fix(glossary): 绑定 QA identity 与人工 override
bat-rust / Build and test Rust (push) Canceled after 0s
bat-rust / Build and test Go API (push) Canceled after 0s

This commit is contained in:
2026-09-09 00:25:49 +08:00
parent 0275a890bc
commit 68c6c91b1e
22 changed files with 844 additions and 92 deletions
+2 -1
View File
@@ -42,7 +42,8 @@ typed 管理转发。当前下载实现使用默认 8 个独立 worker,完成
Rust `bat` 持有 term/alias/recommended/allowed/category/priority、全局或
TextUnit scope、source history 和 approved review。worker、TM 复用、人工 task
结果和 workbench publish 都执行确定性 QAblocking deviation 必须携带
reviewer/reason/provenance 的显式 override`translation.glossary.*` 已通过
稳定 `qa_identity` 以及 reviewer/reason/provenance 的显式 override,所有接受路径都会
按当前 QA 精确校验 identity。`translation.glossary.*` 已通过
`bat.sock` 暴露,Go `bat-api` 仅做鉴权 typed forwarding。
---
Generated
+1
View File
@@ -118,6 +118,7 @@ version = "1.0.0"
dependencies = [
"anyhow",
"async-trait",
"blake3",
"serde",
"serde_json",
"tempfile",
+5 -4
View File
@@ -154,7 +154,7 @@ BAT_API_SKIP_ENV_FILE=1 go run ./cmd/bat-api \
| `GET /admin/translation/memory/query?source_text=...&source_context=...&limit=100` | 按 raw source/context 查询 Rust TM 记录、复用判定和 provenance;需要管理 token |
| `GET /admin/translation/glossary/summary` | 读取 Rust Glossary schema 和 review-state 计数;需要管理 token |
| `GET /admin/translation/glossary/query?source_text=...&review_status=approved&limit=100` | 查询 Rust term、scope、source provenance 和 history;需要管理 token |
| `GET /admin/translation/glossary/diagnose?source_text=...&context=...` | 执行 deterministic Glossary constraints/diagnostics QA;需要管理 token |
| `GET /admin/translation/glossary/diagnose?source_text=...&context=...` | 执行 deterministic Glossary constraints/diagnostics QA 并返回 `qa_identity`;需要管理 token |
| `GET /admin/translation/status` | 读取当前汉化 release、current 指针和 workflow 状态;需要管理 token |
| `POST /admin/control/{action}` | 经白名单转发 Rust `bat` 控制请求;见下文 |
@@ -186,7 +186,7 @@ launcher 兼容端点只服务启动前资源发现。它们复用 Rust `bat` sn
| `schedule-remove` | `schedule.remove` | `{ "id": "..." }` | `202` + Rust schedule report |
| `schedule-run` | `schedule.run` | 可选 `{ "id": "...", "force": true }` | `202` + 执行报告 |
| `task-cancel` | `task.cancel` | `{ "task_id": "..." }` | `202` + 取消请求结果 |
| `translation-task-update` | `translation.task.update` | `{ "task_id": "...", "status": "completed", "provider": "manual", "provider_run_id": "...", "translation_results": [{ "unit_id": "...", "source_text": "...", "translated_text": "..." }] }` | `202` + 当前任务记录 |
| `translation-task-update` | `translation.task.update` | `{ "task_id": "...", "status": "completed", "provider": "manual", "provider_run_id": "...", "translation_results": [{ "unit_id": "...", "source_text": "...", "translated_text": "...", "glossary_override": { "qa_identity": "...", "reviewer": "...", "reason": "...", "provenance": "...", "confirmed_unix_seconds": 1 } }] }` | `202` + 当前任务记录 |
| `translation-worker-run` | `translation.worker.run` | `{ "provider": "mock", "concurrency": 8, "max_tasks": 2 }` | `202` + worker task |
| `translation-proofread` | `translation.proofread` | 无 | `202` + 汉化状态 |
| `translation-memory-confirm` | `translation.memory.confirm` | `{ "record_id": "...", "reviewer": "...", "reason": "..." }` | `202` + 已确认的 TM 记录 |
@@ -346,8 +346,9 @@ bat i18n glossary delete --glossary-term-id term-sensei \
scope 为空表示全局;同一 TextUnit 内冲突会 blockedpriority、scope specificity、
匹配长度和 term ID 使用确定性排序。允许但非推荐译法只产生 warning,系统不会在译文
生成后自动替换文本。provider、TM、人工 task update、workbench 和 localized publish
均执行相同 QAblocking deviation 需要 `reviewer``reason``provenance` 和确认时间
组成显式 override
均执行相同 QA每个具体 QA 都有稳定的 `qa_identity`。blocking deviation 需要
`qa_identity``reviewer``reason``provenance` 和确认时间组成显式 override
Glossary 相关定义变化会使受影响 override 失效,无关术语变化不会使其失效。
---
+5 -2
View File
@@ -527,7 +527,7 @@ paths:
type: string
responses:
"200":
description: Glossary constraints, diagnostics, and blocked decision.
description: Glossary constraints, diagnostics, blocked decision, and stable qa_identity.
"400":
description: Missing source text or invalid context.
"401":
@@ -621,9 +621,12 @@ paths:
type: string
glossary_override:
type: object
required: [reviewer, reason, provenance, confirmed_unix_seconds]
required: [qa_identity, reviewer, reason, provenance, confirmed_unix_seconds]
additionalProperties: false
properties:
qa_identity:
type: string
minLength: 1
reviewer:
type: string
reason:
+1
View File
@@ -12,6 +12,7 @@ serde.workspace = true
serde_json.workspace = true
async-trait.workspace = true
tokio.workspace = true
blake3.workspace = true
[dev-dependencies]
tokio = { workspace = true, features = ["test-util", "macros"] }
+380 -13
View File
@@ -268,6 +268,9 @@ pub struct GlossaryDiagnostic {
/// Result of applying approved terms to one TextUnit source.
#[derive(Debug, Clone, PartialEq, Eq, serde::Serialize, serde::Deserialize)]
pub struct GlossaryEvaluation {
/// Stable identity of the glossary facts used to produce this evaluation.
#[serde(default)]
pub qa_identity: String,
/// Approved constraints sent to a provider.
#[serde(default, skip_serializing_if = "Vec::is_empty")]
pub constraints: Vec<GlossaryConstraint>,
@@ -326,20 +329,30 @@ impl GlossaryEvaluation {
});
}
}
GlossaryQaReport {
status: if blocked {
GlossaryQaStatus::Blocked
} else if diagnostics
.iter()
.any(|diagnostic| diagnostic.kind == GlossaryDiagnosticKind::NonRecommended)
{
GlossaryQaStatus::Warning
} else {
GlossaryQaStatus::Pass
},
let status = if blocked {
GlossaryQaStatus::Blocked
} else if diagnostics
.iter()
.any(|diagnostic| diagnostic.kind == GlossaryDiagnosticKind::NonRecommended)
{
GlossaryQaStatus::Warning
} else {
GlossaryQaStatus::Pass
};
let mut report = GlossaryQaReport {
qa_identity: String::new(),
status,
constraints: self.constraints.clone(),
diagnostics,
}
};
report.qa_identity = output_qa_identity(
&self.qa_identity,
translated_text,
&report.status,
&report.constraints,
&report.diagnostics,
);
report
}
}
@@ -374,6 +387,9 @@ impl GlossaryQaStatus {
/// Persisted glossary QA attached to a translation result.
#[derive(Debug, Clone, PartialEq, Eq, serde::Serialize, serde::Deserialize)]
pub struct GlossaryQaReport {
/// Stable identity of the complete blocking-QA result.
#[serde(default)]
pub qa_identity: String,
/// QA status.
pub status: GlossaryQaStatus,
/// Constraints evaluated.
@@ -387,6 +403,9 @@ pub struct GlossaryQaReport {
/// Explicit human approval to deviate from a blocking glossary result.
#[derive(Debug, Clone, PartialEq, Eq, serde::Serialize, serde::Deserialize)]
pub struct GlossaryOverride {
/// Exact blocking-QA identity reviewed by the human.
#[serde(default)]
pub qa_identity: String,
/// Reviewer identity.
pub reviewer: String,
/// Required reason.
@@ -407,6 +426,7 @@ pub fn evaluate_glossary(
struct Candidate {
term_id: String,
matched_source: String,
definition: GlossaryTermSnapshot,
recommendation: String,
allowed: Vec<String>,
category: Option<String>,
@@ -436,6 +456,7 @@ pub fn evaluate_glossary(
candidates.push(Candidate {
term_id: term.term_id.clone(),
matched_source: spelling.clone(),
definition: term.definition.clone(),
recommendation: term.definition.recommended_translation.clone(),
allowed: term.definition.allowed_translations.clone(),
category: term.definition.category.clone(),
@@ -458,6 +479,18 @@ pub fn evaluate_glossary(
.then_with(|| left.term_id.cmp(&right.term_id))
.then_with(|| left.matched_source.cmp(&right.matched_source))
});
let identity_candidates = candidates
.iter()
.map(|candidate| GlossaryIdentityCandidate {
term_id: candidate.term_id.clone(),
matched_source: candidate.matched_source.clone(),
definition: candidate.definition.clone(),
priority: candidate.priority,
start: candidate.start,
end: candidate.end,
specificity: candidate.specificity,
})
.collect::<Vec<_>>();
let mut selected = Vec::new();
let mut diagnostics = Vec::new();
@@ -505,7 +538,7 @@ pub fn evaluate_glossary(
.then_with(|| left.matched_source.cmp(&right.matched_source))
});
let mut seen = BTreeSet::new();
let constraints = selected
let constraints: Vec<GlossaryConstraint> = selected
.into_iter()
.filter(|candidate| {
seen.insert((
@@ -524,13 +557,242 @@ pub fn evaluate_glossary(
scope: candidate.scope,
})
.collect();
let qa_identity = evaluation_identity(
source_text,
&identity_candidates,
&constraints,
&diagnostics,
);
GlossaryEvaluation {
qa_identity,
constraints,
diagnostics,
blocked,
}
}
/// Validates that an explicit human confirmation authorizes the current
/// blocking glossary result, rather than merely looking structurally complete.
pub fn validate_glossary_override(
qa: &GlossaryQaReport,
glossary_override: Option<&GlossaryOverride>,
) -> crate::Result<()> {
if !qa.status.is_blocked() {
return Err(crate::Error::InvalidArgument(
"Glossary override 只能用于 blocking QA".to_string(),
));
}
if qa.qa_identity.trim().is_empty() {
return Err(crate::Error::InvalidArgument(
"当前 Glossary QA 缺少 qa_identity,不能接受 override".to_string(),
));
}
let Some(glossary_override) = glossary_override else {
return Err(crate::Error::InvalidArgument(
"Glossary QA blocked;需要 reviewer、reason、provenance 和 qa_identity 显式确认"
.to_string(),
));
};
if glossary_override.reviewer.trim().is_empty()
|| glossary_override.reason.trim().is_empty()
|| glossary_override.provenance.trim().is_empty()
|| glossary_override.confirmed_unix_seconds == 0
{
return Err(crate::Error::InvalidArgument(
"Glossary override 的 reviewer、reason、provenance 和 confirmed_unix_seconds 必须有效"
.to_string(),
));
}
if glossary_override.qa_identity.trim().is_empty() {
return Err(crate::Error::InvalidArgument(
"Glossary override 缺少 qa_identity;旧 override 不能自动复用".to_string(),
));
}
if glossary_override.qa_identity != qa.qa_identity {
return Err(crate::Error::InvalidArgument(format!(
"Glossary override 的 qa_identity={} 与当前 QA={} 不一致",
glossary_override.qa_identity, qa.qa_identity
)));
}
Ok(())
}
const GLOSSARY_QA_IDENTITY_VERSION: &[u8] = b"bat-glossary-qa-v1";
fn evaluation_identity(
source_text: &str,
candidates: &[GlossaryIdentityCandidate],
constraints: &[GlossaryConstraint],
diagnostics: &[GlossaryDiagnostic],
) -> String {
let mut hasher = blake3::Hasher::new();
hasher.update(GLOSSARY_QA_IDENTITY_VERSION);
hash_string(&mut hasher, source_text);
let mut candidates = candidates.iter().collect::<Vec<_>>();
candidates.sort_by_key(|candidate| candidate.identity_sort_key());
hash_len(&mut hasher, candidates.len());
for candidate in candidates {
candidate.hash_identity(&mut hasher);
}
hash_constraints(&mut hasher, constraints);
hash_diagnostics(&mut hasher, diagnostics);
format!("gqa-v1-{}", hasher.finalize().to_hex())
}
fn output_qa_identity(
evaluation_identity: &str,
translated_text: &str,
status: &GlossaryQaStatus,
constraints: &[GlossaryConstraint],
diagnostics: &[GlossaryDiagnostic],
) -> String {
let mut hasher = blake3::Hasher::new();
hasher.update(GLOSSARY_QA_IDENTITY_VERSION);
hash_string(&mut hasher, evaluation_identity);
hash_string(&mut hasher, translated_text);
hash_string(&mut hasher, status.as_str());
hash_constraints(&mut hasher, constraints);
hash_diagnostics(&mut hasher, diagnostics);
format!("gqa-v1-{}", hasher.finalize().to_hex())
}
struct GlossaryIdentityCandidate {
term_id: String,
matched_source: String,
definition: GlossaryTermSnapshot,
priority: i32,
start: usize,
end: usize,
specificity: usize,
}
impl GlossaryIdentityCandidate {
fn identity_sort_key(&self) -> (String, String, usize, usize, i32, usize) {
(
self.term_id.clone(),
self.matched_source.clone(),
self.start,
self.end,
self.priority,
self.specificity,
)
}
fn hash_identity(&self, hasher: &mut blake3::Hasher) {
hash_string(hasher, &self.term_id);
hash_string(hasher, &self.matched_source);
hash_usize(hasher, self.start);
hash_usize(hasher, self.end);
hash_i32(hasher, self.priority);
hash_usize(hasher, self.specificity);
hash_snapshot(hasher, &self.definition);
}
}
fn hash_len(hasher: &mut blake3::Hasher, value: usize) {
hasher.update(&(value as u64).to_le_bytes());
}
fn hash_usize(hasher: &mut blake3::Hasher, value: usize) {
hasher.update(&(value as u64).to_le_bytes());
}
fn hash_i32(hasher: &mut blake3::Hasher, value: i32) {
hasher.update(&value.to_le_bytes());
}
fn hash_bool(hasher: &mut blake3::Hasher, value: bool) {
hasher.update(&[value as u8]);
}
fn hash_string(hasher: &mut blake3::Hasher, value: &str) {
hash_len(hasher, value.len());
hasher.update(value.as_bytes());
}
fn hash_option_string(hasher: &mut blake3::Hasher, value: Option<&str>) {
match value {
Some(value) => {
hash_bool(hasher, true);
hash_string(hasher, value);
}
None => hash_bool(hasher, false),
}
}
fn hash_string_list(hasher: &mut blake3::Hasher, values: &[String]) {
let mut values = values.to_vec();
values.sort();
values.dedup();
hash_len(hasher, values.len());
for value in values {
hash_string(hasher, &value);
}
}
fn hash_map(hasher: &mut blake3::Hasher, values: &BTreeMap<String, String>) {
hash_len(hasher, values.len());
for (key, value) in values {
hash_string(hasher, key);
hash_string(hasher, value);
}
}
fn hash_snapshot(hasher: &mut blake3::Hasher, snapshot: &GlossaryTermSnapshot) {
hash_string(hasher, &snapshot.source_term);
hash_string_list(hasher, &snapshot.aliases);
hash_string(hasher, &snapshot.recommended_translation);
hash_string_list(hasher, &snapshot.allowed_translations);
hash_option_string(hasher, snapshot.source_language.as_deref());
hash_option_string(hasher, snapshot.target_language.as_deref());
hash_option_string(hasher, snapshot.category.as_deref());
hash_i32(hasher, snapshot.priority);
hash_map(hasher, &snapshot.scope);
}
fn hash_constraints(hasher: &mut blake3::Hasher, constraints: &[GlossaryConstraint]) {
let mut constraints = constraints.to_vec();
constraints.sort_by(|left, right| {
left.term_id
.cmp(&right.term_id)
.then_with(|| left.matched_source.cmp(&right.matched_source))
.then_with(|| left.priority.cmp(&right.priority))
.then_with(|| {
left.recommended_translation
.cmp(&right.recommended_translation)
})
});
hash_len(hasher, constraints.len());
for constraint in constraints {
hash_string(hasher, &constraint.term_id);
hash_string(hasher, &constraint.matched_source);
hash_string(hasher, &constraint.recommended_translation);
hash_string_list(hasher, &constraint.allowed_translations);
hash_option_string(hasher, constraint.category.as_deref());
hash_i32(hasher, constraint.priority);
hash_map(hasher, &constraint.scope);
}
}
fn hash_diagnostics(hasher: &mut blake3::Hasher, diagnostics: &[GlossaryDiagnostic]) {
let mut diagnostics = diagnostics.to_vec();
diagnostics.sort_by(|left, right| {
left.kind
.as_str()
.cmp(right.kind.as_str())
.then_with(|| left.term_id.cmp(&right.term_id))
.then_with(|| left.value.cmp(&right.value))
.then_with(|| left.message.cmp(&right.message))
});
hash_len(hasher, diagnostics.len());
for diagnostic in diagnostics {
hash_string(hasher, diagnostic.kind.as_str());
hash_option_string(hasher, diagnostic.term_id.as_deref());
hash_option_string(hasher, diagnostic.value.as_deref());
hash_string(hasher, &diagnostic.message);
}
}
/// Validates and normalizes a term draft without choosing a review status.
pub fn validate_glossary_draft(draft: &GlossaryTermDraft) -> crate::Result<()> {
if draft.term_id.trim().is_empty()
@@ -681,6 +943,111 @@ mod tests {
.diagnostics
.iter()
.any(|diagnostic| diagnostic.kind == GlossaryDiagnosticKind::Violation));
assert!(!blocked.qa_identity.is_empty());
}
#[test]
fn qa_identity_is_stable_for_input_order_and_unrelated_terms() {
let mut first = term("first", "Sensei", "老师", 1, BTreeMap::new());
first.definition.aliases = vec!["Teacher".to_string(), "Sensei".to_string()];
let mut second = first.clone();
second.definition.aliases.reverse();
let unrelated = term("unrelated", "Other", "其他", 1, BTreeMap::new());
let source = "Teacher";
let context = BTreeMap::new();
let first_qa = evaluate_glossary(&[first.clone(), unrelated.clone()], source, &context)
.check_translation("先生");
let reordered_qa =
evaluate_glossary(&[unrelated, second], source, &context).check_translation("先生");
assert_eq!(first_qa.qa_identity, reordered_qa.qa_identity);
let mut changed_unrelated = term("unrelated", "Other", "别的译法", 1, BTreeMap::new());
changed_unrelated.definition.allowed_translations = vec!["其他".to_string()];
let unrelated_changed_qa = evaluate_glossary(&[first, changed_unrelated], source, &context)
.check_translation("先生");
assert_eq!(first_qa.qa_identity, unrelated_changed_qa.qa_identity);
}
#[test]
fn qa_identity_changes_when_relevant_glossary_facts_change() {
let base = term("sensei", "Sensei", "老师", 1, BTreeMap::new());
let base_qa = evaluate_glossary(std::slice::from_ref(&base), "Sensei", &BTreeMap::new())
.check_translation("先生");
let mut recommended = base.clone();
recommended.definition.recommended_translation = "教师".to_string();
let recommended_qa =
evaluate_glossary(&[recommended], "Sensei", &BTreeMap::new()).check_translation("先生");
assert_ne!(base_qa.qa_identity, recommended_qa.qa_identity);
let mut allowed = base.clone();
allowed.definition.allowed_translations = vec!["先生".to_string()];
let allowed_qa =
evaluate_glossary(&[allowed], "Sensei", &BTreeMap::new()).check_translation("先生");
assert_ne!(base_qa.qa_identity, allowed_qa.qa_identity);
let mut scoped = base.clone();
scoped
.definition
.scope
.insert("destination".to_string(), "story".to_string());
let scoped_qa =
evaluate_glossary(&[scoped], "Sensei", &BTreeMap::new()).check_translation("先生");
assert_ne!(base_qa.qa_identity, scoped_qa.qa_identity);
let mut aliased = base.clone();
aliased.definition.source_term = "Instructor".to_string();
aliased.definition.aliases = vec!["Sensei".to_string()];
let aliased_qa =
evaluate_glossary(&[aliased], "Sensei", &BTreeMap::new()).check_translation("先生");
assert_ne!(base_qa.qa_identity, aliased_qa.qa_identity);
let conflict = vec![
term("a", "Sensei", "老师", 1, BTreeMap::new()),
term("b", "Sensei", "导师", 1, BTreeMap::new()),
];
let conflict_qa =
evaluate_glossary(&conflict, "Sensei", &BTreeMap::new()).check_translation("先生");
let mut priority_changed = conflict;
priority_changed[1].definition.priority = 2;
let priority_qa = evaluate_glossary(&priority_changed, "Sensei", &BTreeMap::new())
.check_translation("先生");
assert_ne!(conflict_qa.qa_identity, priority_qa.qa_identity);
}
#[test]
fn override_must_match_current_blocking_qa_identity() {
let evaluation = evaluate_glossary(
&[term("sensei", "Sensei", "老师", 1, BTreeMap::new())],
"Sensei",
&BTreeMap::new(),
);
let qa = evaluation.check_translation("先生");
let mut override_record = GlossaryOverride {
qa_identity: qa.qa_identity.clone(),
reviewer: "reviewer".to_string(),
reason: "manual review".to_string(),
provenance: "workbench".to_string(),
confirmed_unix_seconds: 1,
};
assert!(validate_glossary_override(&qa, Some(&override_record)).is_ok());
override_record.qa_identity.clear();
assert!(validate_glossary_override(&qa, Some(&override_record)).is_err());
override_record.qa_identity = "gqa-v1-old".to_string();
assert!(validate_glossary_override(&qa, Some(&override_record)).is_err());
let old_override: GlossaryOverride = serde_json::from_str(
r#"{
"reviewer": "reviewer",
"reason": "manual review",
"provenance": "workbench",
"confirmed_unix_seconds": 1
}"#,
)
.unwrap();
assert!(validate_glossary_override(&qa, Some(&old_override)).is_err());
}
#[test]
+3 -3
View File
@@ -10,9 +10,9 @@ pub mod translation_memory;
pub use game_client::{ClientStatus, GameClient, GameRegion};
pub use game_version::{GameVersion, UnityVersion};
pub use glossary::{
evaluate_glossary, validate_glossary_draft, GlossaryConstraint, GlossaryDiagnostic,
GlossaryDiagnosticKind, GlossaryEvaluation, GlossaryHistoryRecord, GlossaryOverride,
GlossaryQaReport, GlossaryQaStatus, GlossaryReviewStatus, GlossarySourceKind,
evaluate_glossary, validate_glossary_draft, validate_glossary_override, GlossaryConstraint,
GlossaryDiagnostic, GlossaryDiagnosticKind, GlossaryEvaluation, GlossaryHistoryRecord,
GlossaryOverride, GlossaryQaReport, GlossaryQaStatus, GlossaryReviewStatus, GlossarySourceKind,
GlossarySourceRecord, GlossarySummary, GlossaryTerm, GlossaryTermDraft, GlossaryTermSnapshot,
};
pub use resource::{
+24 -5
View File
@@ -89,7 +89,23 @@ bat i18n set \
--translated-text '中文文本'
```
也可以使用 `--translated-file` 读取 UTF-8 文本。需要复核单条内容时:
也可以使用 `--translated-file` 读取 UTF-8 文本。
如果译文触发 blocking Glossary QA,需先从 diagnose/任务结果取得当前
`qa_identity`,并与 reviewer、reason、provenance 一起提交;系统不会根据 workbench
中旧的 QA 自动补填:
```bash
bat i18n set \
--translation-file /tmp/bat-workbench.json \
--translation-id <text-unit-id> \
--translated-text '人工确认的译文' \
--glossary-qa-identity <qa-identity> \
--glossary-reviewer operator \
--glossary-reason '人工确认术语偏离' \
--glossary-provenance workbench
```
需要复核单条内容时:
```bash
bat i18n get \
@@ -166,7 +182,8 @@ Glossary V1 是 Rust `bat` 持有的独立项目级 SQLite 资产,默认位于
`BAT_GLOSSARY_PATH``[translation.worker].glossary_path` 指定。worker 只把
`approved` term 转成 provider-neutral constraints,并在 TM 复用、provider 返回
和人工工作台/任务回写时执行相同的确定性 QA。冲突或不符合推荐/允许译法的结果会
阻止自动完成;必须提交带 reviewer、reason 和 provenance 的显式 override。
阻止自动完成;必须提交带当前 `qa_identity`reviewer、reason 和 provenance 的显式
override。
常用 Glossary 操作:
@@ -303,7 +320,9 @@ Rust `translation.tasks` / `translation.handoff`,不在 Go 侧维护状态。
`translation.task.update`。人工校对流程提交译文时必须使用 `status=completed`
并为每个 `translation_results[]` 提供 `unit_id``source_text`
`translated_text`Rust 会用当前 `official-textunit-index.json` 校验 unit、
source text、destination 和 archive entry 后再落库。
source text、destination 和 archive entry 后再落库。blocking Glossary QA 还必须提交
与当前 QA 完全相等的 `glossary_override.qa_identity`;旧或缺少 identity 的 override
不会授权。
`POST /admin/control/translation-worker-run` 会触发 Rust 侧
`translation.worker.run`,请求字段为 `provider``fixture_path`
@@ -320,8 +339,8 @@ source text、destination 和 archive entry 后再落库。
release,校验工作台与当前 TextUnit 索引的 source/location 一致后,写入已有支持
范围内的 TextAsset、TypeTree string field 和 managed-reference string field
patch。校验通过后才原子切换 `localized/current`,并在 release manifest 中记录
源/目标 BLAKE3、字节数、patch kind、TextUnit、provider、review 和 rollback
信息。ZIP 内 bundle 不会被静默改写。
源/目标 BLAKE3、字节数、patch kind、TextUnit、provider、review、发布时重新计算的
Glossary QA/override 和 rollback 信息。ZIP 内 bundle 不会被静默改写。
使用人工编辑的工作台发布:
+10 -3
View File
@@ -287,8 +287,10 @@ offset 和 error。TypeTree-covered managed reference 字段会进入结构化
`status=completed` 时额外提交 `provider``provider_run_id`
`translation_results[]`,每个结果必须包含 `unit_id``source_text`
`translated_text`;结果也可以提交完整的 `glossary_override``reviewer`
`reason``provenance``confirmed_unix_seconds`),用于人工确认 Glossary
blocking deviation。Rust 会用当前 `official-textunit-index.json` 校验 unit、
`reason``provenance``confirmed_unix_seconds` 和当前 blocking QA 的
`qa_identity`),用于人工确认 Glossary blocking deviation。`qa_identity` 必须与
Rust 重新计算的当前 QA 完全相等;缺失或过期的 override 不授权。Rust 会用当前
`official-textunit-index.json` 校验 unit、
source text、destination 和 archive entry 后再落库。因此 worker 或人工校对流程
消费 handoff 后,bat-api 可通过 `translation.tasks` 查询单项任务,也可通过
`translation.handoff` 获取完整 job/unit/provider run 状态。`translation.handoff`
@@ -326,7 +328,9 @@ Glossary 只把 `approved` term 发送为 provider constraints。worker 会在 t
复用前、provider 返回后、人工 `translation.task.update` 和 workbench publish 前执行
同一套确定性 QA;冲突或未使用推荐/允许译法的结果不会自动完成或发布。允许但非推荐译法
产生 warningblocking deviation 必须在对应结果中提交 `glossary_override`,并包含
`reviewer``reason``provenance` 和确认时间。系统不会在译文生成后做静默字符串替换。
`qa_identity``reviewer``reason``provenance` 和确认时间。Glossary 定义变化会
只使受影响 QA 的旧 override 失效;无关术语变化不会改变该 QA identity。系统不会在
译文生成后做静默字符串替换。
### localized
@@ -351,6 +355,9 @@ Glossary 只把 `approved` term 发送为 provider constraints。worker 会在 t
`patch_manifest_path``patch_manifest_available`
`patch_manifest_matches_release``patch_file_count`
`patch_text_asset_operation_count``rollback_previous_current_target`
每个 localized patch operation 的 manifest metadata 记录发布时重新计算的
`glossary_qa`(包括 `qa_identity`)及对应 `glossary_override`,不会复用 workbench
中已经过期的 QA 快照。
### catalog
+1 -1
View File
@@ -117,7 +117,7 @@ Rust `bat` 已提供独立项目级 SQLite TM,记录 raw source/hash、完整
### G-013Glossary V1 已实现,协作视图仍缺失
Rust `bat` 已提供独立项目级 `glossary.sqlite`term/alias/recommended/allowed/category/priority、全局与 TextUnit scope、source history、approved review、冲突诊断、provider-neutral constraints 和确定性 QA 均由 Rust 持有。trusted TM 复用会先经过 Glossary QAprovider、TM、人工 task/workbench 结果都记录 QAblocking deviation 必须显式提交 reviewer/reason/provenance`translation.glossary.*` 已通过 `bat.sock` 暴露,Go 仅提供鉴权后的 typed forwarding。剩余缺口是完整 Web 术语协作视图和更丰富的导入/搜索能力。
Rust `bat` 已提供独立项目级 `glossary.sqlite`term/alias/recommended/allowed/category/priority、全局与 TextUnit scope、source history、approved review、冲突诊断、provider-neutral constraints 和确定性 QA 均由 Rust 持有。trusted TM 复用会先经过 Glossary QAprovider、TM、人工 task/workbench 结果都记录 QAblocking deviation 必须显式提交与当前 QA 精确绑定的 `qa_identity` 及 reviewer/reason/provenance。localized publish 会把发布时重算的 QA 写入 manifest`translation.glossary.*` 已通过 `bat.sock` 暴露,Go 仅提供鉴权后的 typed forwarding。剩余缺口是完整 Web 术语协作视图和更丰富的导入/搜索能力。
### G-014:完整 Provider 扩展体系未实现
+10 -5
View File
@@ -10,10 +10,10 @@ use bat_infrastructure::{
apply_unityfs_text_asset_patch_file, changed_endpoint_urls,
completed_worker_translation_workbench, diff_extended_snapshot, export_translation_workbench,
gc_orphan_staging_with_cas_root, get_translation_entry, lexical_absolute,
localized_patch_operations, open_append_file, read_download_manifest_at, read_file_no_symlink,
read_localized_patch_manifest_at, read_localized_version_state, read_parse_cache_at,
read_snapshot, read_textunit_index_at, read_translation_workbench, read_version_state,
redact_proxy_url, repack_bundle, resolve_curl_proxy, set_translation,
localized_patch_operations_with_glossary_path, open_append_file, read_download_manifest_at,
read_file_no_symlink, read_localized_patch_manifest_at, read_localized_version_state,
read_parse_cache_at, read_snapshot, read_textunit_index_at, read_translation_workbench,
read_version_state, redact_proxy_url, repack_bundle, resolve_curl_proxy, set_translation,
set_translation_checked_with_glossary_path, unset_translation, validate_output_root,
validate_runtime_state_dir, validate_translation_workbench_with_glossary_path,
write_file_atomic, write_official_textunit_queues, CurlProxyConfig, CurlProxyMode,
@@ -430,6 +430,7 @@ struct CliOptions {
glossary_reviewer: Option<String>,
glossary_reason: Option<String>,
glossary_override_provenance: Option<String>,
glossary_qa_identity: Option<String>,
glossary_source_text: Option<String>,
glossary_context_json: Option<String>,
glossary_review_status: Option<String>,
@@ -555,6 +556,7 @@ impl Default for CliOptions {
glossary_reviewer: None,
glossary_reason: None,
glossary_override_provenance: None,
glossary_qa_identity: None,
glossary_source_text: None,
glossary_context_json: None,
glossary_review_status: None,
@@ -6949,6 +6951,9 @@ fn parse_args_with_env(
"--glossary-provenance" | "--glossary-override-provenance" => {
options.glossary_override_provenance = Some(next_option_value(&mut args, &flag)?);
}
"--glossary-qa-identity" => {
options.glossary_qa_identity = Some(next_option_value(&mut args, &flag)?);
}
"--glossary-source-text" => {
options.glossary_source_text = Some(next_option_value(&mut args, &flag)?);
}
@@ -7871,7 +7876,7 @@ fn parse_args_with_env(
|| tools_are_non_default(&options.config, &options.env_baseline_config)
{
return Err(anyhow::anyhow!(
"translation workbench 编辑命令只接受 --translation-file、--translation-id、译文字段、--state-dir --json/--human"
"translation workbench 编辑命令只接受 --translation-file、--translation-id、译文和 Glossary override 字段、--state-dir 以及 --json/--human"
));
}
options.progress = false;
@@ -448,6 +448,7 @@ Sync:
--glossary-reviewer <ID> Reviewer for Glossary updates/reviews/delete
--glossary-reason <TEXT> Reason for Glossary review/delete or override
--glossary-provenance <TEXT> Provenance for an explicit Glossary override
--glossary-qa-identity <ID> Current blocking Glossary QA identity for an override
--worker-concurrency <N> Translation worker concurrency (default: 8, range 1..=256)
--worker-max-attempts <N> Maximum claims per translation task
--worker-lease-seconds <N> Lease seconds for one claimed task
+10 -15
View File
@@ -1,6 +1,6 @@
use super::report_output::print_json_value;
use super::*;
use bat_core::domain::{GlossaryOverride, TranslationMemoryContext};
use bat_core::domain::{validate_glossary_override, GlossaryOverride, TranslationMemoryContext};
use bat_core::repositories::TranslationMemoryRepository;
use std::collections::BTreeMap;
@@ -377,24 +377,19 @@ async fn build_manual_translation_results(
} else {
None
};
if let Some(qa) = glossary_qa.as_ref().filter(|qa| qa.status.is_blocked()) {
let Some(override_record) = param.glossary_override.as_ref() else {
if let Some(qa) = glossary_qa.as_ref() {
if qa.status.is_blocked() {
validate_glossary_override(qa, param.glossary_override.as_ref()).map_err(
|error| {
anyhow::anyhow!("TextUnit {} 的 glossary_override 无效:{error}", unit_id)
},
)?;
} else if param.glossary_override.is_some() {
return Err(anyhow::anyhow!(
"TextUnit {} 的 Glossary QA blocked;必须提供 glossary_override",
unit_id
));
};
if override_record.reviewer.trim().is_empty()
|| override_record.reason.trim().is_empty()
|| override_record.provenance.trim().is_empty()
|| override_record.confirmed_unix_seconds == 0
{
return Err(anyhow::anyhow!(
"TextUnit {} 的 glossary_override 不完整或 confirmed_unix_seconds 无效",
"TextUnit {} 不能为非 blocking Glossary QA 指定 override",
unit_id
));
}
let _ = qa;
} else if param.glossary_override.is_some() {
return Err(anyhow::anyhow!(
"TextUnit {} 不能为非 blocking Glossary QA 指定 override",
@@ -150,10 +150,12 @@ pub(super) fn run_translation_set(options: &CliOptions) -> anyhow::Result<()> {
options.glossary_reviewer.as_deref(),
options.glossary_reason.as_deref(),
options.glossary_override_provenance.as_deref(),
options.glossary_qa_identity.as_deref(),
) {
(None, None, None) => None,
(Some(reviewer), Some(reason), Some(provenance)) => Some(
(None, None, None, None) => None,
(Some(reviewer), Some(reason), Some(provenance), Some(qa_identity)) => Some(
bat_core::domain::GlossaryOverride {
qa_identity: qa_identity.to_string(),
reviewer: reviewer.to_string(),
reason: reason.to_string(),
provenance: provenance.to_string(),
@@ -162,7 +164,7 @@ pub(super) fn run_translation_set(options: &CliOptions) -> anyhow::Result<()> {
),
_ => {
return Err(anyhow::anyhow!(
"Glossary override 必须同时指定 --glossary-reviewer、--glossary-reason 和 --glossary-provenance"
"Glossary override 必须同时指定 --glossary-qa-identity、--glossary-reviewer、--glossary-reason 和 --glossary-provenance"
))
}
};
@@ -371,7 +373,11 @@ pub(super) fn publish_localized_report(
&workbench,
options.glossary_path.as_deref(),
)?;
let operations = localized_patch_operations(&resource_root, &workbench)?;
let operations = localized_patch_operations_with_glossary_path(
&resource_root,
&workbench,
options.glossary_path.as_deref(),
)?;
let localized_release_id = options.localized_release_id.clone().or_else(|| {
options
.config
+6 -6
View File
@@ -173,12 +173,12 @@ pub use translation_worker::{
pub use translation_workflow::{
completed_worker_translation_workbench, export_completed_worker_translation_workbench,
export_translation_workbench, get_translation_entry, localized_patch_operations,
localized_text_asset_patches, read_translation_workbench, repack_bundle, set_translation,
set_translation_checked, set_translation_checked_with_glossary_path, unset_translation,
validate_translation_workbench, validate_translation_workbench_with_glossary_path,
write_translation_workbench, RepackOperation, RepackReport, RepackSpec, TranslationWorkbench,
TranslationWorkbenchEntry, TranslationWorkbenchValidationReport, REPACK_SPEC_VERSION,
TRANSLATION_WORKBENCH_VERSION,
localized_patch_operations_with_glossary_path, localized_text_asset_patches,
read_translation_workbench, repack_bundle, set_translation, set_translation_checked,
set_translation_checked_with_glossary_path, unset_translation, validate_translation_workbench,
validate_translation_workbench_with_glossary_path, write_translation_workbench,
RepackOperation, RepackReport, RepackSpec, TranslationWorkbench, TranslationWorkbenchEntry,
TranslationWorkbenchValidationReport, REPACK_SPEC_VERSION, TRANSLATION_WORKBENCH_VERSION,
};
/// Infrastructure 版本号
+51
View File
@@ -352,6 +352,12 @@ pub struct LocalizedPatchOperation {
/// Review state used by the publication input.
#[serde(default, skip_serializing_if = "Option::is_none")]
pub review_status: Option<String>,
/// Glossary QA recomputed for this publication.
#[serde(default, skip_serializing_if = "Option::is_none")]
pub glossary_qa: Option<bat_core::domain::GlossaryQaReport>,
/// Human confirmation bound to the published Glossary QA identity.
#[serde(default, skip_serializing_if = "Option::is_none")]
pub glossary_override: Option<bat_core::domain::GlossaryOverride>,
}
/// Rollback information recorded for a localized publication.
@@ -962,6 +968,8 @@ impl LocalizedPatchOperation {
translation_source_kind: None,
translation_memory_record_id: None,
review_status: None,
glossary_qa: None,
glossary_override: None,
},
metadata,
)
@@ -989,6 +997,8 @@ impl LocalizedPatchOperation {
translation_source_kind: None,
translation_memory_record_id: None,
review_status: None,
glossary_qa: None,
glossary_override: None,
},
metadata,
)
@@ -1015,6 +1025,8 @@ impl LocalizedPatchOperation {
translation_source_kind: None,
translation_memory_record_id: None,
review_status: None,
glossary_qa: None,
glossary_override: None,
},
metadata,
))
@@ -1032,6 +1044,8 @@ impl LocalizedPatchOperation {
operation.translation_source_kind = metadata.translation_source_kind.clone();
operation.translation_memory_record_id = metadata.translation_memory_record_id.clone();
operation.review_status = Some(metadata.review_status.clone());
operation.glossary_qa = metadata.glossary_qa.clone();
operation.glossary_override = metadata.glossary_override.clone();
}
operation
}
@@ -1058,6 +1072,8 @@ impl Default for LocalizedPatchOperation {
translation_source_kind: None,
translation_memory_record_id: None,
review_status: None,
glossary_qa: None,
glossary_override: None,
}
}
}
@@ -1432,6 +1448,8 @@ mod tests {
translation_source_kind: Some("provider".to_string()),
translation_memory_record_id: None,
review_status: Some("provider_completed".to_string()),
glossary_qa: None,
glossary_override: None,
}],
}],
rollback: LocalizedPatchRollbackInfo {
@@ -1461,6 +1479,39 @@ mod tests {
);
}
#[test]
fn manifest_operation_keeps_current_glossary_qa_provenance() {
let qa = bat_core::domain::GlossaryQaReport {
qa_identity: "gqa-v1-current".to_string(),
status: bat_core::domain::GlossaryQaStatus::Pass,
constraints: Vec::new(),
diagnostics: Vec::new(),
};
let input = LocalizedPatchInput::TextAsset(LocalizedTextAssetPatch {
bundle_path: "Bundles/file.bundle".to_string(),
text_asset: TextAssetPatch::new("CAB-asset", 1, b"target".to_vec()),
metadata: Some(LocalizedPatchOperationMetadata {
text_unit_id: "unit-1".to_string(),
source_text_blake3: "source-hash".to_string(),
translation_provider: None,
provider_run_id: None,
translation_source_kind: Some("manual".to_string()),
translation_memory_record_id: None,
review_status: "manual_reviewed".to_string(),
glossary_qa: Some(qa.clone()),
glossary_override: None,
}),
});
let operation = input.manifest_operation().unwrap();
assert_eq!(
operation
.glossary_qa
.as_ref()
.map(|report| report.qa_identity.as_str()),
Some("gqa-v1-current")
);
}
#[cfg(unix)]
#[test]
fn publishes_a_separate_localized_release_atomically() {
+5
View File
@@ -968,6 +968,7 @@ async fn process_claimed_task(
}
} else {
bat_core::domain::GlossaryEvaluation {
qa_identity: String::new(),
constraints: Vec::new(),
diagnostics: Vec::new(),
blocked: false,
@@ -1760,6 +1761,10 @@ mod tests {
.map(|qa| qa.status),
Some(bat_core::domain::GlossaryQaStatus::Pass)
);
assert!(task.translation_results[0]
.glossary_qa
.as_ref()
.is_some_and(|qa| !qa.qa_identity.is_empty()));
}
#[tokio::test]
+309 -26
View File
@@ -16,7 +16,10 @@ use bat_assetbundle::{
patch_unityfs_field, patch_unityfs_string_field, patch_unityfs_text_asset, FieldPatch,
StringFieldPatch, TextAssetPatch, UnitySerializedReplacementValue,
};
use bat_core::domain::{GlossaryOverride, GlossaryQaReport};
use bat_core::domain::{
validate_glossary_override as validate_core_glossary_override, GlossaryOverride,
GlossaryQaReport,
};
use serde::{Deserialize, Serialize};
use std::collections::{BTreeMap, BTreeSet, HashMap};
use std::path::{Path, PathBuf};
@@ -333,9 +336,8 @@ pub fn set_translation_checked_with_glossary_path(
.as_ref()
.map(|glossary| evaluate_glossary_entry(glossary, &current, &translated_text))
.transpose()?;
if let Some(qa) = qa.as_ref().filter(|qa| qa.status.is_blocked()) {
validate_glossary_override(glossary_override.as_ref())?;
let _ = qa;
if let Some(qa) = qa.as_ref() {
validate_current_glossary_qa(qa, glossary_override.as_ref())?;
} else if glossary_override.is_some() {
return Err(anyhow::anyhow!(
"Glossary override 只能用于存在 blocking QA 的译文"
@@ -464,9 +466,12 @@ pub fn validate_translation_workbench_with_glossary_path(
}
if let Some(glossary) = glossary.as_ref() {
let qa = evaluate_glossary_entry(glossary, current, translated_text)?;
if qa.status.is_blocked() {
validate_glossary_override(entry.glossary_override.as_ref())?;
}
validate_current_glossary_qa(&qa, entry.glossary_override.as_ref())?;
} else if entry.glossary_override.is_some() {
return Err(anyhow::anyhow!(
"TextUnit {} 存在 Glossary override,但当前没有 Glossary 数据库",
entry.id
));
}
changed_entries += 1;
let source_kind = normalized_text_source_kind(entry.text_source_kind.as_deref());
@@ -560,20 +565,15 @@ fn evaluate_glossary_entry(
.map_err(|error| anyhow::anyhow!("执行 Glossary QA 失败:{error}"))
}
fn validate_glossary_override(glossary_override: Option<&GlossaryOverride>) -> anyhow::Result<()> {
let Some(glossary_override) = glossary_override else {
return Err(anyhow::anyhow!(
"Glossary QA blocked;需要 reviewer、reason 和 provenance 显式确认"
));
};
if glossary_override.reviewer.trim().is_empty()
|| glossary_override.reason.trim().is_empty()
|| glossary_override.provenance.trim().is_empty()
|| glossary_override.confirmed_unix_seconds == 0
{
return Err(anyhow::anyhow!(
"Glossary override 的 reviewer、reason、provenance 和 confirmed_unix_seconds 必须有效"
));
fn validate_current_glossary_qa(
qa: &GlossaryQaReport,
glossary_override: Option<&GlossaryOverride>,
) -> anyhow::Result<()> {
if qa.status.is_blocked() {
validate_core_glossary_override(qa, glossary_override)
.map_err(|error| anyhow::anyhow!("{error}"))?;
} else if glossary_override.is_some() {
return Err(anyhow::anyhow!("Glossary override 只能用于 blocking QA"));
}
Ok(())
}
@@ -587,6 +587,23 @@ pub fn localized_patch_operations(
resource_root: &Path,
workbench: &TranslationWorkbench,
) -> anyhow::Result<Vec<LocalizedPatchInput>> {
localized_patch_operations_with_glossary_path(resource_root, workbench, None)
}
/// Converts reviewed entries to patch operations using QA recomputed from the
/// current Glossary. The recomputed report is the one persisted to the
/// localized release manifest.
pub fn localized_patch_operations_with_glossary_path(
resource_root: &Path,
workbench: &TranslationWorkbench,
configured_glossary_path: Option<&Path>,
) -> anyhow::Result<Vec<LocalizedPatchInput>> {
validate_translation_workbench_with_glossary_path(
resource_root,
&workbench.official_release_id,
workbench,
configured_glossary_path,
)?;
let index = read_textunit_index_at(resource_root)
.map_err(anyhow::Error::msg)?
.ok_or_else(|| anyhow::anyhow!("缺少当前官方 TextUnit 索引"))?;
@@ -595,6 +612,10 @@ pub fn localized_patch_operations(
.iter()
.map(|unit| (unit.id.as_str(), unit))
.collect::<std::collections::HashMap<_, _>>();
let glossary_path = configured_glossary_path
.map(PathBuf::from)
.unwrap_or_else(|| SqliteGlossaryRepository::repository_path(resource_root));
let glossary = open_glossary_if_present(&glossary_path)?;
let mut seen = BTreeSet::new();
let mut operations = Vec::new();
@@ -609,6 +630,18 @@ pub fn localized_patch_operations(
if translated_text == &entry.source_text {
continue;
}
let current_qa = glossary
.as_ref()
.map(|glossary| evaluate_glossary_entry(glossary, current, translated_text))
.transpose()?;
if let Some(qa) = current_qa.as_ref() {
validate_current_glossary_qa(qa, entry.glossary_override.as_ref())?;
} else if entry.glossary_override.is_some() {
return Err(anyhow::anyhow!(
"TextUnit {} 存在 Glossary override,但当前没有 Glossary 数据库",
entry.id
));
}
let Some(serialized_file) = entry.serialized_file.clone() else {
return Err(anyhow::anyhow!(
"TextUnit {} 没有 serialized_file,当前不能生成重打包 patch",
@@ -640,7 +673,11 @@ pub fn localized_patch_operations(
entry.id
));
}
let metadata = Some(localized_patch_metadata(entry));
let metadata = Some(localized_patch_metadata(
entry,
current_qa.as_ref(),
entry.glossary_override.as_ref(),
));
match source_kind.as_deref() {
Some("textasset") => {
let mut patch = TextAssetPatch::new(
@@ -706,6 +743,8 @@ pub fn localized_text_asset_patches(
.iter()
.map(|unit| (unit.id.as_str(), unit))
.collect::<std::collections::HashMap<_, _>>();
let glossary =
open_glossary_if_present(&SqliteGlossaryRepository::repository_path(resource_root))?;
let mut seen = BTreeSet::new();
let mut patches = Vec::new();
@@ -758,10 +797,26 @@ pub fn localized_text_asset_patches(
translated_text.as_bytes().to_vec(),
);
patch.expected_name = entry.asset_name.clone();
let current_qa = glossary
.as_ref()
.map(|glossary| evaluate_glossary_entry(glossary, current, translated_text))
.transpose()?;
if let Some(qa) = current_qa.as_ref() {
validate_current_glossary_qa(qa, entry.glossary_override.as_ref())?;
} else if entry.glossary_override.is_some() {
return Err(anyhow::anyhow!(
"TextUnit {} 存在 Glossary override,但当前没有 Glossary 数据库",
entry.id
));
}
patches.push(LocalizedTextAssetPatch {
bundle_path: entry.destination.clone(),
text_asset: patch,
metadata: Some(localized_patch_metadata(entry)),
metadata: Some(localized_patch_metadata(
entry,
current_qa.as_ref(),
entry.glossary_override.as_ref(),
)),
});
}
@@ -773,7 +828,11 @@ pub fn localized_text_asset_patches(
Ok(patches)
}
fn localized_patch_metadata(entry: &TranslationWorkbenchEntry) -> LocalizedPatchOperationMetadata {
fn localized_patch_metadata(
entry: &TranslationWorkbenchEntry,
glossary_qa: Option<&GlossaryQaReport>,
glossary_override: Option<&GlossaryOverride>,
) -> LocalizedPatchOperationMetadata {
LocalizedPatchOperationMetadata {
text_unit_id: entry.id.clone(),
source_text_blake3: blake3::hash(entry.source_text.as_bytes())
@@ -787,8 +846,8 @@ fn localized_patch_metadata(entry: &TranslationWorkbenchEntry) -> LocalizedPatch
.review_status
.clone()
.unwrap_or_else(|| "manual_reviewed".to_string()),
glossary_qa: entry.glossary_qa.clone(),
glossary_override: entry.glossary_override.clone(),
glossary_qa: glossary_qa.cloned(),
glossary_override: glossary_override.cloned(),
}
}
@@ -1251,6 +1310,172 @@ mod tests {
assert_eq!(report.unreviewed_entries, 0);
}
#[test]
fn checked_translation_rejects_override_after_glossary_change() {
let temp = tempfile::TempDir::new().unwrap();
let index = crate::official_parse::OfficialTextUnitIndex {
version: crate::official_parse::OFFICIAL_TEXTUNIT_INDEX_VERSION,
generated_unix_seconds: 1,
resource_root: temp.path().to_path_buf(),
summary: Default::default(),
units: vec![OfficialTextUnitIndexUnit {
id: "unit-1".to_string(),
parse_entry_key: "bundle".to_string(),
source_url: "https://example.invalid/bundle".to_string(),
destination: "bundles/test.bundle".to_string(),
archive_entry: None,
source_kind: crate::official_parse::OfficialParseSourceKind::DirectBundle,
unity_version: None,
source_text: "原文".to_string(),
serialized_file: Some("CAB-test".to_string()),
path_id: Some(7),
class_id: Some(49),
field_path: None,
field_offset: None,
field_byte_size: None,
format: Some("plain".to_string()),
text_source_kind: Some("text_asset".to_string()),
asset_name: Some("Story".to_string()),
context: Default::default(),
}],
errors: Vec::new(),
};
crate::official_parse::write_textunit_index_at(temp.path(), &index).unwrap();
let workbench_path = temp.path().join("workbench.json");
write_translation_workbench(&workbench_path, &workbench(temp.path())).unwrap();
let glossary_path = temp.path().join("glossary.sqlite");
let runtime = tokio::runtime::Builder::new_current_thread()
.enable_all()
.build()
.unwrap();
let qa = runtime.block_on(async {
let glossary = SqliteGlossaryRepository::new(&glossary_path).await.unwrap();
glossary
.add(bat_core::domain::GlossaryTermDraft {
term_id: "term-source".to_string(),
definition: bat_core::domain::GlossaryTermSnapshot {
source_term: "原文".to_string(),
aliases: Vec::new(),
recommended_translation: "译文".to_string(),
allowed_translations: Vec::new(),
source_language: None,
target_language: None,
category: None,
priority: 1,
scope: BTreeMap::new(),
},
review_status: bat_core::domain::GlossaryReviewStatus::Draft,
source: bat_core::domain::GlossarySourceRecord {
source_kind: bat_core::domain::GlossarySourceKind::Manual,
source_ref: None,
source_author: None,
source_note: None,
observed_unix_seconds: 1,
},
})
.await
.unwrap();
glossary
.review(
"term-source",
bat_core::domain::GlossaryReviewStatus::Approved,
"reviewer",
Some("approve".to_string()),
)
.await
.unwrap();
let context = crate::translation_memory::translation_memory_context(
"bundles/test.bundle",
None,
Some("CAB-test"),
Some(7),
Some(49),
None,
Some("plain"),
Some("Story"),
Some("text_asset"),
&BTreeMap::new(),
);
let evaluation = glossary.diagnose("原文", &context).await.unwrap();
(glossary, evaluation.check_translation("错误"))
});
let (glossary, qa) = qa;
let override_record = GlossaryOverride {
qa_identity: qa.qa_identity.clone(),
reviewer: "reviewer".to_string(),
reason: "manual review".to_string(),
provenance: "workbench".to_string(),
confirmed_unix_seconds: 1,
};
let updated = set_translation_checked_with_glossary_path(
temp.path(),
&workbench_path,
"unit-1",
"错误".to_string(),
Some(override_record.clone()),
Some(&glossary_path),
)
.unwrap();
assert_eq!(
updated
.glossary_qa
.as_ref()
.map(|report| report.qa_identity.as_str()),
Some(qa.qa_identity.as_str())
);
let loaded = read_translation_workbench(&workbench_path).unwrap();
let validation = validate_translation_workbench_with_glossary_path(
temp.path(),
"release-1",
&loaded,
Some(&glossary_path),
)
.unwrap();
assert_eq!(validation.status, "valid");
runtime.block_on(async {
let current = glossary.find("term-source").await.unwrap();
glossary
.update(
bat_core::domain::GlossaryTermDraft {
term_id: current.term_id,
definition: bat_core::domain::GlossaryTermSnapshot {
recommended_translation: "新译文".to_string(),
..current.definition
},
review_status: bat_core::domain::GlossaryReviewStatus::Draft,
source: bat_core::domain::GlossarySourceRecord {
observed_unix_seconds: 2,
..current.source
},
},
"reviewer",
Some("change recommendation".to_string()),
)
.await
.unwrap();
glossary
.review(
"term-source",
bat_core::domain::GlossaryReviewStatus::Approved,
"reviewer",
Some("approve changed term".to_string()),
)
.await
.unwrap();
});
let error = set_translation_checked_with_glossary_path(
temp.path(),
&workbench_path,
"unit-1",
"错误".to_string(),
Some(override_record),
Some(&glossary_path),
)
.unwrap_err();
assert!(error.to_string().contains("qa_identity"));
}
#[test]
fn localized_operations_preserve_type_tree_field_traceability() {
let temp = tempfile::TempDir::new().unwrap();
@@ -1290,6 +1515,59 @@ mod tests {
entry.translated_text = Some("译文".to_string());
entry.translation_provider = Some("mock".to_string());
entry.provider_run_id = Some("run-1".to_string());
entry.glossary_qa = Some(GlossaryQaReport {
qa_identity: "stale-workbench-qa".to_string(),
status: bat_core::domain::GlossaryQaStatus::Pass,
constraints: Vec::new(),
diagnostics: Vec::new(),
});
let glossary_path = temp.path().join("glossary.sqlite");
let runtime = tokio::runtime::Builder::new_current_thread()
.enable_all()
.build()
.unwrap();
let expected_qa = runtime.block_on(async {
let glossary = SqliteGlossaryRepository::new(&glossary_path).await.unwrap();
glossary
.add(bat_core::domain::GlossaryTermDraft {
term_id: "term-source".to_string(),
definition: bat_core::domain::GlossaryTermSnapshot {
source_term: "原文".to_string(),
aliases: Vec::new(),
recommended_translation: "译文".to_string(),
allowed_translations: Vec::new(),
source_language: None,
target_language: None,
category: None,
priority: 1,
scope: BTreeMap::new(),
},
review_status: bat_core::domain::GlossaryReviewStatus::Draft,
source: bat_core::domain::GlossarySourceRecord {
source_kind: bat_core::domain::GlossarySourceKind::Manual,
source_ref: None,
source_author: None,
source_note: None,
observed_unix_seconds: 1,
},
})
.await
.unwrap();
glossary
.review(
"term-source",
bat_core::domain::GlossaryReviewStatus::Approved,
"reviewer",
Some("approve".to_string()),
)
.await
.unwrap();
glossary
.diagnose("原文", &BTreeMap::new())
.await
.unwrap()
.check_translation("译文")
});
let operations = localized_patch_operations(temp.path(), &workbench).unwrap();
assert_eq!(operations.len(), 1);
@@ -1305,6 +1583,11 @@ mod tests {
assert_eq!(metadata.text_unit_id, "unit-1");
assert_eq!(metadata.translation_provider.as_deref(), Some("mock"));
assert_eq!(metadata.provider_run_id.as_deref(), Some("run-1"));
assert_eq!(
metadata.glossary_qa.as_ref().map(|qa| qa.status),
Some(bat_core::domain::GlossaryQaStatus::Pass)
);
assert_eq!(metadata.glossary_qa.as_ref(), Some(&expected_qa));
}
other => panic!("unexpected localized operation: {other:?}"),
}
+3 -2
View File
@@ -1944,7 +1944,7 @@ func TestAdminControlForwardsAllowlistedActions(t *testing.T) {
{name: "repair", action: "repair", rpcMethod: "resource.repair", call: "resource.repair"},
{name: "catalog refresh", action: "catalog-refresh", rpcMethod: "catalog.refresh", call: "catalog.refresh"},
{name: "task cancel", action: "task-cancel", body: `{"task_id":"task-sync-1"}`, rpcMethod: "task.cancel", call: "task.cancel"},
{name: "translation task update", action: "translation-task-update", body: `{"task_id":"textunit/v-current/Scenario","status":"completed","provider":"manual","provider_run_id":"manual-run-1","translation_results":[{"unit_id":"direct:a#unit:0","source_text":"source","translated_text":"译文","glossary_override":{"reviewer":"reviewer","reason":"approved deviation","provenance":"manual-review","confirmed_unix_seconds":100}}]}`, rpcMethod: "translation.task.update", call: "translation.task.update"},
{name: "translation task update", action: "translation-task-update", body: `{"task_id":"textunit/v-current/Scenario","status":"completed","provider":"manual","provider_run_id":"manual-run-1","translation_results":[{"unit_id":"direct:a#unit:0","source_text":"source","translated_text":"译文","glossary_override":{"qa_identity":"gqa-v1-test","reviewer":"reviewer","reason":"approved deviation","provenance":"manual-review","confirmed_unix_seconds":100}}]}`, rpcMethod: "translation.task.update", call: "translation.task.update"},
{name: "translation worker run", action: "translation-worker-run", body: `{"provider":"mock","concurrency":8,"max_tasks":2,"retry_backoff_seconds":0,"worker_id":"dashboard-worker"}`, rpcMethod: "translation.worker.run", call: "translation.worker.run"},
{name: "translation proofread", action: "translation-proofread", rpcMethod: "translation.proofread", call: "translation.proofread"},
{name: "translation memory confirm", action: "translation-memory-confirm", body: `{"record_id":"tm-record-1","reviewer":"reviewer","reason":"reviewed"}`, rpcMethod: "translation.memory.confirm", call: "translation.memory.confirm"},
@@ -1983,7 +1983,8 @@ func TestAdminControlForwardsAllowlistedActions(t *testing.T) {
len(backend.translationTaskUpdates[0].TranslationResults) != 1 ||
backend.translationTaskUpdates[0].TranslationResults[0].TranslatedText != "译文" ||
backend.translationTaskUpdates[0].TranslationResults[0].GlossaryOverride == nil ||
backend.translationTaskUpdates[0].TranslationResults[0].GlossaryOverride.Reviewer != "reviewer" {
backend.translationTaskUpdates[0].TranslationResults[0].GlossaryOverride.Reviewer != "reviewer" ||
backend.translationTaskUpdates[0].TranslationResults[0].GlossaryOverride.QAIdentity != "gqa-v1-test" {
t.Fatalf("translation task updates=%#v", backend.translationTaskUpdates)
}
if len(backend.translationMemoryConfirmParams) != 1 ||
+5 -2
View File
@@ -534,7 +534,7 @@ paths:
type: string
responses:
"200":
description: Glossary constraints, diagnostics, and blocked decision.
description: Glossary constraints, diagnostics, blocked decision, and stable qa_identity.
"400":
description: Missing source text or invalid context.
"401":
@@ -628,9 +628,12 @@ paths:
type: string
glossary_override:
type: object
required: [reviewer, reason, provenance, confirmed_unix_seconds]
required: [qa_identity, reviewer, reason, provenance, confirmed_unix_seconds]
additionalProperties: false
properties:
qa_identity:
type: string
minLength: 1
reviewer:
type: string
reason:
+1
View File
@@ -513,6 +513,7 @@ type GlossaryTermSnapshot struct {
// GlossaryOverride records explicit human approval for a deviation.
type GlossaryOverride struct {
QAIdentity string `json:"qa_identity"`
Reviewer string `json:"reviewer"`
Reason string `json:"reason"`
Provenance string `json:"provenance"`
+1
View File
@@ -464,6 +464,7 @@ func TestTranslationTaskUpdateSendsWorkerParams(t *testing.T) {
SourceText: "source",
TranslatedText: "译文",
GlossaryOverride: &GlossaryOverride{
QAIdentity: "gqa-v1-test",
Reviewer: "reviewer",
Reason: "approved deviation",
Provenance: "manual-review",