{{ message }}
Repository navigation
Expand file tree
/
Copy pathsemantic-diagnostics.ts
More file actions
709 lines (658 loc) · 22.3 KB
/
Copy pathsemantic-diagnostics.ts
File metadata and controls
709 lines (658 loc) · 22.3 KB
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
import type { SQLNamespace } from "@codemirror/lang-sql";
import { type Diagnostic, linter } from "@codemirror/lint";
import type { Extension, Text } from "@codemirror/state";
import type { EditorView } from "@codemirror/view";
import {
findNamespaceItemByEndMatch,
isArrayNamespace,
isObjectNamespace,
isSelfChildrenNamespace,
traverseNamespacePath,
} from "./namespace-utils.js";
import { NodeSqlParser } from "./parser.js";
import { resolveSqlSchema, type SqlSchemaSource } from "./schema-facet.js";
import { type SqlStatement, SqlStructureAnalyzer } from "./structure-analyzer.js";
import type { SqlParser } from "./types.js";
const DEFAULT_DELAY = 750;
/** Severity for a semantic check; "off" disables the check entirely */
export type SemanticSeverity = "error" | "warning" | "off";
/**
* Configuration options for the semantic (schema-aware) SQL linter
*/
export interface SqlSemanticLinterConfig {
/**
* Database schema to validate queries against. Falls back to the shared
* `sqlSchemaFacet` when not provided. Without a schema (or with an empty
* one) the linter is inert.
*
* Function sources are invoked on every lint pass and should be
* cheap/memoized.
*/
schema?: SqlSchemaSource;
/** Custom SQL parser instance to use for analysis */
parser?: SqlParser;
/**
* Structure analyzer used to split the document into statements.
* Pass a shared instance (e.g. the one used by the gutter) to reuse its cache.
*/
structureAnalyzer?: SqlStructureAnalyzer;
/** Delay in milliseconds before running validation (default: 750) */
delay?: number;
/** Per-check severity overrides (default: "warning" for every check) */
severity?: {
/** A referenced table is not present in the schema */
unknownTable?: SemanticSeverity;
/** A column is not present in its resolved table */
unknownColumn?: SemanticSeverity;
/** An unqualified column exists in two or more referenced tables */
ambiguousColumn?: SemanticSeverity;
};
}
type FindingKind = "unknownTable" | "unknownColumn" | "ambiguousColumn";
interface SemanticFinding {
kind: FindingKind;
/** Identifier to highlight in the statement text */
identifier: string;
message: string;
}
/** Loosely-typed node-sql-parser AST node */
interface AstNode {
[key: string]: unknown;
}
/** A table-like source in a query scope (FROM entry, DML target, CTE, subquery) */
interface TableSource {
/** Identifier segments, preserving dots inside quoted names (empty for subqueries) */
path: readonly string[];
/** Unqualified table name */
name: string;
alias: string | null;
/** References a CTE defined in the statement */
isCte: boolean;
/** Subquery/table-function source whose columns we cannot know */
opaque: boolean;
}
/**
* A single query scope (a SELECT, or the target/WHERE portion of a DML
* statement). Column references are checked against the scope's sources.
*/
interface QueryScope {
sources: TableSource[];
columnRefs: AstNode[];
/** Lowercased SELECT-list aliases, referable from GROUP BY/ORDER BY etc. */
selectAliases: Set<string>;
/**
* Whether unqualified column refs may be checked. Disabled for scopes that
* can see an outer scope (correlated subqueries), where an unqualified name
* may legally resolve to an outer table.
*/
checkUnqualified: boolean;
/** A USING/NATURAL join makes shared columns unambiguous; skip that check */
hasUsingJoin: boolean;
}
function isSelectNode(node: unknown): node is AstNode & { type: "select" } {
return (
typeof node === "object" &&
node !== null &&
!Array.isArray(node) &&
(node as AstNode).type === "select"
);
}
function asArray(value: unknown): AstNode[] {
return Array.isArray(value) ? (value as AstNode[]) : [];
}
function newScope(checkUnqualified: boolean): QueryScope {
return {
sources: [],
columnRefs: [],
selectAliases: new Set(),
checkUnqualified,
hasUsingJoin: false,
};
}
/**
* Registers CTE names defined on a statement node and collects scopes from
* their bodies. Names are registered before bodies are walked so a CTE
* referencing another (or itself, when recursive) is never flagged.
*/
function registerCtes(node: AstNode, ctes: Set<string>, scopes: QueryScope[]): void {
const withList = asArray(node.with);
for (const cte of withList) {
const name = cte.name as AstNode | string | undefined;
const value = typeof name === "string" ? name : name?.value;
if (typeof value === "string") {
ctes.add(value.toLowerCase());
}
}
for (const cte of withList) {
const stmt = cte.stmt as AstNode | undefined;
const body = stmt?.ast ?? stmt;
if (isSelectNode(body)) {
collectSelect(body, ctes, scopes, false);
}
}
}
/**
* Adds a FROM/target entry to the scope. Entries are either named tables
* (`{db, schema, table, as}`) or expression sources (subqueries, table
* functions), which are treated as opaque.
*/
function addTableEntry(
entry: AstNode,
scope: QueryScope,
ctes: Set<string>,
scopes: QueryScope[],
): void {
if (entry.using != null) {
scope.hasUsingJoin = true;
}
if (typeof entry.join === "string" && entry.join.toLowerCase().includes("natural")) {
scope.hasUsingJoin = true;
}
if (entry.on != null) {
walkExpr(entry.on, scope, ctes, scopes);
}
const expr = entry.expr as AstNode | undefined;
if (expr != null) {
const sub = expr.ast ?? expr;
if (isSelectNode(sub)) {
collectSelect(sub, ctes, scopes, false);
}
scope.sources.push({
path: [],
name: "",
alias: typeof entry.as === "string" ? entry.as : null,
isCte: false,
opaque: true,
});
return;
}
if (typeof entry.table === "string") {
const qualifier = entry.db ?? entry.catalog;
const parts = [qualifier, entry.schema, entry.table].filter(
(part): part is string => typeof part === "string" && part.length > 0,
);
scope.sources.push({
path: parts,
name: entry.table,
alias: typeof entry.as === "string" ? entry.as : null,
isCte: parts.length === 1 && ctes.has(entry.table.toLowerCase()),
opaque: false,
});
}
}
/**
* Generic walk collecting column refs into the current scope and opening a
* new (correlated) scope for any nested SELECT.
*/
function walkExpr(value: unknown, scope: QueryScope, ctes: Set<string>, scopes: QueryScope[]): void {
if (value == null || typeof value !== "object") {
return;
}
if (Array.isArray(value)) {
for (const item of value) {
walkExpr(item, scope, ctes, scopes);
}
return;
}
const node = value as AstNode;
if (node.type === "column_ref") {
scope.columnRefs.push(node);
return;
}
if (isSelectNode(node)) {
collectSelect(node, ctes, scopes, true);
return;
}
if (isSelectNode(node.ast)) {
collectSelect(node.ast, ctes, scopes, true);
return;
}
for (const [key, child] of Object.entries(node)) {
if (key === "loc" || key === "tableList" || key === "columnList") {
continue;
}
walkExpr(child, scope, ctes, scopes);
}
}
function collectSelect(
node: AstNode,
inheritedCtes: Set<string>,
scopes: QueryScope[],
correlated: boolean,
): void {
const ctes = new Set(inheritedCtes);
registerCtes(node, ctes, scopes);
const scope = newScope(!correlated);
scopes.push(scope);
for (const entry of asArray(node.from)) {
addTableEntry(entry, scope, ctes, scopes);
}
for (const col of asArray(node.columns)) {
if (typeof col.as === "string") {
scope.selectAliases.add(col.as.toLowerCase());
}
}
for (const [key, child] of Object.entries(node)) {
// `with` and `from` were handled above; walking them again would create
// duplicate scopes (and `on` clauses are walked by addTableEntry)
if (key === "with" || key === "from" || key === "loc") {
continue;
}
walkExpr(child, scope, ctes, scopes);
}
}
/**
* Collects query scopes from a single parsed statement. DDL targets (CREATE
* TABLE x) define tables rather than read them, so they are never added as
* sources; DML targets (INSERT/UPDATE/DELETE) must exist and are.
*/
function collectStatement(node: AstNode, scopes: QueryScope[]): void {
const ctes = new Set<string>();
switch (node.type) {
case "select": {
collectSelect(node, ctes, scopes, false);
return;
}
case "update": {
registerCtes(node, ctes, scopes);
const scope = newScope(true);
scopes.push(scope);
for (const entry of asArray(node.table)) {
addTableEntry(entry, scope, ctes, scopes);
}
for (const entry of asArray(node.from)) {
addTableEntry(entry, scope, ctes, scopes);
}
// SET entries are `{column, table, value}` where `column` is a plain
// string in some dialect grammars (no column_ref type), so the generic
// walk would miss them
for (const entry of asArray(node.set)) {
if (entry.column != null) {
scope.columnRefs.push({
type: "column_ref",
table: entry.table ?? null,
column: entry.column,
});
}
walkExpr(entry.value, scope, ctes, scopes);
}
for (const [key, child] of Object.entries(node)) {
if (
key === "with" ||
key === "table" ||
key === "from" ||
key === "set" ||
key === "loc"
) {
continue;
}
walkExpr(child, scope, ctes, scopes);
}
return;
}
case "delete": {
registerCtes(node, ctes, scopes);
const scope = newScope(true);
scopes.push(scope);
// node.table duplicates node.from entries (with `addition: true`), so
// only FROM is used as the source list
for (const entry of asArray(node.from)) {
addTableEntry(entry, scope, ctes, scopes);
}
for (const [key, child] of Object.entries(node)) {
if (key === "with" || key === "table" || key === "from" || key === "loc") {
continue;
}
walkExpr(child, scope, ctes, scopes);
}
return;
}
case "insert":
case "replace": {
registerCtes(node, ctes, scopes);
// The insert target must exist, but its scope has no readable columns
// context, so unqualified column checks stay off
const scope = newScope(false);
scopes.push(scope);
for (const entry of asArray(node.table)) {
addTableEntry(entry, scope, ctes, scopes);
}
for (const [key, child] of Object.entries(node)) {
// `columns` holds the target column names as plain strings
if (key === "with" || key === "table" || key === "columns" || key === "loc") {
continue;
}
if (isSelectNode(child)) {
// INSERT INTO t SELECT ... — the select is a top-level query
collectSelect(child, ctes, scopes, false);
} else {
walkExpr(child, scope, ctes, scopes);
}
}
return;
}
case "create":
case "alter":
case "drop": {
const scope = newScope(false);
scopes.push(scope);
for (const [key, child] of Object.entries(node)) {
if (key === "table" || key === "loc") {
continue;
}
if (isSelectNode(child)) {
// CREATE TABLE/VIEW ... AS SELECT — top-level query, not correlated
collectSelect(child, ctes, scopes, false);
} else {
walkExpr(child, scope, ctes, scopes);
}
}
return;
}
default: {
const scope = newScope(false);
scopes.push(scope);
for (const [key, child] of Object.entries(node)) {
if (key === "loc") {
continue;
}
walkExpr(child, scope, ctes, scopes);
}
}
}
}
/**
* Extracts the column name from a column_ref node. Depending on the dialect
* grammar the column is either a plain string or `{expr: {value}}`.
*/
function getColumnName(ref: AstNode): string | null {
const column = ref.column;
if (typeof column === "string") {
return column;
}
if (typeof column === "object" && column !== null) {
const expr = (column as AstNode).expr as AstNode | undefined;
const value = expr?.value ?? (column as AstNode).value;
if (typeof value === "string") {
return value;
}
}
return null;
}
function getColumnQualifier(ref: AstNode): readonly string[] | null {
if (typeof ref.table !== "string" || ref.table.length === 0) {
return null;
}
return [ref.db, ref.schema, ref.table].filter(
(segment): segment is string => typeof segment === "string" && segment.length > 0,
);
}
/** Extracts column names when the namespace node is a column array */
function columnsOf(namespace: SQLNamespace | undefined): string[] | null {
if (namespace == null) {
return null;
}
const resolved = isSelfChildrenNamespace(namespace) ? namespace.children : namespace;
if (isArrayNamespace(resolved)) {
return resolved.map((col) => (typeof col === "string" ? col : col.label));
}
return null;
}
interface ResolvedTable {
exists: boolean;
/** Column names when the table resolves to a column array; null otherwise */
columns: string[] | null;
}
/**
* Resolves a (possibly qualified) table path against the schema,
* case-insensitively. An unqualified or partially-qualified name also matches
* a table nested deeper in the namespace (e.g. `users` matches `mydb.users`)
* so that under-qualified references are never flagged.
*/
function resolveTable(schema: SQLNamespace, path: readonly string[]): ResolvedTable {
const exact = traverseNamespacePath(schema, path, { caseSensitive: false });
if (exact) {
return { exists: true, columns: columnsOf(exact.namespace) };
}
const segments = path.map((segment) => segment.toLowerCase());
const lastSegment = segments[segments.length - 1];
if (!lastSegment) {
return { exists: false, columns: null };
}
const matches = findNamespaceItemByEndMatch(schema, lastSegment).filter((item) => {
if (columnsOf(item.namespace) === null) {
return false; // only table-like matches (nodes holding a column array)
}
if (item.path.length < segments.length) {
return false;
}
const suffix = item.path.slice(-segments.length).map((segment) => segment.toLowerCase());
return segments.every((segment, i) => suffix[i] === segment);
});
if (matches.length === 0) {
return { exists: false, columns: null };
}
// Only trust the column list when the match is unambiguous
return {
exists: true,
columns: matches.length === 1 ? columnsOf(matches[0]?.namespace) : null,
};
}
function includesIgnoreCase(haystack: string[], needle: string): boolean {
const lower = needle.toLowerCase();
return haystack.some((item) => item.toLowerCase() === lower);
}
/**
* Runs the semantic checks for one parsed statement and returns findings,
* deduplicated by kind + identifier.
*/
function analyzeStatementAst(ast: AstNode, schema: SQLNamespace): SemanticFinding[] {
const scopes: QueryScope[] = [];
collectStatement(ast, scopes);
const findings = new Map<string, SemanticFinding>();
const addFinding = (finding: SemanticFinding) => {
const key = `${finding.kind}::${finding.identifier.toLowerCase()}`;
if (!findings.has(key)) {
findings.set(key, finding);
}
};
for (const scope of scopes) {
const resolved = scope.sources.map((source) => {
if (source.opaque || source.isCte) {
return { source, exists: true, columns: null };
}
const table = resolveTable(schema, source.path);
return { source, exists: table.exists, columns: table.columns };
});
for (const { source, exists } of resolved) {
if (!source.opaque && !source.isCte && !exists) {
addFinding({
kind: "unknownTable",
identifier: source.name,
message: `Unknown table '${source.path.join(".")}'`,
});
}
}
for (const ref of scope.columnRefs) {
const name = getColumnName(ref);
if (!name || name === "*") {
continue;
}
const qualifier = getColumnQualifier(ref);
if (qualifier) {
const segments = qualifier.map((segment) => segment.toLowerCase());
const unqualifiedName = segments.length === 1 ? segments[0] : undefined;
const aliasMatch = unqualifiedName
? resolved.find((entry) => entry.source.alias?.toLowerCase() === unqualifiedName)
: undefined;
const match =
aliasMatch ??
resolved.find(
(entry) =>
entry.source.name.toLowerCase() === unqualifiedName ||
(entry.source.path.length === segments.length &&
entry.source.path.every((segment, i) => segment.toLowerCase() === segments[i])),
);
// An unresolved qualifier may be an outer-scope alias; skip it
if (match?.columns && !includesIgnoreCase(match.columns, name)) {
addFinding({
kind: "unknownColumn",
identifier: name,
message: `Column '${name}' not found in table '${match.source.path.join(".") || qualifier.join(".")}'`,
});
}
continue;
}
if (!scope.checkUnqualified || scope.selectAliases.has(name.toLowerCase())) {
continue;
}
const containing = resolved.filter(
(entry) => entry.columns !== null && includesIgnoreCase(entry.columns, name),
);
if (containing.length >= 2 && !scope.hasUsingJoin) {
addFinding({
kind: "ambiguousColumn",
identifier: name,
message: `Column '${name}' is ambiguous; it exists in ${containing
.map((entry) => entry.source.name)
.join(", ")}`,
});
continue;
}
// Only flag unknown columns when the scope reads exactly one table
// whose columns are fully known — anything else is not confidently
// resolvable and false positives are worse than under-reporting
const only = resolved.length === 1 ? resolved[0] : undefined;
if (only && only.columns !== null && containing.length === 0) {
addFinding({
kind: "unknownColumn",
identifier: name,
message: `Column '${name}' not found in table '${only.source.path.join(".")}'`,
});
}
}
}
return [...findings.values()];
}
/**
* Finds an identifier occurrence (word-bounded, case-insensitive) in text.
* Used to position diagnostics, since node-sql-parser's name lists and AST
* nodes for tables/columns carry no location information.
*/
function findIdentifier(text: string, name: string): number {
const lowerText = text.toLowerCase();
const needle = name.toLowerCase();
let index = lowerText.indexOf(needle);
while (index !== -1) {
const before = index > 0 ? (lowerText[index - 1] ?? "") : "";
const after = lowerText[index + needle.length] ?? "";
if (!/[\w$]/.test(before) && !/[\w$]/.test(after)) {
return index;
}
index = lowerText.indexOf(needle, index + 1);
}
return -1;
}
function findingToDiagnostic(
finding: SemanticFinding,
stmt: SqlStatement,
doc: Text,
severity: "error" | "warning",
): Diagnostic {
const statementText = doc.sliceString(stmt.from, stmt.to);
const offset = findIdentifier(statementText, finding.identifier);
const from = offset === -1 ? stmt.from : stmt.from + offset;
const to =
offset === -1
? Math.min(stmt.from + (statementText.match(/^[\w"'`.]+/)?.[0].length ?? 1), stmt.to)
: Math.min(from + finding.identifier.length, stmt.to);
return {
from,
to,
severity,
message: finding.message,
source: "sql-schema",
};
}
/** True when the schema has nothing to validate against */
function isEmptySchema(schema: SQLNamespace): boolean {
if (isArrayNamespace(schema)) {
return schema.length === 0;
}
if (isObjectNamespace(schema)) {
return Object.keys(schema).length === 0;
}
return false;
}
function createSemanticLintSource(config: SqlSemanticLinterConfig = {}) {
const parser = config.parser || new NodeSqlParser();
const analyzer = config.structureAnalyzer || new SqlStructureAnalyzer(parser);
const severities = {
unknownTable: config.severity?.unknownTable ?? "warning",
unknownColumn: config.severity?.unknownColumn ?? "warning",
ambiguousColumn: config.severity?.ambiguousColumn ?? "warning",
} as const;
return async (view: EditorView): Promise<Diagnostic[]> => {
const schema = resolveSqlSchema(config.schema, view);
// No schema (or an empty placeholder while it loads) → stay inert rather
// than flagging every table as unknown
if (schema == null || isEmptySchema(schema)) {
return [];
}
const doc = view.state.doc;
if (!doc.toString().trim()) {
return [];
}
const diagnostics: Diagnostic[] = [];
const statements = await analyzer.analyzeDocument(view.state);
for (const stmt of statements) {
// Never stack semantic noise on top of syntax errors
if (!stmt.isValid) {
continue;
}
const result = await parser.parse(stmt.content, { state: view.state });
if (!result.success || result.ast == null) {
continue;
}
const asts = Array.isArray(result.ast) ? result.ast : [result.ast];
for (const ast of asts) {
for (const finding of analyzeStatementAst(ast as AstNode, schema)) {
const severity = severities[finding.kind];
if (severity === "off") {
continue;
}
diagnostics.push(findingToDiagnostic(finding, stmt, doc, severity));
}
}
}
return diagnostics;
};
}
/**
* Creates a schema-aware SQL linter extension that validates queries against
* a database schema, reporting unknown tables, unknown columns, and ambiguous
* column references.
*
* Checks only run on statements that parse cleanly, and skip anything that is
* not confidently resolvable (CTEs, subquery outputs, aliases from outer
* scopes, expressions), preferring under-reporting over false positives.
*
* @param config Configuration options for the semantic linter
* @returns A CodeMirror linter extension
*
* @example
* ```ts
* import { sqlSemanticLinter } from '@marimo-team/codemirror-sql';
*
* const linter = sqlSemanticLinter({
* schema: { users: ['id', 'name'], posts: ['id', 'user_id'] },
* severity: { unknownTable: 'error' },
* });
* ```
*/
export function sqlSemanticLinter(config: SqlSemanticLinterConfig = {}): Extension {
return linter(createSemanticLintSource(config), {
delay: config.delay || DEFAULT_DELAY,
});
}
export const exportedForTesting = { createSemanticLintSource };
You can’t perform that action at this time.