	if filepath.IsAbs(value) {
		if _, err := os.Stat(value); err == nil {
			return value
		}
		return ""
	}
	candidate := resolveAutoReplyPath(value)
	if _, err := os.Stat(candidate); err == nil {
		return candidate
	}
	return ""
}

func extractAfterSalesFileContent(path string) (string, string) {
	if strings.TrimSpace(path) == "" {
		return "", afterSalesFileStatusMissing
	}
	info, err := os.Stat(path)
	if err != nil {
		return "", afterSalesFileStatusMissing
	}
	if info.IsDir() {
		return "", afterSalesFileStatusUnsupported
	}
	ext := strings.ToLower(filepath.Ext(path))
	if ext == ".zip" {
		return "", "zip_not_parsed"
	}
	switch ext {
	case ".txt", ".md", ".csv", ".xlsx", ".docx", ".pdf":
	default:
		return "", afterSalesFileStatusUnsupported
	}
	blocks, err := parseKnowledgeFile(path, filepath.Dir(path))
	if err != nil {
		var warning knowledgeParseWarning
		if !errorAs(err, &warning) {
			return "", afterSalesFileStatusExtractFailed + ": " + truncateText(err.Error(), 160)
		}
	}
	parts := make([]string, 0, len(blocks))
	for _, block := range blocks {
		text := strings.TrimSpace(block.Content)
		if text == "" {
			continue
		}
		if strings.TrimSpace(block.Title) != "" {
			text = strings.TrimSpace(block.Title) + "\n" + text
		}
		parts = append(parts, text)
	}
	content := truncateText(strings.TrimSpace(strings.Join(parts, "\n\n")), afterSalesFileContentLimit)
	if content == "" {
		return "", afterSalesFileStatusSaved
	}
	return content, afterSalesFileStatusReady
}

func normalizeAfterSalesFileAttachment(item AfterSalesFileAttachment) AfterSalesFileAttachment {
	item.Name = strings.TrimSpace(item.Name)
	item.Path = strings.Trim(strings.TrimSpace(item.Path), "\"'")
	item.Ref = strings.TrimSpace(item.Ref)
	item.Content = truncateText(strings.TrimSpace(item.Content), afterSalesFileContentLimit)
	item.ExtractStatus = strings.TrimSpace(item.ExtractStatus)
	item.SourceMessageID = strings.TrimSpace(item.SourceMessageID)
	if item.Name == "" {
		item.Name = firstNonEmpty(filepath.Base(item.Path), filepath.Base(item.Ref), "客户附件")
	}
	if item.ExtractStatus == "" {
		if item.Content != "" {
			item.ExtractStatus = afterSalesFileStatusReady
		} else if item.Path != "" {
			item.ExtractStatus = afterSalesFileStatusSaved
		} else {
			item.ExtractStatus = afterSalesFileStatusDownloadFail
		}
	}
	return item
}

func normalizeAfterSalesFileAttachments(items []AfterSalesFileAttachment) []AfterSalesFileAttachment {
	seen := make(map[string]struct{})
	result := make([]AfterSalesFileAttachment, 0, len(items))
	for _, item := range items {
		item = normalizeAfterSalesFileAttachment(item)
		if item.Path == "" && item.Ref == "" && item.Name == "" && item.Content == "" {
			continue
		}
		key := firstNonEmpty(item.SourceMessageID, item.Path, item.Ref, item.Name+"|"+item.Content)
		if _, exists := seen[key]; exists {
			continue
		}
		seen[key] = struct{}{}
		result = append(result, item)
	}
	return result
}

func afterSalesFilePromptText(files []AfterSalesFileAttachment, limit int) string {
	if limit <= 0 {
		limit = afterSalesFilePromptLimit
	}
	var b strings.Builder
	for _, file := range normalizeAfterSalesFileAttachments(files) {
		if b.Len() > 0 {
			b.WriteString("\n")
		}
		b.WriteString("文件:")
		b.WriteString(firstNonEmpty(file.Name, filepath.Base(file.Path), file.Ref))
		if file.ExtractStatus != "" {
			b.WriteString(" 状态:")
			b.WriteString(file.ExtractStatus)
		}
		if file.Content != "" {
			b.WriteString("\n内容:\n")
			b.WriteString(truncateText(file.Content, limit))
		}
	}
	return strings.TrimSpace(b.String())
}

func formatAfterSalesFileAttachmentsForExcel(files []AfterSalesFileAttachment, includeContent bool) string {
	files = normalizeAfterSalesFileAttachments(files)
	if len(files) == 0 {
		return ""
	}
	lines := make([]string, 0, len(files))
	for _, file := range files {
		line := firstNonEmpty(file.Name, filepath.Base(file.Path), file.Ref, "客户附件")
		if file.Path != "" {
			line += " | " + file.Path
		} else if file.Ref != "" {
			line += " | " + file.Ref
		}
		if file.ExtractStatus != "" {
			line += " | " + file.ExtractStatus
		}
		if includeContent && file.Content != "" {
			line += fmt.Sprintf("\n%s", truncateText(file.Content, 1200))
		}
		lines = append(lines, line)
	}
	return strings.Join(lines, "\n\n")
}