// Copyright 2022 E99p1ant. All rights reserved. // Use of this source code is governed by a MIT-style // license that can be found in the LICENSE file. package censor import ( "context" "strings" "sync" "time" "github.com/pkg/errors" "github.com/sirupsen/logrus" "github.com/tamakyi/TamaBox/internal/conf" "github.com/tamakyi/TamaBox/internal/db" ) var ( censorCacheNoMoreThan = 31 * 24 * time.Hour // 1 month // warnNoCensorCredentialsOnce 保证「没有可用凭据」的告警只打一次,避免按请求刷屏。 warnNoCensorCredentialsOnce sync.Once ) // censorChain 返回按 [security] text_censor_source 选定的通道列表,顺序即优先级: // 靠前通道判定通过即返回,未通过则交给后面的通道复核。 // qiniu / aliyun 只用对应通道(凭据齐备性已由 conf.Init 启动校验); // 其它取值(含留空 / auto)为七牛优先、阿里云兜底,与引入该配置前一致。 func censorChain() []TextCensor { qiniu := NewQiniuTextCensor(conf.App.QiniuAccessKey, conf.App.QiniuAccessSecret) aliyun := NewAliyunTextCensor(conf.App.AliyunAccessKey, conf.App.AliyunAccessKeySecret) switch strings.ToLower(strings.TrimSpace(conf.Security.TextCensorSource)) { case conf.CensorSourceQiniu: return []TextCensor{qiniu} case conf.CensorSourceAliyun: return []TextCensor{aliyun} default: return []TextCensor{qiniu, aliyun} } } // warnIfNoCensorCredentials 在一条可用凭据都没有时告警一次:此时审核必然失败, // 而调用方是 fail-closed(拒绝提交),表现为全站发不出内容/私信且只在日志可见。 // 显式选定通道但缺凭据已由 conf.Init 挡住,这里兜的是 auto 的漏网情况。 func warnIfNoCensorCredentials(ctx context.Context) { qiniuReady := conf.App.QiniuAccessKey != "" && conf.App.QiniuAccessSecret != "" aliyunReady := conf.App.AliyunAccessKey != "" && conf.App.AliyunAccessKeySecret != "" if qiniuReady || aliyunReady { return } warnNoCensorCredentialsOnce.Do(func() { logrus.WithContext(ctx).WithField("text_censor_source", conf.Security.TextCensorSource). Error("Text censor is enabled but no usable credentials are configured; all submissions will be rejected (fail-closed)") }) } // Text 检查文本是否含敏感内容,并把审核结果落库(31 天内可回放复用)。 func Text(ctx context.Context, text string) (*TextCensorResponse, error) { if !conf.Security.EnableTextCensor { return &TextCensorResponse{Pass: true}, nil } // 缓存键与送审内容统一用归一化文本(已剔除可信域名链接):否则缓存按原文命中、 // 实际送审的却是归一化结果 —— 审计对不上,且归一化后相同的原文会各占一条缓存、重复调 API。 normalized := RemoveTrustedURL(text) var responses []*TextCensorResponse warnIfNoCensorCredentials(ctx) for _, censor := range censorChain() { sourceName := censor.String() censorLog, err := db.CensorLogs.GetByText(ctx, sourceName, normalized, time.Now().Add(-censorCacheNoMoreThan)) if err != nil { if !errors.Is(err, db.ErrCensorLogsNotFound) { logrus.WithContext(ctx).WithError(err).Error("Failed to get censor log") } } else { raw := censorLog.RawResponse switch censorLog.SourceName { case "qiniu": // 只有「通过」才直接复用(省一次 API 调用);未通过则交给链上下一个通道复核 // (auto 模式下即阿里云,引入通道配置前就是如此)。 // text_censor_source=qiniu 时链上只有七牛,continue 后循环结束, // 仍由函数末尾返回这份结果,判定不会丢失。 response, err := QiniuTextCensorParser(raw) if err == nil { if response.Pass { return response, nil } responses = append(responses, response) continue } case "aliyun": // 能解析就复用;解析失败必须退回实时调用而不是把错误往上抛: // 最典型的失败源是迁到内容安全 2.0 前落库的 1.0 缓存,无法按 2.0 语义解释, // 若在此报错,fail-closed 会让这些历史文本在缓存过期前一直发不出去。 // 与七牛一致:仅「通过」才直接返回,未通过则交给链上下一个通道复核 // (阿里云现为链尾,continue 后循环结束,仍由函数末尾返回同一份结果)。 response, err := AliyunTextCensorParser(raw) if err == nil { if response.Pass { return response, nil } responses = append(responses, response) continue } logrus.WithContext(ctx).WithError(err).WithField("censor_source", sourceName). Warn("Failed to replay cached aliyun censor result; falling back to a live call") default: logrus.WithContext(ctx).WithField("censor_source", censorLog.SourceName).Error("Unknown censor source") } } response, err := censor.Censor(ctx, normalized) if err != nil { logrus.WithContext(ctx).WithError(err).WithField("censor_source", sourceName).Error("Failed to censor text") continue } responses = append(responses, response) if err := db.CensorLogs.Create(ctx, db.CreateCensorLogOptions{ SourceName: sourceName, Input: normalized, Pass: response.Pass, RawResponse: response.RawResponse, }); err != nil { logrus.WithContext(ctx).WithError(err).Error("Failed to save censor log") } if response.Pass { return response, nil } } if len(responses) == 0 { return nil, errors.New("no censor source available") } return responses[len(responses)-1], nil }