Agent-Mode mit Aufgabenteilung und automatischer Recherche
This commit is contained in:
@@ -0,0 +1,698 @@
|
||||
package sourceagent
|
||||
|
||||
import (
|
||||
"context"
|
||||
"crypto/rand"
|
||||
"crypto/sha256"
|
||||
"database/sql"
|
||||
"encoding/hex"
|
||||
"encoding/json"
|
||||
"errors"
|
||||
"fmt"
|
||||
"hash/fnv"
|
||||
"math/bits"
|
||||
"net/url"
|
||||
"path/filepath"
|
||||
"sort"
|
||||
"strings"
|
||||
"sync"
|
||||
"time"
|
||||
"unicode"
|
||||
|
||||
_ "modernc.org/sqlite"
|
||||
)
|
||||
|
||||
type Store struct {
|
||||
db *sql.DB
|
||||
mu sync.Mutex
|
||||
}
|
||||
|
||||
func OpenStore(dataDir string) (*Store, error) {
|
||||
path := filepath.Join(dataDir, "source-agents.db")
|
||||
db, err := sql.Open("sqlite", "file:"+filepath.ToSlash(path)+"?_pragma=busy_timeout(5000)&_pragma=foreign_keys(1)&_pragma=journal_mode(WAL)")
|
||||
if err != nil {
|
||||
return nil, err
|
||||
}
|
||||
db.SetMaxOpenConns(4)
|
||||
store := &Store{db: db}
|
||||
if err := store.init(context.Background()); err != nil {
|
||||
_ = db.Close()
|
||||
return nil, err
|
||||
}
|
||||
return store, nil
|
||||
}
|
||||
|
||||
func (s *Store) Close() error {
|
||||
if s == nil || s.db == nil {
|
||||
return nil
|
||||
}
|
||||
return s.db.Close()
|
||||
}
|
||||
|
||||
func (s *Store) init(ctx context.Context) error {
|
||||
statements := []string{
|
||||
`CREATE TABLE IF NOT EXISTS source_agents (
|
||||
id TEXT PRIMARY KEY, name TEXT NOT NULL, token_hash TEXT NOT NULL, enabled INTEGER NOT NULL DEFAULT 1,
|
||||
created_at_ns INTEGER NOT NULL, updated_at_ns INTEGER NOT NULL, last_seen_ns INTEGER NOT NULL DEFAULT 0,
|
||||
last_error TEXT NOT NULL DEFAULT '', version TEXT NOT NULL DEFAULT ''
|
||||
) WITHOUT ROWID`,
|
||||
`CREATE TABLE IF NOT EXISTS source_tasks (
|
||||
id TEXT PRIMARY KEY, agent_id TEXT NOT NULL REFERENCES source_agents(id) ON DELETE CASCADE,
|
||||
name TEXT NOT NULL, type TEXT NOT NULL, url TEXT NOT NULL, enabled INTEGER NOT NULL DEFAULT 1,
|
||||
poll_interval TEXT NOT NULL DEFAULT '4h', categories_json TEXT NOT NULL DEFAULT '[]', max_items INTEGER NOT NULL DEFAULT 20,
|
||||
config_json TEXT NOT NULL DEFAULT '{}', created_at_ns INTEGER NOT NULL, updated_at_ns INTEGER NOT NULL
|
||||
) WITHOUT ROWID`,
|
||||
`CREATE INDEX IF NOT EXISTS idx_source_tasks_agent ON source_tasks(agent_id, enabled, updated_at_ns)`,
|
||||
`CREATE TABLE IF NOT EXISTS source_inbox (
|
||||
id TEXT PRIMARY KEY, agent_id TEXT NOT NULL, task_id TEXT NOT NULL, external_id TEXT NOT NULL DEFAULT '',
|
||||
url TEXT NOT NULL, canonical_url TEXT NOT NULL, title TEXT NOT NULL, published_at_ns INTEGER NOT NULL DEFAULT 0,
|
||||
discovered_at_ns INTEGER NOT NULL DEFAULT 0, received_at_ns INTEGER NOT NULL, updated_at_ns INTEGER NOT NULL,
|
||||
language TEXT NOT NULL DEFAULT '', content_type TEXT NOT NULL DEFAULT '', text_content TEXT NOT NULL,
|
||||
content_sha256 TEXT NOT NULL, source_name TEXT NOT NULL DEFAULT '', source_base_url TEXT NOT NULL DEFAULT '',
|
||||
categories_json TEXT NOT NULL DEFAULT '[]', metadata_json TEXT NOT NULL DEFAULT '{}', signature INTEGER NOT NULL DEFAULT 0,
|
||||
status TEXT NOT NULL DEFAULT 'received', relevance REAL NOT NULL DEFAULT 0, matched_node_id TEXT NOT NULL DEFAULT '',
|
||||
UNIQUE(agent_id, task_id, canonical_url, content_sha256)
|
||||
) WITHOUT ROWID`,
|
||||
`CREATE INDEX IF NOT EXISTS idx_source_inbox_status_received ON source_inbox(status, received_at_ns)`,
|
||||
`CREATE INDEX IF NOT EXISTS idx_source_inbox_candidate_updated ON source_inbox(status, updated_at_ns DESC)`,
|
||||
}
|
||||
for _, statement := range statements {
|
||||
if _, err := s.db.ExecContext(ctx, statement); err != nil {
|
||||
return err
|
||||
}
|
||||
}
|
||||
_, _ = s.db.ExecContext(ctx, `UPDATE source_inbox SET status='received' WHERE status='processing' AND updated_at_ns<?`, time.Now().UTC().Add(-10*time.Minute).UnixNano())
|
||||
return nil
|
||||
}
|
||||
|
||||
func GenerateToken() (string, string, error) {
|
||||
buf := make([]byte, 32)
|
||||
if _, err := rand.Read(buf); err != nil {
|
||||
return "", "", err
|
||||
}
|
||||
token := "brain_agent_" + hex.EncodeToString(buf)
|
||||
return token, hashToken(token), nil
|
||||
}
|
||||
|
||||
func hashToken(token string) string {
|
||||
sum := sha256.Sum256([]byte(strings.TrimSpace(token)))
|
||||
return hex.EncodeToString(sum[:])
|
||||
}
|
||||
|
||||
func randomID(prefix string) string {
|
||||
buf := make([]byte, 12)
|
||||
_, _ = rand.Read(buf)
|
||||
return prefix + "-" + hex.EncodeToString(buf)
|
||||
}
|
||||
|
||||
func normalizeID(value, prefix string) string {
|
||||
value = strings.ToLower(strings.TrimSpace(value))
|
||||
var b strings.Builder
|
||||
for _, r := range value {
|
||||
if unicode.IsLetter(r) || unicode.IsDigit(r) || r == '-' || r == '_' {
|
||||
b.WriteRune(r)
|
||||
}
|
||||
}
|
||||
value = strings.Trim(b.String(), "-_")
|
||||
if value == "" {
|
||||
return randomID(prefix)
|
||||
}
|
||||
return value
|
||||
}
|
||||
|
||||
func (s *Store) CreateAgent(ctx context.Context, id, name string) (Agent, string, error) {
|
||||
id = normalizeID(id, "agent")
|
||||
name = strings.TrimSpace(name)
|
||||
if name == "" {
|
||||
name = id
|
||||
}
|
||||
token, tokenHash, err := GenerateToken()
|
||||
if err != nil {
|
||||
return Agent{}, "", err
|
||||
}
|
||||
now := time.Now().UTC()
|
||||
_, err = s.db.ExecContext(ctx, `INSERT INTO source_agents(id,name,token_hash,enabled,created_at_ns,updated_at_ns) VALUES(?,?,?,?,?,?)`, id, name, tokenHash, 1, now.UnixNano(), now.UnixNano())
|
||||
if err != nil {
|
||||
return Agent{}, "", err
|
||||
}
|
||||
return Agent{ID: id, Name: name, Enabled: true, CreatedAt: now, UpdatedAt: now}, token, nil
|
||||
}
|
||||
|
||||
func (s *Store) RotateToken(ctx context.Context, id string) (string, error) {
|
||||
token, tokenHash, err := GenerateToken()
|
||||
if err != nil {
|
||||
return "", err
|
||||
}
|
||||
result, err := s.db.ExecContext(ctx, `UPDATE source_agents SET token_hash=?, updated_at_ns=? WHERE id=?`, tokenHash, time.Now().UTC().UnixNano(), id)
|
||||
if err != nil {
|
||||
return "", err
|
||||
}
|
||||
n, _ := result.RowsAffected()
|
||||
if n == 0 {
|
||||
return "", sql.ErrNoRows
|
||||
}
|
||||
return token, nil
|
||||
}
|
||||
|
||||
func (s *Store) SetAgentEnabled(ctx context.Context, id string, enabled bool) error {
|
||||
v := 0
|
||||
if enabled {
|
||||
v = 1
|
||||
}
|
||||
result, err := s.db.ExecContext(ctx, `UPDATE source_agents SET enabled=?,updated_at_ns=? WHERE id=?`, v, time.Now().UTC().UnixNano(), id)
|
||||
if err != nil {
|
||||
return err
|
||||
}
|
||||
n, _ := result.RowsAffected()
|
||||
if n == 0 {
|
||||
return sql.ErrNoRows
|
||||
}
|
||||
return nil
|
||||
}
|
||||
|
||||
func (s *Store) DeleteAgent(ctx context.Context, id string) error {
|
||||
_, err := s.db.ExecContext(ctx, `DELETE FROM source_agents WHERE id=?`, id)
|
||||
return err
|
||||
}
|
||||
|
||||
func (s *Store) ListAgents(ctx context.Context) ([]Agent, error) {
|
||||
rows, err := s.db.QueryContext(ctx, `SELECT a.id,a.name,a.enabled,a.created_at_ns,a.updated_at_ns,a.last_seen_ns,a.last_error,a.version,(SELECT COUNT(*) FROM source_tasks t WHERE t.agent_id=a.id) FROM source_agents a ORDER BY a.name,a.id`)
|
||||
if err != nil {
|
||||
return nil, err
|
||||
}
|
||||
defer rows.Close()
|
||||
var out []Agent
|
||||
for rows.Next() {
|
||||
var a Agent
|
||||
var en int
|
||||
var c, u, seen int64
|
||||
if err := rows.Scan(&a.ID, &a.Name, &en, &c, &u, &seen, &a.LastError, &a.Version, &a.TaskCount); err != nil {
|
||||
return nil, err
|
||||
}
|
||||
a.Enabled = en != 0
|
||||
a.CreatedAt = nsTime(c)
|
||||
a.UpdatedAt = nsTime(u)
|
||||
a.LastSeen = nsTime(seen)
|
||||
out = append(out, a)
|
||||
}
|
||||
return out, rows.Err()
|
||||
}
|
||||
|
||||
func (s *Store) Authenticate(ctx context.Context, token string) (Agent, error) {
|
||||
if strings.TrimSpace(token) == "" {
|
||||
return Agent{}, errors.New("empty agent token")
|
||||
}
|
||||
var a Agent
|
||||
var en int
|
||||
var c, u, seen int64
|
||||
err := s.db.QueryRowContext(ctx, `SELECT id,name,enabled,created_at_ns,updated_at_ns,last_seen_ns,last_error,version FROM source_agents WHERE token_hash=?`, hashToken(token)).Scan(&a.ID, &a.Name, &en, &c, &u, &seen, &a.LastError, &a.Version)
|
||||
if err != nil {
|
||||
return Agent{}, err
|
||||
}
|
||||
a.Enabled = en != 0
|
||||
if !a.Enabled {
|
||||
return Agent{}, errors.New("agent disabled")
|
||||
}
|
||||
a.CreatedAt = nsTime(c)
|
||||
a.UpdatedAt = nsTime(u)
|
||||
a.LastSeen = nsTime(seen)
|
||||
return a, nil
|
||||
}
|
||||
|
||||
func validateTask(t Task) (Task, error) {
|
||||
t.ID = normalizeID(t.ID, "task")
|
||||
t.Name = strings.TrimSpace(t.Name)
|
||||
if t.Name == "" {
|
||||
t.Name = t.ID
|
||||
}
|
||||
t.Type = strings.ToLower(strings.TrimSpace(t.Type))
|
||||
switch t.Type {
|
||||
case "rss", "atom", "sitemap", "web":
|
||||
default:
|
||||
return t, fmt.Errorf("unsupported source task type %q", t.Type)
|
||||
}
|
||||
u, err := url.Parse(strings.TrimSpace(t.URL))
|
||||
if err != nil || u.Host == "" || (u.Scheme != "http" && u.Scheme != "https") {
|
||||
return t, fmt.Errorf("task URL must be absolute http(s)")
|
||||
}
|
||||
if u.User != nil {
|
||||
return t, fmt.Errorf("task URL must not contain userinfo")
|
||||
}
|
||||
t.URL = u.String()
|
||||
if strings.TrimSpace(t.PollInterval) == "" {
|
||||
t.PollInterval = "4h"
|
||||
}
|
||||
d, err := time.ParseDuration(t.PollInterval)
|
||||
if err != nil || d < 5*time.Minute || d > 30*24*time.Hour {
|
||||
return t, fmt.Errorf("poll_interval must be between 5m and 720h")
|
||||
}
|
||||
if t.MaxItems <= 0 {
|
||||
t.MaxItems = 20
|
||||
}
|
||||
if t.MaxItems > 500 {
|
||||
return t, fmt.Errorf("max_items must be <=500")
|
||||
}
|
||||
if t.Config == nil {
|
||||
t.Config = map[string]string{}
|
||||
}
|
||||
return t, nil
|
||||
}
|
||||
|
||||
func (s *Store) UpsertTask(ctx context.Context, t Task) (Task, error) {
|
||||
var err error
|
||||
t, err = validateTask(t)
|
||||
if err != nil {
|
||||
return Task{}, err
|
||||
}
|
||||
if strings.TrimSpace(t.AgentID) == "" {
|
||||
return Task{}, errors.New("agent_id required")
|
||||
}
|
||||
cats, _ := json.Marshal(uniqueStrings(t.Categories))
|
||||
cfg, _ := json.Marshal(t.Config)
|
||||
now := time.Now().UTC()
|
||||
_, err = s.db.ExecContext(ctx, `INSERT INTO source_tasks(id,agent_id,name,type,url,enabled,poll_interval,categories_json,max_items,config_json,created_at_ns,updated_at_ns) VALUES(?,?,?,?,?,?,?,?,?,?,?,?) ON CONFLICT(id) DO UPDATE SET agent_id=excluded.agent_id,name=excluded.name,type=excluded.type,url=excluded.url,enabled=excluded.enabled,poll_interval=excluded.poll_interval,categories_json=excluded.categories_json,max_items=excluded.max_items,config_json=excluded.config_json,updated_at_ns=excluded.updated_at_ns`, t.ID, t.AgentID, t.Name, t.Type, t.URL, boolInt(t.Enabled), t.PollInterval, string(cats), t.MaxItems, string(cfg), now.UnixNano(), now.UnixNano())
|
||||
if err != nil {
|
||||
return Task{}, err
|
||||
}
|
||||
t.UpdatedAt = now
|
||||
if t.CreatedAt.IsZero() {
|
||||
t.CreatedAt = now
|
||||
}
|
||||
return t, nil
|
||||
}
|
||||
|
||||
func (s *Store) DeleteTask(ctx context.Context, id string) error {
|
||||
_, err := s.db.ExecContext(ctx, `DELETE FROM source_tasks WHERE id=?`, id)
|
||||
return err
|
||||
}
|
||||
func (s *Store) ListTasks(ctx context.Context, agentID string) ([]Task, error) {
|
||||
rows, err := s.db.QueryContext(ctx, `SELECT id,agent_id,name,type,url,enabled,poll_interval,categories_json,max_items,config_json,created_at_ns,updated_at_ns FROM source_tasks WHERE (?='' OR agent_id=?) ORDER BY name,id`, agentID, agentID)
|
||||
if err != nil {
|
||||
return nil, err
|
||||
}
|
||||
defer rows.Close()
|
||||
var out []Task
|
||||
for rows.Next() {
|
||||
var t Task
|
||||
var en int
|
||||
var cats, cfg string
|
||||
var c, u int64
|
||||
if err := rows.Scan(&t.ID, &t.AgentID, &t.Name, &t.Type, &t.URL, &en, &t.PollInterval, &cats, &t.MaxItems, &cfg, &c, &u); err != nil {
|
||||
return nil, err
|
||||
}
|
||||
t.Enabled = en != 0
|
||||
_ = json.Unmarshal([]byte(cats), &t.Categories)
|
||||
_ = json.Unmarshal([]byte(cfg), &t.Config)
|
||||
t.CreatedAt = nsTime(c)
|
||||
t.UpdatedAt = nsTime(u)
|
||||
out = append(out, t)
|
||||
}
|
||||
return out, rows.Err()
|
||||
}
|
||||
|
||||
func (s *Store) RemoteConfig(ctx context.Context, agent Agent) (RemoteConfig, error) {
|
||||
tasks, err := s.ListTasks(ctx, agent.ID)
|
||||
if err != nil {
|
||||
return RemoteConfig{}, err
|
||||
}
|
||||
return RemoteConfig{SchemaVersion: SchemaVersion, Agent: agent, Tasks: tasks, IssuedAt: time.Now().UTC()}, nil
|
||||
}
|
||||
|
||||
func (s *Store) Heartbeat(ctx context.Context, agentID string, h Heartbeat) error {
|
||||
_, err := s.db.ExecContext(ctx, `UPDATE source_agents SET last_seen_ns=?,last_error=?,version=?,updated_at_ns=? WHERE id=?`, time.Now().UTC().UnixNano(), strings.TrimSpace(h.LastError), strings.TrimSpace(h.Version), time.Now().UTC().UnixNano(), agentID)
|
||||
return err
|
||||
}
|
||||
|
||||
func normalizeDocument(d Document) (Document, error) {
|
||||
d.URL = strings.TrimSpace(d.URL)
|
||||
if d.URL == "" {
|
||||
return d, errors.New("document url required")
|
||||
}
|
||||
u, err := url.Parse(d.URL)
|
||||
if err != nil || u.Host == "" || (u.Scheme != "http" && u.Scheme != "https") {
|
||||
return d, errors.New("invalid document url")
|
||||
}
|
||||
d.CanonicalURL = strings.TrimSpace(d.CanonicalURL)
|
||||
if d.CanonicalURL == "" {
|
||||
d.CanonicalURL = d.URL
|
||||
}
|
||||
d.Title = strings.TrimSpace(d.Title)
|
||||
d.Text = strings.TrimSpace(d.Text)
|
||||
if len([]rune(d.Text)) < 80 {
|
||||
return d, errors.New("document text too short")
|
||||
}
|
||||
if d.DiscoveredAt.IsZero() {
|
||||
d.DiscoveredAt = time.Now().UTC()
|
||||
}
|
||||
if d.ExternalID == "" {
|
||||
d.ExternalID = d.CanonicalURL
|
||||
}
|
||||
d.Categories = uniqueStrings(d.Categories)
|
||||
if d.Metadata == nil {
|
||||
d.Metadata = map[string]any{}
|
||||
}
|
||||
if d.ContentSHA256 == "" {
|
||||
sum := sha256.Sum256([]byte(d.Text))
|
||||
d.ContentSHA256 = hex.EncodeToString(sum[:])
|
||||
}
|
||||
return d, nil
|
||||
}
|
||||
|
||||
func (s *Store) Ingest(ctx context.Context, agentID, taskID string, docs []Document) (IngestResult, error) {
|
||||
var taskOwner string
|
||||
if err := s.db.QueryRowContext(ctx, `SELECT agent_id FROM source_tasks WHERE id=? AND enabled=1`, taskID).Scan(&taskOwner); err != nil {
|
||||
return IngestResult{}, fmt.Errorf("unknown or disabled source task: %w", err)
|
||||
}
|
||||
if taskOwner != agentID {
|
||||
return IngestResult{}, errors.New("source task does not belong to authenticated agent")
|
||||
}
|
||||
result := IngestResult{BatchID: randomID("batch")}
|
||||
now := time.Now().UTC()
|
||||
s.mu.Lock()
|
||||
defer s.mu.Unlock()
|
||||
tx, err := s.db.BeginTx(ctx, nil)
|
||||
if err != nil {
|
||||
return result, err
|
||||
}
|
||||
defer tx.Rollback()
|
||||
for _, raw := range docs {
|
||||
d, e := normalizeDocument(raw)
|
||||
if e != nil {
|
||||
result.Rejected++
|
||||
continue
|
||||
}
|
||||
cats, _ := json.Marshal(d.Categories)
|
||||
meta, _ := json.Marshal(d.Metadata)
|
||||
idSum := sha256.Sum256([]byte(agentID + "\x00" + taskID + "\x00" + d.CanonicalURL + "\x00" + d.ContentSHA256))
|
||||
id := hex.EncodeToString(idSum[:12])
|
||||
sig := int64(simhash64(d.Title + "\n" + d.Text))
|
||||
res, e := tx.ExecContext(ctx, `INSERT OR IGNORE INTO source_inbox(id,agent_id,task_id,external_id,url,canonical_url,title,published_at_ns,discovered_at_ns,received_at_ns,updated_at_ns,language,content_type,text_content,content_sha256,source_name,source_base_url,categories_json,metadata_json,signature,status) VALUES(?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?)`, id, agentID, taskID, d.ExternalID, d.URL, d.CanonicalURL, d.Title, timeNS(d.PublishedAt), timeNS(d.DiscoveredAt), now.UnixNano(), now.UnixNano(), d.Language, d.ContentType, d.Text, d.ContentSHA256, d.SourceName, d.SourceBaseURL, string(cats), string(meta), sig, "received")
|
||||
if e != nil {
|
||||
result.Rejected++
|
||||
continue
|
||||
}
|
||||
n, _ := res.RowsAffected()
|
||||
if n == 0 {
|
||||
result.Duplicates++
|
||||
} else {
|
||||
result.Accepted++
|
||||
}
|
||||
}
|
||||
if err := tx.Commit(); err != nil {
|
||||
return result, err
|
||||
}
|
||||
return result, nil
|
||||
}
|
||||
|
||||
func (s *Store) ClaimInbox(ctx context.Context, limit int) ([]InboxDocument, error) {
|
||||
if limit < 1 {
|
||||
limit = 1
|
||||
}
|
||||
if limit > 100 {
|
||||
limit = 100
|
||||
}
|
||||
s.mu.Lock()
|
||||
defer s.mu.Unlock()
|
||||
tx, err := s.db.BeginTx(ctx, nil)
|
||||
if err != nil {
|
||||
return nil, err
|
||||
}
|
||||
defer tx.Rollback()
|
||||
rows, err := tx.QueryContext(ctx, `SELECT id FROM source_inbox WHERE status='received' ORDER BY received_at_ns LIMIT ?`, limit)
|
||||
if err != nil {
|
||||
return nil, err
|
||||
}
|
||||
var ids []string
|
||||
for rows.Next() {
|
||||
var id string
|
||||
if err := rows.Scan(&id); err != nil {
|
||||
rows.Close()
|
||||
return nil, err
|
||||
}
|
||||
ids = append(ids, id)
|
||||
}
|
||||
rows.Close()
|
||||
if len(ids) == 0 {
|
||||
return nil, tx.Commit()
|
||||
}
|
||||
now := time.Now().UTC().UnixNano()
|
||||
for _, id := range ids {
|
||||
if _, err := tx.ExecContext(ctx, `UPDATE source_inbox SET status='processing',updated_at_ns=? WHERE id=? AND status='received'`, now, id); err != nil {
|
||||
return nil, err
|
||||
}
|
||||
}
|
||||
if err := tx.Commit(); err != nil {
|
||||
return nil, err
|
||||
}
|
||||
return s.GetInboxByIDs(ctx, ids)
|
||||
}
|
||||
|
||||
func (s *Store) CompleteClassification(ctx context.Context, id, status string, relevance float64, matchedNodeID string, meta map[string]any) error {
|
||||
if status != "candidate" && status != "archived" {
|
||||
return errors.New("invalid inbox classification status")
|
||||
}
|
||||
data, _ := json.Marshal(meta)
|
||||
_, err := s.db.ExecContext(ctx, `UPDATE source_inbox SET status=?,relevance=?,matched_node_id=?,metadata_json=?,updated_at_ns=? WHERE id=?`, status, relevance, matchedNodeID, string(data), time.Now().UTC().UnixNano(), id)
|
||||
return err
|
||||
}
|
||||
|
||||
func (s *Store) MarkUsed(ctx context.Context, canonicalURLs []string) error {
|
||||
now := time.Now().UTC().UnixNano()
|
||||
for _, u := range uniqueStrings(canonicalURLs) {
|
||||
if _, err := s.db.ExecContext(ctx, `UPDATE source_inbox SET status='used',updated_at_ns=? WHERE canonical_url=?`, now, u); err != nil {
|
||||
return err
|
||||
}
|
||||
}
|
||||
return nil
|
||||
}
|
||||
|
||||
func (s *Store) GetInboxByIDs(ctx context.Context, ids []string) ([]InboxDocument, error) {
|
||||
if len(ids) == 0 {
|
||||
return nil, nil
|
||||
}
|
||||
out := make([]InboxDocument, 0, len(ids))
|
||||
for _, id := range ids {
|
||||
doc, err := s.getInbox(ctx, id)
|
||||
if err == nil {
|
||||
out = append(out, doc)
|
||||
}
|
||||
}
|
||||
return out, nil
|
||||
}
|
||||
func (s *Store) getInbox(ctx context.Context, id string) (InboxDocument, error) {
|
||||
row := s.db.QueryRowContext(ctx, `SELECT id,agent_id,task_id,external_id,url,canonical_url,title,published_at_ns,discovered_at_ns,received_at_ns,updated_at_ns,language,content_type,text_content,content_sha256,source_name,source_base_url,categories_json,metadata_json,status,relevance,matched_node_id FROM source_inbox WHERE id=?`, id)
|
||||
return scanInbox(row)
|
||||
}
|
||||
|
||||
type rowScanner interface{ Scan(...any) error }
|
||||
|
||||
func scanInbox(row rowScanner) (InboxDocument, error) {
|
||||
var d InboxDocument
|
||||
var ext, urlv, canon, title, lang, ctype, text, sha, source, base, cats, meta, status, matched string
|
||||
var pub, disc, recv, upd int64
|
||||
var rel float64
|
||||
err := row.Scan(&d.ID, &d.AgentID, &d.TaskID, &ext, &urlv, &canon, &title, &pub, &disc, &recv, &upd, &lang, &ctype, &text, &sha, &source, &base, &cats, &meta, &status, &rel, &matched)
|
||||
if err != nil {
|
||||
return d, err
|
||||
}
|
||||
d.Document = Document{ExternalID: ext, URL: urlv, CanonicalURL: canon, Title: title, PublishedAt: nsTime(pub), DiscoveredAt: nsTime(disc), Language: lang, ContentType: ctype, Text: text, ContentSHA256: sha, SourceName: source, SourceBaseURL: base}
|
||||
_ = json.Unmarshal([]byte(cats), &d.Document.Categories)
|
||||
_ = json.Unmarshal([]byte(meta), &d.Metadata)
|
||||
d.Status = status
|
||||
d.Relevance = rel
|
||||
d.MatchedNodeID = matched
|
||||
d.ReceivedAt = nsTime(recv)
|
||||
d.UpdatedAt = nsTime(upd)
|
||||
return d, nil
|
||||
}
|
||||
|
||||
func (s *Store) ListInbox(ctx context.Context, status string, limit int) ([]InboxDocument, error) {
|
||||
if limit < 1 {
|
||||
limit = 100
|
||||
}
|
||||
if limit > 1000 {
|
||||
limit = 1000
|
||||
}
|
||||
rows, err := s.db.QueryContext(ctx, `SELECT id,agent_id,task_id,external_id,url,canonical_url,title,published_at_ns,discovered_at_ns,received_at_ns,updated_at_ns,language,content_type,text_content,content_sha256,source_name,source_base_url,categories_json,metadata_json,status,relevance,matched_node_id FROM source_inbox WHERE (?='' OR status=?) ORDER BY received_at_ns DESC LIMIT ?`, status, status, limit)
|
||||
if err != nil {
|
||||
return nil, err
|
||||
}
|
||||
defer rows.Close()
|
||||
var out []InboxDocument
|
||||
for rows.Next() {
|
||||
d, err := scanInbox(rows)
|
||||
if err != nil {
|
||||
return nil, err
|
||||
}
|
||||
out = append(out, d)
|
||||
}
|
||||
return out, rows.Err()
|
||||
}
|
||||
|
||||
func (s *Store) Stats(ctx context.Context) (InboxStats, error) {
|
||||
rows, err := s.db.QueryContext(ctx, `SELECT status,COUNT(*) FROM source_inbox GROUP BY status`)
|
||||
if err != nil {
|
||||
return InboxStats{}, err
|
||||
}
|
||||
defer rows.Close()
|
||||
var st InboxStats
|
||||
for rows.Next() {
|
||||
var status string
|
||||
var n int
|
||||
if err := rows.Scan(&status, &n); err != nil {
|
||||
return st, err
|
||||
}
|
||||
st.Total += n
|
||||
switch status {
|
||||
case "received":
|
||||
st.Received = n
|
||||
case "processing":
|
||||
st.Processing = n
|
||||
case "candidate":
|
||||
st.Candidate = n
|
||||
case "archived":
|
||||
st.Archived = n
|
||||
case "used":
|
||||
st.Used = n
|
||||
}
|
||||
}
|
||||
return st, rows.Err()
|
||||
}
|
||||
|
||||
func (s *Store) SearchCandidates(ctx context.Context, query string, limit int, maxAge time.Duration) ([]ScoredDocument, error) {
|
||||
if limit < 1 {
|
||||
limit = 3
|
||||
}
|
||||
if limit > 50 {
|
||||
limit = 50
|
||||
}
|
||||
since := int64(0)
|
||||
if maxAge > 0 {
|
||||
since = time.Now().UTC().Add(-maxAge).UnixNano()
|
||||
}
|
||||
rows, err := s.db.QueryContext(ctx, `SELECT id,agent_id,task_id,external_id,url,canonical_url,title,published_at_ns,discovered_at_ns,received_at_ns,updated_at_ns,language,content_type,text_content,content_sha256,source_name,source_base_url,categories_json,metadata_json,status,relevance,matched_node_id,signature FROM source_inbox WHERE status IN ('candidate','used') AND (?=0 OR COALESCE(NULLIF(published_at_ns,0),received_at_ns)>=?) ORDER BY updated_at_ns DESC LIMIT 5000`, since, since)
|
||||
if err != nil {
|
||||
return nil, err
|
||||
}
|
||||
defer rows.Close()
|
||||
qsig := simhash64(query)
|
||||
qterms := termSet(query)
|
||||
var out []ScoredDocument
|
||||
for rows.Next() {
|
||||
var d InboxDocument
|
||||
var ext, urlv, canon, title, lang, ctype, text, sha, source, base, cats, meta, status, matched string
|
||||
var pub, disc, recv, upd, sig int64
|
||||
var rel float64
|
||||
if err := rows.Scan(&d.ID, &d.AgentID, &d.TaskID, &ext, &urlv, &canon, &title, &pub, &disc, &recv, &upd, &lang, &ctype, &text, &sha, &source, &base, &cats, &meta, &status, &rel, &matched, &sig); err != nil {
|
||||
return nil, err
|
||||
}
|
||||
d.Document = Document{ExternalID: ext, URL: urlv, CanonicalURL: canon, Title: title, PublishedAt: nsTime(pub), DiscoveredAt: nsTime(disc), Language: lang, ContentType: ctype, Text: text, ContentSHA256: sha, SourceName: source, SourceBaseURL: base}
|
||||
_ = json.Unmarshal([]byte(cats), &d.Document.Categories)
|
||||
_ = json.Unmarshal([]byte(meta), &d.Metadata)
|
||||
d.Status = status
|
||||
d.Relevance = rel
|
||||
d.MatchedNodeID = matched
|
||||
d.ReceivedAt = nsTime(recv)
|
||||
d.UpdatedAt = nsTime(upd)
|
||||
hamming := bits.OnesCount64(qsig ^ uint64(sig))
|
||||
hashScore := 1 - float64(hamming)/64.0
|
||||
lex := jaccard(qterms, termSet(title+" "+text))
|
||||
score := 0.45*hashScore + 0.45*lex + 0.10*rel
|
||||
if lex < 0.03 && hashScore < 0.58 {
|
||||
continue
|
||||
}
|
||||
out = append(out, ScoredDocument{InboxDocument: d, QueryScore: score})
|
||||
}
|
||||
sort.Slice(out, func(i, j int) bool { return out[i].QueryScore > out[j].QueryScore })
|
||||
if len(out) > limit {
|
||||
out = out[:limit]
|
||||
}
|
||||
return out, nil
|
||||
}
|
||||
|
||||
func boolInt(v bool) int {
|
||||
if v {
|
||||
return 1
|
||||
}
|
||||
return 0
|
||||
}
|
||||
func timeNS(t time.Time) int64 {
|
||||
if t.IsZero() {
|
||||
return 0
|
||||
}
|
||||
return t.UTC().UnixNano()
|
||||
}
|
||||
func nsTime(v int64) time.Time {
|
||||
if v <= 0 {
|
||||
return time.Time{}
|
||||
}
|
||||
return time.Unix(0, v).UTC()
|
||||
}
|
||||
func uniqueStrings(values []string) []string {
|
||||
seen := map[string]bool{}
|
||||
out := make([]string, 0, len(values))
|
||||
for _, v := range values {
|
||||
v = strings.TrimSpace(v)
|
||||
if v == "" || seen[v] {
|
||||
continue
|
||||
}
|
||||
seen[v] = true
|
||||
out = append(out, v)
|
||||
}
|
||||
return out
|
||||
}
|
||||
func termSet(value string) map[string]bool {
|
||||
out := map[string]bool{}
|
||||
for _, r := range strings.FieldsFunc(strings.ToLower(value), func(r rune) bool { return !unicode.IsLetter(r) && !unicode.IsDigit(r) }) {
|
||||
if len([]rune(r)) >= 3 {
|
||||
out[r] = true
|
||||
}
|
||||
}
|
||||
return out
|
||||
}
|
||||
func jaccard(a, b map[string]bool) float64 {
|
||||
if len(a) == 0 || len(b) == 0 {
|
||||
return 0
|
||||
}
|
||||
inter := 0
|
||||
union := len(a)
|
||||
for k := range b {
|
||||
if a[k] {
|
||||
inter++
|
||||
} else {
|
||||
union++
|
||||
}
|
||||
}
|
||||
if union == 0 {
|
||||
return 0
|
||||
}
|
||||
return float64(inter) / float64(union)
|
||||
}
|
||||
func simhash64(value string) uint64 {
|
||||
weights := [64]int{}
|
||||
for token := range termSet(value) {
|
||||
h := fnv.New64a()
|
||||
_, _ = h.Write([]byte(token))
|
||||
x := h.Sum64()
|
||||
for i := 0; i < 64; i++ {
|
||||
if x&(1<<i) != 0 {
|
||||
weights[i]++
|
||||
} else {
|
||||
weights[i]--
|
||||
}
|
||||
}
|
||||
}
|
||||
var sig uint64
|
||||
for i, w := range weights {
|
||||
if w >= 0 {
|
||||
sig |= 1 << i
|
||||
}
|
||||
}
|
||||
return sig
|
||||
}
|
||||
|
||||
func (s *Store) ReleaseInbox(ctx context.Context, id, message string) error {
|
||||
meta, _ := json.Marshal(map[string]any{"classification_error": strings.TrimSpace(message)})
|
||||
_, err := s.db.ExecContext(ctx, `UPDATE source_inbox SET status='received',metadata_json=?,updated_at_ns=? WHERE id=?`, string(meta), time.Now().UTC().UnixNano(), id)
|
||||
return err
|
||||
}
|
||||
Reference in New Issue
Block a user