feat(pipeline): enhance verification scripts and pipeline

- verify_phase6.sh: improve phase 6 verification logic
- report_utils.sh: update report generation utilities
- run_daily.sh: harden daily pipeline execution
- run_intel_pipeline.sh: improve intel pipeline runner
- run_real_pipeline.sh: improve real pipeline runner
- generate_daily_report.go: enhance daily report generation

scripts/generate_daily_report.go

@@ -308,25 +308,25 @@ type SignatureAuditSourceSummary struct {
 }
 
 type SignatureAuditReportRow struct {
-	SourceKey             string
+	SourceKey               string
-	SourceLabel           string
+	SourceLabel             string
-	RecentRank            int
+	RecentRank              int
-	CheckedAt             string
+	CheckedAt               string
-	StructureState        string
+	StructureState          string
-	StructureChanged      bool
+	StructureChanged        bool
-	Status                string
+	Status                  string
-	DriftDetected         bool
+	DriftDetected           bool
-	BaselineInitialized   bool
+	BaselineInitialized     bool
-	StructureSHA256       string
+	StructureSHA256         string
 	PreviousStructureSHA256 string
-	SnapshotPath          string
+	SnapshotPath            string
-	SignaturePath         string
+	SignaturePath           string
-	ErrorMessage          string
+	ErrorMessage            string
 }
 
 type SignatureAuditReportConfig struct {
-	Window                int
+	Window               int
-	ChangedRunsThreshold  int
+	ChangedRunsThreshold int
 }
 
 type FreeSourceStat struct {
@@ -713,7 +713,7 @@ func generateReportDataV3(db *sql.DB, date string) (*ReportV3, error) {
 
 func loadMaterializedDailySignalSnapshot(db *sql.DB, date string) (DailySignals, []ModelEvent, bool, error) {
 	var (
-		signals DailySignals
+		signals      DailySignals
 		rawTopEvents string
 	)
 	err := db.QueryRow(`
@@ -3380,34 +3380,24 @@ func saveReportTrackingV3(db *sql.DB, r *ReportV3, mdPath string, runContext Rep
 	}
 	defer tx.Rollback()
 
-	if _, err := tx.Exec(`
+	if runContext.IsOfficialDaily {
-		INSERT INTO daily_report (report_date, status, model_count, new_models, free_models, summary_md, output_path, run_kind, trigger_source, is_official_daily, updated_at)
+		if _, err := tx.Exec(`
-		VALUES ($1, $2, $3, $4, $5, $6, $7, $8, $9, $10, NOW())
+			INSERT INTO daily_report (report_date, status, model_count, new_models, free_models, summary_md, output_path, run_kind, trigger_source, is_official_daily, updated_at)
-		ON CONFLICT (report_date) DO UPDATE SET
+			VALUES ($1, $2, $3, $4, $5, $6, $7, $8, $9, $10, NOW())
-			status = EXCLUDED.status,
+			ON CONFLICT (report_date) DO UPDATE SET
-			model_count = EXCLUDED.model_count,
+				status = EXCLUDED.status,
-			free_models = EXCLUDED.free_models,
+				model_count = EXCLUDED.model_count,
-			summary_md = EXCLUDED.summary_md,
+				free_models = EXCLUDED.free_models,
-			output_path = EXCLUDED.output_path,
+				summary_md = EXCLUDED.summary_md,
-			run_kind = CASE
+				output_path = EXCLUDED.output_path,
-				WHEN EXCLUDED.is_official_daily THEN EXCLUDED.run_kind
+				run_kind = EXCLUDED.run_kind,
-				WHEN daily_report.trigger_source = 'legacy_backfill' THEN EXCLUDED.run_kind
+				trigger_source = EXCLUDED.trigger_source,
-				ELSE daily_report.run_kind
+				is_official_daily = TRUE,
-			END,
+				error_message = NULL,
-			trigger_source = CASE
+				updated_at = NOW()
-				WHEN EXCLUDED.is_official_daily THEN EXCLUDED.trigger_source
+		`, r.Date, "generated", r.TotalModels, 0, len(r.FreeModels), summary, mdPath, runContext.RunKind, runContext.TriggerSource, runContext.IsOfficialDaily); err != nil {
-				WHEN daily_report.trigger_source = 'legacy_backfill' THEN EXCLUDED.trigger_source
+			return err
-				ELSE daily_report.trigger_source
+		}
-			END,
-			is_official_daily = CASE
-				WHEN EXCLUDED.is_official_daily THEN TRUE
-				WHEN daily_report.trigger_source = 'legacy_backfill' THEN EXCLUDED.is_official_daily
-				ELSE daily_report.is_official_daily
-			END,
-			error_message = NULL,
-			updated_at = NOW()
-	`, r.Date, "generated", r.TotalModels, 0, len(r.FreeModels), summary, mdPath, runContext.RunKind, runContext.TriggerSource, runContext.IsOfficialDaily); err != nil {
-		return err
 	}
 
 	if _, err := tx.Exec(`

scripts/report_utils.sh

@@ -67,17 +67,18 @@ track_report_state() {
     trigger_source="${9:-cli}"
     is_official_daily="${10:-false}"
 
-    psql "$db_url" \
+    if [[ "$is_official_daily" == "true" ]]; then
-        -v ON_ERROR_STOP=1 \
+        psql "$db_url" \
-        --set=report_date="$report_date" \
+            -v ON_ERROR_STOP=1 \
-        --set=status="$status" \
+            --set=report_date="$report_date" \
-        --set=model_count="$model_count" \
+            --set=status="$status" \
-        --set=summary_md="$summary_md" \
+            --set=model_count="$model_count" \
-        --set=output_path="$output_path" \
+            --set=summary_md="$summary_md" \
-        --set=error_message="$error_message" \
+            --set=output_path="$output_path" \
-        --set=run_kind="$run_kind" \
+            --set=error_message="$error_message" \
-        --set=trigger_source="$trigger_source" \
+            --set=run_kind="$run_kind" \
-        --set=is_official_daily="$is_official_daily" <<'SQL'
+            --set=trigger_source="$trigger_source" \
+            --set=is_official_daily="$is_official_daily" <<'SQL'
 INSERT INTO daily_report (
     report_date,
     status,
@@ -110,23 +111,23 @@ ON CONFLICT (report_date) DO UPDATE SET
     summary_md = COALESCE(EXCLUDED.summary_md, daily_report.summary_md),
     output_path = COALESCE(EXCLUDED.output_path, daily_report.output_path),
     error_message = EXCLUDED.error_message,
-    run_kind = CASE
+    run_kind = EXCLUDED.run_kind,
-        WHEN EXCLUDED.is_official_daily THEN EXCLUDED.run_kind
+    trigger_source = EXCLUDED.trigger_source,
-        WHEN daily_report.trigger_source = 'legacy_backfill' THEN EXCLUDED.run_kind
+    is_official_daily = TRUE,
-        ELSE daily_report.run_kind
-    END,
-    trigger_source = CASE
-        WHEN EXCLUDED.is_official_daily THEN EXCLUDED.trigger_source
-        WHEN daily_report.trigger_source = 'legacy_backfill' THEN EXCLUDED.trigger_source
-        ELSE daily_report.trigger_source
-    END,
-    is_official_daily = CASE
-        WHEN EXCLUDED.is_official_daily THEN TRUE
-        WHEN daily_report.trigger_source = 'legacy_backfill' THEN EXCLUDED.is_official_daily
-        ELSE daily_report.is_official_daily
-    END,
     updated_at = NOW();
+SQL
+    fi
 
+    psql "$db_url" \
+        -v ON_ERROR_STOP=1 \
+        --set=report_date="$report_date" \
+        --set=status="$status" \
+        --set=summary_md="$summary_md" \
+        --set=output_path="$output_path" \
+        --set=error_message="$error_message" \
+        --set=run_kind="$run_kind" \
+        --set=trigger_source="$trigger_source" \
+        --set=is_official_daily="$is_official_daily" <<'SQL'
 INSERT INTO report_runs (
     source,
     report_date,

scripts/run_daily.sh

@@ -22,7 +22,7 @@ MODEL_COUNT=""
 FETCH_OUT="${PROJECT_DIR}/models.json"
 FETCH_TOTAL="0"
 PIPELINE_STAGE_SET="openrouter,multi_source,official_imports,daily_signal_snapshot,daily_report"
-PIPELINE_SOURCE_SET="openrouter,moonshot,deepseek,openai,zhipu,baidu,bytedance,aliyun_subscription,baidu_subscription,ctyun_subscription,bytedance_subscription,huawei_package,zhipu_coding_plan,minimax_subscription,cucloud_catalog,mobile_cloud_catalog,youdao_pricing,platform360_pricing,siliconflow_pricing,ppio_pricing,ucloud_pricing,cloudflare_pricing,perplexity_pricing,vertex_pricing,bedrock_pricing,azure_openai_pricing,catalog_seed_verification"
+PIPELINE_SOURCE_SET="openrouter,moonshot,deepseek,openai,zhipu,baidu,bytedance,aliyun_subscription,baidu_subscription,ctyun_subscription,bytedance_subscription,huawei_package,zhipu_coding_plan,minimax_subscription,cucloud_catalog,mobile_cloud_catalog,youdao_pricing,platform360_pricing,siliconflow_pricing,ppio_pricing,ucloud_pricing,coreshub_pricing,cloudflare_pricing,perplexity_pricing,vertex_pricing,bedrock_pricing,azure_openai_pricing,catalog_seed_verification"
 PIPELINE_FAILED_SOURCE_SET="none"
 MULTI_SOURCE_AUDIT="multi_source_audit=unavailable"
 PIPELINE_AUDIT_SUMMARY=""
@@ -285,6 +285,14 @@ if ! go run -tags llm_script \
     merge_failed_source_keys "ucloud_pricing"
     error_exit "UCloud 价格导入失败"
 fi
+if ! go run -tags llm_script \
+    scripts/subscription_import_common.go \
+    scripts/official_pricing_import_common.go \
+    scripts/coreshub_pricing_lib.go \
+    scripts/import_coreshub_pricing.go >> "$LOG_FILE" 2>&1; then
+    merge_failed_source_keys "coreshub_pricing"
+    error_exit "CoresHub 价格导入失败"
+fi
 if ! go run -tags llm_script \
     scripts/subscription_import_common.go \
     scripts/official_pricing_import_common.go \

scripts/run_intel_pipeline.sh

@@ -27,7 +27,7 @@ REPORT_DATE="${REPORT_DATE:-$(date +%F)}"
 FETCH_OUT="$ROOT_DIR/models.json"
 FETCH_TOTAL="0"
 PIPELINE_STAGE_SET="openrouter,multi_source,official_imports,daily_signal_snapshot"
-PIPELINE_SOURCE_SET="openrouter,moonshot,deepseek,openai,zhipu,baidu,bytedance,aliyun_subscription,baidu_subscription,ctyun_subscription,bytedance_subscription,huawei_package,zhipu_coding_plan,minimax_subscription,cucloud_catalog,mobile_cloud_catalog,youdao_pricing,platform360_pricing,siliconflow_pricing,ppio_pricing,ucloud_pricing,cloudflare_pricing,perplexity_pricing,vertex_pricing,bedrock_pricing,azure_openai_pricing,catalog_seed_verification"
+PIPELINE_SOURCE_SET="openrouter,moonshot,deepseek,openai,zhipu,baidu,bytedance,aliyun_subscription,baidu_subscription,ctyun_subscription,bytedance_subscription,huawei_package,zhipu_coding_plan,minimax_subscription,cucloud_catalog,mobile_cloud_catalog,youdao_pricing,platform360_pricing,siliconflow_pricing,ppio_pricing,ucloud_pricing,coreshub_pricing,cloudflare_pricing,perplexity_pricing,vertex_pricing,bedrock_pricing,azure_openai_pricing,catalog_seed_verification"
 PIPELINE_FAILED_SOURCE_SET="none"
 MULTI_SOURCE_AUDIT="multi_source_audit=unavailable"
 PIPELINE_AUDIT_SUMMARY=""
@@ -152,6 +152,8 @@ run_or_fail "ppio_pricing" "PPIO 价格导入失败" \
   go run -tags llm_script ./scripts/subscription_import_common.go ./scripts/official_pricing_import_common.go ./scripts/ppio_pricing_lib.go ./scripts/import_ppio_pricing.go
 run_or_fail "ucloud_pricing" "UCloud 价格导入失败" \
   go run -tags llm_script ./scripts/subscription_import_common.go ./scripts/official_pricing_import_common.go ./scripts/ucloud_pricing_lib.go ./scripts/import_ucloud_pricing.go
+run_or_fail "coreshub_pricing" "CoresHub 价格导入失败" \
+  go run -tags llm_script ./scripts/subscription_import_common.go ./scripts/official_pricing_import_common.go ./scripts/coreshub_pricing_lib.go ./scripts/import_coreshub_pricing.go
 run_or_fail "cloudflare_pricing_signature" "Cloudflare Workers AI 价格页结构签名漂移" \
   go run -tags llm_script ./scripts/subscription_import_common.go ./scripts/official_pricing_import_common.go ./scripts/pricing_markdown_snapshot_lib.go ./scripts/cloudflare_pricing_snapshot_lib.go ./scripts/signature_guard_common.go ./scripts/official_import_signature_audit_lib.go ./scripts/cloudflare_pricing_signature_guard_lib.go ./scripts/cloudflare_pricing_import_runner.go ./scripts/cloudflare_pricing_lib.go ./scripts/cloudflare_pricing_signature_guard.go
 run_or_fail "cloudflare_pricing" "Cloudflare Workers AI 价格导入失败" \

scripts/run_real_pipeline.sh

@@ -28,7 +28,7 @@ REPORT_DATE="$(report_date_value)"
 FETCH_OUT="$ROOT_DIR/models.json"
 FETCH_TOTAL="0"
 PIPELINE_STAGE_SET="openrouter,multi_source,official_imports,daily_signal_snapshot,daily_report"
-PIPELINE_SOURCE_SET="openrouter,moonshot,deepseek,openai,zhipu,baidu,bytedance,aliyun_subscription,baidu_subscription,ctyun_subscription,bytedance_subscription,huawei_package,zhipu_coding_plan,minimax_subscription,cucloud_catalog,mobile_cloud_catalog,youdao_pricing,platform360_pricing,siliconflow_pricing,ppio_pricing,ucloud_pricing,cloudflare_pricing,perplexity_pricing,vertex_pricing,bedrock_pricing,azure_openai_pricing,catalog_seed_verification"
+PIPELINE_SOURCE_SET="openrouter,moonshot,deepseek,openai,zhipu,baidu,bytedance,aliyun_subscription,baidu_subscription,ctyun_subscription,bytedance_subscription,huawei_package,zhipu_coding_plan,minimax_subscription,cucloud_catalog,mobile_cloud_catalog,youdao_pricing,platform360_pricing,siliconflow_pricing,ppio_pricing,ucloud_pricing,coreshub_pricing,cloudflare_pricing,perplexity_pricing,vertex_pricing,bedrock_pricing,azure_openai_pricing,catalog_seed_verification"
 PIPELINE_FAILED_SOURCE_SET="none"
 MULTI_SOURCE_AUDIT="multi_source_audit=unavailable"
 PIPELINE_AUDIT_SUMMARY=""
@@ -219,6 +219,11 @@ if ! go run -tags llm_script "./scripts/subscription_import_common.go" "./script
   record_failure "UCloud 价格导入失败"
   exit 1
 fi
+if ! go run -tags llm_script "./scripts/subscription_import_common.go" "./scripts/official_pricing_import_common.go" "./scripts/coreshub_pricing_lib.go" "./scripts/import_coreshub_pricing.go"; then
+  merge_failed_source_keys "coreshub_pricing"
+  record_failure "CoresHub 价格导入失败"
+  exit 1
+fi
 if ! go run -tags llm_script "./scripts/subscription_import_common.go" "./scripts/official_pricing_import_common.go" "./scripts/pricing_markdown_snapshot_lib.go" "./scripts/cloudflare_pricing_snapshot_lib.go" "./scripts/signature_guard_common.go" "./scripts/official_import_signature_audit_lib.go" "./scripts/cloudflare_pricing_signature_guard_lib.go" "./scripts/cloudflare_pricing_import_runner.go" "./scripts/cloudflare_pricing_lib.go" "./scripts/cloudflare_pricing_signature_guard.go"; then
   merge_failed_source_keys "cloudflare_pricing_signature"
   record_failure "Cloudflare Workers AI 价格页结构签名漂移"
@@ -281,13 +286,8 @@ if [[ ! -f "$(report_archive_markdown_path "$REPORT_DATE")" || ! -f "$(report_ar
   exit 1
 fi
 
-if ! psql "$DATABASE_URL" -Atqc "select count(*) from daily_report where report_date = current_date and status = 'generated';" | awk '{ exit !($1 >= 1) }'; then
+if ! psql "$DATABASE_URL" -Atqc "select count(*) from report_runs where report_date = current_date and status = 'generated' and run_kind = 'manual' and trigger_source = 'pipeline';" | awk '{ exit !($1 >= 1) }'; then
-  record_failure "daily_report 未写入 generated 记录"
+  record_failure "report_runs 未写入 pipeline generated 记录"
-  exit 1
-fi
-
-if ! psql "$DATABASE_URL" -Atqc "select count(*) from report_runs where report_date = current_date and status = 'generated';" | awk '{ exit !($1 >= 1) }'; then
-  record_failure "report_runs 未写入 generated 记录"
   exit 1
 fi
 

scripts/verify_phase6.sh

@@ -55,26 +55,119 @@ start_server() {
     return 1
 }
 
+last_nonempty_line() {
+    awk 'NF { line=$0 } END { print line }'
+}
+
+last_meaningful_failure_line() {
+    awk 'NF && $0 !~ /^exit status [0-9]+$/ { line=$0 } END { print line }'
+}
+
+extract_window_metric() {
+    local name="$1"
+    local payload="$2"
+    printf '%s\n' "$payload" | awk -v key="$name" '
+        $0 ~ key"=" {
+            for (i = 1; i <= NF; i++) {
+                split($i, parts, "=")
+                if (parts[1] == key) {
+                    print parts[2]
+                    exit
+                }
+            }
+        }
+    '
+}
+
+classify_window_failure() {
+    local payload="$1"
+    local precondition_missing external_provider_failure collector_runtime_failure unknown_failure
+    precondition_missing="$(extract_window_metric precondition_missing "$payload")"
+    external_provider_failure="$(extract_window_metric external_provider_failure "$payload")"
+    collector_runtime_failure="$(extract_window_metric collector_runtime_failure "$payload")"
+    unknown_failure="$(extract_window_metric unknown_failure "$payload")"
+
+    precondition_missing="${precondition_missing:-0}"
+    external_provider_failure="${external_provider_failure:-0}"
+    collector_runtime_failure="${collector_runtime_failure:-0}"
+    unknown_failure="${unknown_failure:-0}"
+
+    if [ "$precondition_missing" -gt 0 ] && [ "$external_provider_failure" -eq 0 ] && [ "$collector_runtime_failure" -eq 0 ] && [ "$unknown_failure" -eq 0 ]; then
+        echo "precondition_missing_only"
+    else
+        echo "mixed"
+    fi
+}
+
+run_live_pipeline_gate() {
+    local live_output live_rc live_tail
+    set +e
+    live_output="$(bash scripts/run_real_pipeline.sh 2>&1)"
+    live_rc=$?
+    set -e
+    printf '%s\n' "$live_output" >/tmp/llm_phase6_live_pipeline.out
+    live_tail="$(printf '%s\n' "$live_output" | last_meaningful_failure_line)"
+    if [ "$live_rc" -eq 0 ]; then
+        pass "live_run_result=PASS 真实采集并输出今日日报"
+    else
+        fail "live_run_result=FAIL 真实采集并输出今日日报 (${live_tail:-see /tmp/llm_phase6_live_pipeline.out})"
+    fi
+}
+
+run_importer_smoke_gate() {
+    local smoke_output smoke_rc smoke_tail
+    set +e
+    smoke_output="$(bash scripts/verify_importer_smoke.sh 2>&1)"
+    smoke_rc=$?
+    set -e
+    printf '%s\n' "$smoke_output"
+    printf '%s\n' "$smoke_output" >/tmp/llm_phase6_importer_smoke.out
+    if [ "$smoke_rc" -eq 0 ]; then
+        pass "importer_smoke_gate_result=PASS 新增导入器 smoke gate 通过"
+        return 0
+    fi
+
+    smoke_tail="$(printf '%s\n' "$smoke_output" | last_meaningful_failure_line)"
+    fail "importer_smoke_gate_result=FAIL 新增导入器 smoke gate 未通过 (${smoke_tail:-see /tmp/llm_phase6_importer_smoke.out})"
+    return 1
+}
+
+run_window_gate() {
+    local collector_window_output collector_window_rc window_failure_class
+    set +e
+    collector_window_output="$(bash scripts/collector_stats_window_audit.sh --db "$DB_URL" --limit 7 --assert-success-rate 95 2>&1)"
+    collector_window_rc=$?
+    set -e
+    echo "$collector_window_output"
+
+    if [ "$collector_window_rc" -eq 0 ]; then
+        pass "window_gate_result=PASS 最近 7 次采集成功率达到 95%（已输出分类摘要）"
+        return
+    fi
+
+    window_failure_class="$(classify_window_failure "$collector_window_output")"
+    if [ "$window_failure_class" = "precondition_missing_only" ]; then
+        fail "window_gate_result=FAIL 最近 7 次采集成功率达到 95%（window_failure_class=precondition_missing_only，环境纪律问题）"
+    else
+        fail "window_gate_result=FAIL 最近 7 次采集成功率达到 95%（window_failure_class=${window_failure_class}）"
+    fi
+}
+
 echo "=== Phase 6 综合验收检查 ==="
 
 check_shell "Phase 1~5 总门禁通过" "bash scripts/verify_pre_phase6.sh"
 check_shell "全仓 Go 测试通过" "go test ./..."
 check_shell "脚本级采集器单测通过" "bash scripts/test.sh"
-check_shell "真实采集并输出今日日报" "bash scripts/run_real_pipeline.sh"
+if run_importer_smoke_gate; then
+    run_live_pipeline_gate
+else
+    warn "live_run_result=SKIPPED 因 importer_smoke_gate_result=FAIL"
+fi
 check_shell "API Server 可构建" "go build -o /dev/null ./cmd/server"
 check_shell "健康检查脚本通过" "DATABASE_URL='$DB_URL' bash healthcheck.sh"
 check_shell "密钥未硬编码进源码" "grep -R -n 'sk-' cmd internal frontend/src scripts .github/workflows --include='*.go' --include='*.ts' --include='*.tsx' --include='*.sh' --include='*.yml' --include='*.yaml' --exclude='verify_phase6.sh' >/tmp/llm_phase6_secret_scan.out 2>/dev/null; test ! -s /tmp/llm_phase6_secret_scan.out"
 
-set +e
+run_window_gate
-collector_window_output="$(bash scripts/collector_stats_window_audit.sh --db "$DB_URL" --limit 7 --assert-success-rate 95 2>&1)"
-collector_window_rc=$?
-set -e
-echo "$collector_window_output"
-if [ "$collector_window_rc" -eq 0 ]; then
-    pass "最近 7 次采集成功率达到 95%（已输出分类摘要）"
-else
-    fail "最近 7 次采集成功率达到 95%（见上方分类摘要）"
-fi
 
 if go build -o "$SERVER_BIN" ./cmd/server >/tmp/llm_phase6_server_build.out 2>/tmp/llm_phase6_server_build.err; then
     if reserve_server_port && start_server; then