From c014c0335493c95b4eb3e00ef20abc5917059f0c Mon Sep 17 00:00:00 2001 From: hfxia Date: Wed, 27 May 2026 15:45:43 +0800 Subject: [PATCH] docs: correct built-in KB dump list and collection-name guidance The install form's "Built-in KnowledgeBase File" selector listed two obsolete dumps (docvec_gte_cs2000/cs3000, described by chunk size). The v1.4 chart (plugin-config.yaml) actually offers three dumps keyed to ACP documentation version, defaulting to acp_4_3: - docvec_gte_acp_4_1_20260508.dump (ACP 4.1) - docvec_gte_acp_4_2_20260508.dump (ACP 4.2) - docvec_gte_acp_4_3_20260508.dump (ACP 4.3, default) Also fix the "PG collection name" guidance: it must match the selected dump's internal collection name (filename without .dump). Since the first-start restore loads the dump as-is without renaming the collection (init-database.sh), selecting a non-default dump while leaving the collection name at its acp_4_3 default silently breaks retrieval. Added a NOTE covering this. Fix the same stale cs2000 name in the upgrade doc's example swap log (final name is PG_COLLECTION_NAME, default docvec_gte_acp_4_3_20260508). --- docs/en/install.mdx | 11 +++++++---- docs/en/upgrade/index.mdx | 2 +- 2 files changed, 8 insertions(+), 5 deletions(-) diff --git a/docs/en/install.mdx b/docs/en/install.mdx index b12b295..5a16c15 100644 --- a/docs/en/install.mdx +++ b/docs/en/install.mdx @@ -104,10 +104,13 @@ The install form is grouped below by topic. Required fields are marked **(requir ``` Then enter the secret name in **pg database secret name**. - **PG database name** — the chat-history database name (default `docvec`). Created on first start if missing. -- **PG collection name** — the LangChain PGVector collection name used by the server. The default value matches the bundled gte-multilingual-base dump and **should not be changed** unless you are deploying a custom KB built with [Build a Custom Knowledge Base](build-custom-kb.mdx). -- **Built-in KnowledgeBase File** — choose which built-in dump to restore on first start. The selector currently offers: - - `docvec_gte_cs2000_20260410.dump` (default) — chunk size 2000, balanced recall and answer focus. - - `docvec_gte_cs3000_20260410.dump` — chunk size 3000, slightly better recall on long-form documents at the cost of larger context per hit. +- **PG collection name** — the LangChain PGVector collection name the server reads from `docvec_sys_kb`. It **must** match the internal collection name of the selected **Built-in KnowledgeBase File**, which by convention is the dump filename without the `.dump` suffix. The default `docvec_gte_acp_4_3_20260508` matches the default dump, so leave it unchanged unless you either (a) select a non-default built-in dump below, or (b) deploy a custom KB built with [Build a Custom Knowledge Base](build-custom-kb.mdx) — in both cases set this to the corresponding dump filename without `.dump`. +- **Built-in KnowledgeBase File** — choose which built-in dump to restore on first start. Each dump is the Alauda Container Platform (ACP) product-documentation knowledge base captured at a specific ACP release, all embedded with the gte-multilingual-base model. Pick the one matching the ACP version you run; the default is the latest. The selector offers: + - `docvec_gte_acp_4_1_20260508.dump` — ACP 4.1 documentation corpus. + - `docvec_gte_acp_4_2_20260508.dump` — ACP 4.2 documentation corpus. + - `docvec_gte_acp_4_3_20260508.dump` (default) — ACP 4.3 documentation corpus (latest). + + > **NOTE:** The first-start restore loads the dump as-is and keeps its internal collection name. If you select a dump other than the default, you **must** also set **PG collection name** to that dump's filename without `.dump` (e.g. `docvec_gte_acp_4_1_20260508`), otherwise the server queries a collection that does not exist and retrieval returns nothing. - **Enable builtin Redis** — when enabled, the chart provisions a single Redis instance used by the rate limiter. When disabled, supply a Redis credentials secret in **redis database secret name**. ### Scheduling diff --git a/docs/en/upgrade/index.mdx b/docs/en/upgrade/index.mdx index 2a17495..f74701b 100644 --- a/docs/en/upgrade/index.mdx +++ b/docs/en/upgrade/index.mdx @@ -101,7 +101,7 @@ kubectl -n cpaas-system logs -l app=smart-doc -c init-database | grep '\[upgrade A successful swap prints lines like: ``` -[upgrade] swap docvec_mpnet_acp_4_1_1215 (in docvec_sys_kb) -> dump docvec_gte_acp_4_1_20260508.dump; final name=docvec_gte_cs2000_20260410 +[upgrade] swap docvec_mpnet_acp_4_1_1215 (in docvec_sys_kb) -> dump docvec_gte_acp_4_1_20260508.dump; final name=docvec_gte_acp_4_3_20260508 [upgrade] 003_data_swap_docvec_mpnet_acp_4_1_1215_20260508-gte-acp-4-x applied. ```