@intentius/chant-lexicon-sql 0.124.46 → 0.124.49

package/dist/sqlserver/lexer.js

@@ -0,0 +1,254 @@
+/**
+ * A T-SQL tokenizer for the statements `chant import` reads from SQL Server.
+ *
+ * T-SQL differs from the lexers the other dialects use in four places: an
+ * identifier can be `[bracketed]` (a `]` inside is written `]]`) or
+ * "double quoted", a string can carry an `N` prefix, a line holding only `GO`
+ * ends a batch (it is a client command, not a statement), and statements need
+ * no terminating `;`. Comments (`--` and nestable `/* *\/`) are dropped; each
+ * token keeps its 1-based line and column in the source.
+ */
+export class TsqlLexError extends Error {
+    constructor(message, line, column) {
+        super(`${message} (line ${String(line)}, column ${String(column)})`);
+        this.line = line;
+        this.column = column;
+        this.name = "TsqlLexError";
+    }
+}
+const WORD_START = /[A-Za-z_#@\u0080-￿]/;
+const WORD_PART = /[A-Za-z0-9_#$@\u0080-￿]/;
+/** Tokenize a file. A `GO` on a line of its own (optionally with a count) becomes one `go` token. */
+export function tokenize(text) {
+    const tokens = [];
+    let i = 0;
+    let line = 1;
+    let lineStart = 0;
+    const n = text.length;
+    // A byte order mark starts files saved by SSMS and Visual Studio.
+    if (text.charCodeAt(0) === 0xfeff) {
+        i = 1;
+        lineStart = 1;
+    }
+    const push = (kind, start, end, value, startLine, startCol) => tokens.push({ kind, text: text.slice(start, end), value, line: startLine, column: startCol, start, end });
+    while (i < n) {
+        const c = text[i];
+        if (c === "\n") {
+            line++;
+            i++;
+            lineStart = i;
+            continue;
+        }
+        if (/\s/.test(c)) {
+            i++;
+            continue;
+        }
+        const col = i - lineStart + 1;
+        if (c === "-" && text[i + 1] === "-") {
+            while (i < n && text[i] !== "\n")
+                i++;
+            continue;
+        }
+        if (c === "/" && text[i + 1] === "*") {
+            let depth = 1;
+            const startLine = line;
+            const startCol = col;
+            i += 2;
+            while (i < n && depth > 0) {
+                if (text[i] === "/" && text[i + 1] === "*") {
+                    depth++;
+                    i += 2;
+                }
+                else if (text[i] === "*" && text[i + 1] === "/") {
+                    depth--;
+                    i += 2;
+                }
+                else {
+                    if (text[i] === "\n") {
+                        line++;
+                        lineStart = i + 1;
+                    }
+                    i++;
+                }
+            }
+            if (depth > 0)
+                throw new TsqlLexError("unterminated /* comment", startLine, startCol);
+            continue;
+        }
+        if (c === "[") {
+            const start = i;
+            let value = "";
+            i++;
+            for (;;) {
+                if (i >= n)
+                    throw new TsqlLexError("unterminated [identifier]", line, col);
+                if (text[i] === "]") {
+                    if (text[i + 1] === "]") {
+                        value += "]";
+                        i += 2;
+                        continue;
+                    }
+                    i++;
+                    break;
+                }
+                if (text[i] === "\n") {
+                    line++;
+                    lineStart = i + 1;
+                }
+                value += text[i];
+                i++;
+            }
+            push("qident", start, i, value, line, col);
+            continue;
+        }
+        if (c === '"') {
+            const start = i;
+            let value = "";
+            i++;
+            for (;;) {
+                if (i >= n)
+                    throw new TsqlLexError('unterminated "identifier"', line, col);
+                if (text[i] === '"') {
+                    if (text[i + 1] === '"') {
+                        value += '"';
+                        i += 2;
+                        continue;
+                    }
+                    i++;
+                    break;
+                }
+                value += text[i];
+                i++;
+            }
+            push("qident", start, i, value, line, col);
+            continue;
+        }
+        if (c === "'" || ((c === "N" || c === "n") && text[i + 1] === "'")) {
+            const start = i;
+            const startLine = line;
+            if (c !== "'")
+                i++;
+            i++;
+            let value = "";
+            for (;;) {
+                if (i >= n)
+                    throw new TsqlLexError("unterminated string", startLine, col);
+                if (text[i] === "'") {
+                    if (text[i + 1] === "'") {
+                        value += "'";
+                        i += 2;
+                        continue;
+                    }
+                    i++;
+                    break;
+                }
+                if (text[i] === "\n") {
+                    line++;
+                    lineStart = i + 1;
+                }
+                value += text[i];
+                i++;
+            }
+            push("string", start, i, value, startLine, col);
+            continue;
+        }
+        if (/[0-9]/.test(c) || (c === "." && /[0-9]/.test(text[i + 1] ?? ""))) {
+            const start = i;
+            if (c === "0" && (text[i + 1] === "x" || text[i + 1] === "X")) {
+                i += 2;
+                while (i < n && /[0-9A-Fa-f]/.test(text[i]))
+                    i++;
+            }
+            else {
+                while (i < n && /[0-9]/.test(text[i]))
+                    i++;
+                if (text[i] === ".") {
+                    i++;
+                    while (i < n && /[0-9]/.test(text[i]))
+                        i++;
+                }
+                if ((text[i] === "e" || text[i] === "E") && /[0-9+-]/.test(text[i + 1] ?? "")) {
+                    i += 2;
+                    while (i < n && /[0-9]/.test(text[i]))
+                        i++;
+                }
+            }
+            push("number", start, i, text.slice(start, i), line, col);
+            continue;
+        }
+        if (c === "@" && text[i + 1] !== undefined && WORD_PART.test(text[i + 1])) {
+            const start = i;
+            i++;
+            while (i < n && WORD_PART.test(text[i]))
+                i++;
+            push("variable", start, i, text.slice(start, i), line, col);
+            continue;
+        }
+        if (WORD_START.test(c)) {
+            const start = i;
+            i++;
+            while (i < n && WORD_PART.test(text[i]))
+                i++;
+            const word = text.slice(start, i);
+            push("word", start, i, word, line, col);
+            continue;
+        }
+        // Two-character operators are one token so a `<>` is not a `<` and a `>`.
+        const two = text.slice(i, i + 2);
+        if (["<>", ">=", "<=", "!=", "+=", "-=", "*=", "/=", "||", "::"].includes(two)) {
+            push("punct", i, i + 2, two, line, col);
+            i += 2;
+            continue;
+        }
+        push("punct", i, i + 1, c, line, col);
+        i++;
+    }
+    return markBatchSeparators(tokens, text);
+}
+/** `GO` alone on its line (or `GO 5`) is a separator, not a word. */
+function markBatchSeparators(tokens, text) {
+    const out = [];
+    for (let k = 0; k < tokens.length; k++) {
+        const t = tokens[k];
+        if (t.kind === "word" && t.value.toUpperCase() === "GO") {
+            const prev = tokens[k - 1];
+            const next = tokens[k + 1];
+            const alone = (a) => a === undefined || a.line !== t.line;
+            const count = next && next.kind === "number" && next.line === t.line ? next : undefined;
+            const after = count ? tokens[k + 2] : next;
+            // The line before `GO` ends the earlier statement, and a `GO` mid-line is a name.
+            if (alone(prev) && alone(after) && !/[A-Za-z0-9_]/.test(text.slice(t.end, t.end + 1))) {
+                out.push({ ...t, kind: "go" });
+                if (count)
+                    k++;
+                continue;
+            }
+        }
+        out.push(t);
+    }
+    return out;
+}
+/** Tokens split into batches at `go` tokens; empty batches are dropped. */
+export function batches(tokens) {
+    const out = [];
+    let cur = [];
+    for (const t of tokens) {
+        if (t.kind === "go") {
+            if (cur.length > 0)
+                out.push(cur);
+            cur = [];
+        }
+        else
+            cur.push(t);
+    }
+    if (cur.length > 0)
+        out.push(cur);
+    return out;
+}
+export const upper = (t) => (t && t.kind === "word" ? t.value.toUpperCase() : "");
+/** The source text of a run of tokens, from the first token's start to the last token's end. */
+export function sourceOf(text, tokens) {
+    if (tokens.length === 0)
+        return "";
+    return text.slice(tokens[0].start, tokens[tokens.length - 1].end);
+}

package/dist/sqlserver/parse.d.ts

@@ -0,0 +1,185 @@
+/**
+ * T-SQL DDL into a syntax tree: `CREATE TABLE`, `CREATE [UNIQUE] [CLUSTERED |
+ * NONCLUSTERED] INDEX`, `CREATE [CLUSTERED | NONCLUSTERED] COLUMNSTORE
+ * INDEX ... [ORDER (...)]`, `CREATE VIEW`, `CREATE SCHEMA`, `ALTER TABLE ...
+ * ADD CONSTRAINT`, and `sp_addextendedproperty` descriptions.
+ *
+ * Grammar sources (Microsoft Learn, Transact-SQL reference):
+ * - CREATE TABLE: https://learn.microsoft.com/en-us/sql/t-sql/statements/create-table-transact-sql
+ * - CREATE INDEX: https://learn.microsoft.com/en-us/sql/t-sql/statements/create-index-transact-sql
+ * - CREATE COLUMNSTORE INDEX (the `ORDER (column [, ...n])` clause, which the
+ *   SQL Server grammar places before or after `WITH`):
+ *   https://learn.microsoft.com/en-us/sql/t-sql/statements/create-columnstore-index-transact-sql
+ *
+ * Every statement that is not one of these comes back as `refused` (a kind of
+ * object this source cannot read, named: `trigger [dbo].[trg_audit]`) or
+ * `ignored` (a session setting or a `DROP`, which declares nothing). Nothing
+ * is dropped without being returned.
+ */
+import { type Token } from "./lexer.js";
+export interface Loc {
+    line: number;
+    column: number;
+}
+/** A multi-part name, unquoted: `[Sales].[Orders]` is `["Sales", "Orders"]`. */
+export interface QName {
+    parts: string[];
+    loc: Loc;
+}
+export interface KeyColumn {
+    name: string;
+    descending: boolean;
+}
+export type ConstraintKind = "primary-key" | "unique" | "foreign-key" | "check" | "default";
+export interface ConstraintAst {
+    kind: ConstraintKind;
+    name?: string;
+    /** `CLUSTERED` or `NONCLUSTERED` when written. */
+    clustered?: boolean;
+    columns: KeyColumn[];
+    references?: {
+        table: QName;
+        columns: string[];
+    };
+    /** A check's or a default's expression, as written (outer parentheses kept). */
+    expression?: string;
+    expressionTokens?: Token[];
+    /** The column a column-level constraint was written on; the default's column. */
+    onColumn?: string;
+    /** `WITH (...)` or `ON ...` text after the key, which the translation names but does not carry. */
+    options: string[];
+    loc: Loc;
+}
+export interface ColumnAst {
+    name: string;
+    loc: Loc;
+    /** Lower-cased type name; `sysname` and alias types appear as written. */
+    type?: string;
+    /** Qualified type (`[dbo].[Flag]`) when the name has a schema. */
+    typeSchema?: string;
+    /** `max` is `"max"`. */
+    typeArgs: Array<number | "max">;
+    nullable?: boolean;
+    identity?: {
+        seed: string;
+        increment: string;
+    };
+    /** `AS (expr) [PERSISTED]` and no type. */
+    computed?: {
+        expression: string;
+        persisted: boolean;
+    };
+    /** `GENERATED ALWAYS AS ROW START | END | TRANSACTION_ID_* ...` as written. */
+    generated?: string;
+    collate?: string;
+    rowguid: boolean;
+    sparse: boolean;
+    /** Other column options carried as text and named in the review: `MASKED WITH`, `ENCRYPTED WITH`, `FILESTREAM`, `NOT FOR REPLICATION`. */
+    extras: string[];
+}
+export interface TableAst {
+    kind: "table";
+    name: QName;
+    columns: ColumnAst[];
+    constraints: ConstraintAst[];
+    /** `INDEX name [UNIQUE] [CLUSTERED|NONCLUSTERED] (cols)` inside the table. */
+    indexes: IndexAst[];
+    /** `PERIOD FOR SYSTEM_TIME (a, b)`. */
+    period?: {
+        start: string;
+        end: string;
+    };
+    /** Text of each clause after the column list: `WITH (...)`, `ON [PRIMARY]`, `TEXTIMAGE_ON ...`. */
+    options: string[];
+    /** `AS FileTable` / `AS NODE` / `AS EDGE`. */
+    asKind?: string;
+    loc: Loc;
+}
+export interface IndexAst {
+    kind: "index";
+    name: string;
+    /** Absent for an index written inside `CREATE TABLE`. */
+    table?: QName;
+    unique: boolean;
+    clustered: boolean;
+    columnstore: boolean;
+    columns: KeyColumn[];
+    include: string[];
+    /** The columnstore `ORDER (...)` columns. */
+    order: string[];
+    where?: string;
+    options: string[];
+    loc: Loc;
+}
+export interface ViewAst {
+    kind: "view";
+    name: QName;
+    columns: string[];
+    /** `WITH SCHEMABINDING, ...` */
+    options: string[];
+    /** The SELECT, as written, without the trailing `;`. */
+    body: string;
+    bodyTokens: Token[];
+    loc: Loc;
+}
+export interface SchemaAst {
+    kind: "schema";
+    name: string;
+    loc: Loc;
+}
+export interface AlterConstraintAst {
+    kind: "alter-constraint";
+    table: QName;
+    constraint: ConstraintAst;
+    loc: Loc;
+}
+export interface ExtendedPropertyAst {
+    kind: "extended-property";
+    /** `description` or the property's name. */
+    property: string;
+    value: string;
+    schema?: string;
+    table?: string;
+    column?: string;
+    /** `INDEX`, `CONSTRAINT`, `VIEW`, ... when level 2 is not a column. */
+    level2?: {
+        type: string;
+        name: string;
+    };
+    level1Type?: string;
+    loc: Loc;
+}
+export interface RefusedAst {
+    kind: "refused";
+    /** `trigger`, `procedure`, `function`, ... */
+    objectKind: string;
+    name: string;
+    reason: string;
+    loc: Loc;
+}
+export interface IgnoredAst {
+    kind: "ignored";
+    /** What the statement was, e.g. `SET ANSI_NULLS ON`. */
+    text: string;
+    reason: string;
+    loc: Loc;
+}
+export type StatementAst = TableAst | IndexAst | ViewAst | SchemaAst | AlterConstraintAst | ExtendedPropertyAst | RefusedAst | IgnoredAst;
+export declare class TsqlParseError extends Error {
+    readonly loc: Loc;
+    constructor(message: string, loc: Loc);
+}
+/** A statement that began and could not be read: its first words, and why. */
+export interface UnreadStatement {
+    text: string;
+    reason: string;
+    loc: Loc;
+}
+export interface ParseResult {
+    statements: StatementAst[];
+    /** Statements that begin like one this source reads and are malformed or use syntax it does not read. */
+    unread: UnreadStatement[];
+}
+/** Parse a file of T-SQL. Lexing errors throw `TsqlLexError`. */
+export declare function parseTsql(text: string): ParseResult;
+//# sourceMappingURL=parse.d.ts.map

package/dist/sqlserver/parse.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"parse.d.ts","sourceRoot":"","sources":["../../src/sqlserver/parse.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;GAiBG;AAEH,OAAO,EAAsC,KAAK,KAAK,EAAE,MAAM,SAAS,CAAC;AAEzE,MAAM,WAAW,GAAG;IAClB,IAAI,EAAE,MAAM,CAAC;IACb,MAAM,EAAE,MAAM,CAAC;CAChB;AAED,gFAAgF;AAChF,MAAM,WAAW,KAAK;IACpB,KAAK,EAAE,MAAM,EAAE,CAAC;IAChB,GAAG,EAAE,GAAG,CAAC;CACV;AAED,MAAM,WAAW,SAAS;IACxB,IAAI,EAAE,MAAM,CAAC;IACb,UAAU,EAAE,OAAO,CAAC;CACrB;AAED,MAAM,MAAM,cAAc,GAAG,aAAa,GAAG,QAAQ,GAAG,aAAa,GAAG,OAAO,GAAG,SAAS,CAAC;AAE5F,MAAM,WAAW,aAAa;IAC5B,IAAI,EAAE,cAAc,CAAC;IACrB,IAAI,CAAC,EAAE,MAAM,CAAC;IACd,kDAAkD;IAClD,SAAS,CAAC,EAAE,OAAO,CAAC;IACpB,OAAO,EAAE,SAAS,EAAE,CAAC;IACrB,UAAU,CAAC,EAAE;QAAE,KAAK,EAAE,KAAK,CAAC;QAAC,OAAO,EAAE,MAAM,EAAE,CAAA;KAAE,CAAC;IACjD,gFAAgF;IAChF,UAAU,CAAC,EAAE,MAAM,CAAC;IACpB,gBAAgB,CAAC,EAAE,KAAK,EAAE,CAAC;IAC3B,iFAAiF;IACjF,QAAQ,CAAC,EAAE,MAAM,CAAC;IAClB,mGAAmG;IACnG,OAAO,EAAE,MAAM,EAAE,CAAC;IAClB,GAAG,EAAE,GAAG,CAAC;CACV;AAED,MAAM,WAAW,SAAS;IACxB,IAAI,EAAE,MAAM,CAAC;IACb,GAAG,EAAE,GAAG,CAAC;IACT,0EAA0E;IAC1E,IAAI,CAAC,EAAE,MAAM,CAAC;IACd,kEAAkE;IAClE,UAAU,CAAC,EAAE,MAAM,CAAC;IACpB,wBAAwB;IACxB,QAAQ,EAAE,KAAK,CAAC,MAAM,GAAG,KAAK,CAAC,CAAC;IAChC,QAAQ,CAAC,EAAE,OAAO,CAAC;IACnB,QAAQ,CAAC,EAAE;QAAE,IAAI,EAAE,MAAM,CAAC;QAAC,SAAS,EAAE,MAAM,CAAA;KAAE,CAAC;IAC/C,2CAA2C;IAC3C,QAAQ,CAAC,EAAE;QAAE,UAAU,EAAE,MAAM,CAAC;QAAC,SAAS,EAAE,OAAO,CAAA;KAAE,CAAC;IACtD,+EAA+E;IAC/E,SAAS,CAAC,EAAE,MAAM,CAAC;IACnB,OAAO,CAAC,EAAE,MAAM,CAAC;IACjB,OAAO,EAAE,OAAO,CAAC;IACjB,MAAM,EAAE,OAAO,CAAC;IAChB,0IAA0I;IAC1I,MAAM,EAAE,MAAM,EAAE,CAAC;CAClB;AAED,MAAM,WAAW,QAAQ;IACvB,IAAI,EAAE,OAAO,CAAC;IACd,IAAI,EAAE,KAAK,CAAC;IACZ,OAAO,EAAE,SAAS,EAAE,CAAC;IACrB,WAAW,EAAE,aAAa,EAAE,CAAC;IAC7B,8EAA8E;IAC9E,OAAO,EAAE,QAAQ,EAAE,CAAC;IACpB,uCAAuC;IACvC,MAAM,CAAC,EAAE;QAAE,KAAK,EAAE,MAAM,CAAC;QAAC,GAAG,EAAE,MAAM,CAAA;KAAE,CAAC;IACxC,mGAAmG;IACnG,OAAO,EAAE,MAAM,EAAE,CAAC;IAClB,8CAA8C;IAC9C,MAAM,CAAC,EAAE,MAAM,CAAC;IAChB,GAAG,EAAE,GAAG,CAAC;CACV;AAED,MAAM,WAAW,QAAQ;IACvB,IAAI,EAAE,OAAO,CAAC;IACd,IAAI,EAAE,MAAM,CAAC;IACb,yDAAyD;IACzD,KAAK,CAAC,EAAE,KAAK,CAAC;IACd,MAAM,EAAE,OAAO,CAAC;IAChB,SAAS,EAAE,OAAO,CAAC;IACnB,WAAW,EAAE,OAAO,CAAC;IACrB,OAAO,EAAE,SAAS,EAAE,CAAC;IACrB,OAAO,EAAE,MAAM,EAAE,CAAC;IAClB,6CAA6C;IAC7C,KAAK,EAAE,MAAM,EAAE,CAAC;IAChB,KAAK,CAAC,EAAE,MAAM,CAAC;IACf,OAAO,EAAE,MAAM,EAAE,CAAC;IAClB,GAAG,EAAE,GAAG,CAAC;CACV;AAED,MAAM,WAAW,OAAO;IACtB,IAAI,EAAE,MAAM,CAAC;IACb,IAAI,EAAE,KAAK,CAAC;IACZ,OAAO,EAAE,MAAM,EAAE,CAAC;IAClB,gCAAgC;IAChC,OAAO,EAAE,MAAM,EAAE,CAAC;IAClB,wDAAwD;IACxD,IAAI,EAAE,MAAM,CAAC;IACb,UAAU,EAAE,KAAK,EAAE,CAAC;IACpB,GAAG,EAAE,GAAG,CAAC;CACV;AAED,MAAM,WAAW,SAAS;IACxB,IAAI,EAAE,QAAQ,CAAC;IACf,IAAI,EAAE,MAAM,CAAC;IACb,GAAG,EAAE,GAAG,CAAC;CACV;AAED,MAAM,WAAW,kBAAkB;IACjC,IAAI,EAAE,kBAAkB,CAAC;IACzB,KAAK,EAAE,KAAK,CAAC;IACb,UAAU,EAAE,aAAa,CAAC;IAC1B,GAAG,EAAE,GAAG,CAAC;CACV;AAED,MAAM,WAAW,mBAAmB;IAClC,IAAI,EAAE,mBAAmB,CAAC;IAC1B,4CAA4C;IAC5C,QAAQ,EAAE,MAAM,CAAC;IACjB,KAAK,EAAE,MAAM,CAAC;IACd,MAAM,CAAC,EAAE,MAAM,CAAC;IAChB,KAAK,CAAC,EAAE,MAAM,CAAC;IACf,MAAM,CAAC,EAAE,MAAM,CAAC;IAChB,uEAAuE;IACvE,MAAM,CAAC,EAAE;QAAE,IAAI,EAAE,MAAM,CAAC;QAAC,IAAI,EAAE,MAAM,CAAA;KAAE,CAAC;IACxC,UAAU,CAAC,EAAE,MAAM,CAAC;IACpB,GAAG,EAAE,GAAG,CAAC;CACV;AAED,MAAM,WAAW,UAAU;IACzB,IAAI,EAAE,SAAS,CAAC;IAChB,8CAA8C;IAC9C,UAAU,EAAE,MAAM,CAAC;IACnB,IAAI,EAAE,MAAM,CAAC;IACb,MAAM,EAAE,MAAM,CAAC;IACf,GAAG,EAAE,GAAG,CAAC;CACV;AAED,MAAM,WAAW,UAAU;IACzB,IAAI,EAAE,SAAS,CAAC;IAChB,wDAAwD;IACxD,IAAI,EAAE,MAAM,CAAC;IACb,MAAM,EAAE,MAAM,CAAC;IACf,GAAG,EAAE,GAAG,CAAC;CACV;AAED,MAAM,MAAM,YAAY,GAAG,QAAQ,GAAG,QAAQ,GAAG,OAAO,GAAG,SAAS,GAAG,kBAAkB,GAAG,mBAAmB,GAAG,UAAU,GAAG,UAAU,CAAC;AAE1I,qBAAa,cAAe,SAAQ,KAAK;IAGrC,QAAQ,CAAC,GAAG,EAAE,GAAG;gBADjB,OAAO,EAAE,MAAM,EACN,GAAG,EAAE,GAAG;CAKpB;AAID,8EAA8E;AAC9E,MAAM,WAAW,eAAe;IAC9B,IAAI,EAAE,MAAM,CAAC;IACb,MAAM,EAAE,MAAM,CAAC;IACf,GAAG,EAAE,GAAG,CAAC;CACV;AAED,MAAM,WAAW,WAAW;IAC1B,UAAU,EAAE,YAAY,EAAE,CAAC;IAC3B,yGAAyG;IACzG,MAAM,EAAE,eAAe,EAAE,CAAC;CAC3B;AAypBD,iEAAiE;AACjE,wBAAgB,SAAS,CAAC,IAAI,EAAE,MAAM,GAAG,WAAW,CAMnD"}