@huggingface/transformers-structured-output 4.3.0

package/types/StructuredOutputProcessor.d.ts

@@ -0,0 +1,22 @@
+import { LogitsProcessorList } from '@huggingface/transformers';
+import { type JSONSchema, type TokenizerSource } from './engine';
+export type ResponseFormat = {
+    type: 'json_object';
+} | {
+    type: 'json_schema';
+    json_schema: JSONSchema;
+} | {
+    type: 'regex';
+    regex: string;
+};
+export declare class StructuredOutputProcessor extends LogitsProcessorList {
+    /**
+     * Precomputes the tokenizer-derived data structures used by every
+     * constraint. The first processor per tokenizer otherwise pays this cost
+     * (hundreds of milliseconds for large vocabularies) in its constructor;
+     * call this once after loading the model to pay it early instead.
+     */
+    static warmup(tokenizer: TokenizerSource): void;
+    constructor(tokenizer: TokenizerSource, responseFormat: ResponseFormat);
+}
+//# sourceMappingURL=StructuredOutputProcessor.d.ts.map

package/types/StructuredOutputProcessor.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"StructuredOutputProcessor.d.ts","sourceRoot":"","sources":["../src/StructuredOutputProcessor.ts"],"names":[],"mappings":"AAAA,OAAO,EAAmB,mBAAmB,EAAe,MAAM,2BAA2B,CAAC;AAE9F,OAAO,EAGH,KAAK,UAAU,EAEf,KAAK,eAAe,EACvB,MAAM,UAAU,CAAC;AAGlB,MAAM,MAAM,cAAc,GACpB;IAAE,IAAI,EAAE,aAAa,CAAA;CAAE,GACvB;IAAE,IAAI,EAAE,aAAa,CAAC;IAAC,WAAW,EAAE,UAAU,CAAA;CAAE,GAChD;IAAE,IAAI,EAAE,OAAO,CAAC;IAAC,KAAK,EAAE,MAAM,CAAA;CAAE,CAAC;AAWvC,qBAAa,yBAA0B,SAAQ,mBAAmB;IAC9D;;;;;OAKG;IACH,MAAM,CAAC,MAAM,CAAC,SAAS,EAAE,eAAe,GAAG,IAAI;gBAInC,SAAS,EAAE,eAAe,EAAE,cAAc,EAAE,cAAc;CAOzE"}

package/types/engine/constraint.d.ts

@@ -0,0 +1,30 @@
+import type { JSONSchema, TokenizerSource } from './types';
+type ResponseFormat = {
+    type: 'json_object';
+} | {
+    type: 'json_schema';
+    json_schema: JSONSchema;
+} | {
+    type: 'regex';
+    regex: string;
+};
+export type TokenConstraint = {
+    vocabSize: number;
+    fillMask(target: Uint32Array): boolean;
+    commit(tokenId: number): boolean;
+    repeatedWhitespace(): {
+        tokenIds: readonly number[];
+        count: number;
+    } | undefined;
+};
+/**
+ * Builds and caches the tokenizer-derived data structures (token tries, string
+ * masks) ahead of time. This is the expensive part of creating the first
+ * constraint for a tokenizer (hundreds of milliseconds for a 256k vocabulary),
+ * so calling this right after loading a model moves that cost off the first
+ * generation. Subsequent calls with the same tokenizer are free.
+ */
+export declare function prepareTokenizer(tokenizerSource: TokenizerSource): void;
+export declare function createTokenConstraint(tokenizerSource: TokenizerSource, responseFormat: ResponseFormat): TokenConstraint;
+export {};
+//# sourceMappingURL=constraint.d.ts.map

package/types/engine/constraint.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"constraint.d.ts","sourceRoot":"","sources":["../../src/engine/constraint.ts"],"names":[],"mappings":"AAGA,OAAO,KAAK,EAAmB,UAAU,EAAE,eAAe,EAAE,MAAM,SAAS,CAAC;AAmB5E,KAAK,cAAc,GACb;IAAE,IAAI,EAAE,aAAa,CAAA;CAAE,GACvB;IAAE,IAAI,EAAE,aAAa,CAAC;IAAC,WAAW,EAAE,UAAU,CAAA;CAAE,GAChD;IAAE,IAAI,EAAE,OAAO,CAAC;IAAC,KAAK,EAAE,MAAM,CAAA;CAAE,CAAC;AAEvC,MAAM,MAAM,eAAe,GAAG;IAC1B,SAAS,EAAE,MAAM,CAAC;IAClB,QAAQ,CAAC,MAAM,EAAE,WAAW,GAAG,OAAO,CAAC;IACvC,MAAM,CAAC,OAAO,EAAE,MAAM,GAAG,OAAO,CAAC;IACjC,kBAAkB,IAAI;QAAE,QAAQ,EAAE,SAAS,MAAM,EAAE,CAAC;QAAC,KAAK,EAAE,MAAM,CAAA;KAAE,GAAG,SAAS,CAAC;CACpF,CAAC;AAKF;;;;;;GAMG;AACH,wBAAgB,gBAAgB,CAAC,eAAe,EAAE,eAAe,GAAG,IAAI,CAEvE;AAED,wBAAgB,qBAAqB,CACjC,eAAe,EAAE,eAAe,EAChC,cAAc,EAAE,cAAc,GAC/B,eAAe,CA+FjB"}

package/types/engine/index.d.ts

@@ -0,0 +1,3 @@
+export { createTokenConstraint, prepareTokenizer, type TokenConstraint } from './constraint';
+export type { JSONSchema, TokenizerSource } from './types';
+//# sourceMappingURL=index.d.ts.map

package/types/engine/index.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"index.d.ts","sourceRoot":"","sources":["../../src/engine/index.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,qBAAqB,EAAE,gBAAgB,EAAE,KAAK,eAAe,EAAE,MAAM,cAAc,CAAC;AAC7F,YAAY,EAAE,UAAU,EAAE,eAAe,EAAE,MAAM,SAAS,CAAC"}

package/types/engine/json.d.ts

@@ -0,0 +1,70 @@
+import type { ConstraintState, JSONSchema } from './types';
+type Schema = boolean | Record<string, unknown>;
+type JsonNode = {
+    kind: 'null';
+    value: null;
+} | {
+    kind: 'boolean';
+    value: boolean;
+} | {
+    kind: 'number';
+    value: number;
+    raw: string;
+} | {
+    kind: 'string';
+    value: string;
+} | {
+    kind: 'array';
+    value: unknown[];
+    items: JsonNode[];
+} | {
+    kind: 'object';
+    value: Record<string, unknown>;
+    entries: Array<{
+        key: string;
+        node: JsonNode;
+    }>;
+};
+type Guidance = {
+    itemSeparator: string;
+    keySeparator: string;
+    itemBytes: Uint8Array;
+    keyBytes: Uint8Array;
+    whitespaceFlexible: boolean;
+};
+type Mode = 'value' | 'array-value' | 'object-key' | 'colon' | 'item-separator' | 'after-value' | 'string' | 'key-string' | 'escape' | 'key-escape' | 'unicode' | 'key-unicode' | 'number' | 'literal' | 'done' | 'dead';
+type ObjectFrame = {
+    kind: 'object';
+    schema: Schema;
+    seen: ReadonlySet<string>;
+    entries: ReadonlyArray<{
+        key: string;
+        node: JsonNode;
+    }>;
+    key?: string;
+    childSchema?: Schema;
+};
+type ArrayFrame = {
+    kind: 'array';
+    schema: Schema;
+    length: number;
+    items: readonly JsonNode[];
+};
+type Frame = ObjectFrame | ArrayFrame;
+type JsonState = {
+    mode: Mode;
+    schema: Schema;
+    stack: readonly Frame[];
+    bytes?: readonly number[];
+    stringLength?: number;
+    stringPending?: number;
+    text?: string;
+    literal?: string;
+    literalValue?: null | boolean;
+    index?: number;
+    highSurrogate?: number;
+    guidance: Guidance;
+};
+export declare function compileJsonSchema(schema: JSONSchema, stringKeyClamp?: number): ConstraintState<JsonState>;
+export {};
+//# sourceMappingURL=json.d.ts.map

package/types/engine/json.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"json.d.ts","sourceRoot":"","sources":["../../src/engine/json.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,eAAe,EAAE,UAAU,EAAE,MAAM,SAAS,CAAC;AAE3D,KAAK,MAAM,GAAG,OAAO,GAAG,MAAM,CAAC,MAAM,EAAE,OAAO,CAAC,CAAC;AAEhD,KAAK,QAAQ,GACP;IAAE,IAAI,EAAE,MAAM,CAAC;IAAC,KAAK,EAAE,IAAI,CAAA;CAAE,GAC7B;IAAE,IAAI,EAAE,SAAS,CAAC;IAAC,KAAK,EAAE,OAAO,CAAA;CAAE,GACnC;IAAE,IAAI,EAAE,QAAQ,CAAC;IAAC,KAAK,EAAE,MAAM,CAAC;IAAC,GAAG,EAAE,MAAM,CAAA;CAAE,GAC9C;IAAE,IAAI,EAAE,QAAQ,CAAC;IAAC,KAAK,EAAE,MAAM,CAAA;CAAE,GACjC;IAAE,IAAI,EAAE,OAAO,CAAC;IAAC,KAAK,EAAE,OAAO,EAAE,CAAC;IAAC,KAAK,EAAE,QAAQ,EAAE,CAAA;CAAE,GACtD;IAAE,IAAI,EAAE,QAAQ,CAAC;IAAC,KAAK,EAAE,MAAM,CAAC,MAAM,EAAE,OAAO,CAAC,CAAC;IAAC,OAAO,EAAE,KAAK,CAAC;QAAE,GAAG,EAAE,MAAM,CAAC;QAAC,IAAI,EAAE,QAAQ,CAAA;KAAE,CAAC,CAAA;CAAE,CAAC;AAC1G,KAAK,QAAQ,GAAG;IACZ,aAAa,EAAE,MAAM,CAAC;IACtB,YAAY,EAAE,MAAM,CAAC;IACrB,SAAS,EAAE,UAAU,CAAC;IACtB,QAAQ,EAAE,UAAU,CAAC;IACrB,kBAAkB,EAAE,OAAO,CAAC;CAC/B,CAAC;AACF,KAAK,IAAI,GACH,OAAO,GACP,aAAa,GACb,YAAY,GACZ,OAAO,GACP,gBAAgB,GAChB,aAAa,GACb,QAAQ,GACR,YAAY,GACZ,QAAQ,GACR,YAAY,GACZ,SAAS,GACT,aAAa,GACb,QAAQ,GACR,SAAS,GACT,MAAM,GACN,MAAM,CAAC;AAEb,KAAK,WAAW,GAAG;IACf,IAAI,EAAE,QAAQ,CAAC;IACf,MAAM,EAAE,MAAM,CAAC;IACf,IAAI,EAAE,WAAW,CAAC,MAAM,CAAC,CAAC;IAC1B,OAAO,EAAE,aAAa,CAAC;QAAE,GAAG,EAAE,MAAM,CAAC;QAAC,IAAI,EAAE,QAAQ,CAAA;KAAE,CAAC,CAAC;IACxD,GAAG,CAAC,EAAE,MAAM,CAAC;IACb,WAAW,CAAC,EAAE,MAAM,CAAC;CACxB,CAAC;AAEF,KAAK,UAAU,GAAG;IACd,IAAI,EAAE,OAAO,CAAC;IACd,MAAM,EAAE,MAAM,CAAC;IACf,MAAM,EAAE,MAAM,CAAC;IACf,KAAK,EAAE,SAAS,QAAQ,EAAE,CAAC;CAC9B,CAAC;AAEF,KAAK,KAAK,GAAG,WAAW,GAAG,UAAU,CAAC;AAEtC,KAAK,SAAS,GAAG;IACb,IAAI,EAAE,IAAI,CAAC;IACX,MAAM,EAAE,MAAM,CAAC;IACf,KAAK,EAAE,SAAS,KAAK,EAAE,CAAC;IACxB,KAAK,CAAC,EAAE,SAAS,MAAM,EAAE,CAAC;IAK1B,YAAY,CAAC,EAAE,MAAM,CAAC;IACtB,aAAa,CAAC,EAAE,MAAM,CAAC;IACvB,IAAI,CAAC,EAAE,MAAM,CAAC;IACd,OAAO,CAAC,EAAE,MAAM,CAAC;IACjB,YAAY,CAAC,EAAE,IAAI,GAAG,OAAO,CAAC;IAC9B,KAAK,CAAC,EAAE,MAAM,CAAC;IACf,aAAa,CAAC,EAAE,MAAM,CAAC;IACvB,QAAQ,EAAE,QAAQ,CAAC;CACtB,CAAC;AA2CF,wBAAgB,iBAAiB,CAAC,MAAM,EAAE,UAAU,EAAE,cAAc,SAAW,GAAG,eAAe,CAAC,SAAS,CAAC,CAa3G"}

package/types/engine/regex.d.ts

@@ -0,0 +1,3 @@
+import type { ConstraintState } from './types';
+export declare function compileRegex(source: string): ConstraintState<number>;
+//# sourceMappingURL=regex.d.ts.map

package/types/engine/regex.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"regex.d.ts","sourceRoot":"","sources":["../../src/engine/regex.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,eAAe,EAAE,MAAM,SAAS,CAAC;AAY/C,wBAAgB,YAAY,CAAC,MAAM,EAAE,MAAM,GAAG,eAAe,CAAC,MAAM,CAAC,CASpE"}

package/types/engine/tokenizer.d.ts

@@ -0,0 +1,8 @@
+import type { TokenizerSource } from './types';
+export type TokenizerData = {
+    tokens: Uint8Array[];
+    eosTokenId: number;
+    specialTokenIds: Set<number>;
+};
+export declare function extractTokenizer(tokenizer: TokenizerSource): TokenizerData;
+//# sourceMappingURL=tokenizer.d.ts.map

package/types/engine/tokenizer.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"tokenizer.d.ts","sourceRoot":"","sources":["../../src/engine/tokenizer.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,eAAe,EAAE,MAAM,SAAS,CAAC;AAI/C,MAAM,MAAM,aAAa,GAAG;IACxB,MAAM,EAAE,UAAU,EAAE,CAAC;IACrB,UAAU,EAAE,MAAM,CAAC;IACnB,eAAe,EAAE,GAAG,CAAC,MAAM,CAAC,CAAC;CAChC,CAAC;AAKF,wBAAgB,gBAAgB,CAAC,SAAS,EAAE,eAAe,GAAG,aAAa,CAyD1E"}

package/types/engine/types.d.ts

@@ -0,0 +1,16 @@
+export type JSONValue = null | boolean | number | string | JSONValue[] | {
+    [key: string]: JSONValue;
+};
+export type JSONSchema = boolean | {
+    [key: string]: JSONValue;
+};
+export type TokenizerSource = object | ((...args: unknown[]) => unknown);
+export interface ConstraintState<State> {
+    readonly initial: State;
+    transition(state: State, byte: number): State;
+    viable(state: State): boolean;
+    accepting(state: State): boolean;
+    stringCapacity?(state: State): number | undefined;
+    maskKey?(state: State): string | undefined;
+}
+//# sourceMappingURL=types.d.ts.map

package/types/engine/types.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"types.d.ts","sourceRoot":"","sources":["../../src/engine/types.ts"],"names":[],"mappings":"AAAA,MAAM,MAAM,SAAS,GAAG,IAAI,GAAG,OAAO,GAAG,MAAM,GAAG,MAAM,GAAG,SAAS,EAAE,GAAG;IAAE,CAAC,GAAG,EAAE,MAAM,GAAG,SAAS,CAAA;CAAE,CAAC;AACtG,MAAM,MAAM,UAAU,GAAG,OAAO,GAAG;IAAE,CAAC,GAAG,EAAE,MAAM,GAAG,SAAS,CAAA;CAAE,CAAC;AAChE,MAAM,MAAM,eAAe,GAAG,MAAM,GAAG,CAAC,CAAC,GAAG,IAAI,EAAE,OAAO,EAAE,KAAK,OAAO,CAAC,CAAC;AAEzE,MAAM,WAAW,eAAe,CAAC,KAAK;IAClC,QAAQ,CAAC,OAAO,EAAE,KAAK,CAAC;IACxB,UAAU,CAAC,KAAK,EAAE,KAAK,EAAE,IAAI,EAAE,MAAM,GAAG,KAAK,CAAC;IAC9C,MAAM,CAAC,KAAK,EAAE,KAAK,GAAG,OAAO,CAAC;IAC9B,SAAS,CAAC,KAAK,EAAE,KAAK,GAAG,OAAO,CAAC;IACjC,cAAc,CAAC,CAAC,KAAK,EAAE,KAAK,GAAG,MAAM,GAAG,SAAS,CAAC;IAClD,OAAO,CAAC,CAAC,KAAK,EAAE,KAAK,GAAG,MAAM,GAAG,SAAS,CAAC;CAC9C"}

package/types/index.d.ts

@@ -0,0 +1,3 @@
+export { StructuredOutputProcessor } from './StructuredOutputProcessor';
+export type { ResponseFormat } from './StructuredOutputProcessor';
+//# sourceMappingURL=index.d.ts.map

package/types/index.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"index.d.ts","sourceRoot":"","sources":["../src/index.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,yBAAyB,EAAE,MAAM,6BAA6B,CAAC;AACxE,YAAY,EAAE,cAAc,EAAE,MAAM,6BAA6B,CAAC"}

package/types/utils/mask.d.ts

@@ -0,0 +1,3 @@
+import { type Tensor } from '@huggingface/transformers';
+export declare function applyMask(logits: Tensor, mask: Uint32Array, vocabSize: number): void;
+//# sourceMappingURL=mask.d.ts.map

package/types/utils/mask.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"mask.d.ts","sourceRoot":"","sources":["../../src/utils/mask.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,KAAK,MAAM,EAAE,MAAM,2BAA2B,CAAC;AAIxD,wBAAgB,SAAS,CAAC,MAAM,EAAE,MAAM,EAAE,IAAI,EAAE,WAAW,EAAE,SAAS,EAAE,MAAM,GAAG,IAAI,CAyBpF"}