d3c03b2f34
A hardening + finalization pass over the agent-first CLI: correctness fixes,
richer machine-readable signals, flag/naming consistency, and a symmetric
config surface. Pre-1.0, so it includes breaking renames.
Correctness:
- agent update: resolve/validate --model/--rerank-model (was storing a bogus
name verbatim, corrupting config.model_id).
- doctor: honor WEKNORA_HOST / WEKNORA_API_KEY (headless path no longer reports
"no host configured").
- session ask / MCP session_ask: text answer was empty on non-TTY — the agent
stream sets Done=true on an intermediate agent_query frame before the answer,
and AgentAccumulator treated the first Done as terminal. Terminate on the
`complete` event (new sdk.AgentResponseTypeComplete), not a per-frame Done.
- batch exit codes: any per-item failure collapses to operation.failed (exit 1),
including `doc upload --recursive` partial failures — a permanent per-file
failure (e.g. a duplicate) no longer surfaces as a retryable exit 7 an agent
would loop on; per-item typed errors stay in the envelope.
Agent-first signals & discovery:
- error.exit_code in the envelope (type + exit_code disambiguate the
input.invalid_argument exit-2-vs-5 split in one JSON read).
- meta.hint on empty content search and on draft doc create; doc wait fails fast
on a never-parsing draft instead of hanging to --timeout.
- retrieval-readiness is visible in the natural flow: kb status / kb check emit
retrieval_ready, and kb create hints the fix when no embedding model is bound
— an unconfigured KB no longer looks silently healthy.
- schema contract completeness: every leaf declares output + >=1 example
(drift-guarded); output strings match the meta actually emitted; chunk list
and search docs now emit meta.total_count (both previously dropped it).
- schema tolerates a quoted multi-word command label; zero-state auth — and
`link` with no profile — point at profile setup / the headless WEKNORA_KB_ID
path instead of looping on `auth login`.
- id-addressed reads tolerate a redundant --kb (doc view/wait, chunk list/view
accept and ignore it, declared in schema) so a carried-over --kb doesn't
exit 2; streaming commands warn that --jq does not apply to an NDJSON stream.
- keep JSON-always as the default; --jq hints spell out the .data path.
Consistency & gating:
- doc create: drop the deprecated --name alias (--title only; pre-1.0 break).
- chunk list --limit aligned to 1..10000; model list --limit/-L with
has_more/total_count; api write-gates -X PUT/PATCH (exit 10); skills install
expands a leading ~.
- docs corrected: search docs / doc list --keyword help is case-insensitive
(server does LOWER LIKE); AGENTS.md risk-action list (no phantom kb.init; add
model.update / kb.config.set) and batch example (failed item carries `error`);
session resume --message id comes from `message list`, not the stream.
- auth/profile ergonomics: env credentials are now first-class — `auth token`
prints the active WEKNORA_API_KEY / WEKNORA_TOKEN, and auth login/logout/refresh
give an env-aware message instead of looping on "run auth login". `auth logout`
clears credentials but keeps the profile registered (host preserved for
re-login); deleting a profile is `profile remove`'s job (clean logout/remove
separation, matching gh / lark).
Config surface (symmetric read/write, in-place model edits):
- kb config now returns a secret-free KBModelConfigView (was {}); `kb config`
reads, new `kb config set` writes; `kb init` removed (misnomer).
- kb create --chat-model: retrieval-ready in one step.
- model update: edit a model in place (id preserved, references survive) —
rotate --api-key-stdin, change base-url / display-name / etc.
- session continue-stream renamed to session resume.
Docs: AGENTS.md is the single wire-contract source; CHANGELOG slimmed; stale
kb-init / continue-stream references removed; skill wire-vocab guard extended.
AGENTS.md / weknora-shared SKILL document retrieval_ready (a KB needs an
embedding model to be searchable), that --jq does not apply to NDJSON
streams, and the env-credential-first auth path; the KB quickstart example
now creates a retrieval-ready KB.
167 lines
6.8 KiB
Go
167 lines
6.8 KiB
Go
package kb
|
|
|
|
import (
|
|
"context"
|
|
"fmt"
|
|
"strings"
|
|
|
|
"github.com/spf13/cobra"
|
|
|
|
"github.com/Tencent/WeKnora/cli/internal/cmdutil"
|
|
"github.com/Tencent/WeKnora/cli/internal/iostreams"
|
|
sdk "github.com/Tencent/WeKnora/client"
|
|
)
|
|
|
|
type ConfigSetOptions struct {
|
|
ChatModel string
|
|
EmbeddingModel string
|
|
Yes bool
|
|
DryRun bool
|
|
}
|
|
|
|
// ConfigSetService is the narrow SDK surface this command depends on. SetKBModelConfig
|
|
// points the KB at already-registered models; GetInitializationConfig re-reads
|
|
// the server's resulting state so the success envelope reflects what stuck.
|
|
type ConfigSetService interface {
|
|
SetKBModelConfig(ctx context.Context, kbID string, cfg *sdk.KBModelConfig) error
|
|
GetInitializationConfig(ctx context.Context, kbID string) (*sdk.KBModelConfigView, error)
|
|
}
|
|
|
|
// newKBModelWriteCmd builds the `kb config set` model-binding write command.
|
|
// head is the argv prefix used for the risk action and retry_argv (weknora kb
|
|
// config set).
|
|
func newKBModelWriteCmd(f *cmdutil.Factory, use string, head []string) *cobra.Command {
|
|
opts := &ConfigSetOptions{}
|
|
action := strings.Join(head[1:], ".") // e.g. "kb.config.set"
|
|
cmd := &cobra.Command{
|
|
Use: use,
|
|
Short: "Bind embedding + chat models to a knowledge base (make it usable)",
|
|
Long: `Bind already-registered models to a knowledge base so it can embed, retrieve,
|
|
and generate. Both --chat-model (LLM, used for generation/summary) and
|
|
--embedding-model (used for retrieval) are required; register models first with
|
|
'weknora model create' and discover ids with 'weknora model list'.
|
|
|
|
High-risk write: changing a KB's embedding model affects how its content is
|
|
indexed and searched (the server refuses to CHANGE it once the KB has
|
|
documents; setting it on an unconfigured KB is allowed). Without -y/--yes in a
|
|
non-TTY / JSON context it exits 10 (input.confirmation_required) without
|
|
applying the change.`,
|
|
Args: cobra.ExactArgs(1),
|
|
RunE: func(c *cobra.Command, args []string) error {
|
|
fopts, err := cmdutil.CheckFormatFlag(c)
|
|
if err != nil {
|
|
return err
|
|
}
|
|
fopts.ResolveDefault(iostreams.IO.IsStdoutTTY())
|
|
opts.Yes, _ = c.Flags().GetBool("yes")
|
|
kbID := args[0]
|
|
// Validate required flags before the dry-run gate so --dry-run rejects
|
|
// identically to the live path.
|
|
if err := validateConfigSetFlags(opts); err != nil {
|
|
return err
|
|
}
|
|
if handled, err := cmdutil.HandleDryRun(c, opts.DryRun, cmdutil.DryRunPlan{
|
|
Action: action,
|
|
Args: map[string]any{"kb": kbID, "chat_model": opts.ChatModel, "embedding_model": opts.EmbeddingModel},
|
|
}); handled {
|
|
return err
|
|
}
|
|
cli, err := f.Client()
|
|
if err != nil {
|
|
return err
|
|
}
|
|
if err := cmdutil.ConfirmDestructive(f.Prompter(), opts.Yes, fopts.WantsJSON(),
|
|
"configure", "knowledge base", kbID, action,
|
|
cmdutil.BuildRetryArgv(c, append(append([]string{}, head...), kbID), "chat-model", "embedding-model", "format")); err != nil {
|
|
return err
|
|
}
|
|
// Resolve name-or-id for the model flags (a UUID passes through; a
|
|
// name is looked up among models of the expected type). Network read
|
|
// on the live path only — the dry-run above shows the raw refs.
|
|
if opts.ChatModel, err = cmdutil.ResolveModelRef(c.Context(), cli, opts.ChatModel, "KnowledgeQA"); err != nil {
|
|
return err
|
|
}
|
|
if opts.EmbeddingModel, err = cmdutil.ResolveModelRef(c.Context(), cli, opts.EmbeddingModel, "Embedding"); err != nil {
|
|
return err
|
|
}
|
|
return runConfigSet(c.Context(), opts, fopts, cli, kbID)
|
|
},
|
|
}
|
|
cmd.Flags().StringVar(&opts.ChatModel, "chat-model", "", "Chat / LLM model id or name for generation & summary (required) — see `weknora model list`")
|
|
cmd.Flags().StringVar(&opts.EmbeddingModel, "embedding-model", "", "Embedding model id or name for retrieval (required) — see `weknora model list`")
|
|
cmdutil.AddFormatFlag(cmd, kbConfigFields...)
|
|
cmdutil.AddDryRunFlag(cmd, &opts.DryRun)
|
|
cmdutil.SetRisk(cmd, action)
|
|
cmdutil.SetAgentHelp(cmd, cmdutil.AgentHelp{
|
|
UsedFor: "bind models to a KB so it becomes retrieval-ready. --chat-model and --embedding-model are required and accept a model id or name; discover them with `weknora model list`. Read the result back with `weknora kb config`.",
|
|
RequiredFlags: []string{"<kb-id> (positional)", "--chat-model", "--embedding-model"},
|
|
Examples: []string{
|
|
"weknora kb config set kb_abc --chat-model model_llm --embedding-model model_emb -y",
|
|
},
|
|
Output: "envelope.data is the resulting secret-free config view {retrieval_ready, embedding, llm, rerank, multimodal}",
|
|
Warnings: []string{
|
|
"Requires explicit user approval (exit 10 / input.confirmation_required); never auto-add -y.",
|
|
"The server refuses to CHANGE the embedding model of a KB that already has documents (setting it on an unconfigured KB is fine).",
|
|
},
|
|
})
|
|
return cmd
|
|
}
|
|
|
|
func validateConfigSetFlags(opts *ConfigSetOptions) error {
|
|
var missing []string
|
|
if strings.TrimSpace(opts.ChatModel) == "" {
|
|
missing = append(missing, "--chat-model")
|
|
}
|
|
if strings.TrimSpace(opts.EmbeddingModel) == "" {
|
|
missing = append(missing, "--embedding-model")
|
|
}
|
|
if len(missing) == 0 {
|
|
return nil
|
|
}
|
|
return &cmdutil.Error{
|
|
Code: cmdutil.CodeInputMissingFlag,
|
|
Message: "kb config set requires " + strings.Join(missing, " and "),
|
|
Hint: "discover model ids with `weknora model list` (or register one with `weknora model create`), then pass --chat-model <id> --embedding-model <id>",
|
|
}
|
|
}
|
|
|
|
func runConfigSet(ctx context.Context, opts *ConfigSetOptions, fopts *cmdutil.FormatOptions, svc ConfigSetService, kbID string) error {
|
|
if err := validateConfigSetFlags(opts); err != nil {
|
|
return err
|
|
}
|
|
cfg := &sdk.KBModelConfig{
|
|
LLMModelID: opts.ChatModel,
|
|
EmbeddingModelID: opts.EmbeddingModel,
|
|
}
|
|
if err := svc.SetKBModelConfig(ctx, kbID, cfg); err != nil {
|
|
return cmdutil.WrapHTTP(err, "configure knowledge base %q", kbID)
|
|
}
|
|
// Re-read the server's resulting state (secret-free view) so the envelope
|
|
// reflects what stuck — the same shape `kb config` returns.
|
|
result, err := svc.GetInitializationConfig(ctx, kbID)
|
|
if err != nil || result == nil {
|
|
// The write succeeded; surface what we applied if the read-back failed.
|
|
result = &sdk.KBModelConfigView{
|
|
RetrievalReady: opts.EmbeddingModel != "",
|
|
Embedding: sdk.ModelSlotView{Configured: opts.EmbeddingModel != "", ModelName: opts.EmbeddingModel},
|
|
LLM: sdk.ModelSlotView{Configured: opts.ChatModel != "", ModelName: opts.ChatModel},
|
|
}
|
|
}
|
|
if fopts.WantsJSON() {
|
|
return fopts.Emit(iostreams.IO.Out, result, nil)
|
|
}
|
|
fmt.Fprintf(iostreams.IO.Out, "✓ Configured knowledge base %s (chat: %s, embedding: %s)\n",
|
|
kbID, orUnset(result.LLM.ModelName), orUnset(result.Embedding.ModelName))
|
|
return nil
|
|
}
|
|
|
|
func orUnset(s string) string {
|
|
if s == "" {
|
|
return "(unset)"
|
|
}
|
|
return s
|
|
}
|
|
|
|
// compile-time check: the production SDK client implements ConfigSetService.
|
|
var _ ConfigSetService = (*sdk.Client)(nil)
|