sift/internal/fleetctx/fleetctx.go

// Package fleetctx is Honcho retrieval for sift. Read-only.
// The restricted workspace (guardian-star) is reachable only when the
// authenticated caller is in Client.Permit and the caller's mapped workspace
// is that name. Workspace is never taken from the request.
package fleetctx

import (
	"bytes"
	"context"
	"encoding/json"
	"fmt"
	"log/slog"
	"net/http"
	"net/url"
	"regexp"
	"strings"
	"sync/atomic"
	"time"
	"unicode/utf8"

	"git.wntrmute.dev/kyle/sift/internal/auth"
)

const (
	Restricted = "guardian-star"
	Timeout    = 3 * time.Second
	Limit      = 5
	MaxChars   = 300
)

type Item struct {
	Peer      string `json:"peer"`
	Content   string `json:"content"`
	CreatedAt string `json:"created_at"`
	Session   string `json:"session"`
}

type Client struct {
	Base     string
	Allow    []string
	Permit   []string // caller names allowed to query Restricted
	HTTP     *http.Client
	Log      *slog.Logger
	Calls    atomic.Uint64
	Hits     atomic.Uint64
	Failures atomic.Uint64
}

func restricted(ws string) bool {
	return strings.EqualFold(strings.TrimSpace(ws), Restricted)
}

func permitted(caller string, permit []string) bool {
	if caller == "" {
		return false
	}
	for _, p := range permit {
		if p == caller {
			return true
		}
	}
	return false
}

var workspaceName = regexp.MustCompile(`^[A-Za-z0-9_-]+$`)

func workspaceOK(ws string) bool {
	return workspaceName.MatchString(ws)
}

func allowed(ws string, allow []string) bool {
	for _, a := range allow {
		if a == ws {
			return true
		}
	}
	return false
}

// Search returns prior knowledge for workspace. Failure → (nil, warning). Never
// an error that fails the fetch.
// Restricted workspace requires caller ∈ Permit. A permitted caller whose mapped
// workspace is not Restricted is refused and never queried (no fallback).
func (c *Client) Search(ctx context.Context, workspace, caller, query string) (items []Item, warn string) {
	if c == nil || query == "" {
		return nil, ""
	}
	if permitted(caller, c.Permit) && !restricted(workspace) {
		if c.Log != nil {
			c.Log.Error("restricted caller mapped off guardian-star", "caller", caller, "workspace", workspace)
		}
		c.Failures.Add(1)
		return nil, "context refused: workspace"
	}
	if restricted(workspace) && !permitted(caller, c.Permit) {
		if c.Log != nil {
			c.Log.Warn("context refused", "workspace", workspace, "caller", caller)
		}
		c.Failures.Add(1)
		return nil, "context refused: workspace"
	}
	if !workspaceOK(workspace) {
		c.Failures.Add(1)
		return nil, "context skipped: workspace"
	}
	if !allowed(workspace, c.Allow) {
		if c.Log != nil {
			c.Log.Warn("context workspace not allowlisted", "workspace", workspace)
		}
		return nil, "context skipped: workspace"
	}
	c.Calls.Add(1)
	sctx, cancel := context.WithTimeout(ctx, Timeout)
	defer cancel()
	path := strings.TrimRight(c.Base, "/") + "/v3/workspaces/" + url.PathEscape(workspace) + "/search"
	payload, _ := json.Marshal(map[string]any{"query": query, "limit": Limit})
	req, err := http.NewRequestWithContext(sctx, http.MethodPost, path, bytes.NewReader(payload))
	if err != nil {
		c.Failures.Add(1)
		return nil, "context failed"
	}
	req.Header.Set("Content-Type", "application/json")
	httpClient := c.HTTP
	if httpClient == nil {
		httpClient = &http.Client{Timeout: Timeout}
	}
	resp, err := httpClient.Do(req)
	if err != nil {
		c.Failures.Add(1)
		return nil, "context failed"
	}
	defer resp.Body.Close()
	if resp.StatusCode != 200 {
		c.Failures.Add(1)
		return nil, "context failed"
	}
	var raw []struct {
		Content     string `json:"content"`
		CreatedAt   string `json:"created_at"`
		PeerID      string `json:"peer_id"`
		SessionID   string `json:"session_id"`
		WorkspaceID string `json:"workspace_id"`
	}
	if err := json.NewDecoder(resp.Body).Decode(&raw); err != nil {
		c.Failures.Add(1)
		return nil, "context failed"
	}
	for _, m := range raw {
		if m.WorkspaceID != workspace {
			continue
		}
		content := m.Content
		if utf8.RuneCountInString(content) > MaxChars {
			r := []rune(content)
			content = string(r[:MaxChars])
		}
		items = append(items, Item{
			Peer:      m.PeerID,
			Content:   content,
			CreatedAt: m.CreatedAt,
			Session:   m.SessionID,
		})
	}
	if len(items) > 0 {
		c.Hits.Add(1)
	}
	return items, ""
}

func Query(question, title string, headings []string) string {
	if q := strings.TrimSpace(question); q != "" {
		return q
	}
	var b strings.Builder
	b.WriteString(title)
	n := 0
	for _, h := range headings {
		if n >= 3 {
			break
		}
		if strings.TrimSpace(h) == "" {
			continue
		}
		b.WriteByte(' ')
		b.WriteString(h)
		n++
	}
	return strings.TrimSpace(b.String())
}

func (c *Client) String() string { return fmt.Sprintf("fleetctx base=%s", c.Base) }

// CallerWS is a keys.toml name+workspace pair for startup/reload checks.
type CallerWS struct {
	Name, Workspace string
}

// CheckMappings logs ERROR for every permitted caller whose workspace is not
// Restricted, and WARN for every permit_restricted name with no key at all.
func (c *Client) CheckMappings(callers []CallerWS) {
	if c == nil {
		return
	}
	have := make(map[string]struct{}, len(callers))
	for _, a := range callers {
		have[a.Name] = struct{}{}
		if a.Name == auth.NameMetrics {
			continue
		}
		if permitted(a.Name, c.Permit) && !restricted(a.Workspace) {
			if c.Log != nil {
				c.Log.Error("restricted caller mapped off guardian-star", "caller", a.Name, "workspace", a.Workspace)
			}
		}
	}
	for _, name := range c.Permit {
		if _, ok := have[name]; !ok {
			if c.Log != nil {
				c.Log.Warn("permit_restricted caller has no key", "caller", name)
			}
		}
	}
}