@hatk/hatk 0.0.1-alpha.6 → 0.0.1-alpha.61
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- package/dist/adapter.d.ts +19 -0
- package/dist/adapter.d.ts.map +1 -0
- package/dist/adapter.js +108 -0
- package/dist/backfill.d.ts +2 -2
- package/dist/backfill.d.ts.map +1 -1
- package/dist/backfill.js +83 -41
- package/dist/car.d.ts +42 -10
- package/dist/car.d.ts.map +1 -1
- package/dist/car.js +154 -14
- package/dist/cli.js +243 -1043
- package/dist/config.d.ts +31 -1
- package/dist/config.d.ts.map +1 -1
- package/dist/config.js +40 -9
- package/dist/database/adapter-factory.d.ts +6 -0
- package/dist/database/adapter-factory.d.ts.map +1 -0
- package/dist/database/adapter-factory.js +20 -0
- package/dist/database/adapters/duckdb-search.d.ts +12 -0
- package/dist/database/adapters/duckdb-search.d.ts.map +1 -0
- package/dist/database/adapters/duckdb-search.js +27 -0
- package/dist/database/adapters/duckdb.d.ts +25 -0
- package/dist/database/adapters/duckdb.d.ts.map +1 -0
- package/dist/database/adapters/duckdb.js +161 -0
- package/dist/database/adapters/sqlite-search.d.ts +23 -0
- package/dist/database/adapters/sqlite-search.d.ts.map +1 -0
- package/dist/database/adapters/sqlite-search.js +74 -0
- package/dist/database/adapters/sqlite.d.ts +18 -0
- package/dist/database/adapters/sqlite.d.ts.map +1 -0
- package/dist/database/adapters/sqlite.js +88 -0
- package/dist/{db.d.ts → database/db.d.ts} +57 -6
- package/dist/database/db.d.ts.map +1 -0
- package/dist/{db.js → database/db.js} +730 -549
- package/dist/database/dialect.d.ts +45 -0
- package/dist/database/dialect.d.ts.map +1 -0
- package/dist/database/dialect.js +72 -0
- package/dist/{fts.d.ts → database/fts.d.ts} +7 -0
- package/dist/database/fts.d.ts.map +1 -0
- package/dist/{fts.js → database/fts.js} +116 -32
- package/dist/database/index.d.ts +7 -0
- package/dist/database/index.d.ts.map +1 -0
- package/dist/database/index.js +6 -0
- package/dist/database/ports.d.ts +50 -0
- package/dist/database/ports.d.ts.map +1 -0
- package/dist/database/ports.js +1 -0
- package/dist/{schema.d.ts → database/schema.d.ts} +14 -3
- package/dist/database/schema.d.ts.map +1 -0
- package/dist/{schema.js → database/schema.js} +81 -41
- package/dist/dev-entry.d.ts +8 -0
- package/dist/dev-entry.d.ts.map +1 -0
- package/dist/dev-entry.js +113 -0
- package/dist/feeds.d.ts +12 -8
- package/dist/feeds.d.ts.map +1 -1
- package/dist/feeds.js +51 -6
- package/dist/hooks.d.ts +85 -0
- package/dist/hooks.d.ts.map +1 -0
- package/dist/hooks.js +161 -0
- package/dist/hydrate.d.ts +7 -6
- package/dist/hydrate.d.ts.map +1 -1
- package/dist/hydrate.js +4 -16
- package/dist/indexer.d.ts +23 -0
- package/dist/indexer.d.ts.map +1 -1
- package/dist/indexer.js +181 -34
- package/dist/labels.d.ts +36 -0
- package/dist/labels.d.ts.map +1 -1
- package/dist/labels.js +71 -6
- package/dist/lexicon-resolve.d.ts.map +1 -1
- package/dist/lexicon-resolve.js +27 -112
- package/dist/lexicons/com/atproto/label/defs.json +75 -0
- package/dist/lexicons/com/atproto/moderation/defs.json +30 -0
- package/dist/lexicons/com/atproto/repo/strongRef.json +24 -0
- package/dist/lexicons/dev/hatk/applyWrites.json +87 -0
- package/dist/lexicons/dev/hatk/createRecord.json +40 -0
- package/dist/lexicons/dev/hatk/createReport.json +48 -0
- package/dist/lexicons/dev/hatk/deleteRecord.json +25 -0
- package/dist/lexicons/dev/hatk/describeCollections.json +41 -0
- package/dist/lexicons/dev/hatk/describeFeeds.json +29 -0
- package/dist/lexicons/dev/hatk/describeLabels.json +45 -0
- package/dist/lexicons/dev/hatk/getFeed.json +30 -0
- package/dist/lexicons/dev/hatk/getPreferences.json +19 -0
- package/dist/lexicons/dev/hatk/getRecord.json +26 -0
- package/dist/lexicons/dev/hatk/getRecords.json +32 -0
- package/dist/lexicons/dev/hatk/putPreference.json +28 -0
- package/dist/lexicons/dev/hatk/putRecord.json +41 -0
- package/dist/lexicons/dev/hatk/searchRecords.json +32 -0
- package/dist/lexicons/dev/hatk/uploadBlob.json +23 -0
- package/dist/logger.d.ts +29 -0
- package/dist/logger.d.ts.map +1 -1
- package/dist/logger.js +29 -0
- package/dist/main.js +138 -67
- package/dist/mst.d.ts +18 -1
- package/dist/mst.d.ts.map +1 -1
- package/dist/mst.js +19 -8
- package/dist/oauth/db.d.ts +3 -1
- package/dist/oauth/db.d.ts.map +1 -1
- package/dist/oauth/db.js +48 -19
- package/dist/oauth/server.d.ts +24 -0
- package/dist/oauth/server.d.ts.map +1 -1
- package/dist/oauth/server.js +198 -22
- package/dist/oauth/session.d.ts +11 -0
- package/dist/oauth/session.d.ts.map +1 -0
- package/dist/oauth/session.js +65 -0
- package/dist/opengraph.d.ts +10 -0
- package/dist/opengraph.d.ts.map +1 -1
- package/dist/opengraph.js +80 -40
- package/dist/pds-proxy.d.ts +60 -0
- package/dist/pds-proxy.d.ts.map +1 -0
- package/dist/pds-proxy.js +277 -0
- package/dist/push.d.ts +34 -0
- package/dist/push.d.ts.map +1 -0
- package/dist/push.js +184 -0
- package/dist/renderer.d.ts +27 -0
- package/dist/renderer.d.ts.map +1 -0
- package/dist/renderer.js +46 -0
- package/dist/resolve-hatk.d.ts +6 -0
- package/dist/resolve-hatk.d.ts.map +1 -0
- package/dist/resolve-hatk.js +20 -0
- package/dist/response.d.ts +16 -0
- package/dist/response.d.ts.map +1 -0
- package/dist/response.js +69 -0
- package/dist/scanner.d.ts +21 -0
- package/dist/scanner.d.ts.map +1 -0
- package/dist/scanner.js +88 -0
- package/dist/seed.d.ts +19 -0
- package/dist/seed.d.ts.map +1 -1
- package/dist/seed.js +43 -4
- package/dist/server-init.d.ts +8 -0
- package/dist/server-init.d.ts.map +1 -0
- package/dist/server-init.js +62 -0
- package/dist/server.d.ts +26 -3
- package/dist/server.d.ts.map +1 -1
- package/dist/server.js +629 -635
- package/dist/setup.d.ts +28 -1
- package/dist/setup.d.ts.map +1 -1
- package/dist/setup.js +50 -3
- package/dist/templates/feed.tpl +14 -0
- package/dist/templates/hook.tpl +5 -0
- package/dist/templates/label.tpl +15 -0
- package/dist/templates/og.tpl +17 -0
- package/dist/templates/seed.tpl +11 -0
- package/dist/templates/setup.tpl +5 -0
- package/dist/templates/test-feed.tpl +19 -0
- package/dist/templates/test-xrpc.tpl +19 -0
- package/dist/templates/xrpc.tpl +41 -0
- package/dist/test.d.ts +1 -1
- package/dist/test.d.ts.map +1 -1
- package/dist/test.js +39 -32
- package/dist/views.js +1 -1
- package/dist/vite-plugin.d.ts +1 -1
- package/dist/vite-plugin.d.ts.map +1 -1
- package/dist/vite-plugin.js +254 -66
- package/dist/xrpc.d.ts +75 -11
- package/dist/xrpc.d.ts.map +1 -1
- package/dist/xrpc.js +189 -39
- package/package.json +14 -7
- package/public/admin.html +133 -54
- package/dist/db.d.ts.map +0 -1
- package/dist/fts.d.ts.map +0 -1
- package/dist/oauth/hooks.d.ts +0 -10
- package/dist/oauth/hooks.d.ts.map +0 -1
- package/dist/oauth/hooks.js +0 -40
- package/dist/schema.d.ts.map +0 -1
- package/dist/test-browser.d.ts +0 -14
- package/dist/test-browser.d.ts.map +0 -1
- package/dist/test-browser.js +0 -26
|
@@ -1,136 +1,59 @@
|
|
|
1
|
-
import {
|
|
2
|
-
import {
|
|
3
|
-
import {
|
|
4
|
-
import {
|
|
5
|
-
import {
|
|
6
|
-
let
|
|
7
|
-
let
|
|
8
|
-
let readCon;
|
|
1
|
+
import { toSnakeCase, q } from "./schema.js";
|
|
2
|
+
import { getSearchColumns, stripStopWords, getSearchPort, updateFtsRecord, deleteFtsRecord } from "./fts.js";
|
|
3
|
+
import { emit, timer } from "../logger.js";
|
|
4
|
+
import { OAUTH_DDL } from "../oauth/db.js";
|
|
5
|
+
import { getDialect } from "./dialect.js";
|
|
6
|
+
let port;
|
|
7
|
+
let dialect;
|
|
9
8
|
const schemas = new Map();
|
|
10
|
-
export function
|
|
11
|
-
|
|
12
|
-
readCon?.closeSync();
|
|
13
|
-
}
|
|
14
|
-
catch { }
|
|
15
|
-
try {
|
|
16
|
-
con?.closeSync();
|
|
17
|
-
}
|
|
18
|
-
catch { }
|
|
19
|
-
try {
|
|
20
|
-
instance?.closeSync();
|
|
21
|
-
}
|
|
22
|
-
catch { }
|
|
23
|
-
}
|
|
24
|
-
let writeQueue = Promise.resolve();
|
|
25
|
-
let readQueue = Promise.resolve();
|
|
26
|
-
function enqueue(queue, fn) {
|
|
27
|
-
if (queue === 'write') {
|
|
28
|
-
const p = writeQueue.then(fn);
|
|
29
|
-
writeQueue = p.then(() => { }, () => { });
|
|
30
|
-
return p;
|
|
31
|
-
}
|
|
32
|
-
else {
|
|
33
|
-
const p = readQueue.then(fn);
|
|
34
|
-
readQueue = p.then(() => { }, () => { });
|
|
35
|
-
return p;
|
|
36
|
-
}
|
|
9
|
+
export function getDatabasePort() {
|
|
10
|
+
return port;
|
|
37
11
|
}
|
|
38
|
-
function
|
|
39
|
-
|
|
40
|
-
const idx = i + 1;
|
|
41
|
-
const value = params[i];
|
|
42
|
-
if (value === null || value === undefined) {
|
|
43
|
-
prepared.bindNull(idx);
|
|
44
|
-
}
|
|
45
|
-
else if (typeof value === 'string') {
|
|
46
|
-
prepared.bindVarchar(idx, value);
|
|
47
|
-
}
|
|
48
|
-
else if (typeof value === 'number') {
|
|
49
|
-
if (Number.isInteger(value)) {
|
|
50
|
-
prepared.bindInteger(idx, value);
|
|
51
|
-
}
|
|
52
|
-
else {
|
|
53
|
-
prepared.bindDouble(idx, value);
|
|
54
|
-
}
|
|
55
|
-
}
|
|
56
|
-
else if (typeof value === 'boolean') {
|
|
57
|
-
prepared.bindBoolean(idx, value);
|
|
58
|
-
}
|
|
59
|
-
else if (typeof value === 'bigint') {
|
|
60
|
-
prepared.bindBigInt(idx, value);
|
|
61
|
-
}
|
|
62
|
-
else if (value instanceof Uint8Array) {
|
|
63
|
-
prepared.bindBlob(idx, value);
|
|
64
|
-
}
|
|
65
|
-
else {
|
|
66
|
-
prepared.bindVarchar(idx, String(value));
|
|
67
|
-
}
|
|
68
|
-
}
|
|
12
|
+
export function getSqlDialect() {
|
|
13
|
+
return dialect;
|
|
69
14
|
}
|
|
70
|
-
|
|
71
|
-
|
|
72
|
-
await con.run(sql);
|
|
73
|
-
return;
|
|
74
|
-
}
|
|
75
|
-
const prepared = await con.prepare(sql);
|
|
76
|
-
bindParams(prepared, params);
|
|
77
|
-
await prepared.run();
|
|
15
|
+
export function closeDatabase() {
|
|
16
|
+
port?.close();
|
|
78
17
|
}
|
|
79
|
-
async function run(sql,
|
|
80
|
-
return
|
|
18
|
+
async function run(sql, params = []) {
|
|
19
|
+
return port.execute(sql, params);
|
|
81
20
|
}
|
|
82
21
|
export async function runBatch(operations) {
|
|
83
|
-
|
|
84
|
-
|
|
22
|
+
await port.beginTransaction();
|
|
23
|
+
try {
|
|
85
24
|
for (const op of operations) {
|
|
86
25
|
try {
|
|
87
|
-
|
|
88
|
-
await con.run(op.sql);
|
|
89
|
-
}
|
|
90
|
-
else {
|
|
91
|
-
const prepared = await con.prepare(op.sql);
|
|
92
|
-
bindParams(prepared, op.params);
|
|
93
|
-
await prepared.run();
|
|
94
|
-
}
|
|
26
|
+
await port.execute(op.sql, op.params);
|
|
95
27
|
}
|
|
96
28
|
catch {
|
|
97
29
|
// Skip bad records, continue with rest of batch
|
|
98
30
|
}
|
|
99
31
|
}
|
|
100
|
-
await
|
|
101
|
-
}
|
|
102
|
-
|
|
103
|
-
|
|
104
|
-
if (params.length === 0) {
|
|
105
|
-
const reader = await readCon.runAndReadAll(sql);
|
|
106
|
-
return reader.getRowObjects();
|
|
32
|
+
await port.commit();
|
|
33
|
+
}
|
|
34
|
+
catch {
|
|
35
|
+
await port.rollback();
|
|
107
36
|
}
|
|
108
|
-
const prepared = await readCon.prepare(sql);
|
|
109
|
-
bindParams(prepared, params);
|
|
110
|
-
const reader = await prepared.runAndReadAll();
|
|
111
|
-
return reader.getRowObjects();
|
|
112
37
|
}
|
|
113
|
-
async function all(sql,
|
|
114
|
-
return
|
|
38
|
+
async function all(sql, params = []) {
|
|
39
|
+
return port.query(sql, params);
|
|
115
40
|
}
|
|
116
|
-
export async function initDatabase(dbPath, tableSchemas, ddlStatements) {
|
|
117
|
-
|
|
118
|
-
|
|
119
|
-
|
|
41
|
+
export async function initDatabase(adapter, dbPath, tableSchemas, ddlStatements) {
|
|
42
|
+
port = adapter;
|
|
43
|
+
dialect = getDialect(adapter.dialect);
|
|
44
|
+
await port.open(dbPath);
|
|
120
45
|
for (const schema of tableSchemas) {
|
|
121
46
|
schemas.set(schema.collection, schema);
|
|
122
47
|
}
|
|
123
48
|
for (const ddl of ddlStatements) {
|
|
124
|
-
|
|
125
|
-
await run(statement);
|
|
126
|
-
}
|
|
49
|
+
await port.executeMultiple(ddl);
|
|
127
50
|
}
|
|
128
51
|
// Internal tables for backfill state
|
|
129
52
|
await run(`CREATE TABLE IF NOT EXISTS _repos (
|
|
130
53
|
did TEXT PRIMARY KEY,
|
|
131
54
|
status TEXT NOT NULL DEFAULT 'pending',
|
|
132
55
|
handle TEXT,
|
|
133
|
-
backfilled_at
|
|
56
|
+
backfilled_at ${dialect.timestampType},
|
|
134
57
|
rev TEXT,
|
|
135
58
|
retry_count INTEGER NOT NULL DEFAULT 0,
|
|
136
59
|
retry_after INTEGER NOT NULL DEFAULT 0
|
|
@@ -147,74 +70,337 @@ export async function initDatabase(dbPath, tableSchemas, ddlStatements) {
|
|
|
147
70
|
value TEXT NOT NULL
|
|
148
71
|
)`);
|
|
149
72
|
// Labels table (atproto-compatible)
|
|
150
|
-
|
|
151
|
-
|
|
152
|
-
|
|
153
|
-
|
|
154
|
-
|
|
155
|
-
|
|
156
|
-
|
|
157
|
-
|
|
158
|
-
|
|
159
|
-
|
|
73
|
+
if (dialect.supportsSequences) {
|
|
74
|
+
await run(`CREATE SEQUENCE IF NOT EXISTS _labels_seq START 1`);
|
|
75
|
+
await run(`CREATE TABLE IF NOT EXISTS _labels (
|
|
76
|
+
id INTEGER PRIMARY KEY DEFAULT nextval('_labels_seq'),
|
|
77
|
+
src TEXT NOT NULL,
|
|
78
|
+
uri TEXT NOT NULL,
|
|
79
|
+
val TEXT NOT NULL,
|
|
80
|
+
neg ${dialect.typeMap.boolean} DEFAULT FALSE,
|
|
81
|
+
cts ${dialect.timestampType} NOT NULL,
|
|
82
|
+
exp ${dialect.timestampType}
|
|
83
|
+
)`);
|
|
84
|
+
}
|
|
85
|
+
else {
|
|
86
|
+
await run(`CREATE TABLE IF NOT EXISTS _labels (
|
|
87
|
+
id INTEGER PRIMARY KEY AUTOINCREMENT,
|
|
88
|
+
src TEXT NOT NULL,
|
|
89
|
+
uri TEXT NOT NULL,
|
|
90
|
+
val TEXT NOT NULL,
|
|
91
|
+
neg INTEGER DEFAULT 0,
|
|
92
|
+
cts TEXT NOT NULL,
|
|
93
|
+
exp TEXT
|
|
94
|
+
)`);
|
|
95
|
+
}
|
|
160
96
|
await run(`CREATE INDEX IF NOT EXISTS idx_labels_uri ON _labels(uri)`);
|
|
161
97
|
await run(`CREATE INDEX IF NOT EXISTS idx_labels_src ON _labels(src)`);
|
|
162
98
|
// Preferences table (generic key-value per user)
|
|
163
99
|
await run(`CREATE TABLE IF NOT EXISTS _preferences (
|
|
164
100
|
did TEXT NOT NULL,
|
|
165
101
|
key TEXT NOT NULL,
|
|
166
|
-
value
|
|
167
|
-
updated_at
|
|
102
|
+
value ${dialect.jsonType} NOT NULL,
|
|
103
|
+
updated_at ${dialect.timestampType} DEFAULT ${dialect.currentTimestamp},
|
|
168
104
|
PRIMARY KEY (did, key)
|
|
105
|
+
)`);
|
|
106
|
+
// Reports table (user-submitted moderation reports)
|
|
107
|
+
if (dialect.supportsSequences) {
|
|
108
|
+
await run(`CREATE SEQUENCE IF NOT EXISTS _reports_seq START 1`);
|
|
109
|
+
await run(`CREATE TABLE IF NOT EXISTS _reports (
|
|
110
|
+
id INTEGER PRIMARY KEY DEFAULT nextval('_reports_seq'),
|
|
111
|
+
subject_uri TEXT NOT NULL,
|
|
112
|
+
subject_did TEXT NOT NULL,
|
|
113
|
+
label TEXT NOT NULL,
|
|
114
|
+
reason TEXT,
|
|
115
|
+
reported_by TEXT NOT NULL,
|
|
116
|
+
status TEXT NOT NULL DEFAULT 'open',
|
|
117
|
+
resolved_by TEXT,
|
|
118
|
+
resolved_at ${dialect.timestampType},
|
|
119
|
+
created_at ${dialect.timestampType} NOT NULL
|
|
120
|
+
)`);
|
|
121
|
+
}
|
|
122
|
+
else {
|
|
123
|
+
await run(`CREATE TABLE IF NOT EXISTS _reports (
|
|
124
|
+
id INTEGER PRIMARY KEY AUTOINCREMENT,
|
|
125
|
+
subject_uri TEXT NOT NULL,
|
|
126
|
+
subject_did TEXT NOT NULL,
|
|
127
|
+
label TEXT NOT NULL,
|
|
128
|
+
reason TEXT,
|
|
129
|
+
reported_by TEXT NOT NULL,
|
|
130
|
+
status TEXT NOT NULL DEFAULT 'open',
|
|
131
|
+
resolved_by TEXT,
|
|
132
|
+
resolved_at TEXT,
|
|
133
|
+
created_at TEXT NOT NULL
|
|
134
|
+
)`);
|
|
135
|
+
}
|
|
136
|
+
await run(`CREATE INDEX IF NOT EXISTS idx_reports_status ON _reports(status)`);
|
|
137
|
+
await run(`CREATE INDEX IF NOT EXISTS idx_reports_subject_uri ON _reports(subject_uri)`);
|
|
138
|
+
// Push notification tokens
|
|
139
|
+
await run(`CREATE TABLE IF NOT EXISTS _push_tokens (
|
|
140
|
+
did TEXT NOT NULL,
|
|
141
|
+
token TEXT NOT NULL,
|
|
142
|
+
platform TEXT NOT NULL,
|
|
143
|
+
created_at TEXT NOT NULL,
|
|
144
|
+
PRIMARY KEY (did, token)
|
|
169
145
|
)`);
|
|
170
146
|
// OAuth tables
|
|
171
|
-
|
|
172
|
-
|
|
147
|
+
await port.executeMultiple(OAUTH_DDL);
|
|
148
|
+
// Migrations: add pds_auth_server to existing sessions tables
|
|
149
|
+
try {
|
|
150
|
+
await run(`ALTER TABLE _oauth_sessions ADD COLUMN pds_auth_server TEXT`);
|
|
151
|
+
}
|
|
152
|
+
catch { }
|
|
153
|
+
}
|
|
154
|
+
/** Normalize SQL type names to handle dialect differences (e.g. VARCHAR → TEXT) */
|
|
155
|
+
function normalizeType(type) {
|
|
156
|
+
const upper = type.toUpperCase();
|
|
157
|
+
if (upper === 'VARCHAR' || upper === 'CHARACTER VARYING')
|
|
158
|
+
return 'TEXT';
|
|
159
|
+
if (upper === 'TIMESTAMP WITH TIME ZONE')
|
|
160
|
+
return 'TIMESTAMPTZ';
|
|
161
|
+
if (upper === 'BOOLEAN' || upper === 'BOOL')
|
|
162
|
+
return 'BOOLEAN';
|
|
163
|
+
if (upper === 'INT' || upper === 'INT4' || upper === 'INT8' || upper === 'BIGINT' || upper === 'SMALLINT')
|
|
164
|
+
return 'INTEGER';
|
|
165
|
+
return upper;
|
|
166
|
+
}
|
|
167
|
+
async function getExistingColumns(tableName) {
|
|
168
|
+
if (!/^[a-zA-Z0-9._]+$/.test(tableName)) {
|
|
169
|
+
throw new Error(`Invalid table name for introspection: ${tableName}`);
|
|
170
|
+
}
|
|
171
|
+
const cols = new Map();
|
|
172
|
+
try {
|
|
173
|
+
const query = dialect.introspectColumnsQuery(tableName);
|
|
174
|
+
const rows = await all(query);
|
|
175
|
+
for (const row of rows) {
|
|
176
|
+
// SQLite PRAGMA returns { name, type }, DuckDB returns { column_name, data_type }
|
|
177
|
+
const name = (row.column_name || row.name);
|
|
178
|
+
const type = normalizeType((row.data_type || row.type || 'TEXT'));
|
|
179
|
+
cols.set(name, type);
|
|
180
|
+
}
|
|
181
|
+
}
|
|
182
|
+
catch {
|
|
183
|
+
// Table doesn't exist yet
|
|
184
|
+
}
|
|
185
|
+
return cols;
|
|
186
|
+
}
|
|
187
|
+
function diffColumns(tableName, existingCols, expectedCols, changes) {
|
|
188
|
+
for (const [colName, colType] of expectedCols) {
|
|
189
|
+
if (!existingCols.has(colName)) {
|
|
190
|
+
changes.push({ table: tableName, action: 'add', column: colName, type: colType });
|
|
191
|
+
}
|
|
192
|
+
}
|
|
193
|
+
for (const [colName] of existingCols) {
|
|
194
|
+
if (!expectedCols.has(colName)) {
|
|
195
|
+
changes.push({ table: tableName, action: 'drop', column: colName });
|
|
196
|
+
}
|
|
197
|
+
}
|
|
198
|
+
for (const [colName, colType] of expectedCols) {
|
|
199
|
+
const existingType = existingCols.get(colName);
|
|
200
|
+
if (existingType && normalizeType(existingType) !== normalizeType(colType)) {
|
|
201
|
+
changes.push({ table: tableName, action: 'retype', column: colName, type: colType });
|
|
202
|
+
}
|
|
203
|
+
}
|
|
204
|
+
}
|
|
205
|
+
/** Build expected columns map for a child/union table */
|
|
206
|
+
function buildChildExpectedCols(columns) {
|
|
207
|
+
const expected = new Map();
|
|
208
|
+
expected.set('parent_uri', 'TEXT');
|
|
209
|
+
expected.set('parent_did', 'TEXT');
|
|
210
|
+
for (const col of columns) {
|
|
211
|
+
expected.set(col.name, normalizeType(col.sqlType));
|
|
212
|
+
}
|
|
213
|
+
return expected;
|
|
214
|
+
}
|
|
215
|
+
export async function migrateSchema(tableSchemas) {
|
|
216
|
+
const changes = [];
|
|
217
|
+
const newCollections = new Set();
|
|
218
|
+
for (const schema of tableSchemas) {
|
|
219
|
+
if (schema.columns.length === 0)
|
|
220
|
+
continue; // generic JSON storage, skip
|
|
221
|
+
const tableName = schema.collection;
|
|
222
|
+
const existingCols = await getExistingColumns(tableName);
|
|
223
|
+
if (existingCols.size === 0) {
|
|
224
|
+
newCollections.add(schema.collection);
|
|
225
|
+
continue; // table just created, nothing to migrate
|
|
226
|
+
}
|
|
227
|
+
// Expected columns: base columns (uri, cid, did, indexed_at) + schema columns
|
|
228
|
+
const expectedCols = new Map();
|
|
229
|
+
expectedCols.set('uri', 'TEXT');
|
|
230
|
+
expectedCols.set('cid', 'TEXT');
|
|
231
|
+
expectedCols.set('did', 'TEXT');
|
|
232
|
+
expectedCols.set('indexed_at', normalizeType(dialect.timestampType));
|
|
233
|
+
for (const col of schema.columns) {
|
|
234
|
+
expectedCols.set(col.name, normalizeType(col.sqlType));
|
|
235
|
+
}
|
|
236
|
+
diffColumns(tableName, existingCols, expectedCols, changes);
|
|
237
|
+
// Diff child tables
|
|
238
|
+
for (const child of schema.children) {
|
|
239
|
+
const childTable = child.tableName.replace(/"/g, '');
|
|
240
|
+
const existingChildCols = await getExistingColumns(childTable);
|
|
241
|
+
if (existingChildCols.size === 0)
|
|
242
|
+
continue;
|
|
243
|
+
diffColumns(childTable, existingChildCols, buildChildExpectedCols(child.columns), changes);
|
|
244
|
+
}
|
|
245
|
+
// Diff union branch tables
|
|
246
|
+
for (const union of schema.unions) {
|
|
247
|
+
for (const branch of union.branches) {
|
|
248
|
+
const branchTable = branch.tableName.replace(/"/g, '');
|
|
249
|
+
const existingBranchCols = await getExistingColumns(branchTable);
|
|
250
|
+
if (existingBranchCols.size === 0)
|
|
251
|
+
continue;
|
|
252
|
+
diffColumns(branchTable, existingBranchCols, buildChildExpectedCols(branch.columns), changes);
|
|
253
|
+
}
|
|
254
|
+
}
|
|
255
|
+
}
|
|
256
|
+
// Detect and drop orphaned child/union tables (query table list once)
|
|
257
|
+
let allTableNames = null;
|
|
258
|
+
try {
|
|
259
|
+
const rows = await all(dialect.listTablesQuery);
|
|
260
|
+
allTableNames = rows.map((r) => r.table_name);
|
|
261
|
+
}
|
|
262
|
+
catch { }
|
|
263
|
+
if (allTableNames) {
|
|
264
|
+
for (const schema of tableSchemas) {
|
|
265
|
+
if (schema.columns.length === 0)
|
|
266
|
+
continue;
|
|
267
|
+
const expectedTables = new Set();
|
|
268
|
+
for (const child of schema.children) {
|
|
269
|
+
expectedTables.add(child.tableName.replace(/"/g, ''));
|
|
270
|
+
}
|
|
271
|
+
for (const union of schema.unions) {
|
|
272
|
+
for (const branch of union.branches) {
|
|
273
|
+
expectedTables.add(branch.tableName.replace(/"/g, ''));
|
|
274
|
+
}
|
|
275
|
+
}
|
|
276
|
+
for (const name of allTableNames) {
|
|
277
|
+
if (name.startsWith(schema.collection + '__') && !expectedTables.has(name)) {
|
|
278
|
+
await run(`DROP TABLE IF EXISTS "${name}"`);
|
|
279
|
+
emit('migration', 'drop_table', { table: name });
|
|
280
|
+
}
|
|
281
|
+
}
|
|
282
|
+
}
|
|
283
|
+
}
|
|
284
|
+
if (changes.length > 0) {
|
|
285
|
+
await applyMigrationChanges(changes);
|
|
286
|
+
}
|
|
287
|
+
// Trigger backfill only for genuinely new collections (tables created this startup)
|
|
288
|
+
// Previously this checked ALL empty tables, which caused infinite resync loops
|
|
289
|
+
// for collections that are legitimately empty (e.g. blocks when nobody has blocked)
|
|
290
|
+
if (newCollections.size > 0) {
|
|
291
|
+
const [hasRepos] = await all(`SELECT 1 FROM _repos LIMIT 1`);
|
|
292
|
+
if (hasRepos) {
|
|
293
|
+
await run(`UPDATE _repos SET status = 'pending' WHERE status = 'active'`);
|
|
294
|
+
for (const collection of newCollections) {
|
|
295
|
+
emit('migration', 'new_collection', { collection });
|
|
296
|
+
}
|
|
297
|
+
}
|
|
298
|
+
}
|
|
299
|
+
return changes;
|
|
300
|
+
}
|
|
301
|
+
async function applyMigrationChanges(changes) {
|
|
302
|
+
for (const change of changes) {
|
|
303
|
+
const quotedTable = `"${change.table}"`;
|
|
304
|
+
const quotedColumn = `"${change.column}"`;
|
|
305
|
+
try {
|
|
306
|
+
switch (change.action) {
|
|
307
|
+
case 'add': {
|
|
308
|
+
await run(`ALTER TABLE ${quotedTable} ADD COLUMN ${quotedColumn} ${change.type}`);
|
|
309
|
+
emit('migration', 'add_column', { table: change.table, column: change.column, type: change.type });
|
|
310
|
+
const schema = schemas.get(change.table);
|
|
311
|
+
if (schema?.refColumns.includes(change.column)) {
|
|
312
|
+
const prefix = change.table.replace(/\./g, '_');
|
|
313
|
+
await run(`CREATE INDEX IF NOT EXISTS idx_${prefix}_${change.column} ON ${quotedTable}(${quotedColumn})`);
|
|
314
|
+
}
|
|
315
|
+
break;
|
|
316
|
+
}
|
|
317
|
+
case 'drop':
|
|
318
|
+
await run(`ALTER TABLE ${quotedTable} DROP COLUMN ${quotedColumn}`);
|
|
319
|
+
emit('migration', 'drop_column', { table: change.table, column: change.column });
|
|
320
|
+
break;
|
|
321
|
+
case 'retype':
|
|
322
|
+
await run(`ALTER TABLE ${quotedTable} DROP COLUMN ${quotedColumn}`);
|
|
323
|
+
await run(`ALTER TABLE ${quotedTable} ADD COLUMN ${quotedColumn} ${change.type}`);
|
|
324
|
+
emit('migration', 'retype_column', { table: change.table, column: change.column, type: change.type });
|
|
325
|
+
break;
|
|
326
|
+
}
|
|
327
|
+
}
|
|
328
|
+
catch (err) {
|
|
329
|
+
console.warn(`[migration] failed to ${change.action} column "${change.column}" on "${change.table}": ${err.message}`);
|
|
330
|
+
emit('migration', 'error', {
|
|
331
|
+
action: change.action,
|
|
332
|
+
table: change.table,
|
|
333
|
+
column: change.column,
|
|
334
|
+
error: err.message,
|
|
335
|
+
});
|
|
336
|
+
}
|
|
173
337
|
}
|
|
174
338
|
}
|
|
175
339
|
export async function getCursor(key) {
|
|
176
|
-
const rows = await all(`SELECT value FROM _cursor WHERE key = $1`, key);
|
|
340
|
+
const rows = await all(`SELECT value FROM _cursor WHERE key = $1`, [key]);
|
|
177
341
|
return rows[0]?.value || null;
|
|
178
342
|
}
|
|
179
343
|
export async function setCursor(key, value) {
|
|
180
|
-
await run(`INSERT OR REPLACE INTO _cursor (key, value) VALUES ($1, $2)`, key, value);
|
|
344
|
+
await run(`INSERT OR REPLACE INTO _cursor (key, value) VALUES ($1, $2)`, [key, value]);
|
|
181
345
|
}
|
|
182
346
|
export async function getRepoStatus(did) {
|
|
183
|
-
const rows = await all(`SELECT status FROM _repos WHERE did = $1`, did);
|
|
347
|
+
const rows = await all(`SELECT status FROM _repos WHERE did = $1`, [did]);
|
|
184
348
|
return rows[0]?.status || null;
|
|
185
349
|
}
|
|
186
350
|
export async function setRepoStatus(did, status, rev, opts) {
|
|
187
351
|
if (status === 'active') {
|
|
188
352
|
// Update existing row preserving handle if not provided
|
|
189
|
-
await run(`UPDATE _repos SET status = $1, handle = COALESCE($2, handle), backfilled_at = $3, rev = COALESCE($4, rev), retry_count = 0, retry_after = 0 WHERE did = $5`, status, opts?.handle || null, new Date().toISOString(), rev || null, did);
|
|
353
|
+
await run(`UPDATE _repos SET status = $1, handle = COALESCE($2, handle), backfilled_at = $3, rev = COALESCE($4, rev), retry_count = 0, retry_after = 0 WHERE did = $5`, [status, opts?.handle || null, new Date().toISOString(), rev || null, did]);
|
|
190
354
|
// Insert if row didn't exist yet
|
|
191
|
-
await run(`INSERT OR IGNORE INTO _repos (did, status, handle, backfilled_at, rev, retry_count, retry_after) VALUES ($1, $2, $3, $4, $5, 0, 0)`, did, status, opts?.handle || null, new Date().toISOString(), rev || null);
|
|
355
|
+
await run(`INSERT OR IGNORE INTO _repos (did, status, handle, backfilled_at, rev, retry_count, retry_after) VALUES ($1, $2, $3, $4, $5, 0, 0)`, [did, status, opts?.handle || null, new Date().toISOString(), rev || null]);
|
|
192
356
|
}
|
|
193
357
|
else if (status === 'failed' && opts) {
|
|
194
|
-
await run(`UPDATE _repos SET status = $1, retry_count = $2, retry_after = $3, handle = COALESCE($4, handle) WHERE did = $5`, status, opts.retryCount ?? 0, opts.retryAfter ?? 0, opts.handle || null, did);
|
|
358
|
+
await run(`UPDATE _repos SET status = $1, retry_count = $2, retry_after = $3, handle = COALESCE($4, handle) WHERE did = $5`, [status, opts.retryCount ?? 0, opts.retryAfter ?? 0, opts.handle || null, did]);
|
|
195
359
|
// If row didn't exist yet, insert it
|
|
196
|
-
await run(`INSERT OR IGNORE INTO _repos (did, status, handle, retry_count, retry_after) VALUES ($1, $2, $3, $4, $5)`, did, status, opts.handle || null, opts.retryCount ?? 0, opts.retryAfter ?? 0);
|
|
360
|
+
await run(`INSERT OR IGNORE INTO _repos (did, status, handle, retry_count, retry_after) VALUES ($1, $2, $3, $4, $5)`, [did, status, opts.handle || null, opts.retryCount ?? 0, opts.retryAfter ?? 0]);
|
|
197
361
|
}
|
|
198
362
|
else {
|
|
199
|
-
await run(`UPDATE _repos SET status = $1 WHERE did = $2`, status, did);
|
|
200
|
-
await run(`INSERT OR IGNORE INTO _repos (did, status) VALUES ($1, $2)`, did, status);
|
|
363
|
+
await run(`UPDATE _repos SET status = $1 WHERE did = $2`, [status, did]);
|
|
364
|
+
await run(`INSERT OR IGNORE INTO _repos (did, status) VALUES ($1, $2)`, [did, status]);
|
|
201
365
|
}
|
|
202
366
|
}
|
|
367
|
+
/** Update the handle for a DID if it exists in _repos. */
|
|
368
|
+
export async function updateRepoHandle(did, handle) {
|
|
369
|
+
await run(`UPDATE _repos SET handle = $1 WHERE did = $2`, [handle, did]);
|
|
370
|
+
}
|
|
371
|
+
export async function getRepoRev(did) {
|
|
372
|
+
const rows = await all(`SELECT rev FROM _repos WHERE did = $1`, [did]);
|
|
373
|
+
return rows[0]?.rev ?? null;
|
|
374
|
+
}
|
|
203
375
|
export async function getRepoRetryInfo(did) {
|
|
204
|
-
const rows = await all(`SELECT retry_count, retry_after FROM _repos WHERE did = $1`, did);
|
|
376
|
+
const rows = await all(`SELECT retry_count, retry_after FROM _repos WHERE did = $1`, [did]);
|
|
205
377
|
if (rows.length === 0)
|
|
206
378
|
return null;
|
|
207
379
|
return { retryCount: Number(rows[0].retry_count), retryAfter: Number(rows[0].retry_after) };
|
|
208
380
|
}
|
|
209
381
|
export async function listRetryEligibleRepos(maxRetries) {
|
|
210
382
|
const now = Math.floor(Date.now() / 1000);
|
|
211
|
-
const rows = await all(`SELECT did FROM _repos WHERE status = 'failed' AND retry_after <= $1 AND retry_count < $2`,
|
|
383
|
+
const rows = await all(`SELECT did FROM _repos WHERE status = 'failed' AND retry_after <= $1 AND retry_count < $2`, [
|
|
384
|
+
now,
|
|
385
|
+
maxRetries,
|
|
386
|
+
]);
|
|
212
387
|
return rows.map((r) => r.did);
|
|
213
388
|
}
|
|
214
389
|
export async function listPendingRepos() {
|
|
215
390
|
const rows = await all(`SELECT did FROM _repos WHERE status = 'pending'`);
|
|
216
391
|
return rows.map((r) => r.did);
|
|
217
392
|
}
|
|
393
|
+
export async function listActiveRepoDids() {
|
|
394
|
+
const rows = await all(`SELECT did FROM _repos WHERE status = 'active'`);
|
|
395
|
+
return rows.map((r) => r.did);
|
|
396
|
+
}
|
|
397
|
+
export async function removeRepo(did) {
|
|
398
|
+
await run(`DELETE FROM _repos WHERE did = $1`, [did]);
|
|
399
|
+
}
|
|
400
|
+
export async function getRepoHandle(did) {
|
|
401
|
+
const rows = await all(`SELECT handle FROM _repos WHERE did = $1`, [did]);
|
|
402
|
+
return rows[0]?.handle ?? null;
|
|
403
|
+
}
|
|
218
404
|
export async function listAllRepoStatuses() {
|
|
219
405
|
return (await all(`SELECT did, status FROM _repos`));
|
|
220
406
|
}
|
|
@@ -228,27 +414,96 @@ export async function listReposPaginated(opts = {}) {
|
|
|
228
414
|
params.push(status);
|
|
229
415
|
}
|
|
230
416
|
if (q) {
|
|
231
|
-
conditions.push(`(did
|
|
417
|
+
conditions.push(`(did ${dialect.ilike} $${paramIdx} OR handle ${dialect.ilike} $${paramIdx})`);
|
|
232
418
|
params.push(`%${q}%`);
|
|
233
419
|
paramIdx++;
|
|
234
420
|
}
|
|
235
421
|
const where = conditions.length ? ' WHERE ' + conditions.join(' AND ') : '';
|
|
236
|
-
const countRows = await all(`SELECT
|
|
422
|
+
const countRows = await all(`SELECT ${dialect.countAsInteger} as total FROM _repos${where}`, params);
|
|
237
423
|
const total = Number(countRows[0]?.total || 0);
|
|
238
|
-
const rows = await all(`SELECT did, handle, status, backfilled_at, rev FROM _repos${where} ORDER BY backfilled_at
|
|
424
|
+
const rows = await all(`SELECT did, handle, status, backfilled_at, rev FROM _repos${where} ORDER BY CASE WHEN backfilled_at IS NULL THEN 1 ELSE 0 END, backfilled_at DESC, did LIMIT $${paramIdx++} OFFSET $${paramIdx++}`, [...params, limit, offset]);
|
|
239
425
|
return { repos: rows, total };
|
|
240
426
|
}
|
|
241
427
|
export async function getCollectionCounts() {
|
|
242
428
|
const counts = {};
|
|
243
429
|
for (const [collection, schema] of schemas) {
|
|
244
|
-
const rows = await all(`SELECT
|
|
430
|
+
const rows = await all(`SELECT ${dialect.countAsInteger} as count FROM ${schema.tableName}`);
|
|
245
431
|
counts[collection] = Number(rows[0]?.count || 0);
|
|
246
432
|
}
|
|
247
433
|
return counts;
|
|
248
434
|
}
|
|
435
|
+
export async function getRepoStatusCounts() {
|
|
436
|
+
const rows = await all(`SELECT status, ${dialect.countAsInteger} as count FROM _repos GROUP BY status`);
|
|
437
|
+
const counts = {};
|
|
438
|
+
for (const row of rows)
|
|
439
|
+
counts[row.status] = Number(row.count);
|
|
440
|
+
return counts;
|
|
441
|
+
}
|
|
442
|
+
export async function getDatabaseSize() {
|
|
443
|
+
if (dialect.supportsSequences) {
|
|
444
|
+
// DuckDB: pragma_database_size returns pre-formatted strings
|
|
445
|
+
const rows = await all('SELECT database_size, memory_usage, memory_limit FROM pragma_database_size()');
|
|
446
|
+
return rows[0] ?? {};
|
|
447
|
+
}
|
|
448
|
+
// SQLite: compute from page_count * page_size
|
|
449
|
+
const pages = await all('SELECT page_count FROM pragma_page_count()');
|
|
450
|
+
const sizes = await all('SELECT page_size FROM pragma_page_size()');
|
|
451
|
+
const pageCount = Number(pages[0]?.page_count ?? 0);
|
|
452
|
+
const pageSize = Number(sizes[0]?.page_size ?? 0);
|
|
453
|
+
const bytes = pageCount * pageSize;
|
|
454
|
+
const mib = (bytes / 1024 / 1024).toFixed(1);
|
|
455
|
+
return { database_size: `${mib} MiB`, memory_usage: 'N/A', memory_limit: 'N/A' };
|
|
456
|
+
}
|
|
457
|
+
export async function getLabelCount(val) {
|
|
458
|
+
const rows = await all(`SELECT ${dialect.countAsInteger} as count FROM _labels WHERE val = $1`, [val]);
|
|
459
|
+
return Number(rows[0]?.count || 0);
|
|
460
|
+
}
|
|
461
|
+
export async function deleteLabels(val) {
|
|
462
|
+
const count = await getLabelCount(val);
|
|
463
|
+
await run(`DELETE FROM _labels WHERE val = $1`, [val]);
|
|
464
|
+
return count;
|
|
465
|
+
}
|
|
466
|
+
export async function getRecentRecords(collection, limit) {
|
|
467
|
+
const schema = schemas.get(collection);
|
|
468
|
+
if (!schema)
|
|
469
|
+
return [];
|
|
470
|
+
const rows = await all(`SELECT t.* FROM ${schema.tableName} t JOIN _repos r ON t.did = r.did WHERE t.indexed_at > r.backfilled_at ORDER BY t.indexed_at DESC LIMIT $1`, [limit]);
|
|
471
|
+
return rows;
|
|
472
|
+
}
|
|
249
473
|
export async function getSchemaDump() {
|
|
250
|
-
|
|
251
|
-
|
|
474
|
+
let rows;
|
|
475
|
+
if (dialect.supportsSequences) {
|
|
476
|
+
// DuckDB: use duckdb_tables() for full DDL
|
|
477
|
+
rows = await all(`SELECT sql FROM duckdb_tables() ORDER BY table_name`);
|
|
478
|
+
}
|
|
479
|
+
else {
|
|
480
|
+
// SQLite: use sqlite_master, skip FTS shadow/internal tables
|
|
481
|
+
rows = await all(`SELECT sql FROM sqlite_master WHERE type='table' AND name NOT LIKE 'sqlite_%' AND name NOT LIKE '_fts_%' AND sql IS NOT NULL ORDER BY name`);
|
|
482
|
+
}
|
|
483
|
+
// Normalize indentation and formatting
|
|
484
|
+
return rows
|
|
485
|
+
.map((r) => {
|
|
486
|
+
let sql = r.sql.trim();
|
|
487
|
+
// Remove quotes around column names (SQLite adds them for some columns)
|
|
488
|
+
sql = sql.replace(/\n\s*"(\w+)"/g, '\n$1');
|
|
489
|
+
// Ensure closing paren is on its own line
|
|
490
|
+
sql = sql.replace(/([^(\s])\)$/, '$1\n)');
|
|
491
|
+
// Normalize leading-comma columns added by ALTER TABLE into trailing commas
|
|
492
|
+
sql = sql.replace(/\n\s*,\s*/g, ',\n');
|
|
493
|
+
// Split into lines and re-indent consistently
|
|
494
|
+
const lines = sql.split('\n').map((l) => l.trim());
|
|
495
|
+
sql = lines
|
|
496
|
+
.map((line, i) => {
|
|
497
|
+
if (i === 0)
|
|
498
|
+
return line; // CREATE TABLE line
|
|
499
|
+
if (line.startsWith(')'))
|
|
500
|
+
return ')'; // closing paren at top level
|
|
501
|
+
return ' ' + line; // indent columns
|
|
502
|
+
})
|
|
503
|
+
.join('\n');
|
|
504
|
+
return sql + ';';
|
|
505
|
+
})
|
|
506
|
+
.join('\n\n');
|
|
252
507
|
}
|
|
253
508
|
export function buildInsertOp(collection, uri, cid, authorDid, record) {
|
|
254
509
|
const schema = schemas.get(collection);
|
|
@@ -267,12 +522,12 @@ export function buildInsertOp(collection, uri, cid, authorDid, record) {
|
|
|
267
522
|
else if (col.originalName.endsWith('__cid') && record[col.originalName.replace('__cid', '')]) {
|
|
268
523
|
rawValue = record[col.originalName.replace('__cid', '')].cid;
|
|
269
524
|
}
|
|
270
|
-
colNames.push(col.name);
|
|
525
|
+
colNames.push(q(col.name));
|
|
271
526
|
placeholders.push(`$${paramIdx++}`);
|
|
272
527
|
if (rawValue === undefined || rawValue === null) {
|
|
273
528
|
values.push(null);
|
|
274
529
|
}
|
|
275
|
-
else if (col.
|
|
530
|
+
else if (col.isJson) {
|
|
276
531
|
values.push(JSON.stringify(rawValue));
|
|
277
532
|
}
|
|
278
533
|
else {
|
|
@@ -287,34 +542,34 @@ export async function insertRecord(collection, uri, cid, authorDid, record) {
|
|
|
287
542
|
if (!schema)
|
|
288
543
|
throw new Error(`Unknown collection: ${collection}`);
|
|
289
544
|
const { sql, params } = buildInsertOp(collection, uri, cid, authorDid, record);
|
|
290
|
-
await run(sql,
|
|
545
|
+
await run(sql, params);
|
|
291
546
|
// Insert child table rows
|
|
292
547
|
for (const child of schema.children) {
|
|
293
548
|
const items = record[child.fieldName];
|
|
294
549
|
if (!Array.isArray(items))
|
|
295
550
|
continue;
|
|
296
551
|
// Delete existing child rows (handles INSERT OR REPLACE on main table)
|
|
297
|
-
await run(`DELETE FROM ${child.tableName} WHERE parent_uri = $1`, uri);
|
|
552
|
+
await run(`DELETE FROM ${child.tableName} WHERE parent_uri = $1`, [uri]);
|
|
298
553
|
for (const item of items) {
|
|
299
554
|
const colNames = ['parent_uri', 'parent_did'];
|
|
300
555
|
const placeholders = ['$1', '$2'];
|
|
301
556
|
const values = [uri, authorDid];
|
|
302
557
|
let idx = 3;
|
|
303
558
|
for (const col of child.columns) {
|
|
304
|
-
colNames.push(col.name);
|
|
559
|
+
colNames.push(q(col.name));
|
|
305
560
|
placeholders.push(`$${idx++}`);
|
|
306
561
|
const raw = item[col.originalName];
|
|
307
562
|
if (raw === undefined || raw === null) {
|
|
308
563
|
values.push(null);
|
|
309
564
|
}
|
|
310
|
-
else if (col.
|
|
565
|
+
else if (col.isJson) {
|
|
311
566
|
values.push(JSON.stringify(raw));
|
|
312
567
|
}
|
|
313
568
|
else {
|
|
314
569
|
values.push(raw);
|
|
315
570
|
}
|
|
316
571
|
}
|
|
317
|
-
await run(`INSERT INTO ${child.tableName} (${colNames.join(', ')}) VALUES (${placeholders.join(', ')})`,
|
|
572
|
+
await run(`INSERT INTO ${child.tableName} (${colNames.join(', ')}) VALUES (${placeholders.join(', ')})`, values);
|
|
318
573
|
}
|
|
319
574
|
}
|
|
320
575
|
// Insert union branch rows
|
|
@@ -327,7 +582,7 @@ export async function insertRecord(collection, uri, cid, authorDid, record) {
|
|
|
327
582
|
continue;
|
|
328
583
|
// Delete existing branch rows (handles INSERT OR REPLACE)
|
|
329
584
|
for (const b of union.branches) {
|
|
330
|
-
await run(`DELETE FROM ${b.tableName} WHERE parent_uri = $1`, uri);
|
|
585
|
+
await run(`DELETE FROM ${b.tableName} WHERE parent_uri = $1`, [uri]);
|
|
331
586
|
}
|
|
332
587
|
if (branch.isArray && branch.arrayField) {
|
|
333
588
|
// Array branch (e.g., embed.images) — insert one row per array item
|
|
@@ -340,20 +595,20 @@ export async function insertRecord(collection, uri, cid, authorDid, record) {
|
|
|
340
595
|
const values = [uri, authorDid];
|
|
341
596
|
let idx = 3;
|
|
342
597
|
for (const col of branch.columns) {
|
|
343
|
-
colNames.push(col.name);
|
|
598
|
+
colNames.push(q(col.name));
|
|
344
599
|
placeholders.push(`$${idx++}`);
|
|
345
600
|
const raw = item[col.originalName];
|
|
346
601
|
if (raw === undefined || raw === null) {
|
|
347
602
|
values.push(null);
|
|
348
603
|
}
|
|
349
|
-
else if (col.
|
|
604
|
+
else if (col.isJson) {
|
|
350
605
|
values.push(JSON.stringify(raw));
|
|
351
606
|
}
|
|
352
607
|
else {
|
|
353
608
|
values.push(raw);
|
|
354
609
|
}
|
|
355
610
|
}
|
|
356
|
-
await run(`INSERT INTO ${branch.tableName} (${colNames.join(', ')}) VALUES (${placeholders.join(', ')})`,
|
|
611
|
+
await run(`INSERT INTO ${branch.tableName} (${colNames.join(', ')}) VALUES (${placeholders.join(', ')})`, values);
|
|
357
612
|
}
|
|
358
613
|
}
|
|
359
614
|
else {
|
|
@@ -364,22 +619,24 @@ export async function insertRecord(collection, uri, cid, authorDid, record) {
|
|
|
364
619
|
const values = [uri, authorDid];
|
|
365
620
|
let idx = 3;
|
|
366
621
|
for (const col of branch.columns) {
|
|
367
|
-
colNames.push(col.name);
|
|
622
|
+
colNames.push(q(col.name));
|
|
368
623
|
placeholders.push(`$${idx++}`);
|
|
369
624
|
const raw = branchData[col.originalName];
|
|
370
625
|
if (raw === undefined || raw === null) {
|
|
371
626
|
values.push(null);
|
|
372
627
|
}
|
|
373
|
-
else if (col.
|
|
628
|
+
else if (col.isJson) {
|
|
374
629
|
values.push(JSON.stringify(raw));
|
|
375
630
|
}
|
|
376
631
|
else {
|
|
377
632
|
values.push(raw);
|
|
378
633
|
}
|
|
379
634
|
}
|
|
380
|
-
await run(`INSERT INTO ${branch.tableName} (${colNames.join(', ')}) VALUES (${placeholders.join(', ')})`,
|
|
635
|
+
await run(`INSERT INTO ${branch.tableName} (${colNames.join(', ')}) VALUES (${placeholders.join(', ')})`, values);
|
|
381
636
|
}
|
|
382
637
|
}
|
|
638
|
+
// Incrementally update FTS index for this record
|
|
639
|
+
await updateFtsRecord(collection, uri);
|
|
383
640
|
}
|
|
384
641
|
/** Extract branch data from a union value, handling wrapper properties */
|
|
385
642
|
function resolveBranchData(unionValue, branch) {
|
|
@@ -394,32 +651,41 @@ export async function deleteRecord(collection, uri) {
|
|
|
394
651
|
const schema = schemas.get(collection);
|
|
395
652
|
if (!schema)
|
|
396
653
|
return;
|
|
654
|
+
// Remove from FTS index before deleting the record data
|
|
655
|
+
await deleteFtsRecord(collection, uri);
|
|
397
656
|
for (const child of schema.children) {
|
|
398
|
-
await run(`DELETE FROM ${child.tableName} WHERE parent_uri = $1`, uri);
|
|
657
|
+
await run(`DELETE FROM ${child.tableName} WHERE parent_uri = $1`, [uri]);
|
|
399
658
|
}
|
|
400
659
|
for (const union of schema.unions) {
|
|
401
660
|
for (const branch of union.branches) {
|
|
402
|
-
await run(`DELETE FROM ${branch.tableName} WHERE parent_uri = $1`, uri);
|
|
661
|
+
await run(`DELETE FROM ${branch.tableName} WHERE parent_uri = $1`, [uri]);
|
|
403
662
|
}
|
|
404
663
|
}
|
|
405
|
-
await run(`DELETE FROM ${schema.tableName} WHERE uri = $1`, uri);
|
|
664
|
+
await run(`DELETE FROM ${schema.tableName} WHERE uri = $1`, [uri]);
|
|
406
665
|
}
|
|
407
666
|
export async function insertLabels(labels) {
|
|
408
667
|
if (labels.length === 0)
|
|
409
668
|
return;
|
|
410
669
|
for (const label of labels) {
|
|
411
|
-
// Skip if an active (non-negated, non-expired, not-superseded-by-negation) label already exists
|
|
412
|
-
const existing = await all(`SELECT 1 FROM _labels l1 WHERE l1.src = $1 AND l1.uri = $2 AND l1.val = $3 AND l1.neg = false AND (l1.exp IS NULL OR l1.exp > CURRENT_TIMESTAMP) AND NOT EXISTS (SELECT 1 FROM _labels l2 WHERE l2.uri = l1.uri AND l2.val = l1.val AND l2.neg = true AND l2.id > l1.id) LIMIT 1`, label.src, label.uri, label.val);
|
|
670
|
+
// Skip if an active (non-negated, non-expired, not-superseded-by-negation) label already exists for this src+uri+val
|
|
671
|
+
const existing = await all(`SELECT 1 FROM _labels l1 WHERE l1.src = $1 AND l1.uri = $2 AND l1.val = $3 AND l1.neg = false AND (l1.exp IS NULL OR l1.exp > CURRENT_TIMESTAMP) AND NOT EXISTS (SELECT 1 FROM _labels l2 WHERE l2.uri = l1.uri AND l2.val = l1.val AND l2.neg = true AND l2.id > l1.id) LIMIT 1`, [label.src, label.uri, label.val]);
|
|
413
672
|
if (!label.neg && existing.length > 0)
|
|
414
673
|
continue;
|
|
415
|
-
await run(`INSERT INTO _labels (src, uri, val, neg, cts, exp) VALUES ($1, $2, $3, $4, $5, $6)`,
|
|
674
|
+
await run(`INSERT INTO _labels (src, uri, val, neg, cts, exp) VALUES ($1, $2, $3, $4, $5, $6)`, [
|
|
675
|
+
label.src,
|
|
676
|
+
label.uri,
|
|
677
|
+
label.val,
|
|
678
|
+
label.neg || false,
|
|
679
|
+
label.cts || new Date().toISOString(),
|
|
680
|
+
label.exp || null,
|
|
681
|
+
]);
|
|
416
682
|
}
|
|
417
683
|
}
|
|
418
684
|
export async function queryLabelsForUris(uris) {
|
|
419
685
|
if (uris.length === 0)
|
|
420
686
|
return new Map();
|
|
421
687
|
const placeholders = uris.map((_, i) => `$${i + 1}`).join(',');
|
|
422
|
-
const rows = await all(`SELECT src, uri, val, neg, cts, exp FROM _labels l1 WHERE uri IN (${placeholders}) AND (exp IS NULL OR exp > CURRENT_TIMESTAMP) AND neg = false AND NOT EXISTS (SELECT 1 FROM _labels l2 WHERE l2.uri = l1.uri AND l2.val = l1.val AND l2.neg = true AND l2.id > l1.id)`,
|
|
688
|
+
const rows = await all(`SELECT src, uri, val, neg, cts, exp FROM _labels l1 WHERE uri IN (${placeholders}) AND (exp IS NULL OR exp > CURRENT_TIMESTAMP) AND neg = false AND NOT EXISTS (SELECT 1 FROM _labels l2 WHERE l2.uri = l1.uri AND l2.val = l1.val AND l2.neg = true AND l2.id > l1.id) GROUP BY uri, val`, uris);
|
|
423
689
|
const result = new Map();
|
|
424
690
|
for (const row of rows) {
|
|
425
691
|
const key = row.uri;
|
|
@@ -429,9 +695,9 @@ export async function queryLabelsForUris(uris) {
|
|
|
429
695
|
src: row.src,
|
|
430
696
|
uri: row.uri,
|
|
431
697
|
val: row.val,
|
|
432
|
-
neg: row.neg,
|
|
698
|
+
neg: !!row.neg,
|
|
433
699
|
cts: normalizeValue(row.cts),
|
|
434
|
-
|
|
700
|
+
...(row.exp ? { exp: String(row.exp) } : {}),
|
|
435
701
|
});
|
|
436
702
|
}
|
|
437
703
|
return result;
|
|
@@ -452,246 +718,209 @@ export async function bulkInsertRecords(records) {
|
|
|
452
718
|
if (!schema)
|
|
453
719
|
continue;
|
|
454
720
|
const stagingTable = `_staging_${collection.replace(/\./g, '_')}`;
|
|
455
|
-
const allCols = ['uri', 'cid', 'did', 'indexed_at', ...schema.columns.map((c) => c.name)];
|
|
721
|
+
const allCols = ['uri', 'cid', 'did', 'indexed_at', ...schema.columns.map((c) => q(c.name))];
|
|
456
722
|
const colDefs = [
|
|
457
723
|
'uri TEXT',
|
|
458
724
|
'cid TEXT',
|
|
459
725
|
'did TEXT',
|
|
460
726
|
'indexed_at TEXT',
|
|
461
|
-
...schema.columns.map((c) =>
|
|
727
|
+
...schema.columns.map((c) => {
|
|
728
|
+
const t = c.sqlType;
|
|
729
|
+
// Use TEXT for timestamp columns in staging (will cast on merge)
|
|
730
|
+
return `${q(c.name)} ${t === 'TIMESTAMP' || t === 'TIMESTAMPTZ' ? 'TEXT' : t}`;
|
|
731
|
+
}),
|
|
462
732
|
];
|
|
463
|
-
|
|
464
|
-
await
|
|
465
|
-
|
|
466
|
-
|
|
467
|
-
|
|
468
|
-
|
|
733
|
+
await port.execute(`DROP TABLE IF EXISTS ${stagingTable}`, []);
|
|
734
|
+
await port.execute(`CREATE TABLE ${stagingTable} (${colDefs.join(', ')})`, []);
|
|
735
|
+
const inserter = await port.createBulkInserter(stagingTable, allCols);
|
|
736
|
+
const now = new Date().toISOString();
|
|
737
|
+
for (const rec of recs) {
|
|
738
|
+
try {
|
|
739
|
+
const values = [rec.uri, rec.cid, rec.did, now];
|
|
740
|
+
for (const col of schema.columns) {
|
|
741
|
+
values.push(resolveColumnValue(col, rec.record));
|
|
742
|
+
}
|
|
743
|
+
inserter.append(values);
|
|
744
|
+
inserted++;
|
|
745
|
+
}
|
|
746
|
+
catch {
|
|
747
|
+
// Skip bad records
|
|
748
|
+
}
|
|
749
|
+
}
|
|
750
|
+
await inserter.close();
|
|
751
|
+
// Merge into target, filtering rows that would violate NOT NULL
|
|
752
|
+
const selectCols = allCols.map((name) => {
|
|
753
|
+
const col = schema.columns.find((c) => q(c.name) === name);
|
|
754
|
+
if (name === 'indexed_at' || (col && (col.sqlType === 'TIMESTAMP' || col.sqlType === 'TIMESTAMPTZ'))) {
|
|
755
|
+
return `${dialect.tryCastTimestamp(name)} AS ${name}`;
|
|
756
|
+
}
|
|
757
|
+
return name;
|
|
758
|
+
});
|
|
759
|
+
const notNullChecks = ['uri IS NOT NULL', 'did IS NOT NULL'];
|
|
760
|
+
for (const col of schema.columns) {
|
|
761
|
+
if (col.notNull) {
|
|
762
|
+
if (col.sqlType === 'TIMESTAMP' || col.sqlType === 'TIMESTAMPTZ') {
|
|
763
|
+
notNullChecks.push(`${dialect.tryCastTimestamp(q(col.name))} IS NOT NULL`);
|
|
764
|
+
}
|
|
765
|
+
else {
|
|
766
|
+
notNullChecks.push(`${q(col.name)} IS NOT NULL`);
|
|
767
|
+
}
|
|
768
|
+
}
|
|
769
|
+
}
|
|
770
|
+
const whereClause = notNullChecks.length ? ` WHERE ${notNullChecks.join(' AND ')}` : '';
|
|
771
|
+
await port.execute(`INSERT OR REPLACE INTO ${schema.tableName} (${allCols.join(', ')}) SELECT ${selectCols.join(', ')} FROM ${stagingTable}${whereClause}`, []);
|
|
772
|
+
await port.execute(`DROP TABLE ${stagingTable}`, []);
|
|
773
|
+
// Populate child tables
|
|
774
|
+
for (const child of schema.children) {
|
|
775
|
+
const childStagingTable = `_staging_${collection.replace(/\./g, '_')}__${child.fieldName}`;
|
|
776
|
+
const childColDefs = [
|
|
777
|
+
'parent_uri TEXT',
|
|
778
|
+
'parent_did TEXT',
|
|
779
|
+
...child.columns.map((c) => {
|
|
780
|
+
const t = c.sqlType;
|
|
781
|
+
return `${q(c.name)} ${t === 'TIMESTAMP' || t === 'TIMESTAMPTZ' ? 'TEXT' : t}`;
|
|
782
|
+
}),
|
|
783
|
+
];
|
|
784
|
+
const childAllCols = ['parent_uri', 'parent_did', ...child.columns.map((c) => q(c.name))];
|
|
785
|
+
await port.execute(`DROP TABLE IF EXISTS ${childStagingTable}`, []);
|
|
786
|
+
await port.execute(`CREATE TABLE ${childStagingTable} (${childColDefs.join(', ')})`, []);
|
|
787
|
+
const childInserter = await port.createBulkInserter(childStagingTable, childAllCols);
|
|
469
788
|
for (const rec of recs) {
|
|
470
|
-
|
|
471
|
-
|
|
472
|
-
|
|
473
|
-
|
|
474
|
-
|
|
475
|
-
|
|
476
|
-
|
|
477
|
-
|
|
478
|
-
rawValue = rawValue.uri;
|
|
479
|
-
}
|
|
480
|
-
else if (col.originalName.endsWith('__cid') && rec.record[col.originalName.replace('__cid', '')]) {
|
|
481
|
-
rawValue = rec.record[col.originalName.replace('__cid', '')].cid;
|
|
482
|
-
}
|
|
483
|
-
if (rawValue === undefined || rawValue === null) {
|
|
484
|
-
appender.appendNull();
|
|
485
|
-
}
|
|
486
|
-
else if (col.duckdbType === 'JSON') {
|
|
487
|
-
appender.appendVarchar(JSON.stringify(rawValue));
|
|
488
|
-
}
|
|
489
|
-
else if (col.duckdbType === 'INTEGER') {
|
|
490
|
-
appender.appendInteger(typeof rawValue === 'number' ? rawValue : parseInt(rawValue));
|
|
491
|
-
}
|
|
492
|
-
else if (col.duckdbType === 'BOOLEAN') {
|
|
493
|
-
appender.appendBoolean(!!rawValue);
|
|
494
|
-
}
|
|
495
|
-
else {
|
|
496
|
-
appender.appendVarchar(String(rawValue));
|
|
789
|
+
const items = rec.record[child.fieldName];
|
|
790
|
+
if (!Array.isArray(items))
|
|
791
|
+
continue;
|
|
792
|
+
for (const item of items) {
|
|
793
|
+
try {
|
|
794
|
+
const values = [rec.uri, rec.did];
|
|
795
|
+
for (const col of child.columns) {
|
|
796
|
+
values.push(resolveRawColumnValue(col, item));
|
|
497
797
|
}
|
|
798
|
+
childInserter.append(values);
|
|
799
|
+
}
|
|
800
|
+
catch {
|
|
801
|
+
// Skip bad items
|
|
498
802
|
}
|
|
499
|
-
appender.endRow();
|
|
500
|
-
inserted++;
|
|
501
|
-
}
|
|
502
|
-
catch {
|
|
503
|
-
// Skip bad records
|
|
504
803
|
}
|
|
505
804
|
}
|
|
506
|
-
|
|
507
|
-
|
|
508
|
-
|
|
509
|
-
|
|
510
|
-
|
|
511
|
-
|
|
512
|
-
|
|
805
|
+
await childInserter.close();
|
|
806
|
+
// Delete existing child rows for these URIs, then merge staging
|
|
807
|
+
const uriPlaceholders = recs.map((_, i) => `$${i + 1}`).join(',');
|
|
808
|
+
await port.execute(`DELETE FROM ${child.tableName} WHERE parent_uri IN (${uriPlaceholders})`, recs.map((r) => r.uri));
|
|
809
|
+
const childSelectCols = childAllCols.map((name) => {
|
|
810
|
+
const col = child.columns.find((c) => q(c.name) === name);
|
|
811
|
+
if (col && (col.sqlType === 'TIMESTAMP' || col.sqlType === 'TIMESTAMPTZ')) {
|
|
812
|
+
return `${dialect.tryCastTimestamp(name)} AS ${name}`;
|
|
513
813
|
}
|
|
514
814
|
return name;
|
|
515
815
|
});
|
|
516
|
-
|
|
517
|
-
|
|
518
|
-
|
|
519
|
-
|
|
520
|
-
|
|
521
|
-
|
|
522
|
-
|
|
523
|
-
|
|
524
|
-
notNullChecks.push(`${col.name} IS NOT NULL`);
|
|
525
|
-
}
|
|
526
|
-
}
|
|
527
|
-
}
|
|
528
|
-
const whereClause = notNullChecks.length ? ` WHERE ${notNullChecks.join(' AND ')}` : '';
|
|
529
|
-
await con.run(`INSERT OR REPLACE INTO ${schema.tableName} (${allCols.join(', ')}) SELECT ${selectCols.join(', ')} FROM ${stagingTable}${whereClause}`);
|
|
530
|
-
await con.run(`DROP TABLE ${stagingTable}`);
|
|
531
|
-
// Populate child tables
|
|
532
|
-
for (const child of schema.children) {
|
|
533
|
-
const childStagingTable = `_staging_${collection.replace(/\./g, '_')}__${child.fieldName}`;
|
|
534
|
-
const childColDefs = [
|
|
816
|
+
await port.execute(`INSERT INTO ${child.tableName} (${childAllCols.join(', ')}) SELECT ${childSelectCols.join(', ')} FROM ${childStagingTable} WHERE parent_uri IS NOT NULL`, []);
|
|
817
|
+
await port.execute(`DROP TABLE ${childStagingTable}`, []);
|
|
818
|
+
}
|
|
819
|
+
// Populate union branch tables
|
|
820
|
+
for (const union of schema.unions) {
|
|
821
|
+
for (const branch of union.branches) {
|
|
822
|
+
const branchStagingTable = `_staging_${collection.replace(/\./g, '_')}__${toSnakeCase(union.fieldName)}_${branch.branchName}`;
|
|
823
|
+
const branchColDefs = [
|
|
535
824
|
'parent_uri TEXT',
|
|
536
825
|
'parent_did TEXT',
|
|
537
|
-
...
|
|
826
|
+
...branch.columns.map((c) => {
|
|
827
|
+
const t = c.sqlType;
|
|
828
|
+
return `${q(c.name)} ${t === 'TIMESTAMP' || t === 'TIMESTAMPTZ' ? 'TEXT' : t}`;
|
|
829
|
+
}),
|
|
538
830
|
];
|
|
539
|
-
const
|
|
540
|
-
await
|
|
541
|
-
await
|
|
542
|
-
const
|
|
831
|
+
const branchAllCols = ['parent_uri', 'parent_did', ...branch.columns.map((c) => q(c.name))];
|
|
832
|
+
await port.execute(`DROP TABLE IF EXISTS ${branchStagingTable}`, []);
|
|
833
|
+
await port.execute(`CREATE TABLE ${branchStagingTable} (${branchColDefs.join(', ')})`, []);
|
|
834
|
+
const branchInserter = await port.createBulkInserter(branchStagingTable, branchAllCols);
|
|
543
835
|
for (const rec of recs) {
|
|
544
|
-
const
|
|
545
|
-
if (!
|
|
836
|
+
const unionValue = rec.record[union.fieldName];
|
|
837
|
+
if (!unionValue || typeof unionValue !== 'object')
|
|
546
838
|
continue;
|
|
547
|
-
|
|
548
|
-
|
|
549
|
-
|
|
550
|
-
|
|
551
|
-
|
|
552
|
-
const rawValue = item[col.originalName];
|
|
553
|
-
if (rawValue === undefined || rawValue === null) {
|
|
554
|
-
childAppender.appendNull();
|
|
555
|
-
}
|
|
556
|
-
else if (col.duckdbType === 'JSON') {
|
|
557
|
-
childAppender.appendVarchar(JSON.stringify(rawValue));
|
|
558
|
-
}
|
|
559
|
-
else if (col.duckdbType === 'INTEGER') {
|
|
560
|
-
childAppender.appendInteger(typeof rawValue === 'number' ? rawValue : parseInt(rawValue));
|
|
561
|
-
}
|
|
562
|
-
else if (col.duckdbType === 'BOOLEAN') {
|
|
563
|
-
childAppender.appendBoolean(!!rawValue);
|
|
564
|
-
}
|
|
565
|
-
else {
|
|
566
|
-
childAppender.appendVarchar(String(rawValue));
|
|
567
|
-
}
|
|
568
|
-
}
|
|
569
|
-
childAppender.endRow();
|
|
570
|
-
}
|
|
571
|
-
catch {
|
|
572
|
-
// Skip bad items
|
|
573
|
-
}
|
|
574
|
-
}
|
|
575
|
-
}
|
|
576
|
-
childAppender.flushSync();
|
|
577
|
-
childAppender.closeSync();
|
|
578
|
-
// Delete existing child rows for these URIs, then merge staging
|
|
579
|
-
const uriPlaceholders = recs.map((_, i) => `$${i + 1}`).join(',');
|
|
580
|
-
const delStmt = await con.prepare(`DELETE FROM ${child.tableName} WHERE parent_uri IN (${uriPlaceholders})`);
|
|
581
|
-
bindParams(delStmt, recs.map((r) => r.uri));
|
|
582
|
-
await delStmt.run();
|
|
583
|
-
const childSelectCols = childAllCols.map((name) => {
|
|
584
|
-
const col = child.columns.find((c) => c.name === name);
|
|
585
|
-
if (col && col.duckdbType === 'TIMESTAMP')
|
|
586
|
-
return `TRY_CAST(${name} AS TIMESTAMP) AS ${name}`;
|
|
587
|
-
return name;
|
|
588
|
-
});
|
|
589
|
-
await con.run(`INSERT INTO ${child.tableName} (${childAllCols.join(', ')}) SELECT ${childSelectCols.join(', ')} FROM ${childStagingTable} WHERE parent_uri IS NOT NULL`);
|
|
590
|
-
await con.run(`DROP TABLE ${childStagingTable}`);
|
|
591
|
-
}
|
|
592
|
-
// Populate union branch tables
|
|
593
|
-
for (const union of schema.unions) {
|
|
594
|
-
for (const branch of union.branches) {
|
|
595
|
-
const branchStagingTable = `_staging_${collection.replace(/\./g, '_')}__${toSnakeCase(union.fieldName)}_${branch.branchName}`;
|
|
596
|
-
const branchColDefs = [
|
|
597
|
-
'parent_uri TEXT',
|
|
598
|
-
'parent_did TEXT',
|
|
599
|
-
...branch.columns.map((c) => `${c.name} ${c.duckdbType === 'TIMESTAMP' ? 'TEXT' : c.duckdbType}`),
|
|
600
|
-
];
|
|
601
|
-
const branchAllCols = ['parent_uri', 'parent_did', ...branch.columns.map((c) => c.name)];
|
|
602
|
-
await con.run(`DROP TABLE IF EXISTS ${branchStagingTable}`);
|
|
603
|
-
await con.run(`CREATE TABLE ${branchStagingTable} (${branchColDefs.join(', ')})`);
|
|
604
|
-
const branchAppender = await con.createAppender(branchStagingTable);
|
|
605
|
-
for (const rec of recs) {
|
|
606
|
-
const unionValue = rec.record[union.fieldName];
|
|
607
|
-
if (!unionValue || typeof unionValue !== 'object')
|
|
608
|
-
continue;
|
|
609
|
-
if (unionValue.$type !== branch.type)
|
|
839
|
+
if (unionValue.$type !== branch.type)
|
|
840
|
+
continue;
|
|
841
|
+
if (branch.isArray && branch.arrayField) {
|
|
842
|
+
const items = resolveBranchData(unionValue, branch)[branch.arrayField];
|
|
843
|
+
if (!Array.isArray(items))
|
|
610
844
|
continue;
|
|
611
|
-
|
|
612
|
-
const items = resolveBranchData(unionValue, branch)[branch.arrayField];
|
|
613
|
-
if (!Array.isArray(items))
|
|
614
|
-
continue;
|
|
615
|
-
for (const item of items) {
|
|
616
|
-
try {
|
|
617
|
-
branchAppender.appendVarchar(rec.uri);
|
|
618
|
-
branchAppender.appendVarchar(rec.did);
|
|
619
|
-
for (const col of branch.columns) {
|
|
620
|
-
const rawValue = item[col.originalName];
|
|
621
|
-
if (rawValue === undefined || rawValue === null) {
|
|
622
|
-
branchAppender.appendNull();
|
|
623
|
-
}
|
|
624
|
-
else if (col.duckdbType === 'JSON') {
|
|
625
|
-
branchAppender.appendVarchar(JSON.stringify(rawValue));
|
|
626
|
-
}
|
|
627
|
-
else if (col.duckdbType === 'INTEGER') {
|
|
628
|
-
branchAppender.appendInteger(typeof rawValue === 'number' ? rawValue : parseInt(rawValue));
|
|
629
|
-
}
|
|
630
|
-
else if (col.duckdbType === 'BOOLEAN') {
|
|
631
|
-
branchAppender.appendBoolean(!!rawValue);
|
|
632
|
-
}
|
|
633
|
-
else {
|
|
634
|
-
branchAppender.appendVarchar(String(rawValue));
|
|
635
|
-
}
|
|
636
|
-
}
|
|
637
|
-
branchAppender.endRow();
|
|
638
|
-
}
|
|
639
|
-
catch {
|
|
640
|
-
// Skip bad items
|
|
641
|
-
}
|
|
642
|
-
}
|
|
643
|
-
}
|
|
644
|
-
else {
|
|
845
|
+
for (const item of items) {
|
|
645
846
|
try {
|
|
646
|
-
const
|
|
647
|
-
branchAppender.appendVarchar(rec.uri);
|
|
648
|
-
branchAppender.appendVarchar(rec.did);
|
|
847
|
+
const values = [rec.uri, rec.did];
|
|
649
848
|
for (const col of branch.columns) {
|
|
650
|
-
|
|
651
|
-
if (rawValue === undefined || rawValue === null) {
|
|
652
|
-
branchAppender.appendNull();
|
|
653
|
-
}
|
|
654
|
-
else if (col.duckdbType === 'JSON') {
|
|
655
|
-
branchAppender.appendVarchar(JSON.stringify(rawValue));
|
|
656
|
-
}
|
|
657
|
-
else if (col.duckdbType === 'INTEGER') {
|
|
658
|
-
branchAppender.appendInteger(typeof rawValue === 'number' ? rawValue : parseInt(rawValue));
|
|
659
|
-
}
|
|
660
|
-
else if (col.duckdbType === 'BOOLEAN') {
|
|
661
|
-
branchAppender.appendBoolean(!!rawValue);
|
|
662
|
-
}
|
|
663
|
-
else {
|
|
664
|
-
branchAppender.appendVarchar(String(rawValue));
|
|
665
|
-
}
|
|
849
|
+
values.push(resolveRawColumnValue(col, item));
|
|
666
850
|
}
|
|
667
|
-
|
|
851
|
+
branchInserter.append(values);
|
|
668
852
|
}
|
|
669
853
|
catch {
|
|
670
|
-
// Skip bad
|
|
854
|
+
// Skip bad items
|
|
671
855
|
}
|
|
672
856
|
}
|
|
673
857
|
}
|
|
674
|
-
|
|
675
|
-
|
|
676
|
-
|
|
677
|
-
|
|
678
|
-
|
|
679
|
-
|
|
680
|
-
|
|
681
|
-
|
|
682
|
-
|
|
683
|
-
|
|
684
|
-
|
|
685
|
-
|
|
686
|
-
}
|
|
687
|
-
await con.run(`INSERT INTO ${branch.tableName} (${branchAllCols.join(', ')}) SELECT ${branchSelectCols.join(', ')} FROM ${branchStagingTable} WHERE parent_uri IS NOT NULL`);
|
|
688
|
-
await con.run(`DROP TABLE ${branchStagingTable}`);
|
|
858
|
+
else {
|
|
859
|
+
try {
|
|
860
|
+
const branchData = resolveBranchData(unionValue, branch);
|
|
861
|
+
const values = [rec.uri, rec.did];
|
|
862
|
+
for (const col of branch.columns) {
|
|
863
|
+
values.push(resolveRawColumnValue(col, branchData));
|
|
864
|
+
}
|
|
865
|
+
branchInserter.append(values);
|
|
866
|
+
}
|
|
867
|
+
catch {
|
|
868
|
+
// Skip bad records
|
|
869
|
+
}
|
|
870
|
+
}
|
|
689
871
|
}
|
|
872
|
+
await branchInserter.close();
|
|
873
|
+
// Delete existing branch rows for these URIs, then merge staging
|
|
874
|
+
const uriPlaceholders = recs.map((_, i) => `$${i + 1}`).join(',');
|
|
875
|
+
await port.execute(`DELETE FROM ${branch.tableName} WHERE parent_uri IN (${uriPlaceholders})`, recs.map((r) => r.uri));
|
|
876
|
+
const branchSelectCols = branchAllCols.map((name) => {
|
|
877
|
+
const col = branch.columns.find((c) => q(c.name) === name);
|
|
878
|
+
if (col && (col.sqlType === 'TIMESTAMP' || col.sqlType === 'TIMESTAMPTZ')) {
|
|
879
|
+
return `${dialect.tryCastTimestamp(name)} AS ${name}`;
|
|
880
|
+
}
|
|
881
|
+
return name;
|
|
882
|
+
});
|
|
883
|
+
await port.execute(`INSERT INTO ${branch.tableName} (${branchAllCols.join(', ')}) SELECT ${branchSelectCols.join(', ')} FROM ${branchStagingTable} WHERE parent_uri IS NOT NULL`, []);
|
|
884
|
+
await port.execute(`DROP TABLE ${branchStagingTable}`, []);
|
|
690
885
|
}
|
|
691
|
-
}
|
|
886
|
+
}
|
|
692
887
|
}
|
|
693
888
|
return inserted;
|
|
694
889
|
}
|
|
890
|
+
/** Extract a column value from a record, handling strongRef expansion and type coercion for bulk insert */
|
|
891
|
+
function resolveColumnValue(col, record) {
|
|
892
|
+
let rawValue = record[col.originalName];
|
|
893
|
+
if (rawValue && typeof rawValue === 'object' && col.name.endsWith('_uri') && col.isRef) {
|
|
894
|
+
rawValue = rawValue.uri;
|
|
895
|
+
}
|
|
896
|
+
else if (col.originalName.endsWith('__cid') && record[col.originalName.replace('__cid', '')]) {
|
|
897
|
+
rawValue = record[col.originalName.replace('__cid', '')].cid;
|
|
898
|
+
}
|
|
899
|
+
return coerceValue(col.sqlType, rawValue);
|
|
900
|
+
}
|
|
901
|
+
/** Extract a raw column value from a data object and coerce for bulk insert */
|
|
902
|
+
function resolveRawColumnValue(col, data) {
|
|
903
|
+
return coerceValue(col.sqlType, data[col.originalName]);
|
|
904
|
+
}
|
|
905
|
+
/** Coerce a value to the appropriate type for insertion */
|
|
906
|
+
function coerceValue(sqlType, rawValue) {
|
|
907
|
+
if (rawValue === undefined || rawValue === null)
|
|
908
|
+
return null;
|
|
909
|
+
// Objects and arrays always need JSON stringification regardless of sqlType
|
|
910
|
+
// (on SQLite, JSON columns map to TEXT but still need stringification)
|
|
911
|
+
if (typeof rawValue === 'object' && !(rawValue instanceof Uint8Array)) {
|
|
912
|
+
return JSON.stringify(rawValue);
|
|
913
|
+
}
|
|
914
|
+
if (sqlType === 'JSON' || sqlType === 'TEXT') {
|
|
915
|
+
return String(rawValue);
|
|
916
|
+
}
|
|
917
|
+
if (sqlType === 'INTEGER' || sqlType === 'BIGINT') {
|
|
918
|
+
return typeof rawValue === 'number' ? rawValue : parseInt(rawValue);
|
|
919
|
+
}
|
|
920
|
+
if (sqlType === 'BOOLEAN')
|
|
921
|
+
return !!rawValue;
|
|
922
|
+
return String(rawValue);
|
|
923
|
+
}
|
|
695
924
|
export async function queryRecords(collection, opts = {}) {
|
|
696
925
|
const schema = schemas.get(collection);
|
|
697
926
|
if (!schema)
|
|
@@ -734,7 +963,7 @@ export async function queryRecords(collection, opts = {}) {
|
|
|
734
963
|
sql += ' AND ' + conditions.join(' AND ');
|
|
735
964
|
sql += ` ORDER BY t.${sortName} ${order.toUpperCase()}, t.cid ${order.toUpperCase()} LIMIT $${paramIdx++}`;
|
|
736
965
|
params.push(limit + 1); // fetch one extra for cursor
|
|
737
|
-
const rows = await all(sql,
|
|
966
|
+
const rows = await all(sql, params);
|
|
738
967
|
const hasMore = rows.length > limit;
|
|
739
968
|
if (hasMore)
|
|
740
969
|
rows.pop();
|
|
@@ -774,7 +1003,7 @@ export async function queryRecords(collection, opts = {}) {
|
|
|
774
1003
|
}
|
|
775
1004
|
export async function getRecordByUri(uri) {
|
|
776
1005
|
for (const [_collection, schema] of schemas) {
|
|
777
|
-
const rows = await all(`SELECT t.*, r.handle FROM ${schema.tableName} t LEFT JOIN _repos r ON t.did = r.did WHERE t.uri = $1 AND (r.status IS NULL OR r.status != 'takendown')`, uri);
|
|
1006
|
+
const rows = await all(`SELECT t.*, r.handle FROM ${schema.tableName} t LEFT JOIN _repos r ON t.did = r.did WHERE t.uri = $1 AND (r.status IS NULL OR r.status != 'takendown')`, [uri]);
|
|
778
1007
|
if (rows.length > 0) {
|
|
779
1008
|
const row = rows[0];
|
|
780
1009
|
if (schema.children.length > 0) {
|
|
@@ -811,7 +1040,7 @@ export async function getRecordsByUris(collection, uris) {
|
|
|
811
1040
|
if (!schema)
|
|
812
1041
|
return [];
|
|
813
1042
|
const placeholders = uris.map((_, i) => `$${i + 1}`).join(',');
|
|
814
|
-
const rows = await all(`SELECT t.*, r.handle FROM ${schema.tableName} t LEFT JOIN _repos r ON t.did = r.did WHERE t.uri IN (${placeholders}) AND (r.status IS NULL OR r.status != 'takendown')`,
|
|
1043
|
+
const rows = await all(`SELECT t.*, r.handle FROM ${schema.tableName} t LEFT JOIN _repos r ON t.did = r.did WHERE t.uri IN (${placeholders}) AND (r.status IS NULL OR r.status != 'takendown')`, uris);
|
|
815
1044
|
// Batch-fetch child rows for all URIs
|
|
816
1045
|
const childData = new Map();
|
|
817
1046
|
for (const child of schema.children) {
|
|
@@ -839,6 +1068,19 @@ export async function getRecordsByUris(collection, uris) {
|
|
|
839
1068
|
const byUri = new Map(rows.map((r) => [r.uri, r]));
|
|
840
1069
|
return uris.map((u) => byUri.get(u)).filter(Boolean);
|
|
841
1070
|
}
|
|
1071
|
+
/** Fetch records by URIs and return as a shaped Map keyed by URI. */
|
|
1072
|
+
export async function getRecordsMap(collection, uris) {
|
|
1073
|
+
if (uris.length === 0)
|
|
1074
|
+
return new Map();
|
|
1075
|
+
const records = await getRecordsByUris(collection, uris);
|
|
1076
|
+
const map = new Map();
|
|
1077
|
+
for (const r of records) {
|
|
1078
|
+
const shaped = reshapeRow(r, r?.__childData, r?.__unionData);
|
|
1079
|
+
if (shaped)
|
|
1080
|
+
map.set(shaped.uri, shaped);
|
|
1081
|
+
}
|
|
1082
|
+
return map;
|
|
1083
|
+
}
|
|
842
1084
|
/**
|
|
843
1085
|
* Multi-phase search across any collection's records.
|
|
844
1086
|
*
|
|
@@ -859,8 +1101,8 @@ export async function searchRecords(collection, query, opts = {}) {
|
|
|
859
1101
|
if (!schema)
|
|
860
1102
|
throw new Error(`Unknown collection: ${collection}`);
|
|
861
1103
|
const elapsed = timer();
|
|
862
|
-
const { limit = 20,
|
|
863
|
-
const textCols = schema.columns.filter((c) => c.
|
|
1104
|
+
const { limit = 20, fuzzy = true } = opts;
|
|
1105
|
+
const textCols = schema.columns.filter((c) => c.sqlType === 'TEXT');
|
|
864
1106
|
// Also check if FTS has indexed any columns (including derived JSON columns)
|
|
865
1107
|
const ftsSearchCols = getSearchColumns(collection);
|
|
866
1108
|
if (textCols.length === 0 && ftsSearchCols.length === 0) {
|
|
@@ -868,149 +1110,67 @@ export async function searchRecords(collection, query, opts = {}) {
|
|
|
868
1110
|
}
|
|
869
1111
|
// FTS shadow table name (dots replaced with underscores)
|
|
870
1112
|
const safeName = '_fts_' + collection.replace(/\./g, '_');
|
|
871
|
-
const ftsSchema = `fts_main_${safeName}`;
|
|
872
1113
|
const phaseErrors = [];
|
|
873
1114
|
const phasesUsed = [];
|
|
874
|
-
// Phase 1: BM25 ranked search
|
|
1115
|
+
// Phase 1: BM25 ranked search via SearchPort
|
|
875
1116
|
let bm25Results = [];
|
|
876
|
-
|
|
877
|
-
|
|
878
|
-
|
|
879
|
-
|
|
880
|
-
|
|
881
|
-
|
|
882
|
-
|
|
883
|
-
|
|
884
|
-
|
|
885
|
-
|
|
886
|
-
|
|
887
|
-
|
|
888
|
-
|
|
889
|
-
|
|
890
|
-
|
|
891
|
-
|
|
892
|
-
|
|
893
|
-
|
|
894
|
-
|
|
895
|
-
|
|
1117
|
+
const sp = getSearchPort();
|
|
1118
|
+
if (sp)
|
|
1119
|
+
try {
|
|
1120
|
+
const ftsQuery = stripStopWords(query);
|
|
1121
|
+
const ftsSearchColNames = getSearchColumns(collection);
|
|
1122
|
+
// Get ranked URIs from the search port
|
|
1123
|
+
const hits = await sp.search(safeName, ftsQuery, ftsSearchColNames, limit + 1, 0);
|
|
1124
|
+
if (hits.length > 0) {
|
|
1125
|
+
const uriList = hits.map((h) => h.uri);
|
|
1126
|
+
const scoreMap = new Map(hits.map((h) => [h.uri, h.score]));
|
|
1127
|
+
// Fetch full records for matched URIs
|
|
1128
|
+
const placeholders = uriList.map((_, i) => `$${i + 1}`).join(', ');
|
|
1129
|
+
const rows = await all(`SELECT m.* FROM ${schema.tableName} m
|
|
1130
|
+
LEFT JOIN _repos r ON m.did = r.did
|
|
1131
|
+
WHERE m.uri IN (${placeholders})
|
|
1132
|
+
AND (r.status IS NULL OR r.status != 'takendown')`, uriList);
|
|
1133
|
+
// Re-attach scores and sort
|
|
1134
|
+
bm25Results = rows
|
|
1135
|
+
.map((r) => ({ ...r, score: scoreMap.get(r.uri) ?? 0 }))
|
|
1136
|
+
.sort((a, b) => b.score - a.score);
|
|
896
1137
|
}
|
|
1138
|
+
phasesUsed.push('bm25');
|
|
1139
|
+
}
|
|
1140
|
+
catch (err) {
|
|
1141
|
+
phaseErrors.push(`bm25: ${err.message}`);
|
|
897
1142
|
}
|
|
898
|
-
sql += ` ORDER BY score, m.cid DESC LIMIT $${paramIdx++}`;
|
|
899
|
-
params.push(limit + 1);
|
|
900
|
-
bm25Results = await all(sql, ...params);
|
|
901
|
-
phasesUsed.push('bm25');
|
|
902
|
-
}
|
|
903
|
-
catch (err) {
|
|
904
|
-
phaseErrors.push(`bm25: ${err.message}`);
|
|
905
|
-
}
|
|
906
1143
|
const bm25Count = bm25Results.length;
|
|
907
1144
|
const hasMore = bm25Results.length > limit;
|
|
908
1145
|
if (hasMore)
|
|
909
1146
|
bm25Results.pop();
|
|
910
|
-
// Phase 2: Exact substring match — boosts phrase matches above BM25 results
|
|
911
|
-
const exactMatchResults = [];
|
|
912
|
-
const bm25Uris = new Set(bm25Results.map((r) => r.uri));
|
|
913
|
-
try {
|
|
914
|
-
const searchParam = `%${query}%`;
|
|
915
|
-
let paramIdx = 1;
|
|
916
|
-
const ilikeConds = [];
|
|
917
|
-
const params = [];
|
|
918
|
-
// TEXT columns — direct ILIKE
|
|
919
|
-
for (const c of textCols) {
|
|
920
|
-
ilikeConds.push(`t.${c.name} ILIKE $${paramIdx++}`);
|
|
921
|
-
params.push(searchParam);
|
|
922
|
-
}
|
|
923
|
-
// JSON columns — cast to text then ILIKE
|
|
924
|
-
const jsonCols = schema.columns.filter((c) => c.duckdbType === 'JSON');
|
|
925
|
-
for (const c of jsonCols) {
|
|
926
|
-
ilikeConds.push(`CAST(t.${c.name} AS TEXT) ILIKE $${paramIdx++}`);
|
|
927
|
-
params.push(searchParam);
|
|
928
|
-
}
|
|
929
|
-
// Handle from _repos table
|
|
930
|
-
ilikeConds.push(`r.handle ILIKE $${paramIdx++}`);
|
|
931
|
-
params.push(searchParam);
|
|
932
|
-
if (ilikeConds.length > 0) {
|
|
933
|
-
const exactSQL = `SELECT t.* FROM ${schema.tableName} t LEFT JOIN _repos r ON t.did = r.did
|
|
934
|
-
WHERE (${ilikeConds.join(' OR ')})
|
|
935
|
-
ORDER BY t.indexed_at DESC
|
|
936
|
-
LIMIT $${paramIdx++}`;
|
|
937
|
-
params.push(limit);
|
|
938
|
-
const rows = await all(exactSQL, ...params);
|
|
939
|
-
phasesUsed.push('exact');
|
|
940
|
-
for (const row of rows) {
|
|
941
|
-
if (!bm25Uris.has(row.uri)) {
|
|
942
|
-
exactMatchResults.push(row);
|
|
943
|
-
bm25Uris.add(row.uri);
|
|
944
|
-
}
|
|
945
|
-
}
|
|
946
|
-
}
|
|
947
|
-
}
|
|
948
|
-
catch (err) {
|
|
949
|
-
phaseErrors.push(`exact: ${err.message}`);
|
|
950
|
-
}
|
|
951
|
-
// Merge: exact matches first, then BM25 results, capped at limit
|
|
952
|
-
const mergedResults = [...exactMatchResults, ...bm25Results].slice(0, limit + (hasMore ? 1 : 0));
|
|
953
|
-
// Replace bm25Results with merged for downstream phases
|
|
954
|
-
bm25Results = mergedResults;
|
|
955
|
-
// Phase 3: ILIKE scan of rows written since last FTS rebuild (immediate searchability)
|
|
956
1147
|
const existingUris = new Set(bm25Results.map((r) => r.uri));
|
|
957
|
-
|
|
958
|
-
|
|
959
|
-
let recentCount = 0;
|
|
960
|
-
if (rebuiltAt && bm25Results.length < limit) {
|
|
961
|
-
const remaining = limit - bm25Results.length;
|
|
962
|
-
const searchParam = `%${query}%`;
|
|
963
|
-
let paramIdx = 1;
|
|
964
|
-
const ilikeParts = textCols.map((c) => `t.${c.name} ILIKE $${paramIdx++}`);
|
|
965
|
-
ilikeParts.push(`r.handle ILIKE $${paramIdx++}`);
|
|
966
|
-
const ilikeConds = ilikeParts.join(' OR ');
|
|
967
|
-
const params = [...textCols.map(() => searchParam), searchParam];
|
|
968
|
-
const recentSQL = `SELECT t.* FROM ${schema.tableName} t LEFT JOIN _repos r ON t.did = r.did
|
|
969
|
-
WHERE t.indexed_at >= $${paramIdx++} AND t.uri NOT IN (SELECT uri FROM ${safeName}) AND (${ilikeConds})
|
|
970
|
-
ORDER BY t.indexed_at DESC
|
|
971
|
-
LIMIT $${paramIdx++}`;
|
|
972
|
-
params.push(rebuiltAt, remaining + existingUris.size);
|
|
973
|
-
try {
|
|
974
|
-
const recentRows = await all(recentSQL, ...params);
|
|
975
|
-
phasesUsed.push('recent');
|
|
976
|
-
for (const row of recentRows) {
|
|
977
|
-
if (bm25Results.length >= limit)
|
|
978
|
-
break;
|
|
979
|
-
if (!existingUris.has(row.uri)) {
|
|
980
|
-
existingUris.add(row.uri);
|
|
981
|
-
bm25Results.push(row);
|
|
982
|
-
recentCount++;
|
|
983
|
-
}
|
|
984
|
-
}
|
|
985
|
-
}
|
|
986
|
-
catch (err) {
|
|
987
|
-
phaseErrors.push(`recent: ${err.message}`);
|
|
988
|
-
}
|
|
989
|
-
}
|
|
990
|
-
// Phase 4: Fuzzy fallback for typo tolerance (if still under limit)
|
|
1148
|
+
// Phase 2: Fuzzy fallback for typo tolerance (if still under limit)
|
|
1149
|
+
// Only available on dialects with jaro_winkler_similarity (DuckDB)
|
|
991
1150
|
let fuzzyCount = 0;
|
|
992
|
-
if (fuzzy && bm25Results.length < limit) {
|
|
1151
|
+
if (fuzzy && dialect.jaroWinklerSimilarity && bm25Results.length < limit) {
|
|
993
1152
|
const remaining = limit - bm25Results.length;
|
|
1153
|
+
const jwFn = dialect.jaroWinklerSimilarity;
|
|
994
1154
|
const simExprs = [
|
|
995
|
-
...textCols.map((c) =>
|
|
996
|
-
|
|
1155
|
+
...textCols.map((c) => `${jwFn}(lower(t.${q(c.name)}), lower($1))`),
|
|
1156
|
+
`${jwFn}(lower(r.handle), lower($1))`,
|
|
997
1157
|
];
|
|
998
1158
|
// Include child table TEXT columns via correlated subquery
|
|
999
1159
|
for (const child of schema.children) {
|
|
1000
1160
|
for (const col of child.columns) {
|
|
1001
|
-
if (col.
|
|
1002
|
-
simExprs.push(`COALESCE((SELECT MAX(
|
|
1161
|
+
if (col.sqlType === 'TEXT') {
|
|
1162
|
+
simExprs.push(`COALESCE((SELECT MAX(${jwFn}(lower(c.${q(col.name)}), lower($1))) FROM ${child.tableName} c WHERE c.parent_uri = t.uri), 0)`);
|
|
1003
1163
|
}
|
|
1004
1164
|
}
|
|
1005
1165
|
}
|
|
1006
|
-
const greatestExpr =
|
|
1166
|
+
const greatestExpr = dialect.greatest(simExprs);
|
|
1007
1167
|
const fuzzySQL = `SELECT t.*, ${greatestExpr} AS fuzzy_score
|
|
1008
1168
|
FROM ${schema.tableName} t LEFT JOIN _repos r ON t.did = r.did
|
|
1009
1169
|
WHERE ${greatestExpr} >= 0.8
|
|
1010
1170
|
ORDER BY fuzzy_score DESC
|
|
1011
1171
|
LIMIT $2`;
|
|
1012
1172
|
try {
|
|
1013
|
-
const fuzzyRows = await all(fuzzySQL, query, remaining + existingUris.size);
|
|
1173
|
+
const fuzzyRows = await all(fuzzySQL, [query, remaining + existingUris.size]);
|
|
1014
1174
|
phasesUsed.push('fuzzy');
|
|
1015
1175
|
for (const row of fuzzyRows) {
|
|
1016
1176
|
if (bm25Results.length >= limit)
|
|
@@ -1025,16 +1185,17 @@ export async function searchRecords(collection, query, opts = {}) {
|
|
|
1025
1185
|
phaseErrors.push(`fuzzy: ${err.message}`);
|
|
1026
1186
|
}
|
|
1027
1187
|
}
|
|
1028
|
-
// Remove score columns
|
|
1029
|
-
const
|
|
1188
|
+
// Remove score columns and reshape into Row<T> with value
|
|
1189
|
+
const rawRecords = bm25Results.map(({ score: _score, fuzzy_score: _fuzzy_score, ...rest }) => rest);
|
|
1190
|
+
const records = rawRecords
|
|
1191
|
+
.map((r) => reshapeRow(r, r?.__childData, r?.__unionData))
|
|
1192
|
+
.filter((r) => r != null);
|
|
1030
1193
|
const lastRow = bm25Results[bm25Results.length - 1];
|
|
1031
1194
|
const nextCursor = hasMore && lastRow?.score != null ? packCursor(lastRow.score, lastRow.cid) : undefined;
|
|
1032
1195
|
emit('search', 'query', {
|
|
1033
1196
|
collection,
|
|
1034
1197
|
query,
|
|
1035
1198
|
bm25_count: bm25Count > limit ? bm25Count - 1 : bm25Count,
|
|
1036
|
-
exact_count: exactMatchResults.length,
|
|
1037
|
-
recent_count: recentCount,
|
|
1038
1199
|
fuzzy_count: fuzzyCount,
|
|
1039
1200
|
total_results: records.length,
|
|
1040
1201
|
duration_ms: elapsed(),
|
|
@@ -1045,10 +1206,13 @@ export async function searchRecords(collection, query, opts = {}) {
|
|
|
1045
1206
|
}
|
|
1046
1207
|
// Raw SQL for script feeds
|
|
1047
1208
|
export async function querySQL(sql, params = []) {
|
|
1048
|
-
return all(sql,
|
|
1209
|
+
return all(sql, params);
|
|
1049
1210
|
}
|
|
1050
|
-
export async function runSQL(sql,
|
|
1051
|
-
return run(sql,
|
|
1211
|
+
export async function runSQL(sql, params = []) {
|
|
1212
|
+
return run(sql, params);
|
|
1213
|
+
}
|
|
1214
|
+
export async function createBulkInserterSQL(table, columns, options) {
|
|
1215
|
+
return port.createBulkInserter(table, columns, options);
|
|
1052
1216
|
}
|
|
1053
1217
|
export function getSchema(collection) {
|
|
1054
1218
|
return schemas.get(collection);
|
|
@@ -1057,7 +1221,7 @@ export async function countByField(collection, field, value) {
|
|
|
1057
1221
|
const schema = schemas.get(collection);
|
|
1058
1222
|
if (!schema)
|
|
1059
1223
|
return 0;
|
|
1060
|
-
const rows = await all(`SELECT COUNT(*) as count FROM ${schema.tableName} WHERE ${field} = $1`, value);
|
|
1224
|
+
const rows = await all(`SELECT COUNT(*) as count FROM ${schema.tableName} WHERE ${field} = $1`, [value]);
|
|
1061
1225
|
return Number(rows[0]?.count || 0);
|
|
1062
1226
|
}
|
|
1063
1227
|
export async function countByFieldBatch(collection, field, values) {
|
|
@@ -1067,7 +1231,7 @@ export async function countByFieldBatch(collection, field, values) {
|
|
|
1067
1231
|
if (!schema)
|
|
1068
1232
|
return new Map();
|
|
1069
1233
|
const placeholders = values.map((_, i) => `$${i + 1}`).join(',');
|
|
1070
|
-
const rows = await all(`SELECT ${field}, COUNT(*) as count FROM ${schema.tableName} WHERE ${field} IN (${placeholders}) GROUP BY ${field}`,
|
|
1234
|
+
const rows = await all(`SELECT ${field}, COUNT(*) as count FROM ${schema.tableName} WHERE ${field} IN (${placeholders}) GROUP BY ${field}`, values);
|
|
1071
1235
|
const result = new Map();
|
|
1072
1236
|
for (const row of rows) {
|
|
1073
1237
|
result.set(row[field], Number(row.count));
|
|
@@ -1078,7 +1242,7 @@ export async function findByField(collection, field, value) {
|
|
|
1078
1242
|
const schema = schemas.get(collection);
|
|
1079
1243
|
if (!schema)
|
|
1080
1244
|
return null;
|
|
1081
|
-
const rows = await all(`SELECT * FROM ${schema.tableName} WHERE ${field} = $1 LIMIT 1`, value);
|
|
1245
|
+
const rows = await all(`SELECT * FROM ${schema.tableName} WHERE ${field} = $1 LIMIT 1`, [value]);
|
|
1082
1246
|
return rows[0] || null;
|
|
1083
1247
|
}
|
|
1084
1248
|
export async function findByFieldBatch(collection, field, values) {
|
|
@@ -1088,7 +1252,7 @@ export async function findByFieldBatch(collection, field, values) {
|
|
|
1088
1252
|
if (!schema)
|
|
1089
1253
|
return new Map();
|
|
1090
1254
|
const placeholders = values.map((_, i) => `$${i + 1}`).join(',');
|
|
1091
|
-
const rows = await all(`SELECT t.*, r.handle FROM ${schema.tableName} t LEFT JOIN _repos r ON t.did = r.did WHERE t.${field} IN (${placeholders})`,
|
|
1255
|
+
const rows = await all(`SELECT t.*, r.handle FROM ${schema.tableName} t LEFT JOIN _repos r ON t.did = r.did WHERE t.${field} IN (${placeholders})`, values);
|
|
1092
1256
|
// Attach child data if this collection has decomposed arrays
|
|
1093
1257
|
if (schema.children.length > 0 && rows.length > 0) {
|
|
1094
1258
|
const uris = rows.map((r) => r.uri);
|
|
@@ -1098,7 +1262,6 @@ export async function findByFieldBatch(collection, field, values) {
|
|
|
1098
1262
|
childData.set(child.fieldName, childRows);
|
|
1099
1263
|
}
|
|
1100
1264
|
for (const row of rows) {
|
|
1101
|
-
;
|
|
1102
1265
|
row.__childData = childData;
|
|
1103
1266
|
}
|
|
1104
1267
|
}
|
|
@@ -1129,7 +1292,7 @@ export async function findUriByFields(collection, conditions) {
|
|
|
1129
1292
|
return null;
|
|
1130
1293
|
const where = conditions.map((c, i) => `${c.field} = $${i + 1}`).join(' AND ');
|
|
1131
1294
|
const params = conditions.map((c) => c.value);
|
|
1132
|
-
const rows = await all(`SELECT uri FROM ${schema.tableName} WHERE ${where} LIMIT 1`,
|
|
1295
|
+
const rows = await all(`SELECT uri FROM ${schema.tableName} WHERE ${where} LIMIT 1`, params);
|
|
1133
1296
|
return rows[0]?.uri || null;
|
|
1134
1297
|
}
|
|
1135
1298
|
const ENVELOPE_KEYS = new Set(['uri', 'cid', 'did', 'handle', 'indexed_at']);
|
|
@@ -1145,7 +1308,7 @@ export async function getChildRows(childTableName, parentUris) {
|
|
|
1145
1308
|
if (parentUris.length === 0)
|
|
1146
1309
|
return new Map();
|
|
1147
1310
|
const placeholders = parentUris.map((_, i) => `$${i + 1}`).join(',');
|
|
1148
|
-
const rows = await all(`SELECT * FROM ${childTableName} WHERE parent_uri IN (${placeholders})`,
|
|
1311
|
+
const rows = await all(`SELECT * FROM ${childTableName} WHERE parent_uri IN (${placeholders})`, parentUris);
|
|
1149
1312
|
const result = new Map();
|
|
1150
1313
|
for (const row of rows) {
|
|
1151
1314
|
const key = row.parent_uri;
|
|
@@ -1167,7 +1330,7 @@ export function reshapeRow(row, childData, unionData) {
|
|
|
1167
1330
|
if (schema) {
|
|
1168
1331
|
for (const col of schema.columns) {
|
|
1169
1332
|
nameMap.set(col.name, col.originalName);
|
|
1170
|
-
if (col.
|
|
1333
|
+
if (col.isJson)
|
|
1171
1334
|
jsonCols.add(col.name);
|
|
1172
1335
|
}
|
|
1173
1336
|
}
|
|
@@ -1273,15 +1436,17 @@ export function unpackCursor(cursor) {
|
|
|
1273
1436
|
}
|
|
1274
1437
|
}
|
|
1275
1438
|
export async function queryLabelsByDid(did) {
|
|
1276
|
-
return all(`SELECT * FROM _labels WHERE uri LIKE $1 AND neg = false AND (exp IS NULL OR exp > CURRENT_TIMESTAMP)`,
|
|
1439
|
+
return all(`SELECT * FROM _labels WHERE uri LIKE $1 AND neg = false AND (exp IS NULL OR exp > CURRENT_TIMESTAMP)`, [
|
|
1440
|
+
`at://${did}/%`,
|
|
1441
|
+
]);
|
|
1277
1442
|
}
|
|
1278
1443
|
export async function searchAccounts(query, limit = 20) {
|
|
1279
|
-
return all(`SELECT did, handle, status FROM _repos WHERE did
|
|
1444
|
+
return all(`SELECT did, handle, status FROM _repos WHERE did ${dialect.ilike} $1 OR handle ${dialect.ilike} $1 ORDER BY handle LIMIT $2`, [`%${query}%`, limit]);
|
|
1280
1445
|
}
|
|
1281
1446
|
export async function getAccountRecordCount(did) {
|
|
1282
1447
|
let total = 0;
|
|
1283
1448
|
for (const [, schema] of schemas) {
|
|
1284
|
-
const rows = await all(`SELECT COUNT(*) as count FROM ${schema.tableName} WHERE did = $1`, did);
|
|
1449
|
+
const rows = await all(`SELECT COUNT(*) as count FROM ${schema.tableName} WHERE did = $1`, [did]);
|
|
1285
1450
|
total += Number(rows[0]?.count || 0);
|
|
1286
1451
|
}
|
|
1287
1452
|
return total;
|
|
@@ -1289,17 +1454,17 @@ export async function getAccountRecordCount(did) {
|
|
|
1289
1454
|
export async function getAllRecordUrisForDid(did) {
|
|
1290
1455
|
const uris = [];
|
|
1291
1456
|
for (const [, schema] of schemas) {
|
|
1292
|
-
const rows = await all(`SELECT uri FROM ${schema.tableName} WHERE did = $1`, did);
|
|
1457
|
+
const rows = await all(`SELECT uri FROM ${schema.tableName} WHERE did = $1`, [did]);
|
|
1293
1458
|
uris.push(...rows.map((r) => r.uri));
|
|
1294
1459
|
}
|
|
1295
1460
|
return uris;
|
|
1296
1461
|
}
|
|
1297
1462
|
export async function isTakendownDid(did) {
|
|
1298
|
-
const rows = await all(`SELECT 1 FROM _repos WHERE did = $1 AND status = 'takendown' LIMIT 1`, did);
|
|
1463
|
+
const rows = await all(`SELECT 1 FROM _repos WHERE did = $1 AND status = 'takendown' LIMIT 1`, [did]);
|
|
1299
1464
|
return rows.length > 0;
|
|
1300
1465
|
}
|
|
1301
1466
|
export async function getPreferences(did) {
|
|
1302
|
-
const rows = await all(`SELECT key, value FROM _preferences WHERE did = $1`, did);
|
|
1467
|
+
const rows = await all(`SELECT key, value FROM _preferences WHERE did = $1`, [did]);
|
|
1303
1468
|
const prefs = {};
|
|
1304
1469
|
for (const row of rows) {
|
|
1305
1470
|
try {
|
|
@@ -1312,50 +1477,66 @@ export async function getPreferences(did) {
|
|
|
1312
1477
|
return prefs;
|
|
1313
1478
|
}
|
|
1314
1479
|
export async function putPreference(did, key, value) {
|
|
1315
|
-
await run(`INSERT OR REPLACE INTO _preferences (did, key, value, updated_at) VALUES ($1, $2, $3, $4)`,
|
|
1480
|
+
await run(`INSERT OR REPLACE INTO _preferences (did, key, value, updated_at) VALUES ($1, $2, $3, $4)`, [
|
|
1481
|
+
did,
|
|
1482
|
+
key,
|
|
1483
|
+
JSON.stringify(value),
|
|
1484
|
+
new Date().toISOString(),
|
|
1485
|
+
]);
|
|
1486
|
+
}
|
|
1487
|
+
export async function resolveHandleToDid(handle) {
|
|
1488
|
+
if (handle.startsWith('did:'))
|
|
1489
|
+
return handle;
|
|
1490
|
+
const rows = await all(`SELECT did FROM _repos WHERE handle = $1 LIMIT 1`, [handle]);
|
|
1491
|
+
return rows[0]?.did ?? null;
|
|
1316
1492
|
}
|
|
1317
1493
|
export async function filterTakendownDids(dids) {
|
|
1318
1494
|
if (dids.length === 0)
|
|
1319
1495
|
return new Set();
|
|
1320
1496
|
const placeholders = dids.map((_, i) => `$${i + 1}`).join(',');
|
|
1321
|
-
const rows = await all(`SELECT did FROM _repos WHERE did IN (${placeholders}) AND status = 'takendown'`,
|
|
1497
|
+
const rows = await all(`SELECT did FROM _repos WHERE did IN (${placeholders}) AND status = 'takendown'`, dids);
|
|
1322
1498
|
return new Set(rows.map((r) => r.did));
|
|
1323
1499
|
}
|
|
1324
|
-
export async function
|
|
1325
|
-
|
|
1326
|
-
|
|
1327
|
-
|
|
1328
|
-
|
|
1329
|
-
|
|
1330
|
-
|
|
1331
|
-
|
|
1332
|
-
|
|
1333
|
-
|
|
1334
|
-
|
|
1335
|
-
|
|
1336
|
-
const childColSelects = child.columns
|
|
1337
|
-
.map((c) => `json_extract_string(item.val, '$.${c.originalName}')`)
|
|
1338
|
-
.join(', ');
|
|
1339
|
-
const childColNames = ['parent_uri', 'parent_did', ...child.columns.map((c) => c.name)];
|
|
1340
|
-
const notNullFilters = child.columns
|
|
1341
|
-
.filter((c) => c.notNull)
|
|
1342
|
-
.map((c) => `json_extract_string(item.val, '$.${c.originalName}') IS NOT NULL`);
|
|
1343
|
-
const whereClause = [`p.${snakeField} IS NOT NULL`, ...notNullFilters].join(' AND ');
|
|
1344
|
-
try {
|
|
1345
|
-
await run(`DELETE FROM ${child.tableName}`);
|
|
1346
|
-
await run(`
|
|
1347
|
-
INSERT INTO ${child.tableName} (${childColNames.join(', ')})
|
|
1348
|
-
SELECT p.uri, p.did, ${childColSelects}
|
|
1349
|
-
FROM ${schema.tableName} p,
|
|
1350
|
-
unnest(from_json(p.${snakeField}::JSON, '["json"]')) AS item(val)
|
|
1351
|
-
WHERE ${whereClause}
|
|
1352
|
-
`);
|
|
1353
|
-
const result = await all(`SELECT COUNT(*)::INTEGER as n FROM ${child.tableName}`);
|
|
1354
|
-
console.log(`[db] Backfilled ${child.tableName}: ${result[0]?.n || 0} rows`);
|
|
1355
|
-
}
|
|
1356
|
-
catch (err) {
|
|
1357
|
-
console.warn(`[db] Backfill skipped for ${child.tableName}: ${err.message}`);
|
|
1358
|
-
}
|
|
1359
|
-
}
|
|
1500
|
+
export async function insertReport(report) {
|
|
1501
|
+
const createdAt = new Date().toISOString();
|
|
1502
|
+
const rows = await all(`INSERT INTO _reports (subject_uri, subject_did, label, reason, reported_by, created_at) VALUES ($1, $2, $3, $4, $5, $6) RETURNING id`, [report.subjectUri, report.subjectDid, report.label, report.reason || null, report.reportedBy, createdAt]);
|
|
1503
|
+
return { id: rows[0].id };
|
|
1504
|
+
}
|
|
1505
|
+
export async function queryReports(opts) {
|
|
1506
|
+
const conditions = [];
|
|
1507
|
+
const params = [];
|
|
1508
|
+
let idx = 1;
|
|
1509
|
+
if (opts.status) {
|
|
1510
|
+
conditions.push(`r.status = $${idx++}`);
|
|
1511
|
+
params.push(opts.status);
|
|
1360
1512
|
}
|
|
1513
|
+
if (opts.label) {
|
|
1514
|
+
conditions.push(`r.label = $${idx++}`);
|
|
1515
|
+
params.push(opts.label);
|
|
1516
|
+
}
|
|
1517
|
+
const where = conditions.length > 0 ? `WHERE ${conditions.join(' AND ')}` : '';
|
|
1518
|
+
const limit = opts.limit || 50;
|
|
1519
|
+
const offset = opts.offset || 0;
|
|
1520
|
+
const countRows = await all(`SELECT ${dialect.countAsInteger} as count FROM _reports r ${where}`, params);
|
|
1521
|
+
const total = Number(countRows[0]?.count || 0);
|
|
1522
|
+
const rows = await all(`SELECT r.*, rp.handle as reported_by_handle FROM _reports r LEFT JOIN _repos rp ON r.reported_by = rp.did ${where} ORDER BY r.created_at DESC LIMIT $${idx++} OFFSET $${idx++}`, [...params, limit, offset]);
|
|
1523
|
+
return { reports: rows, total };
|
|
1524
|
+
}
|
|
1525
|
+
export async function resolveReport(id, action, resolvedBy) {
|
|
1526
|
+
const rows = await all(`SELECT subject_uri, label, status FROM _reports WHERE id = $1`, [id]);
|
|
1527
|
+
if (!rows.length)
|
|
1528
|
+
return null;
|
|
1529
|
+
if (rows[0].status !== 'open')
|
|
1530
|
+
return null;
|
|
1531
|
+
await run(`UPDATE _reports SET status = $1, resolved_by = $2, resolved_at = $3 WHERE id = $4`, [
|
|
1532
|
+
action,
|
|
1533
|
+
resolvedBy,
|
|
1534
|
+
new Date().toISOString(),
|
|
1535
|
+
id,
|
|
1536
|
+
]);
|
|
1537
|
+
return { subjectUri: rows[0].subject_uri, label: rows[0].label };
|
|
1538
|
+
}
|
|
1539
|
+
export async function getOpenReportCount() {
|
|
1540
|
+
const rows = await all(`SELECT ${dialect.countAsInteger} as count FROM _reports WHERE status = 'open'`);
|
|
1541
|
+
return Number(rows[0]?.count || 0);
|
|
1361
1542
|
}
|