dsh-ab-ocr 0.1.0

package/lib/index.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"index.js","sourceRoot":"","sources":["../src/index.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;GAsCG;AAGH,OAAO,EAAE,UAAU,EAAwB,MAAM,wBAAwB,CAAA;AAIzE,OAAO,EAAE,SAAS,EAAE,MAAM,aAAa,CAAA;AACvC,OAAO,EAAE,SAAS,EAAE,MAAM,gBAAgB,CAAA;AAC1C,OAAO,EAAE,SAAS,EAAE,MAAM,cAAc,CAAA;AAIxC,OAAO,EAAE,MAAM,EAAE,MAAM,aAAa,CAAA;AACpC,OAAO,EAAE,aAAa,EAAE,QAAQ,EAAE,MAAM,gBAAgB,CAAA;AACxD,OAAO,EAAE,cAAc,EAAE,aAAa,EAAE,MAAM,gBAAgB,CAAA;AAC9D,OAAO,EAAE,sBAAsB,EAAE,gBAAgB,EAAE,YAAY,EAAE,YAAY,EAAE,YAAY,EAAE,MAAM,eAAe,CAAA;AAClH,OAAO,EAAE,WAAW,EAAE,MAAM,aAAa,CAAA;AACzC,OAAO,EACL,oBAAoB,EACpB,QAAQ,EACR,UAAU,EACV,YAAY,EACZ,gBAAgB,EAChB,aAAa,EACb,mBAAmB,GACpB,MAAM,WAAW,CAAA;AAClB,OAAO,EAAE,aAAa,EAAE,MAAM,aAAa,CAAA;AAC3C,OAAO,EAAE,cAAc,EAAE,SAAS,EAAE,MAAM,aAAa,CAAA;AACvD,OAAO,EAAE,SAAS,EAAE,QAAQ,EAAkB,MAAM,cAAc,CAAA;AAClE,OAAO,EAAE,UAAU,EAAE,UAAU,EAAE,MAAM,aAAa,CAAA;AACpD,OAAO,EAAE,SAAS,EAAE,MAAM,gBAAgB,CAAA;AAE1C,MAAM,CAAC,MAAM,IAAI,GAAG,KAAK,CAAA;AACzB,MAAM,CAAC,MAAM,MAAM,GAAG,CAAC,OAAO,EAAE,cAAc,CAAC,CAAA;AAE/C,uCAAuC;AACvC,MAAM,CAAC,MAAM,SAAS,GAAG,KAAK,CAAA;AAE9B;;;;GAIG;AACH,MAAM,CAAC,MAAM,YAAY,GAAG,UAAU,CAAA;AAEtC,uEAAuE;AACvE,MAAM,CAAC,MAAM,aAAa,GAAG,IAAI,CAAA;AAEjC,MAAM,WAAW,GAAG;IAClB,6EAA6E;IAC7E,2FAA2F;IAC3F,iHAAiH;CAClH,CAAC,IAAI,CAAC,GAAG,CAAC,CAAA;AAEX,mFAAmF;AACnF,MAAM,CAAC,MAAM,YAAY,GAAG;IAC1B,oLAAoL;IACpL,sMAAsM;IACtM,gEAAgE;CACjE,CAAC,IAAI,CAAC,GAAG,CAAC,CAAA;AAEX;;;;;GAKG;AACH,MAAM,UAAU,WAAW,CAAC,IAAa;IACvC,OAAO;QACL,IAAI,EAAE,SAAS;QACf,KAAK,EAAE,yBAAyB;QAChC,IAAI,EAAE,MAAM;QACZ,QAAQ,EAAE,IAAI,CAAC,IAAI,IAAI,CAAC,IAAI,CAAC,KAAK,IAAI,EAAE,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC;KACrD,CAAA;AACH,CAAC;AAED;;;;;;GAMG;AACH,MAAM,UAAU,KAAK,CAAC,GAAY,EAAE,MAAc;IAChD,GAAG,CAAC,YAAY,CAAC,OAAO,CAAC;QACvB,IAAI,EAAE,YAAY;QAClB,KAAK,EAAE,aAAa;QACpB,IAAI,EAAE,CAAC,EAAE,KAAK,EAAE,EAAE,EAAE,CAAC,GAAG,CAAC,KAAK,CAAC,GAAG,CAAC,SAAS,EAAE,KAAK,CAAC,KAAK,SAAS,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,YAAY;KACvF,CAAC,CAAA;IAEF,GAAG,CAAC,KAAK,CAAC,QAAQ,CAAC,UAAU,CAAC;QAC5B,IAAI,EAAE,SAAS;QACf,WAAW,EAAE,WAAW;QACxB,UAAU,EAAE;YACV,IAAI,EAAE;gBACJ,IAAI,EAAE,QAAQ;gBACd,WAAW,EAAE,kFAAkF;aAChG;YACD,KAAK,EAAE;gBACL,IAAI,EAAE,OAAO;gBACb,KAAK,EAAE,EAAE,IAAI,EAAE,QAAQ,EAAE;gBACzB,WAAW,EAAE,yCAAyC;aACvD;YACD,KAAK,EAAE;gBACL,IAAI,EAAE,QAAQ;gBACd,WAAW,EAAE,uHAAuH;aACrI;YACD,SAAS,EAAE;gBACT,IAAI,EAAE,QAAQ;gBACd,WAAW,EAAE,6GAA6G;aAC3H;YACD,MAAM,EAAE;gBACN,IAAI,EAAE,QAAQ;gBACd,WAAW,EAAE,0UAA0U;aACxV;SACF;QACD,MAAM,EAAE;YACN,MAAM,EAAE;gBACN,IAAI,EAAE,QAAQ;gBACd,oBAAoB,EAAE,KAAK;gBAC3B,UAAU,EAAE;oBACV,SAAS,EAAE;wBACT,IAAI,EAAE,OAAO;wBACb,QAAQ,EAAE,IAAI;wBACd,KAAK,EAAE;4BACL,IAAI,EAAE,QAAQ;4BACd,oBAAoB,EAAE,KAAK;4BAC3B,UAAU,EAAE;gCACV,KAAK,EAAE,EAAE,IAAI,EAAE,QAAQ,EAAE,QAAQ,EAAE,IAAI,EAAE;gCACzC,MAAM,EAAE,EAAE,IAAI,EAAE,QAAQ,EAAE,QAAQ,EAAE,IAAI,EAAE;gCAC1C,KAAK,EAAE,EAAE,IAAI,EAAE,QAAQ,EAAE,QAAQ,EAAE,IAAI,EAAE;gCACzC,UAAU,EAAE,EAAE,IAAI,EAAE,QAAQ,EAAE,QAAQ,EAAE,IAAI,EAAE;gCAC9C,KAAK,EAAE,EAAE,IAAI,EAAE,QAAQ,EAAE,QAAQ,EAAE,IAAI,EAAE;gCACzC,QAAQ,EAAE,EAAE,IAAI,EAAE,QAAQ,EAAE,QAAQ,EAAE,IAAI,EAAE;gCAC5C,KAAK,EAAE,EAAE,IAAI,EAAE,QAAQ,EAAE,QAAQ,EAAE,IAAI,EAAE;gCACzC,kBAAkB,EAAE,EAAE,IAAI,EAAE,QAAQ,EAAE,QAAQ,EAAE,IAAI,EAAE;gCACtD,mBAAmB,EAAE,EAAE,IAAI,EAAE,QAAQ,EAAE,QAAQ,EAAE,IAAI,EAAE;gCACvD,iBAAiB,EAAE,EAAE,IAAI,EAAE,QAAQ,EAAE,QAAQ,EAAE,IAAI,EAAE;gCACrD,OAAO,EAAE,EAAE,IAAI,EAAE,QAAQ,EAAE,QAAQ,EAAE,IAAI,EAAE;gCAC3C,OAAO,EAAE,EAAE,IAAI,EAAE,QAAQ,EAAE,QAAQ,EAAE,IAAI,EAAE;gCAC3C,MAAM,EAAE,EAAE,IAAI,EAAE,QAAQ,EAAE,QAAQ,EAAE,IAAI,EAAE;gCAC1C,SAAS,EAAE,EAAE,IAAI,EAAE,OAAO,EAAE,QAAQ,EAAE,IAAI,EAAE,KAAK,EAAE,EAAE,IAAI,EAAE,QAAQ,EAAE,EAAE;gCACvE,WAAW,EAAE,EAAE,IAAI,EAAE,QAAQ,EAAE,QAAQ,EAAE,IAAI,EAAE;gCAC/C,gBAAgB,EAAE,EAAE,IAAI,EAAE,SAAS,EAAE,QAAQ,EAAE,IAAI,EAAE;gCACrD,WAAW,EAAE,EAAE,IAAI,EAAE,QAAQ,EAAE,QAAQ,EAAE,IAAI,EAAE;6BAChD;yBACF;qBACF;oBACD,QAAQ,EAAE;wBACR,IAAI,EAAE,OAAO;wBACb,QAAQ,EAAE,IAAI;wBACd,KAAK,EAAE;4BACL,IAAI,EAAE,QAAQ;4BACd,oBAAoB,EAAE,KAAK;4BAC3B,UAAU,EAAE;gCACV,KAAK,EAAE,EAAE,IAAI,EAAE,QAAQ,EAAE,QAAQ,EAAE,IAAI,EAAE;gCACzC,KAAK,EAAE,EAAE,IAAI,EAAE,QAAQ,EAAE,QAAQ,EAAE,IAAI,EAAE;6BAC1C;yBACF;qBACF;iBACF;aACF;YACD,MAAM,EAAE,CAAC,KAAK,EAAE,KAAK,EAAE,EAAE,CAAC,CAAC,EAAE,IAAI,EAAE,MAAM,EAAE,IAAI,EAAE,SAAS,CAAC,KAAK,CAAC,EAAE,CAAC;YACpE,0EAA0E;YAC1E,wEAAwE;YACxE,2CAA2C;YAC3C,gBAAgB,EAAE,CAAC,KAAK,EAAE,KAAK,EAAE,EAAE,CAAC,CAAC;gBACnC,SAAS,EAAE,KAAK,CAAC,SAAS,CAAC,GAAG,CAAC,QAAQ,CAAC,EAAE,CAAC,CAAC;oBAC1C,MAAM,EAAE,QAAQ,CAAC,MAAM;oBACvB,KAAK,EAAE,QAAQ,CAAC,KAAK;oBACrB,UAAU,EAAE,QAAQ,CAAC,UAAU;oBAC/B,KAAK,EAAE,QAAQ,CAAC,KAAK;oBACrB,QAAQ,EAAE,QAAQ,CAAC,QAAQ;oBAC3B,MAAM,EAAE,QAAQ,CAAC,MAAM;iBACxB,CAAC,CAAC;aACJ,CAAC;SACH;QACD,OAAO,CAAC,IAAI,EAAE,IAAI;YAChB,2EAA2E;YAC3E,yEAAyE;YACzE,yEAAyE;YACzE,mEAAmE;YACnE,0EAA0E;YAC1E,0CAA0C;YAC1C,MAAM,OAAO,GAAG,IAAI,CAAC,KAAK,KAAK,SAAS,CAAC,CAAC,CAAC,SAAS,CAAC,CAAC,CAAC,EAAE,OAAO,EAAE,IAAI,CAAC,KAAK,CAAC,OAAO,EAAE,CAAA;YACtF,MAAM,KAAK,GAAG,SAAS,CAAC,GAAG,EAAE,OAAO,CAAC,CAAA;YACrC,OAAO,SAAS,CACd,GAAG,CAAC,GAAG,CAAC,IAAI,CAAC,EACb,IAAI,EACJ,MAAM,EACN,KAAK,EAAE,aAAa,IAAI,IAAI,CAAC,KAAK,EAAE,OAAO,EAAE,MAAM,EAAE,GAAG,IAAI,OAAO,CAAC,GAAG,EAAE,EACzE,IAAI,CAAC,MAAM,EACX,KAAK,CACN,CAAA;QACH,CAAC;QACD,WAAW;KACZ,CAAC,CAAC,CAAA;AACL,CAAC"}

package/lib/levels.d.ts

@@ -0,0 +1,24 @@
+/**
+ * The `levels` request value: heading-level corrections a caller makes after
+ * reading the outline a recognition saved.
+ *
+ * The value is compact because the model writes it: `"h12=2,h15=0,h20=3"`,
+ * where the id is an outline candidate and the number is the Markdown heading
+ * depth it should take. Zero, or anything below it, takes the line out of the
+ * heading hierarchy and leaves it as body text, which is what makes the pass
+ * able to demote a line the merge promoted. Anything else is rejected rather
+ * than half-applied.
+ * @module @deepseek-ai/dsh-ab-ocr/levels
+ */
+/** Candidate ids mapped to the heading level each should take. */
+export type LevelOverrides = Record<string, number>;
+/**
+ * Parse the corrections one `levels` value carries.
+ * @param value - the model-supplied value.
+ * @returns the candidate id to heading level corrections.
+ * @throws Error when the value is empty, when a piece is not `<id>=<level>`,
+ *   when the level is not a whole number, or when two pieces give one
+ *   candidate different levels.
+ */
+export declare function parseLevels(value: string): LevelOverrides;
+//# sourceMappingURL=levels.d.ts.map

package/lib/levels.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"levels.d.ts","sourceRoot":"","sources":["../src/levels.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;GAWG;AAEH,kEAAkE;AAClE,MAAM,MAAM,cAAc,GAAG,MAAM,CAAC,MAAM,EAAE,MAAM,CAAC,CAAA;AAQnD;;;;;;;GAOG;AACH,wBAAgB,WAAW,CAAC,KAAK,EAAE,MAAM,GAAG,cAAc,CA8BzD"}

package/lib/levels.js

@@ -0,0 +1,52 @@
+/**
+ * The `levels` request value: heading-level corrections a caller makes after
+ * reading the outline a recognition saved.
+ *
+ * The value is compact because the model writes it: `"h12=2,h15=0,h20=3"`,
+ * where the id is an outline candidate and the number is the Markdown heading
+ * depth it should take. Zero, or anything below it, takes the line out of the
+ * heading hierarchy and leaves it as body text, which is what makes the pass
+ * able to demote a line the merge promoted. Anything else is rejected rather
+ * than half-applied.
+ * @module @deepseek-ai/dsh-ab-ocr/levels
+ */
+/** One correction: an outline candidate id, then the level it should take. */
+const CORRECTION = /^h([1-9][0-9]*)=(.+)$/;
+/** A level the worker accepts: a whole number, 0 or below meaning body text. */
+const LEVEL = /^-?[0-9]+$/;
+/**
+ * Parse the corrections one `levels` value carries.
+ * @param value - the model-supplied value.
+ * @returns the candidate id to heading level corrections.
+ * @throws Error when the value is empty, when a piece is not `<id>=<level>`,
+ *   when the level is not a whole number, or when two pieces give one
+ *   candidate different levels.
+ */
+export function parseLevels(value) {
+    const trimmed = value.trim();
+    if (trimmed === '') {
+        throw new Error('invalid ocr: levels is empty; write corrections such as "h12=2,h15=0"');
+    }
+    const overrides = {};
+    for (const piece of trimmed.split(',')) {
+        const entry = piece.trim();
+        const match = CORRECTION.exec(entry);
+        if (match === null) {
+            throw new Error('invalid ocr: levels cannot read "' + entry + '"; write an outline id and its level such as "h12=2"');
+        }
+        const id = 'h' + (match[1] ?? '');
+        const raw = match[2] ?? '';
+        if (!LEVEL.test(raw)) {
+            throw new Error('invalid ocr: levels gives "' + raw + '" to ' + id
+                + '; a level is a whole number, 0 for a line that is not a heading and 1 to 6 for a heading');
+        }
+        const level = Number(raw);
+        const previous = overrides[id];
+        if (previous !== undefined && previous !== level) {
+            throw new Error('invalid ocr: levels gives ' + id + ' both ' + String(previous) + ' and ' + String(level));
+        }
+        overrides[id] = level;
+    }
+    return overrides;
+}
+//# sourceMappingURL=levels.js.map

package/lib/levels.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"levels.js","sourceRoot":"","sources":["../src/levels.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;GAWG;AAKH,8EAA8E;AAC9E,MAAM,UAAU,GAAG,uBAAuB,CAAA;AAE1C,gFAAgF;AAChF,MAAM,KAAK,GAAG,YAAY,CAAA;AAE1B;;;;;;;GAOG;AACH,MAAM,UAAU,WAAW,CAAC,KAAa;IACvC,MAAM,OAAO,GAAG,KAAK,CAAC,IAAI,EAAE,CAAA;IAC5B,IAAI,OAAO,KAAK,EAAE,EAAE,CAAC;QACnB,MAAM,IAAI,KAAK,CAAC,uEAAuE,CAAC,CAAA;IAC1F,CAAC;IACD,MAAM,SAAS,GAAmB,EAAE,CAAA;IACpC,KAAK,MAAM,KAAK,IAAI,OAAO,CAAC,KAAK,CAAC,GAAG,CAAC,EAAE,CAAC;QACvC,MAAM,KAAK,GAAG,KAAK,CAAC,IAAI,EAAE,CAAA;QAC1B,MAAM,KAAK,GAAG,UAAU,CAAC,IAAI,CAAC,KAAK,CAAC,CAAA;QACpC,IAAI,KAAK,KAAK,IAAI,EAAE,CAAC;YACnB,MAAM,IAAI,KAAK,CACb,mCAAmC,GAAG,KAAK,GAAG,sDAAsD,CACrG,CAAA;QACH,CAAC;QACD,MAAM,EAAE,GAAG,GAAG,GAAG,CAAC,KAAK,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC,CAAA;QACjC,MAAM,GAAG,GAAG,KAAK,CAAC,CAAC,CAAC,IAAI,EAAE,CAAA;QAC1B,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC;YACrB,MAAM,IAAI,KAAK,CACb,6BAA6B,GAAG,GAAG,GAAG,OAAO,GAAG,EAAE;kBAChD,0FAA0F,CAC7F,CAAA;QACH,CAAC;QACD,MAAM,KAAK,GAAG,MAAM,CAAC,GAAG,CAAC,CAAA;QACzB,MAAM,QAAQ,GAAG,SAAS,CAAC,EAAE,CAAC,CAAA;QAC9B,IAAI,QAAQ,KAAK,SAAS,IAAI,QAAQ,KAAK,KAAK,EAAE,CAAC;YACjD,MAAM,IAAI,KAAK,CAAC,4BAA4B,GAAG,EAAE,GAAG,QAAQ,GAAG,MAAM,CAAC,QAAQ,CAAC,GAAG,OAAO,GAAG,MAAM,CAAC,KAAK,CAAC,CAAC,CAAA;QAC5G,CAAC;QACD,SAAS,CAAC,EAAE,CAAC,GAAG,KAAK,CAAA;IACvB,CAAC;IACD,OAAO,SAAS,CAAA;AAClB,CAAC"}

package/lib/plan.d.ts

@@ -0,0 +1,103 @@
+/**
+ * Request planning: which files one call recognizes, where each document's
+ * artifacts go, and what the worker is told about each one.
+ *
+ * Everything here is decided before a worker process starts, so a call that
+ * names a file the worker cannot read is reported as a failure without paying
+ * for a process. The one exception is the source's own content digest, which
+ * only the worker can compute; {@link prepareJob} therefore resolves the paths
+ * that do not depend on a recognition, and the entry module names the rest when
+ * the document reports it started.
+ * @module @deepseek-ai/dsh-ab-ocr/plan
+ */
+import type FileSystem from '@deepseek-ai/dsh-fs';
+import { type ArtifactRoot, type SourceStamp } from './artifacts.ts';
+import type { Config } from './config.ts';
+import type { OcrArgs, OcrFailure } from './types.ts';
+import type { WorkerJob } from './worker.ts';
+/** Extensions the worker accepts as a still image or a PDF. */
+export declare const SUPPORTED_EXTENSIONS: readonly string[];
+/**
+ * Resolve the Python executable that carries the OCR dependencies.
+ * @param configured - the row's pythonPath value.
+ * @param packageRoot - this package's directory.
+ * @param platform - the platform the plugin is running on.
+ * @returns the configured path, the package's own environment, or the platform's interpreter name.
+ */
+export declare function resolvePython(configured: string, packageRoot: string, platform: string): string;
+/**
+ * Resolve the worker script this package runs.
+ * @param configured - the row's workerScript value.
+ * @param packageRoot - this package's directory.
+ * @returns the configured path, or the script shipped inside the package.
+ */
+export declare function resolveWorkerScript(configured: string, packageRoot: string): string;
+/**
+ * Resolve a deployment directory value against a base.
+ * @param configured - the configured or requested directory.
+ * @param base - the directory a relative value resolves against.
+ * @returns the absolute directory, or an empty string when the value is empty.
+ */
+export declare function resolveDirectory(configured: string, base: string): string;
+/**
+ * Collect the paths one call named, in request order and without duplicates.
+ * @param args - the model-supplied call arguments.
+ * @param base - the directory a relative path resolves against.
+ * @param maxDocuments - documents one call may name.
+ * @returns the absolute paths to recognize.
+ * @throws Error when the call names no path, or more than the ceiling.
+ */
+export declare function requestPaths(args: OcrArgs, base: string, maxDocuments: number): string[];
+/**
+ * Report why a path cannot be recognized, before any process is started.
+ * @param path - the absolute path to check.
+ * @returns the reason, or undefined when the worker can read the path.
+ */
+export declare function rejectPath(path: string): string | undefined;
+/**
+ * Resolve the directory one call's artifacts are written to: the call's own
+ * request when it made one, otherwise the deployment's configured directory.
+ * @param args - the model-supplied call arguments.
+ * @param config - the deployment's values.
+ * @param base - the directory a relative value resolves against.
+ * @returns the call's output directory, or an empty string for the source's own.
+ */
+export declare function requestedOutputDir(args: OcrArgs, config: Config, base: string): string;
+/** One planned job: the worker spec entry and the source it reads. */
+export interface PlannedJob {
+    /** The worker spec entry. */
+    job: WorkerJob;
+    /** The absolute path the model supplied. */
+    input: string;
+    /** Directory the document's artifacts are written under. */
+    base: string;
+}
+/** One planned job with the artifact paths that do not depend on a recognition. */
+export interface PreparedJob extends PlannedJob {
+    /** Every path that is fixed before the worker names the source's digest. */
+    root: ArtifactRoot;
+    /** What the digest covered about the source file, beyond its content. */
+    stamp: SourceStamp;
+}
+/**
+ * Build the worker spec and the failures that never reach the worker.
+ * @param paths - the absolute paths one call named.
+ * @param args - the model-supplied call arguments.
+ * @param config - the deployment's values.
+ * @param base - the directory a relative output directory resolves against.
+ * @returns the jobs to run and the paths rejected before running.
+ */
+export declare function planJobs(paths: readonly string[], args: OcrArgs, config: Config, base: string): {
+    jobs: PlannedJob[];
+    failures: OcrFailure[];
+};
+/**
+ * Resolve the digest and the artifact paths of one planned job.
+ * @param fs - the mounted filesystem capability.
+ * @param planned - the job {@link planJobs} produced.
+ * @param pageDirName - the page directory's name under the base directory.
+ * @param signal - caller cancellation.
+ * @returns the job with its artifact paths and source metadata.
+ */
+export declare function prepareJob(fs: FileSystem, planned: PlannedJob, pageDirName: string, signal?: AbortSignal): Promise<PreparedJob>;
+//# sourceMappingURL=plan.d.ts.map

package/lib/plan.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"plan.d.ts","sourceRoot":"","sources":["../src/plan.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;GAWG;AAIH,OAAO,KAAK,UAAU,MAAM,qBAAqB,CAAA;AAEjD,OAAO,EAA8B,KAAK,YAAY,EAAE,KAAK,WAAW,EAAE,MAAM,gBAAgB,CAAA;AAChG,OAAO,KAAK,EAAE,MAAM,EAAE,MAAM,aAAa,CAAA;AACzC,OAAO,KAAK,EAAgB,OAAO,EAAE,UAAU,EAAE,MAAM,YAAY,CAAA;AACnE,OAAO,KAAK,EAAE,SAAS,EAAE,MAAM,aAAa,CAAA;AAE5C,+DAA+D;AAC/D,eAAO,MAAM,oBAAoB,EAAE,SAAS,MAAM,EAGjD,CAAA;AAED;;;;;;GAMG;AACH,wBAAgB,aAAa,CAAC,UAAU,EAAE,MAAM,EAAE,WAAW,EAAE,MAAM,EAAE,QAAQ,EAAE,MAAM,GAAG,MAAM,CAM/F;AAED;;;;;GAKG;AACH,wBAAgB,mBAAmB,CAAC,UAAU,EAAE,MAAM,EAAE,WAAW,EAAE,MAAM,GAAG,MAAM,CAEnF;AAED;;;;;GAKG;AACH,wBAAgB,gBAAgB,CAAC,UAAU,EAAE,MAAM,EAAE,IAAI,EAAE,MAAM,GAAG,MAAM,CAGzE;AAED;;;;;;;GAOG;AACH,wBAAgB,YAAY,CAAC,IAAI,EAAE,OAAO,EAAE,IAAI,EAAE,MAAM,EAAE,YAAY,EAAE,MAAM,GAAG,MAAM,EAAE,CAcxF;AAED;;;;GAIG;AACH,wBAAgB,UAAU,CAAC,IAAI,EAAE,MAAM,GAAG,MAAM,GAAG,SAAS,CAO3D;AAsBD;;;;;;;GAOG;AACH,wBAAgB,kBAAkB,CAAC,IAAI,EAAE,OAAO,EAAE,MAAM,EAAE,MAAM,EAAE,IAAI,EAAE,MAAM,GAAG,MAAM,CAGtF;AAED,sEAAsE;AACtE,MAAM,WAAW,UAAU;IACzB,6BAA6B;IAC7B,GAAG,EAAE,SAAS,CAAA;IACd,4CAA4C;IAC5C,KAAK,EAAE,MAAM,CAAA;IACb,4DAA4D;IAC5D,IAAI,EAAE,MAAM,CAAA;CACb;AAED,mFAAmF;AACnF,MAAM,WAAW,WAAY,SAAQ,UAAU;IAC7C,4EAA4E;IAC5E,IAAI,EAAE,YAAY,CAAA;IAClB,yEAAyE;IACzE,KAAK,EAAE,WAAW,CAAA;CACnB;AAED;;;;;;;GAOG;AACH,wBAAgB,QAAQ,CACtB,KAAK,EAAE,SAAS,MAAM,EAAE,EACxB,IAAI,EAAE,OAAO,EACb,MAAM,EAAE,MAAM,EACd,IAAI,EAAE,MAAM,GACX;IAAE,IAAI,EAAE,UAAU,EAAE,CAAC;IAAC,QAAQ,EAAE,UAAU,EAAE,CAAA;CAAE,CA+BhD;AA4BD;;;;;;;GAOG;AACH,wBAAsB,UAAU,CAC9B,EAAE,EAAE,UAAU,EACd,OAAO,EAAE,UAAU,EACnB,WAAW,EAAE,MAAM,EACnB,MAAM,CAAC,EAAE,WAAW,GACnB,OAAO,CAAC,WAAW,CAAC,CAOtB"}

package/lib/plan.js

@@ -0,0 +1,210 @@
+/**
+ * Request planning: which files one call recognizes, where each document's
+ * artifacts go, and what the worker is told about each one.
+ *
+ * Everything here is decided before a worker process starts, so a call that
+ * names a file the worker cannot read is reported as a failure without paying
+ * for a process. The one exception is the source's own content digest, which
+ * only the worker can compute; {@link prepareJob} therefore resolves the paths
+ * that do not depend on a recognition, and the entry module names the rest when
+ * the document reports it started.
+ * @module @deepseek-ai/dsh-ab-ocr/plan
+ */
+import { existsSync } from 'node:fs';
+import { isAbsolute, join, resolve } from 'node:path';
+import { artifactBase, artifactRoot } from "./artifacts.js";
+/** Extensions the worker accepts as a still image or a PDF. */
+export const SUPPORTED_EXTENSIONS = [
+    '.pdf', '.png', '.jpg', '.jpeg', '.jpe', '.bmp', '.dib', '.tif', '.tiff',
+    '.webp', '.gif', '.jp2', '.pbm', '.pgm', '.ppm', '.tga', '.ico',
+];
+/**
+ * Resolve the Python executable that carries the OCR dependencies.
+ * @param configured - the row's pythonPath value.
+ * @param packageRoot - this package's directory.
+ * @param platform - the platform the plugin is running on.
+ * @returns the configured path, the package's own environment, or the platform's interpreter name.
+ */
+export function resolvePython(configured, packageRoot, platform) {
+    if (configured !== '')
+        return configured;
+    const bundled = platform === 'win32'
+        ? join(packageRoot, 'python', '.venv', 'Scripts', 'python.exe')
+        : join(packageRoot, 'python', '.venv', 'bin', 'python');
+    return existsSync(bundled) ? bundled : (platform === 'win32' ? 'python' : 'python3');
+}
+/**
+ * Resolve the worker script this package runs.
+ * @param configured - the row's workerScript value.
+ * @param packageRoot - this package's directory.
+ * @returns the configured path, or the script shipped inside the package.
+ */
+export function resolveWorkerScript(configured, packageRoot) {
+    return configured !== '' ? configured : join(packageRoot, 'python', 'ocr_worker.py');
+}
+/**
+ * Resolve a deployment directory value against a base.
+ * @param configured - the configured or requested directory.
+ * @param base - the directory a relative value resolves against.
+ * @returns the absolute directory, or an empty string when the value is empty.
+ */
+export function resolveDirectory(configured, base) {
+    if (configured === '')
+        return '';
+    return isAbsolute(configured) ? configured : resolve(base, configured);
+}
+/**
+ * Collect the paths one call named, in request order and without duplicates.
+ * @param args - the model-supplied call arguments.
+ * @param base - the directory a relative path resolves against.
+ * @param maxDocuments - documents one call may name.
+ * @returns the absolute paths to recognize.
+ * @throws Error when the call names no path, or more than the ceiling.
+ */
+export function requestPaths(args, base, maxDocuments) {
+    const named = [...(args.path === undefined ? [] : [args.path]), ...(args.paths ?? [])];
+    if (named.length === 0)
+        throw new Error('invalid ocr: give path or paths with at least one file');
+    const paths = [];
+    for (const raw of named) {
+        const trimmed = raw.trim();
+        if (trimmed === '')
+            throw new Error('invalid ocr: path must be a non-empty string');
+        const absolute = isAbsolute(trimmed) ? trimmed : resolve(base, trimmed);
+        if (!paths.includes(absolute))
+            paths.push(absolute);
+    }
+    if (paths.length > maxDocuments) {
+        throw new Error('invalid ocr: one call recognizes at most ' + maxDocuments + ' files');
+    }
+    return paths;
+}
+/**
+ * Report why a path cannot be recognized, before any process is started.
+ * @param path - the absolute path to check.
+ * @returns the reason, or undefined when the worker can read the path.
+ */
+export function rejectPath(path) {
+    if (!existsSync(path))
+        return 'no such file: ' + path;
+    const lower = path.toLowerCase();
+    if (!SUPPORTED_EXTENSIONS.some(extension => lower.endsWith(extension))) {
+        return 'unsupported file type: neither a PDF nor a supported image';
+    }
+    return undefined;
+}
+/**
+ * Collect the merge switches and thresholds one call sends.
+ * @param config - the deployment's values.
+ * @returns the merge options, in the names the worker reads them under.
+ */
+function mergeOptions(config) {
+    return {
+        removePageNumbers: config.removePageNumbers,
+        removeRunningHeads: config.removeRunningHeads,
+        runningHeadRatio: config.runningHeadRatio,
+        runningHeadMinPages: config.runningHeadMinPages,
+        detectHeadings: config.detectHeadings,
+        headingMinRatio: config.headingMinRatio,
+        indentRatio: config.indentRatio,
+        paragraphGapRatio: config.paragraphGapRatio,
+        outlineCandidateRatio: config.outlineCandidateRatio,
+        maxOutlineCandidates: config.maxOutlineCandidates,
+    };
+}
+/**
+ * Resolve the directory one call's artifacts are written to: the call's own
+ * request when it made one, otherwise the deployment's configured directory.
+ * @param args - the model-supplied call arguments.
+ * @param config - the deployment's values.
+ * @param base - the directory a relative value resolves against.
+ * @returns the call's output directory, or an empty string for the source's own.
+ */
+export function requestedOutputDir(args, config, base) {
+    const requested = resolveDirectory(args.outputDir ?? '', base);
+    return requested !== '' ? requested : resolveDirectory(config.outputDir, base);
+}
+/**
+ * Build the worker spec and the failures that never reach the worker.
+ * @param paths - the absolute paths one call named.
+ * @param args - the model-supplied call arguments.
+ * @param config - the deployment's values.
+ * @param base - the directory a relative output directory resolves against.
+ * @returns the jobs to run and the paths rejected before running.
+ */
+export function planJobs(paths, args, config, base) {
+    const outputDir = requestedOutputDir(args, config, base);
+    const jobs = [];
+    const failures = [];
+    for (const input of paths) {
+        const rejection = rejectPath(input);
+        if (rejection !== undefined) {
+            failures.push({ input, error: rejection });
+            continue;
+        }
+        jobs.push({
+            input,
+            // The document and the pages beside it share one base directory, so two
+            // sources never share a page directory and a source with no configured
+            // output directory keeps its artifacts next to itself.
+            base: artifactBase(input, outputDir),
+            job: {
+                mode: 'recognize',
+                id: String(jobs.length + 1),
+                input,
+                pages: args.pages ?? null,
+                dpi: config.dpi,
+                maxPixels: config.maxPixels,
+                textScore: config.textScore,
+                detectColumns: config.detectColumns,
+                maxPages: config.maxPages,
+                merge: mergeOptions(config),
+            },
+        });
+    }
+    return { jobs, failures };
+}
+/**
+ * Read the metadata the digest covers for one source file.
+ * @param fs - the mounted filesystem capability.
+ * @param target - the resolved source target.
+ * @param signal - caller cancellation.
+ * @returns the size and modification time the filesystem reported, or the
+ *   documented fallbacks when it reported none or refused the read.
+ */
+async function sourceStamp(fs, target, signal) {
+    try {
+        const info = await fs.stat(target, signal);
+        if (info === undefined)
+            return { size: '', mtimeMs: 0, unavailable: true };
+        const reported = info;
+        return {
+            size: typeof reported.size === 'number' ? reported.size : '',
+            mtimeMs: typeof reported.mtimeMs === 'number' ? reported.mtimeMs : 0,
+            unavailable: false,
+        };
+    }
+    catch {
+        // A backend that refuses this metadata read still lets the document be
+        // recognized; the digest falls back to the empty stamp and the record says
+        // the metadata was unavailable.
+        return { size: '', mtimeMs: 0, unavailable: true };
+    }
+}
+/**
+ * Resolve the digest and the artifact paths of one planned job.
+ * @param fs - the mounted filesystem capability.
+ * @param planned - the job {@link planJobs} produced.
+ * @param pageDirName - the page directory's name under the base directory.
+ * @param signal - caller cancellation.
+ * @returns the job with its artifact paths and source metadata.
+ */
+export async function prepareJob(fs, planned, pageDirName, signal) {
+    const target = await fs.resolve(planned.input);
+    const stamp = await sourceStamp(fs, target, signal);
+    // The worker reads the source in its own execution world, so the job carries
+    // the path that world can open rather than the one the model named.
+    planned.job.input = fs.processPath(target);
+    return { ...planned, root: artifactRoot(planned.input, planned.base, pageDirName), stamp };
+}
+//# sourceMappingURL=plan.js.map

package/lib/plan.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"plan.js","sourceRoot":"","sources":["../src/plan.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;GAWG;AAEH,OAAO,EAAE,UAAU,EAAE,MAAM,SAAS,CAAA;AACpC,OAAO,EAAE,UAAU,EAAE,IAAI,EAAE,OAAO,EAAE,MAAM,WAAW,CAAA;AAGrD,OAAO,EAAE,YAAY,EAAE,YAAY,EAAuC,MAAM,gBAAgB,CAAA;AAKhG,+DAA+D;AAC/D,MAAM,CAAC,MAAM,oBAAoB,GAAsB;IACrD,MAAM,EAAE,MAAM,EAAE,MAAM,EAAE,OAAO,EAAE,MAAM,EAAE,MAAM,EAAE,MAAM,EAAE,MAAM,EAAE,OAAO;IACxE,OAAO,EAAE,MAAM,EAAE,MAAM,EAAE,MAAM,EAAE,MAAM,EAAE,MAAM,EAAE,MAAM,EAAE,MAAM;CAChE,CAAA;AAED;;;;;;GAMG;AACH,MAAM,UAAU,aAAa,CAAC,UAAkB,EAAE,WAAmB,EAAE,QAAgB;IACrF,IAAI,UAAU,KAAK,EAAE;QAAE,OAAO,UAAU,CAAA;IACxC,MAAM,OAAO,GAAG,QAAQ,KAAK,OAAO;QAClC,CAAC,CAAC,IAAI,CAAC,WAAW,EAAE,QAAQ,EAAE,OAAO,EAAE,SAAS,EAAE,YAAY,CAAC;QAC/D,CAAC,CAAC,IAAI,CAAC,WAAW,EAAE,QAAQ,EAAE,OAAO,EAAE,KAAK,EAAE,QAAQ,CAAC,CAAA;IACzD,OAAO,UAAU,CAAC,OAAO,CAAC,CAAC,CAAC,CAAC,OAAO,CAAC,CAAC,CAAC,CAAC,QAAQ,KAAK,OAAO,CAAC,CAAC,CAAC,QAAQ,CAAC,CAAC,CAAC,SAAS,CAAC,CAAA;AACtF,CAAC;AAED;;;;;GAKG;AACH,MAAM,UAAU,mBAAmB,CAAC,UAAkB,EAAE,WAAmB;IACzE,OAAO,UAAU,KAAK,EAAE,CAAC,CAAC,CAAC,UAAU,CAAC,CAAC,CAAC,IAAI,CAAC,WAAW,EAAE,QAAQ,EAAE,eAAe,CAAC,CAAA;AACtF,CAAC;AAED;;;;;GAKG;AACH,MAAM,UAAU,gBAAgB,CAAC,UAAkB,EAAE,IAAY;IAC/D,IAAI,UAAU,KAAK,EAAE;QAAE,OAAO,EAAE,CAAA;IAChC,OAAO,UAAU,CAAC,UAAU,CAAC,CAAC,CAAC,CAAC,UAAU,CAAC,CAAC,CAAC,OAAO,CAAC,IAAI,EAAE,UAAU,CAAC,CAAA;AACxE,CAAC;AAED;;;;;;;GAOG;AACH,MAAM,UAAU,YAAY,CAAC,IAAa,EAAE,IAAY,EAAE,YAAoB;IAC5E,MAAM,KAAK,GAAG,CAAC,GAAG,CAAC,IAAI,CAAC,IAAI,KAAK,SAAS,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,EAAE,GAAG,CAAC,IAAI,CAAC,KAAK,IAAI,EAAE,CAAC,CAAC,CAAA;IACtF,IAAI,KAAK,CAAC,MAAM,KAAK,CAAC;QAAE,MAAM,IAAI,KAAK,CAAC,wDAAwD,CAAC,CAAA;IACjG,MAAM,KAAK,GAAa,EAAE,CAAA;IAC1B,KAAK,MAAM,GAAG,IAAI,KAAK,EAAE,CAAC;QACxB,MAAM,OAAO,GAAG,GAAG,CAAC,IAAI,EAAE,CAAA;QAC1B,IAAI,OAAO,KAAK,EAAE;YAAE,MAAM,IAAI,KAAK,CAAC,8CAA8C,CAAC,CAAA;QACnF,MAAM,QAAQ,GAAG,UAAU,CAAC,OAAO,CAAC,CAAC,CAAC,CAAC,OAAO,CAAC,CAAC,CAAC,OAAO,CAAC,IAAI,EAAE,OAAO,CAAC,CAAA;QACvE,IAAI,CAAC,KAAK,CAAC,QAAQ,CAAC,QAAQ,CAAC;YAAE,KAAK,CAAC,IAAI,CAAC,QAAQ,CAAC,CAAA;IACrD,CAAC;IACD,IAAI,KAAK,CAAC,MAAM,GAAG,YAAY,EAAE,CAAC;QAChC,MAAM,IAAI,KAAK,CAAC,2CAA2C,GAAG,YAAY,GAAG,QAAQ,CAAC,CAAA;IACxF,CAAC;IACD,OAAO,KAAK,CAAA;AACd,CAAC;AAED;;;;GAIG;AACH,MAAM,UAAU,UAAU,CAAC,IAAY;IACrC,IAAI,CAAC,UAAU,CAAC,IAAI,CAAC;QAAE,OAAO,gBAAgB,GAAG,IAAI,CAAA;IACrD,MAAM,KAAK,GAAG,IAAI,CAAC,WAAW,EAAE,CAAA;IAChC,IAAI,CAAC,oBAAoB,CAAC,IAAI,CAAC,SAAS,CAAC,EAAE,CAAC,KAAK,CAAC,QAAQ,CAAC,SAAS,CAAC,CAAC,EAAE,CAAC;QACvE,OAAO,4DAA4D,CAAA;IACrE,CAAC;IACD,OAAO,SAAS,CAAA;AAClB,CAAC;AAED;;;;GAIG;AACH,SAAS,YAAY,CAAC,MAAc;IAClC,OAAO;QACL,iBAAiB,EAAE,MAAM,CAAC,iBAAiB;QAC3C,kBAAkB,EAAE,MAAM,CAAC,kBAAkB;QAC7C,gBAAgB,EAAE,MAAM,CAAC,gBAAgB;QACzC,mBAAmB,EAAE,MAAM,CAAC,mBAAmB;QAC/C,cAAc,EAAE,MAAM,CAAC,cAAc;QACrC,eAAe,EAAE,MAAM,CAAC,eAAe;QACvC,WAAW,EAAE,MAAM,CAAC,WAAW;QAC/B,iBAAiB,EAAE,MAAM,CAAC,iBAAiB;QAC3C,qBAAqB,EAAE,MAAM,CAAC,qBAAqB;QACnD,oBAAoB,EAAE,MAAM,CAAC,oBAAoB;KAClD,CAAA;AACH,CAAC;AAED;;;;;;;GAOG;AACH,MAAM,UAAU,kBAAkB,CAAC,IAAa,EAAE,MAAc,EAAE,IAAY;IAC5E,MAAM,SAAS,GAAG,gBAAgB,CAAC,IAAI,CAAC,SAAS,IAAI,EAAE,EAAE,IAAI,CAAC,CAAA;IAC9D,OAAO,SAAS,KAAK,EAAE,CAAC,CAAC,CAAC,SAAS,CAAC,CAAC,CAAC,gBAAgB,CAAC,MAAM,CAAC,SAAS,EAAE,IAAI,CAAC,CAAA;AAChF,CAAC;AAoBD;;;;;;;GAOG;AACH,MAAM,UAAU,QAAQ,CACtB,KAAwB,EACxB,IAAa,EACb,MAAc,EACd,IAAY;IAEZ,MAAM,SAAS,GAAG,kBAAkB,CAAC,IAAI,EAAE,MAAM,EAAE,IAAI,CAAC,CAAA;IACxD,MAAM,IAAI,GAAiB,EAAE,CAAA;IAC7B,MAAM,QAAQ,GAAiB,EAAE,CAAA;IACjC,KAAK,MAAM,KAAK,IAAI,KAAK,EAAE,CAAC;QAC1B,MAAM,SAAS,GAAG,UAAU,CAAC,KAAK,CAAC,CAAA;QACnC,IAAI,SAAS,KAAK,SAAS,EAAE,CAAC;YAC5B,QAAQ,CAAC,IAAI,CAAC,EAAE,KAAK,EAAE,KAAK,EAAE,SAAS,EAAE,CAAC,CAAA;YAC1C,SAAQ;QACV,CAAC;QACD,IAAI,CAAC,IAAI,CAAC;YACR,KAAK;YACL,wEAAwE;YACxE,uEAAuE;YACvE,uDAAuD;YACvD,IAAI,EAAE,YAAY,CAAC,KAAK,EAAE,SAAS,CAAC;YACpC,GAAG,EAAE;gBACH,IAAI,EAAE,WAAW;gBACjB,EAAE,EAAE,MAAM,CAAC,IAAI,CAAC,MAAM,GAAG,CAAC,CAAC;gBAC3B,KAAK;gBACL,KAAK,EAAE,IAAI,CAAC,KAAK,IAAI,IAAI;gBACzB,GAAG,EAAE,MAAM,CAAC,GAAG;gBACf,SAAS,EAAE,MAAM,CAAC,SAAS;gBAC3B,SAAS,EAAE,MAAM,CAAC,SAAS;gBAC3B,aAAa,EAAE,MAAM,CAAC,aAAa;gBACnC,QAAQ,EAAE,MAAM,CAAC,QAAQ;gBACzB,KAAK,EAAE,YAAY,CAAC,MAAM,CAAC;aAC5B;SACF,CAAC,CAAA;IACJ,CAAC;IACD,OAAO,EAAE,IAAI,EAAE,QAAQ,EAAE,CAAA;AAC3B,CAAC;AAED;;;;;;;GAOG;AACH,KAAK,UAAU,WAAW,CAAC,EAAc,EAAE,MAAgB,EAAE,MAAoB;IAC/E,IAAI,CAAC;QACH,MAAM,IAAI,GAAG,MAAM,EAAE,CAAC,IAAI,CAAC,MAAM,EAAE,MAAM,CAAC,CAAA;QAC1C,IAAI,IAAI,KAAK,SAAS;YAAE,OAAO,EAAE,IAAI,EAAE,EAAE,EAAE,OAAO,EAAE,CAAC,EAAE,WAAW,EAAE,IAAI,EAAE,CAAA;QAC1E,MAAM,QAAQ,GAAG,IAA2C,CAAA;QAC5D,OAAO;YACL,IAAI,EAAE,OAAO,QAAQ,CAAC,IAAI,KAAK,QAAQ,CAAC,CAAC,CAAC,QAAQ,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE;YAC5D,OAAO,EAAE,OAAO,QAAQ,CAAC,OAAO,KAAK,QAAQ,CAAC,CAAC,CAAC,QAAQ,CAAC,OAAO,CAAC,CAAC,CAAC,CAAC;YACpE,WAAW,EAAE,KAAK;SACnB,CAAA;IACH,CAAC;IAAC,MAAM,CAAC;QACP,uEAAuE;QACvE,2EAA2E;QAC3E,gCAAgC;QAChC,OAAO,EAAE,IAAI,EAAE,EAAE,EAAE,OAAO,EAAE,CAAC,EAAE,WAAW,EAAE,IAAI,EAAE,CAAA;IACpD,CAAC;AACH,CAAC;AAED;;;;;;;GAOG;AACH,MAAM,CAAC,KAAK,UAAU,UAAU,CAC9B,EAAc,EACd,OAAmB,EACnB,WAAmB,EACnB,MAAoB;IAEpB,MAAM,MAAM,GAAG,MAAM,EAAE,CAAC,OAAO,CAAC,OAAO,CAAC,KAAK,CAAC,CAAA;IAC9C,MAAM,KAAK,GAAG,MAAM,WAAW,CAAC,EAAE,EAAE,MAAM,EAAE,MAAM,CAAC,CAAA;IACnD,6EAA6E;IAC7E,oEAAoE;IACpE,OAAO,CAAC,GAAG,CAAC,KAAK,GAAG,EAAE,CAAC,WAAW,CAAC,MAAM,CAAC,CAAA;IAC1C,OAAO,EAAE,GAAG,OAAO,EAAE,IAAI,EAAE,YAAY,CAAC,OAAO,CAAC,KAAK,EAAE,OAAO,CAAC,IAAI,EAAE,WAAW,CAAC,EAAE,KAAK,EAAE,CAAA;AAC5F,CAAC"}

package/lib/recognize.d.ts

@@ -0,0 +1,36 @@
+/**
+ * One `ocr` call: refuse an installation that cannot run, drive the worker
+ * through the plan, write each page as it finishes, and merge the document only
+ * after the last page.
+ *
+ * A later call carrying `levels` does not recognize anything: it reads the
+ * record the recognition left behind and sends its stored geometry to the worker
+ * with the caller's heading-level corrections, which rewrites the document at
+ * the path the record named.
+ *
+ * The worker runs in its own execution world and writes only to stdout, so every
+ * value this module persists travels back as an event and is written through the
+ * mounted filesystem capability rather than by the worker.
+ * @module @deepseek-ai/dsh-ab-ocr/recognize
+ */
+import type FileSystem from '@deepseek-ai/dsh-fs';
+import type { Config } from './config.ts';
+import { type CallFence } from './sandbox.ts';
+import type { OcrArgs, OcrResult } from './types.ts';
+/**
+ * Read the documents one call asked for, writing every artifact through the
+ * filesystem capability the composition mounted.
+ * @param fs - the mounted filesystem capability every write goes through, or
+ *   undefined when the composition mounts none.
+ * @param args - the model-supplied call arguments.
+ * @param config - the deployment's values.
+ * @param base - the directory a relative path resolves against.
+ * @param signal - caller cancellation.
+ * @param fence - the fence every write is stamped with; absent on a bare backend.
+ * @returns the canonical value the call records.
+ * @throws Error when the request is malformed, no filesystem is mounted, the
+ *   environment is unusable, the deployment denies a write, the worker reports
+ *   pages out of order, or the worker fails.
+ */
+export declare function recognize(fs: FileSystem | undefined, args: OcrArgs, config: Config, base: string, signal?: AbortSignal, fence?: CallFence | undefined): Promise<OcrResult>;
+//# sourceMappingURL=recognize.d.ts.map

package/lib/recognize.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"recognize.d.ts","sourceRoot":"","sources":["../src/recognize.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;GAcG;AAKH,OAAO,KAAK,UAAU,MAAM,qBAAqB,CAAA;AAEjD,OAAO,KAAK,EAAE,MAAM,EAAE,MAAM,aAAa,CAAA;AAyBzC,OAAO,EAAY,KAAK,SAAS,EAAE,MAAM,cAAc,CAAA;AACvD,OAAO,KAAK,EAAgB,OAAO,EAAoC,SAAS,EAAE,MAAM,YAAY,CAAA;AA6SpG;;;;;;;;;;;;;;GAcG;AACH,wBAAsB,SAAS,CAC7B,EAAE,EAAE,UAAU,GAAG,SAAS,EAC1B,IAAI,EAAE,OAAO,EACb,MAAM,EAAE,MAAM,EACd,IAAI,EAAE,MAAM,EACZ,MAAM,CAAC,EAAE,WAAW,EACpB,KAAK,CAAC,EAAE,SAAS,GAAG,SAAS,GAC5B,OAAO,CAAC,SAAS,CAAC,CAgJpB"}