Files
pansy/internal/api/api.go
T
steveandClaude Opus 4.8 3f3a5b057c
Build image / build-and-push (push) Successful in 25s
Gadfly review (reusable) / review (pull_request) Canceled after 5m56s
Adversarial Review (Gadfly) / review (pull_request) Canceled after 5m56s
Agent runtime: majordomo in-process, Ollama Cloud config, chat endpoint (#56)
Everything below the run loop already existed. This is the thing that runs a
model.

The build tag is gone, deliberately. internal/agent's doc comment promised two
separations — cmd/pansy not importing the package, and the tool wiring behind
//go:build majordomo — and both have been rewritten rather than left as a stale
aspiration. A tag that keeps the agent out of the binary only earns its keep if
you would ever ship a build without the agent, and the agent is the point;
keeping it meant an untagged CI that never compiled the code that matters.
majordomo is a real dependency now, resolved from the Gitea instance as a
pseudo-version with no replace directive, so the Docker build (which has no
sibling checkout) resolves it the same way this machine does. It is stdlib-first
and pure Go, so CGO_ENABLED=0 and the single static binary survive.

A TURN IS ONE CHANGE SET. That is the whole reason acting without a confirmation
prompt is defensible: "empty the garlic bed and plant cucumbers" is one object
edit and a dozen planting inserts, and it has to undo as one action rather than
thirteen. The scope is opened even for a turn that turns out to be a question,
because a change set with no revisions is never written — so asking costs
nothing and history isn't littered with empty entries.

The model spec goes to majordomo.Parse verbatim. That grammar, including
comma-separated failover chains, is majordomo's; re-implementing any of it here
would only mean two places to update when it grows. The key needs a bridge
though: majordomo's ollama-cloud preset reads OLLAMA_API_KEY while pansy (like
gadfly) is configured with OLLAMA_CLOUD_API_KEY, so the provider is registered
explicitly on a private registry rather than depending on ambient environment.

Runs are bounded by a step cap, a timeout and majordomo's loop guards. This is
loop safety, not cost control — pansy is a personal tool and spend caps are
explicitly not a v2 concern. A capped run does NOT fail: it kept whatever it
managed to do, that work is recorded and undoable, and the reply says it stopped
early rather than going silent.

The chat endpoint streams. A turn that clears a bed and replants it makes a
dozen tool calls over tens of seconds, and without streaming that is a long
silence followed by everything at once — which reads as a hang, and defeats a
design that rests on watching the canvas change as it happens.

Conversations persist per (user, garden). Client-held history would be lost on a
refresh, which is exactly when someone reloads to check whether the agent's
change landed. Only the user/assistant TEXT is stored, not the model's full
transcript: continuity needs what was said and what came back, and replaying a
stored tool call would replay a decision made against a garden that has since
moved on. It also keeps majordomo's message shape out of the schema.

An instance with no key starts, serves the app, and doesn't advertise the agent
— the routes aren't registered at all, the same shape as OIDC 404ing when
unconfigured. A configured-but-unresolvable model logs and disables the
assistant rather than refusing to boot: a garden planner that won't start
because of a chat feature is worse than one without chat.

Tool refusals reach the model as tool results it can explain, not 500s. The ACL
story only works if it can narrate the refusal.

Closes #56

Co-Authored-By: Claude Opus 4.8 (1M context) <[email protected]>
Claude-Session: https://claude.ai/code/session_01H3zbym8Doka2d7D48maSgZ
2026-07-21 02:15:22 -04:00

187 lines
7.9 KiB
Go

// Package api wires pansy's HTTP surface: a gin engine with structured logging
// and panic recovery, the versioned JSON API under /api/v1, and (via spa.go) the
// embedded single-page-app fallback. Handlers stay thin — decode, call the
// service layer, encode — so all business logic and permission checks live in
// internal/service (added by later issues).
package api
import (
"log/slog"
"net/http"
"github.com/gin-gonic/gin"
sloggin "github.com/samber/slog-gin"
"gitea.stevedudenhoeffer.com/steve/pansy/internal/agent"
"gitea.stevedudenhoeffer.com/steve/pansy/internal/config"
"gitea.stevedudenhoeffer.com/steve/pansy/internal/service"
)
// handlers carries the dependencies shared by every HTTP handler. Handlers stay
// thin: decode the request, call a service method, encode the result.
type handlers struct {
cfg *config.Config
svc *service.Service
oidc *oidcClient // nil unless OIDC is configured (see config.OIDCReady)
// agent is nil unless the assistant is configured; the chat routes are only
// registered when it isn't, so a handler never has to check.
agent *agent.Runner
}
// New builds the gin engine with the standard middleware stack and registers the
// API routes against the given service. The embedded SPA fallback is registered
// separately by the caller via RegisterSPA (see spa.go) so the API can be built
// and tested without a web build present.
func New(cfg *config.Config, svc *service.Service) *gin.Engine {
gin.SetMode(gin.ReleaseMode)
r := gin.New()
r.Use(sloggin.New(slog.Default()), gin.Recovery())
if err := r.SetTrustedProxies(cfg.TrustedProxies); err != nil {
// Do not leave gin's trust-everyone default active on a parse failure —
// that would let any client spoof X-Forwarded-For. Fall back to trusting
// no proxies, which is also the behavior when none are configured.
slog.Error("api: invalid trusted proxies, trusting none", "error", err)
_ = r.SetTrustedProxies(nil)
}
h := &handlers{cfg: cfg, svc: svc}
v1 := r.Group("/api/v1")
// CSRF defense for every state-changing API call (no-op unless PANSY_BASE_URL
// is set; see csrfGuard).
v1.Use(h.csrfGuard())
v1.GET("/healthz", healthz)
// Auth endpoints are exempt from requireAuth (you can't be logged in yet);
// /me is the one that needs a session. Feature routers in later issues attach
// h.requireAuth() to their own protected groups.
auth := v1.Group("/auth")
auth.POST("/register", h.register)
auth.POST("/login", h.login)
auth.POST("/logout", h.logout)
auth.GET("/providers", h.providers)
auth.GET("/me", h.requireAuth(), h.me)
// OIDC routes exist only when OIDC can actually be offered, so an unconfigured
// instance 404s them (matching what /auth/providers advertises). Provider
// discovery is lazy (first request), so a briefly-unreachable IdP doesn't stop
// the server — or local auth — from starting.
switch {
case cfg.OIDCReady():
h.oidc = newOIDCClient(cfg)
auth.GET("/oidc/login", h.oidcLogin)
auth.GET("/oidc/callback", h.oidcCallback)
case cfg.OIDC.Enabled():
slog.Warn("api: OIDC is configured but PANSY_BASE_URL is unset; OIDC disabled (an absolute redirect URI is required)")
}
// Feature resources sit behind requireAuth, which resolves the session cookie
// to the actor the service layer's permission checks key off.
gardens := v1.Group("/gardens", h.requireAuth())
gardens.GET("", h.listGardens)
gardens.POST("", h.createGarden)
gardens.GET("/:id", h.getGarden)
gardens.PATCH("/:id", h.updateGarden)
gardens.DELETE("/:id", h.deleteGarden)
gardens.POST("/:id/copy", h.copyGarden) // duplicate a garden the actor owns
gardens.GET("/:id/full", h.getGardenFull) // one-shot editor load
gardens.GET("/:id/history", h.getGardenHistory) // change sets, newest first
gardens.GET("/:id/years", h.getGardenYears) // years with planting data
gardens.POST("/:id/objects", h.createObject)
// The grow journal hangs off a garden even for entries about one bed or one
// plop, so it inherits the ordinary garden-role check.
gardens.GET("/:id/journal", h.listJournal)
gardens.POST("/:id/journal", h.createJournalEntry)
gardens.GET("/:id/journal/counts", h.getJournalCounts)
// Sharing (owner-managed; a recipient may remove their own share).
gardens.GET("/:id/shares", h.listShares)
gardens.POST("/:id/shares", h.addShare)
gardens.PATCH("/:id/shares/:userId", h.updateShare)
gardens.DELETE("/:id/shares/:userId", h.removeShare)
// Public read-only share link (owner-managed): GET reports state, POST
// enables/rotates, DELETE disables. The link itself is served unauthenticated
// below.
gardens.GET("/:id/share-link", h.getShareLink)
gardens.POST("/:id/share-link", h.createShareLink)
gardens.DELETE("/:id/share-link", h.deleteShareLink)
// Objects are addressed by their own id; the service resolves the owning
// garden for the permission check.
objects := v1.Group("/objects", h.requireAuth())
objects.PATCH("/:id", h.updateObject)
objects.DELETE("/:id", h.deleteObject)
objects.POST("/:id/plantings", h.createPlanting) // place a plop in this object
// Plantings ("plops") are addressed by their own id; the service resolves the
// owning object/garden for the permission check.
plantings := v1.Group("/plantings", h.requireAuth())
plantings.PATCH("/:id", h.updatePlanting)
plantings.DELETE("/:id", h.deletePlanting)
// The garden assistant, registered only when it can actually be offered —
// the same shape as OIDC. An instance with no API key serves the app
// normally and simply doesn't have these routes.
if cfg.Agent.Ready() {
runner, err := agent.NewRunner(svc, cfg)
if err != nil {
// Configured but unusable (an unresolvable model spec, say). Log it and
// carry on without the assistant rather than refusing to start: a
// garden planner that won't boot because of a chat feature is worse
// than one without chat.
slog.Error("api: garden assistant disabled", "error", err)
} else {
h.agent = runner
agentGroup := v1.Group("/agent", h.requireAuth())
agentGroup.POST("/chat", h.agentChat)
gardens.GET("/:id/agent/history", h.getAgentHistory)
gardens.DELETE("/:id/agent/history", h.deleteAgentHistory)
slog.Info("api: garden assistant enabled", "model", cfg.Agent.Model)
}
}
// Undo. A change set is addressed by its own id; the service resolves the
// owning garden for the permission check, same as objects and plantings.
changeSets := v1.Group("/change-sets", h.requireAuth())
changeSets.POST("/:id/revert", h.revertChangeSet)
// Journal entries are addressed by their own id; the service resolves the
// owning garden for the permission check, same as objects and plantings.
journal := v1.Group("/journal", h.requireAuth())
journal.PATCH("/:id", h.updateJournalEntry)
journal.DELETE("/:id", h.deleteJournalEntry)
// Plant catalog: built-ins (seeded, read-only) plus the actor's own rows.
plants := v1.Group("/plants", h.requireAuth())
plants.GET("", h.listPlants)
plants.POST("", h.createPlant)
plants.PATCH("/:id", h.updatePlant)
plants.DELETE("/:id", h.deletePlant)
// Seed lots: what the actor bought, and what's left. Private to the buyer,
// so these hang off the session actor rather than off a garden.
seedLots := v1.Group("/seed-lots", h.requireAuth())
seedLots.GET("", h.listSeedLots)
seedLots.POST("", h.createSeedLot)
seedLots.GET("/:id", h.getSeedLot)
seedLots.PATCH("/:id", h.updateSeedLot)
seedLots.DELETE("/:id", h.deleteSeedLot)
// Public, unauthenticated read of a garden by its share token. Deliberately
// NOT behind requireAuth: the token is the capability, so a logged-out visitor
// opens a shared link without being redirected to /login or OIDC. Only GET,
// only the read-only /full payload — never a mutation.
public := v1.Group("/public")
public.GET("/gardens/:token", h.getPublicGarden)
return r
}
// healthz is a liveness probe: always returns {"ok": true} when the server is up.
func healthz(c *gin.Context) {
c.JSON(http.StatusOK, gin.H{"ok": true})
}