Files
Enterprise-Knowledge-Base/cli/internal/cmdutil/factory.go
T
nullkey 2ee9741fa1 refactor(cli): finish context→profile cascade + post-review hardening (BREAKING)
Post-review polish on the v0.7 wire / surface contract. Bundles five
follow-ups that landed after the main BREAKING feat commit:

1. Complete context→profile cascade (internal API + YAML schema)

The prior commit renamed only the user-visible surface (commands /
flags / env / project link / envelope field). The internal Go API
and on-disk config schema were still half-renamed — an L-25
self-consistency violation flagged by post-merge review. Closed here:

Internal Go API:
- config.Context           → config.Profile
- config.Config.CurrentContext → CurrentProfile
- config.Config.Contexts       → Profiles
- LoginOptions.Context     → LoginOptions.Profile
- clearContextSecrets()    → clearProfileSecrets()
- saveContextRef()         → saveProfileRef()
- secrets.Store: param name `context` → `profile` (interface +
  FileStore + KeyringStore + MemStore)
- cmdutil.LoadSecret(store, context, key) → LoadSecret(store, profile, key)
- cmdutil.RefreshAndPersist's ctxName → profileName
- Local var `ctx := &config.Profile{...}` → `prof := &config.Profile{...}`
  in auth/login.go to eliminate the visual collision with Go stdlib
  context.Context that motivated the whole rename in the first place.

On-disk config.yaml schema:
- current_context: → current_profile:
- contexts:       → profiles:
- Pre-1.0 break, no compat alias. Users on v0.6 dogfooded configs
  must delete ~/.config/weknora/config.yaml or hand-rename the two
  keys (CHANGELOG migration note added).

Tests / fixtures / golden files:
- factory_test.go YAML fixture + assertion updated.
- acceptance/e2e/e2e_test.go writeContextYAML → writeProfileYAML,
  fixture YAML keys updated.
- acceptance/testdata/wire/doctor.error_network.json golden updated
  ("active context" → "active profile" in hint string).

User-visible prose sweep:
- cmd/mcp/serve.go --help Long: "active context (or --context)" →
  "active profile (or --profile)" — most-visible miss.
- cmd/{kb/list, search/kb, session/list, api/api} Short/Long help.
- cmd/auth/login.go stdout: `(context=%s)` → `(profile=%s)`.
- cmd/auth/logout.go error: `"no current context"` → `"no current profile"`.
- cmd/doctor/doctor.go hint string (also the wire golden above).
- cmd/auth/refresh.go error: `"refresh token missing for context"` →
  `"refresh token missing for profile"`.
- README.md: `## Multi-context` H2 → `## Multi-profile`; code-block
  comment `# current context` → `# current profile`.

Code-comment / docstring sweep across cli/cmd/auth/ and
cli/internal/cmdutil/. Comments referencing Go stdlib context.Context,
the RAG / LLM "context window" concept, and historical CHANGELOG
entries for v0.4 / v0.5 were left alone.

CHANGELOG v0.7 BREAKING entry gains the on-disk-schema bullet under
the existing "context → profile" item.

2. Profile name validation (shell-injection guard)

`envelope.error.retry_command` is a single shell-string field. An
AI agent that exec()s it via `sh -c <retry_command>` was injectable
through a maliciously-named profile:

  weknora auth logout --name 'x; rm -rf ~'
  # would produce: retry_command = "weknora auth logout --name x; rm -rf ~ -y"

`cmd/profile/add.go` already enforced an alphanumeric + `-_.`
allowlist via `validateName`. The `auth login` and `auth logout`
paths bypassed it.

- Moved validation from `cmd/profile/add.go` to
  `cli/internal/cmdutil/profilename.go` as exported
  `ValidateProfileName` (cmdutil is the import-cycle-safe home;
  internal/config can't depend on cmdutil).
- `auth login` runs the validator before any persist call.
- `auth logout` runs the validator on `opts.Name` before
  constructing `retry_command`.
- Unit tests (`profilename_test.go`) cover the allowlist, empty
  rejection, path-traversal, shell metacharacters (`;`, `&`, `|`,
  `$()`, backticks, quotes, whitespace, glob, redirects), and the
  user-facing hint text. The shell-metachar test exists as a
  regression guard.

Wire shape (`retry_command` string → `retry_command_argv []string`)
remains a v0.8 additive change per ROADMAP — this fix removes the
practical exploit path without touching the wire contract.

3. AI-agent terminology disambiguation

"agent" has three referents in this codebase: (a) WeKnora's
server-side Custom Agent resource, (b) the removed `agent invoke`
verb, (c) external LLM/automation consumers. Per project memory
feedback_no_meta_disambiguation_in_docs, the fix is full-term
naming, not "X has N meanings" prose. Surgical changes at section
headers + ambiguous prose:

- AGENTS.md: "Agent decision shortcuts" → "AI agent decision
  shortcuts"; "agent-callable surface" → "AI-agent-callable
  surface".
- README.md: "Designed to be agent-first" → "AI-agent-first";
  "Other agent ergonomics" → "Other AI-agent ergonomics"; "in
  agent contexts" → "in AI-agent contexts"; "for CI / agents" →
  "for CI / AI agents".

Anaphoric "agents" inside paragraphs that already established
"AI agents" was left alone — full substitution everywhere would
have been prose noise without clarity gain.

4. Wire-contract review follow-ups

Real findings from a second-pass review of the v0.7 envelope /
streaming / surface design. Per project memory
feedback_check_in_domain_anchor_first, candidate findings were
first verified against the in-domain peer CLI explicitly cited as
the envelope anchor; two earlier-flagged issues turned out to be
in-pattern and were withdrawn.

Surviving fixes:

- AGENTS.md success-envelope example rewritten. The prior example
  showed `has_more: false` / `_notice: {}` as if they were always
  present, but both fields are `omitempty` and never serialize
  when zero / nil. Replaced with three realistic shapes (list /
  single resource / mutation with no payload) and added a note
  that optional fields are omitted when empty.

- cmd/chat/chat.go Args: MinimumNArgs(1) → ExactArgs(1).
  v0.6 silently joined `weknora chat hello world` into
  `"hello world"`. v0.7 now rejects multi-arg with exit 2,
  matching `weknora session ask`. BREAKING; CHANGELOG entry
  added under v0.7 BREAKING.

- internal/output/envelope.go extracts NewEnvelope(data, meta,
  profile) constructor. The jq-filter path in
  cmdutil.FormatOptions.Emit was manually rebuilding the
  envelope literal alongside the canonical WriteEnvelope path —
  drift risk when fields are added. Single construction point now.

- internal/cmdutil/factory.go adds AddKBFlag(cmd) helper.
  Five files (chat, doc/list, doc/upload, doc/create, doc/fetch)
  had verbatim-identical `cmd.Flags().String("kb", ...)`
  declarations. Centralised so flag name + help text stay
  in sync with Factory.ResolveKB. Docstring reordering + gofmt
  fixup landed in the same edit to keep ResolveKB's own godoc
  attached to its function.

5. OSS-readiness comment / doc sweep

Pre-publication scrub of code, comments, and shipped Markdown to
remove references that only make sense in the development repo:

- AGENTS.md "Deliberate deviations + mainstream alignments"
  section: removed peer-project name-drops from the comparison
  table; rewrote as five flagged design decisions with rationale
  but no specific competitor named. The four rows that previously
  contrasted against a named peer CLI now state WeKnora's choice
  + rationale directly. Section header renamed to "Design
  decisions worth flagging" since it is no longer a
  deviation/alignment matrix.

- CHANGELOG v0.7 BREAKING rationales: three references to a
  named peer CLI removed; the context→profile rationale now
  cites only mainstream multi-credential CLIs by category (AWS /
  Stripe / OpenAI / Anthropic), and the `api -d/--data` removal
  rationale cites only `gh api` / `curl`. `chat` BREAKING entry
  rationale similarly simplified.

- 35 cross-references to design-spec section numbers (§4.1 /
  §4.5 / §5.3 etc.) removed from Go doc comments and test
  comments across 13 files. The referenced spec lives outside
  the shipped tree; readers of the public repo cannot resolve
  them. Each reference replaced with a self-contained semantic
  description (e.g. "the batch envelope" / "AGENTS.md section
  on the success path").

- Mixed-language strings translated to English:
  - Four Go comments: internal/cmdutil/exit.go:213,215,
    internal/cmdutil/errors.go:156,
    internal/output/batch_test.go:90,
    internal/output/envelope_test.go:27.
  - One CHANGELOG section title:
    `v0.7 — Agent-first wire contract + 命令面集中清理` →
    `... + command-surface cleanup`.
  - CJK test fixtures (internal/text/truncate_test.go CJK
    truncation cases, cmd/session/list_test.go Chinese session
    title, acceptance/e2e/e2e_test.go Chinese RAG corpus)
    retained — they are intentional test inputs, not stray prose.

- Makefile help comment: `golangci-lint added in PR-9` →
  `golangci-lint planned`. Internal PR numbering should not
  surface in shipped Makefile prose.

Build green, 28/28 packages, +5 new ValidateProfileName tests.
go vet / gofmt / go mod verify / go mod tidy all clean.

Rationale for the cascade: pre-1.0 is the cheapest moment to close
L-25 self-consistency (L-26). The half-finished internal rename
would have perpetuated the very `context` vs `context.Context`
ambiguity that motivated v0.7's user-visible rename in the first
place.
2026-05-27 10:56:34 +08:00

301 lines
11 KiB
Go

package cmdutil
import (
"context"
"errors"
"fmt"
"io"
"net/http"
"os"
"sync"
"github.com/spf13/cobra"
"github.com/Tencent/WeKnora/cli/internal/config"
"github.com/Tencent/WeKnora/cli/internal/iostreams"
"github.com/Tencent/WeKnora/cli/internal/projectlink"
"github.com/Tencent/WeKnora/cli/internal/prompt"
"github.com/Tencent/WeKnora/cli/internal/secrets"
sdk "github.com/Tencent/WeKnora/client"
)
// Factory is the dependency container injected at command construction. Each
// closure is lazy: --help / completion / `weknora version` must NOT trigger
// HTTP, keyring access, or filesystem I/O beyond the bare minimum.
//
// Four closures:
// - Config: parses ~/.config/weknora/config.yaml (no network)
// - Client: constructs the SDK client; only Secrets is sync.Once-cached,
// so callers should hold the returned *sdk.Client across
// multiple SDK calls within one invocation
// - Prompter: returns interactive prompter; agent mode returns AgentPrompter
// - Secrets: builds the OS keyring / file fallback credential store the
// first time it is requested (probing the keyring at startup
// would fork+exec on macOS and DBus-touch on Linux,
// defeating the lazy contract above).
//
// IOStreams is intentionally NOT a Factory closure - it is the package singleton
// iostreams.IO. The bar to add a new closure is at least 2 commands sharing the
// same dependency; resist factory bloat.
//
// Client returns a *sdk.Client (the WeKnora SDK). Commands that want narrow
// service interfaces declare them in their own files and let the real SDK
// satisfy them implicitly via duck typing.
type Factory struct {
Config func() (*config.Config, error)
Client func() (*sdk.Client, error)
Prompter func() prompt.Prompter
Secrets func() (secrets.Store, error)
// ProfileOverride, if non-empty, replaces config.CurrentProfile for this
// invocation only - set by the global --profile flag in PersistentPreRun.
// Buildable Config() / Client() honor it without writing to disk.
ProfileOverride string
}
// New constructs a production Factory wired to real config / SDK client.
//
// All closures are lazy: invoking --help, version, or shell completion runs
// none of them. Client and Secrets closures memoize via sync.Once so the
// SDK client is built (and the keyring is probed) at most once per process,
// even when Factory.ResolveKB internally calls f.Client() before the
// command's RunE calls it again - without this, name-resolved --kb paths
// would build two clients with two AuthRetryTransports holding independent
// token state.
func New() *Factory {
var (
secretsOnce sync.Once
secretsStore secrets.Store
secretsErr error
clientOnce sync.Once
client *sdk.Client
clientErr error
)
f := &Factory{}
f.Config = func() (*config.Config, error) {
cfg, err := config.Load()
if err != nil {
// Map raw fs / parse errors to typed codes so the stderr line
// doesn't surface bare `server.error` for what's actually a
// local IO / corrupt-config problem.
if errors.Is(err, config.ErrCorrupt) {
return nil, Wrapf(CodeLocalConfigCorrupt, err, "config malformed")
}
return nil, Wrapf(CodeLocalFileIO, err, "load config")
}
if f.ProfileOverride != "" {
cfg.CurrentProfile = f.ProfileOverride
}
return cfg, nil
}
f.Client = func() (*sdk.Client, error) {
clientOnce.Do(func() { client, clientErr = buildClient(f) })
return client, clientErr
}
f.Prompter = func() prompt.Prompter {
if iostreams.IO.IsStdoutTTY() && iostreams.IO.IsStderrTTY() {
return prompt.NewTTYPrompter()
}
return prompt.AgentPrompter{}
}
f.Secrets = func() (secrets.Store, error) {
secretsOnce.Do(func() {
secretsStore, secretsErr = secrets.NewBestEffortStore()
})
return secretsStore, secretsErr
}
return f
}
// buildClient resolves the active profile, loads the credentials from secrets,
// and constructs a *sdk.Client. Returns CodeAuthUnauthenticated when no
// credentials are available so the user gets the right hint to run
// `weknora auth login`.
func buildClient(f *Factory) (*sdk.Client, error) {
cfg, err := f.Config()
if err != nil {
return nil, err
}
profileName := cfg.CurrentProfile
if profileName == "" {
return nil, NewError(CodeAuthUnauthenticated, "no current profile configured; run `weknora auth login` to set one up")
}
prof, ok := cfg.Profiles[profileName]
if !ok {
// If the user explicitly overrode the profile (via --profile flag or
// WEKNORA_PROFILE env), it's a bad argument - not a corrupt config file.
// The destructive "remove config.yaml" hint would be catastrophic for a typo.
if f.ProfileOverride != "" {
return nil, NewError(CodeInputInvalidArgument,
fmt.Sprintf("profile %q not configured", profileName)).
WithHint("list available profiles with `weknora profile list`").
WithRetryCommand("weknora profile list")
}
// ProfileOverride is empty: config.CurrentProfile points at a missing entry.
// That's a genuinely corrupt config file.
return nil, NewError(CodeLocalConfigCorrupt, fmt.Sprintf("config references unknown profile %q", profileName))
}
if prof.Host == "" {
return nil, NewError(CodeLocalConfigCorrupt, fmt.Sprintf("profile %q has no host", profileName))
}
opts := []sdk.ClientOption{}
store, err := f.Secrets()
if err != nil {
return nil, Wrapf(CodeLocalKeychainDenied, err, "init secrets store")
}
// Only fetch the secrets the profile actually references. Skipping the
// unused fetch avoids a `security` exec (macOS) / DBus call (Linux) per
// authenticated invocation.
var accessToken string
if prof.TokenRef != "" {
if access, err := LoadSecret(store, profileName, "access"); err != nil {
return nil, err
} else if access != "" {
accessToken = access
opts = append(opts, sdk.WithBearerToken(access))
}
}
if prof.APIKeyRef != "" {
if apiKey, err := LoadSecret(store, profileName, "api_key"); err != nil {
return nil, err
} else if apiKey != "" {
opts = append(opts, sdk.WithAPIKey(apiKey))
}
}
// JWT profiles (have both access + refresh refs) get the transparent
// 401-retry transport: on the first 401 from a non-/auth/* endpoint, the
// transport reads the stored refresh token, calls /api/v1/auth/refresh,
// persists the new pair, and replays the original request with the new
// bearer. API-key profiles skip this (no refresh semantic) - a 401 from
// them propagates as auth.unauthenticated for the caller to handle.
if prof.TokenRef != "" && prof.RefreshRef != "" {
refreshFn := func(rctx context.Context) (string, error) {
return refreshAccessToken(rctx, store, prof.Host, profileName)
}
opts = append(opts, sdk.WithTransport(
NewAuthRetryTransport(http.DefaultTransport, accessToken, refreshFn),
))
}
// prof.TenantID is intentionally NOT injected as X-Tenant-ID. Servers derive
// tenant from the credential itself (JWT claim or API key prefix); the
// header is only meaningful for explicit cross-tenant switching by users
// with CanAccessAllTenants. Auto-mirroring the persisted tenant from config
// breaks that contract - explicit cross-tenant flags would be required
// before sending it. `tenant_id` stays in config for `auth status` display only.
return sdk.NewClient(prof.Host, opts...), nil
}
// AddKBFlag registers the standard `--kb` flag that ResolveKB reads. Use this
// instead of duplicating the flag declaration in every command that scopes to
// a knowledge base — one source of truth for flag name and help text.
func AddKBFlag(cmd *cobra.Command) {
cmd.Flags().String("kb", "", "Knowledge base UUID or name (overrides env / project link)")
}
// ResolveKB returns the active KB id for the running command, applying the
// 4-level fallback chain (highest to lowest):
// 1. --kb flag (kb_<...> id passed through; anything else resolved via
// ListKnowledgeBases as a name → id lookup)
// 2. WEKNORA_KB_ID env (always an explicit id)
// 3. .weknora/project.yaml (walk-up from cwd)
// 4. error: kb required
func (f *Factory) ResolveKB(cmd *cobra.Command) (string, error) {
if v, _ := cmd.Flags().GetString("kb"); v != "" {
if IsKBID(v) {
return v, nil
}
c, err := f.Client()
if err != nil {
return "", err
}
return ResolveKBNameToID(cmd.Context(), c, v)
}
if v := os.Getenv("WEKNORA_KB_ID"); v != "" {
return v, nil
}
cwd, err := os.Getwd()
if err == nil {
if path, found, derr := projectlink.Discover(cwd); derr == nil && found {
p, lerr := projectlink.Load(path)
if lerr != nil {
return "", Wrapf(CodeProjectLinkCorrupt, lerr, "read project link")
}
if p.KBID != "" {
return p.KBID, nil
}
}
}
return "", NewError(CodeKBIDRequired, "kb is required")
}
// ApplyLogLevel resolves --log-level / WEKNORA_LOG_LEVEL (in priority order)
// and applies the result to the SDK's debug logger. Intended to be called
// from the root command's PersistentPreRunE so the resolved level is in
// effect before any SDK call.
//
// Returns a typed error if the user passed an explicit --log-level with
// an invalid value — matches the strictness of --format validation
// (env values stay silent-fallthrough; flag values are strict).
func (f *Factory) ApplyLogLevel(cmd *cobra.Command, stderr io.Writer) error {
if cmd != nil {
if fl := cmd.Flags().Lookup("log-level"); fl != nil && fl.Changed {
if !IsValidLogLevel(fl.Value.String()) {
return NewFlagError(fmt.Errorf(
"invalid --log-level %q: must be error | warn | info | debug", fl.Value.String()))
}
}
}
level, _ := ResolveLogLevel(cmd, stderr)
sdk.SetDebugLevel(level)
return nil
}
// LoadSecret fetches a named secret for the given profile from the keyring.
// Returns ("", nil) when the secret is absent (ErrNotFound); a real keyring
// access failure surfaces as CodeLocalKeychainDenied. Used by buildClient
// to assemble SDK auth options and by `auth token` to expose the raw
// credential for shell scripting.
func LoadSecret(store secrets.Store, profile, key string) (string, error) {
v, err := store.Get(profile, key)
if errors.Is(err, secrets.ErrNotFound) {
return "", nil
}
if err != nil {
return "", Wrapf(CodeLocalKeychainDenied, err, "load %s", key)
}
return v, nil
}
// refreshAccessToken is the closure target injected into AuthRetryTransport's
// refreshFn. A fresh SDK Client is built here rather than reusing the one
// being constructed - that one is itself wrapped by the transport, which
// would recurse on refresh. The refresh endpoint is unauthenticated apart
// from the refresh token in the body, so no credential options are needed.
func refreshAccessToken(ctx context.Context, store secrets.Store, host, profileName string) (string, error) {
return RefreshAndPersist(ctx, store, sdk.NewClient(host), profileName)
}
// ActiveProfile returns the resolved profile name for this invocation:
// 1. ProfileOverride (set by --profile flag in root PersistentPreRunE)
// 2. WEKNORA_PROFILE env var
// 3. Config's CurrentProfile (the persisted active profile name)
// 4. Empty string when nothing is configured (envelope omits the field).
func (f *Factory) ActiveProfile() string {
if f.ProfileOverride != "" {
return f.ProfileOverride
}
if v := os.Getenv("WEKNORA_PROFILE"); v != "" {
return v
}
if f.Config == nil {
return ""
}
cfg, err := f.Config()
if err != nil || cfg == nil {
return ""
}
return cfg.CurrentProfile
}