实现 Booking Excel 预处理接入 SuperAgent

2026-07-20 00:32:01 +07:00
parent b7b04cf1ac
commit d1955f5097
45 changed files with 2955 additions and 10 deletions
--- a/server/src/main/java/cn/nianxx/thhotel/platform/debug/common/result/DebugEmlSuperAgentRunResult.java
+++ b/server/src/main/java/cn/nianxx/thhotel/platform/debug/common/result/DebugEmlSuperAgentRunResult.java
@@ -1,6 +1,8 @@
 package cn.nianxx.thhotel.platform.debug.common.result;

 import cn.nianxx.thhotel.integrations.ai.superagent.common.result.SuperAgentOpenApiTraceEvent;
+import cn.nianxx.thhotel.workflows.reservation.excelimport.common.result.BookingExcelAttachmentExtractionResult;
+import com.fasterxml.jackson.annotation.JsonInclude;
 import com.fasterxml.jackson.annotation.JsonProperty;
 import com.fasterxml.jackson.databind.JsonNode;
 import java.util.List;
@@ -22,6 +24,7 @@ import java.util.Map;
 * @param htmlSanitizeRequired 是否要求前端按安全 HTML 规则展示
 * @param htmlRenderMode HTML 渲染建议模式
 * @param agentbusLikePayload 发送给 SuperAgent 的 AgentBus Outlook-like 主 payload
+ * @param attachmentExtractions Booking Excel 附件预处理安全预览
 * @param superagentSessionId SuperAgent session ID
 * @param superagentRunId SuperAgent run ID
 * @param superagentRawAnswer SuperAgent 原始最终回答
@@ -58,6 +61,9 @@ public record DebugEmlSuperAgentRunResult(
        String htmlRenderMode,
        @JsonProperty("agentbus_like_payload")
        Map<String, Object> agentbusLikePayload,
+        @JsonProperty("attachment_extractions")
+        @JsonInclude(JsonInclude.Include.NON_EMPTY)
+        List<BookingExcelAttachmentExtractionResult> attachmentExtractions,
        @JsonProperty("superagent_session_id")
        String superagentSessionId,
        @JsonProperty("superagent_run_id")
--- a/server/src/main/java/cn/nianxx/thhotel/platform/debug/service/impl/DebugEmlSuperAgentProperties.java
+++ b/server/src/main/java/cn/nianxx/thhotel/platform/debug/service/impl/DebugEmlSuperAgentProperties.java
@@ -19,6 +19,8 @@ public class DebugEmlSuperAgentProperties {
    private long maxFileBytes = 10 * 1024 * 1024L;
    /** Debug SSE 心跳间隔，用于避免长时间调用 SuperAgent 时中间链路空闲断开。 */
    private Duration sseHeartbeatInterval = Duration.ofSeconds(15);
+    /** 是否在 Debug EML 调用 SuperAgent 前追加 Booking Excel 附件预处理结果。 */
+    private boolean includeBookingExcelExtractions = false;

    public boolean isEnabled() {
        return enabled;
@@ -51,4 +53,12 @@ public class DebugEmlSuperAgentProperties {
    public void setSseHeartbeatInterval(Duration sseHeartbeatInterval) {
        this.sseHeartbeatInterval = sseHeartbeatInterval;
    }
+
+    public boolean isIncludeBookingExcelExtractions() {
+        return includeBookingExcelExtractions;
+    }
+
+    public void setIncludeBookingExcelExtractions(boolean includeBookingExcelExtractions) {
+        this.includeBookingExcelExtractions = includeBookingExcelExtractions;
+    }
 }
--- a/server/src/main/java/cn/nianxx/thhotel/platform/debug/service/impl/DebugEmlSuperAgentRunServiceImpl.java
+++ b/server/src/main/java/cn/nianxx/thhotel/platform/debug/service/impl/DebugEmlSuperAgentRunServiceImpl.java
@@ -35,6 +35,9 @@ import cn.nianxx.thhotel.platform.message.service.EmlMessageParseService;
 import cn.nianxx.thhotel.platform.message.service.SourceMessageCaptureService;
 import cn.nianxx.thhotel.platform.message.service.SourceMessageHtmlSanitizerService;
 import cn.nianxx.thhotel.platform.message.service.impl.EmlMessageParseException;
+import cn.nianxx.thhotel.workflows.reservation.excelimport.common.request.BookingExcelAttachmentExtractionRequest;
+import cn.nianxx.thhotel.workflows.reservation.excelimport.common.result.BookingExcelAttachmentExtractionResult;
+import cn.nianxx.thhotel.workflows.reservation.excelimport.service.ReservationBookingExcelAttachmentExtractionService;
 import com.fasterxml.jackson.core.JsonProcessingException;
 import com.fasterxml.jackson.core.type.TypeReference;
 import com.fasterxml.jackson.databind.JsonNode;
@@ -49,6 +52,7 @@ import java.security.MessageDigest;
 import java.security.NoSuchAlgorithmException;
 import java.time.Duration;
 import java.time.LocalDateTime;
+import java.time.ZoneId;
 import java.time.ZoneOffset;
 import java.time.format.DateTimeFormatter;
 import java.util.ArrayList;
@@ -84,6 +88,9 @@ public class DebugEmlSuperAgentRunServiceImpl implements DebugEmlSuperAgentRunSe
    private static final String SOURCE_CHANNEL = "EMAIL";
    private static final String REPLY_POLICY_MODE = "manual";
    private static final String SCHEMA_VERSION = "debug-eml-upload-v1";
+    private static final String PAYLOAD_FIELD_ATTACHMENT_EXTRACTIONS = "attachment_extractions";
+    private static final String BOOKING_EXCEL_EXTRACTION_DISABLED_REASON = "BOOKING_EXCEL_EXTRACTION_DISABLED";
+    private static final ZoneId DEFAULT_HOTEL_ZONE = ZoneId.of("Asia/Bangkok");
    private static final DateTimeFormatter DATE_FOLDER_FORMATTER = DateTimeFormatter.BASIC_ISO_DATE;
    private static final Pattern CID_REFERENCE_PATTERN = Pattern.compile(
            "(?i)cid:(?:<[^>]+>|%3c[^\\s\"'>]+%3e|[^\\s\"'>]+)");
@@ -103,6 +110,7 @@ public class DebugEmlSuperAgentRunServiceImpl implements DebugEmlSuperAgentRunSe
    private final EmlMessageParseService parseService;
    private final ObjectStorageService objectStorageService;
    private final SourceMessageCaptureService sourceMessageCaptureService;
+    private final ReservationBookingExcelAttachmentExtractionService bookingExcelExtractionService;
    private final SourceMessageInboxRepository sourceMessageInboxRepository;
    private final SourceMessageHtmlSanitizerService htmlSanitizerService;
    private final SuperAgentOpenApiClient superAgentOpenApiClient;
@@ -119,6 +127,7 @@ public class DebugEmlSuperAgentRunServiceImpl implements DebugEmlSuperAgentRunSe
            EmlMessageParseService parseService,
            ObjectStorageService objectStorageService,
            SourceMessageCaptureService sourceMessageCaptureService,
+            ReservationBookingExcelAttachmentExtractionService bookingExcelExtractionService,
            SourceMessageInboxRepository sourceMessageInboxRepository,
            SourceMessageHtmlSanitizerService htmlSanitizerService,
            SuperAgentOpenApiClient superAgentOpenApiClient,
@@ -130,6 +139,7 @@ public class DebugEmlSuperAgentRunServiceImpl implements DebugEmlSuperAgentRunSe
        this.parseService = parseService;
        this.objectStorageService = objectStorageService;
        this.sourceMessageCaptureService = sourceMessageCaptureService;
+        this.bookingExcelExtractionService = bookingExcelExtractionService;
        this.sourceMessageInboxRepository = sourceMessageInboxRepository;
        this.htmlSanitizerService = htmlSanitizerService;
        this.superAgentOpenApiClient = superAgentOpenApiClient;
@@ -486,7 +496,7 @@ public class DebugEmlSuperAgentRunServiceImpl implements DebugEmlSuperAgentRunSe
                parsed,
                htmlWithOssUrls,
                uploadedMedia);
-        String payloadJson = objectMapper.writeValueAsString(payload);
+        String sourcePayloadJson = objectMapper.writeValueAsString(payload);
        markStage(
                runId,
                DebugEmlSuperAgentRunStatus.CAPTURING_SOURCE_MESSAGE,
@@ -505,9 +515,13 @@ public class DebugEmlSuperAgentRunServiceImpl implements DebugEmlSuperAgentRunSe
                parsed.subject(),
                parsed.textBody(),
                htmlWithOssUrls,
-                payloadJson,
+                sourcePayloadJson,
                SCHEMA_VERSION,
                uploadedMedia.stream().map(UploadedMedia::captureMedia).toList()));
+        List<BookingExcelAttachmentExtractionResult> attachmentExtractions =
+                extractBookingExcelAttachments(runId, parsed.mediaItems(), createdAt, warnings);
+        Map<String, Object> superAgentPayload = payloadWithAttachmentExtractions(payload, attachmentExtractions);
+        String superAgentPayloadJson = objectMapper.writeValueAsString(superAgentPayload);
        updateSourceCapturedRun(
                runId,
                captureResult.inboxId(),
@@ -516,7 +530,7 @@ public class DebugEmlSuperAgentRunServiceImpl implements DebugEmlSuperAgentRunSe
                safeFileName,
                uploadedMedia.get(0).result().externalUrl(),
                sha256,
-                payloadJson);
+                superAgentPayloadJson);

        markStage(
                runId,
@@ -524,7 +538,7 @@ public class DebugEmlSuperAgentRunServiceImpl implements DebugEmlSuperAgentRunSe
                "阶段：调用 SuperAgent Open API，source_message_id=" + captureResult.inboxId() + "。",
                streamSink);
        SuperAgentMailDebugRequest superAgentRequest = new SuperAgentMailDebugRequest(
-                buildSuperAgentMessage(payloadJson),
+                buildSuperAgentMessage(superAgentPayloadJson),
                "debug-eml-" + runId,
                buildSuperAgentMetadata(
                        runId,
@@ -561,7 +575,7 @@ public class DebugEmlSuperAgentRunServiceImpl implements DebugEmlSuperAgentRunSe
                safeFileName,
                uploadedMedia.get(0).result().externalUrl(),
                sha256,
-                payloadJson,
+                superAgentPayloadJson,
                superAgentResult.sessionId(),
                superAgentResult.runId(),
                superAgentResult.profileId(),
@@ -589,7 +603,8 @@ public class DebugEmlSuperAgentRunServiceImpl implements DebugEmlSuperAgentRunSe
                htmlBodySanitized,
                true,
                htmlRenderMode,
-                payload,
+                superAgentPayload,
+                attachmentExtractions,
                superAgentResult.sessionId(),
                superAgentResult.runId(),
                superAgentResult.rawAnswer(),
@@ -607,6 +622,7 @@ public class DebugEmlSuperAgentRunServiceImpl implements DebugEmlSuperAgentRunSe
        SourceMessageOriginalContent originalContent = loadOriginalContent(snapshot.sourceMessageId());
        String htmlBodyWithOssUrls = originalContent == null ? null : originalContent.htmlBody();
        String htmlBodySanitized = htmlBodyWithOssUrls == null ? null : htmlSanitizerService.sanitizeHtml(htmlBodyWithOssUrls);
+        Map<String, Object> payload = parsePayloadJson(snapshot.payloadJson());
        return new DebugEmlSuperAgentRunResult(
                snapshot.id().toString(),
                snapshot.sourceMessageId() == null ? null : snapshot.sourceMessageId().toString(),
@@ -620,7 +636,8 @@ public class DebugEmlSuperAgentRunServiceImpl implements DebugEmlSuperAgentRunSe
                htmlBodySanitized,
                originalContent == null ? null : true,
                originalContent == null ? null : htmlSanitizerService.htmlRenderMode(htmlBodyWithOssUrls),
-                parsePayloadJson(snapshot.payloadJson()),
+                payload,
+                attachmentExtractionsFromPayload(payload),
                snapshot.superagentSessionId(),
                snapshot.superagentRunId(),
                snapshot.superagentRawAnswer(),
@@ -708,6 +725,99 @@ public class DebugEmlSuperAgentRunServiceImpl implements DebugEmlSuperAgentRunSe
        }
    }

+    /**
+     * 从 Debug EML 附件字节中抽取 Booking Excel 高亮行；该结果只追加给 Debug SuperAgent 输入和调试响应。
+     */
+    private List<BookingExcelAttachmentExtractionResult> extractBookingExcelAttachments(
+            Long runId,
+            List<ParsedEmlMediaItem> mediaItems,
+            LocalDateTime createdAt,
+            List<String> warnings) {
+        if (!properties.isIncludeBookingExcelExtractions() || mediaItems == null || mediaItems.isEmpty()) {
+            return List.of();
+        }
+        List<BookingExcelAttachmentExtractionResult> results = new ArrayList<>();
+        for (ParsedEmlMediaItem mediaItem : mediaItems) {
+            if (!SourceMessageMediaType.ATTACHMENT.code().equals(mediaItem.mediaType())) {
+                continue;
+            }
+            String attachmentName = safeFileName(mediaItem.fileName(), "attachment-" + (results.size() + 1));
+            if (!isExcelAttachmentName(attachmentName)) {
+                continue;
+            }
+            try {
+                byte[] content = mediaItem.bytes();
+                Long sizeBytes = mediaItem.sizeBytes() == null && content != null
+                        ? (long) content.length
+                        : mediaItem.sizeBytes();
+                BookingExcelAttachmentExtractionResult extractionResult =
+                        bookingExcelExtractionService.extract(new BookingExcelAttachmentExtractionRequest(
+                                attachmentName,
+                                mediaItem.contentType(),
+                                sizeBytes,
+                                content,
+                                createdAt.atOffset(ZoneOffset.UTC).toInstant(),
+                                DEFAULT_HOTEL_ZONE,
+                                null,
+                                null));
+                if (!BOOKING_EXCEL_EXTRACTION_DISABLED_REASON.equals(extractionResult.skippedReason())) {
+                    results.add(extractionResult);
+                }
+            } catch (RuntimeException exception) {
+                warnings.add("Booking Excel 附件预处理失败：" + attachmentName);
+                log.warn(
+                        "Debug EML Booking Excel extraction failed. debug_run_id={}, attachment_name={}, exception_chain={}",
+                        runId,
+                        attachmentName,
+                        safeExceptionChain(exception));
+            }
+        }
+        return List.copyOf(results);
+    }
+
+    /**
+     * 将附件预处理结果并入发给 SuperAgent 的 payload；无结果时沿用原始 payload。
+     */
+    private Map<String, Object> payloadWithAttachmentExtractions(
+            Map<String, Object> payload,
+            List<BookingExcelAttachmentExtractionResult> attachmentExtractions) {
+        if (attachmentExtractions == null || attachmentExtractions.isEmpty()) {
+            return payload;
+        }
+        Map<String, Object> enrichedPayload = new LinkedHashMap<>(payload);
+        enrichedPayload.put(PAYLOAD_FIELD_ATTACHMENT_EXTRACTIONS, attachmentExtractions);
+        return enrichedPayload;
+    }
+
+    /**
+     * 查询历史 Debug run 时，从已保存的 payload 中还原附件预处理结果。
+     */
+    private List<BookingExcelAttachmentExtractionResult> attachmentExtractionsFromPayload(Map<String, Object> payload) {
+        if (payload == null || !payload.containsKey(PAYLOAD_FIELD_ATTACHMENT_EXTRACTIONS)) {
+            return List.of();
+        }
+        Object value = payload.get(PAYLOAD_FIELD_ATTACHMENT_EXTRACTIONS);
+        if (!(value instanceof List<?>)) {
+            return List.of();
+        }
+        try {
+            return objectMapper.convertValue(
+                    value,
+                    new TypeReference<List<BookingExcelAttachmentExtractionResult>>() {
+                    });
+        } catch (IllegalArgumentException exception) {
+            return List.of();
+        }
+    }
+
+    /**
+     * 判断附件名是否为可预处理的 Excel 文件。
+     */
+    private boolean isExcelAttachmentName(String attachmentName) {
+        String lowerName = attachmentName == null ? "" : attachmentName.toLowerCase(Locale.ROOT);
+        return lowerName.endsWith(".xlsx") || lowerName.endsWith(".xls");
+    }
+
    /**
     * 解析已入库的 SuperAgent JSON；解析失败返回 null，前端仍可查看 raw answer。
     */