Skip to content

World's first open-source enterprise world model.

The single-document extract path gets the same transaction as the batch one (#31)

WaylandYang73d3f47e876a
diff --git a/crates/utopia-server/src/api/graph_routes.rs b/crates/utopia-server/src/api/graph_routes.rs
index 904d9e9e..abc75206 100644
--- a/crates/utopia-server/src/api/graph_routes.rs
+++ b/crates/utopia-server/src/api/graph_routes.rs
@@ -120,15 +120,9 @@ pub async fn extract(
) -> ApiResult<Json<serde_json::Value>> {
let doc = utopia_store::documents::get(&state.pool, document_id).await?;
require_kb(&state, &user, doc.kb_id, Role::Editor).await?;
- // 手动触发 = 强制全量:清掉增量跳过标记,重抽每一块
- utopia_store::documents::clear_extraction_marks(&state.pool, document_id).await?;
- utopia_store::documents::set_graph_status(&state.pool, document_id, "queued").await?;
- let job_id = utopia_store::jobs::enqueue(
- &state.pool,
- "extract_document",
- json!({ "document_id": document_id }),
- )
- .await?;
+ // 手动触发 = 强制全量:清增量标记、解雇在跑的任务、置 queued、建任务,一个事务办完
+ let job_id = utopia_store::documents::queue_extraction_one(&state.pool, document_id).await?;
+ state.emit_document(doc.kb_id, document_id);
Ok(Json(json!({ "job_id": job_id })))
}
diff --git a/crates/utopia-store/src/documents.rs b/crates/utopia-store/src/documents.rs
index 8870bbc7..c7b31e4f 100644
--- a/crates/utopia-store/src/documents.rs
+++ b/crates/utopia-store/src/documents.rs
@@ -393,10 +393,12 @@ pub async fn mark_chunk_extracted(pool: &PgPool, chunk_id: Uuid) -> AppResult<()
Ok(())
}
-/// 清除文档现行分块的抽取标记(手动 Extract = 强制全量重抽)。
-pub async fn clear_extraction_marks(pool: &PgPool, document_id: Uuid) -> AppResult<()> {
- // 两条同事务:清了标记却没换到新 epoch,在跑的旧任务便察觉不到自己已被顶替,
- // 会继续把刚清空的分块重抽一遍。
+/// 单篇排队重抽(手动 Extract = 强制全量):清增量标记、解雇在跑的任务、置
+/// queued、建抽取任务——与 `queue_extraction` 同一套语义,只是作用于一篇。返回 job id。
+///
+/// 整体一个事务。分开做时任一步失败都会留下半截状态:清了标记却没换 epoch,
+/// 旧任务察觉不到自己已被顶替;或者置了 queued 却没建成任务,文档就此无人接手。
+pub async fn queue_extraction_one(pool: &PgPool, document_id: Uuid) -> AppResult<i64> {
let mut tx = pool.begin().await?;
sqlx::query(
"UPDATE chunks SET extracted_at = NULL
@@ -405,13 +407,32 @@ pub async fn clear_extraction_marks(pool: &PgPool, document_id: Uuid) -> AppResu
.bind(document_id)
.execute(&mut *tx)
.await?;
- // 与批量重抽同理:自增 epoch 解雇可能正在跑的旧任务
- sqlx::query("UPDATE documents SET extract_epoch = extract_epoch + 1 WHERE id = $1")
- .bind(document_id)
- .execute(&mut *tx)
- .await?;
+ // 未开跑的旧任务顺手删掉:连点两次 Extract 不该攒出两个任务同抽一篇
+ sqlx::query(
+ "DELETE FROM jobs WHERE kind = 'extract_document' AND status = 'queued'
+ AND payload->>'document_id' = $1",
+ )
+ .bind(document_id.to_string())
+ .execute(&mut *tx)
+ .await?;
+ sqlx::query(
+ "UPDATE documents SET graph_status = 'queued', graph_error = NULL,
+ extract_epoch = extract_epoch + 1
+ WHERE id = $1",
+ )
+ .bind(document_id)
+ .execute(&mut *tx)
+ .await?;
+ let (job_id,): (i64,) = sqlx::query_as(
+ "INSERT INTO jobs (kind, payload)
+ VALUES ('extract_document', jsonb_build_object('document_id', $1::text))
+ RETURNING id",
+ )
+ .bind(document_id)
+ .fetch_one(&mut *tx)
+ .await?;
tx.commit().await?;
- Ok(())
+ Ok(job_id)
}
/// 文档查看器:全部分块(按顺序)。