@eigenpal/cli 0.16.6 → 0.16.8

package/CHANGELOG.md

@@ -1,5 +1,11 @@
 # @eigenpal/cli
 
+## 0.16.7
+
+### Patch Changes
+
+- 35375db: The CLI now uploads macro-enabled Excel workbooks with the correct media type, so original `.xlsm` workflow inputs work without manual conversion.
+
 ## 0.16.4
 
 ### Minor Changes

package/dist/cli.js

@@ -8038,6 +8038,9 @@ function refineLayoutOutputFormat(config, ctx) {
     message: LAYOUT_OUTPUT_PARSE_MODE_ERROR
   });
 }
+function mimeTypeForExtension(extension) {
+  return EXTENSION_TO_MIME[extension.replace(/^\./, "").toLowerCase()];
+}
 var OFFICE_MIME_TYPES, ParserCategory, ParserTypeSchema, ParseModeSchema, PARSE_OUTPUT_FORMATS, ParseOutputFormatSchema, PageProvenanceSourceSchema, PageProvenanceSchema, ProcessingStrategySchema, LAYOUT_OUTPUT_PARSE_MODE_ERROR = 'outputFormat "layout" is native PDF-only. Set parseMode to "native" or "native-or-ocr". Omitted parseMode, OCR, vision, and nativeText: true cannot request layout.', PARSE_OUTPUT_FORMAT_DESCRIPTION = "Format for extracted text. `markdown` (default) keeps structure; `plain` is unstyled text; `djot`/`html` preserve more markup. `layout` is native-PDF spatial text with column gaps preserved as spaces. It requires `parseMode: native` or `parseMode: native-or-ocr` and a PDF — omitted parseMode and `nativeText: true` are not enough. Native mode is layout-only and fail-closed. `native-or-ocr` keeps layout text on accepted native pages and uses OCR markdown on suspect or empty pages, so the document can mix fixed-width layout and markdown. Extractor failure never falls back to another parser. Office, plaintext, and images fail. OCR/vision cannot request `layout`.", DocumentMetadataSchema, ParseUsageSchema, StructuredDocumentSchema, CoordinateUnitSchema, PointSchema, BoundingRegionSchema, PositionedTextSchema, TableCellSchema, ExtractedTableSchema, LayoutElementTypeSchema, LayoutElementSchema, PageResultSchema, ParseResultSchema, ParseOptionsSchema, ParserInputSchema, EXTENSION_TO_MIME, MIME_TO_EXTENSION;
 var init_parser = __esm(() => {
   init_zod();
@@ -8587,7 +8590,7 @@ var init_split = __esm(() => {
     rules: string2().optional().describe("Optional natural-language rules appended to the system prompt"),
     provider: string2().optional().describe('Provider ID from eigenpal.config.yaml (e.g. "openai-gpt5.4-mini"). Falls back to the tenant default LLM provider when omitted.'),
     reasoningEffort: ReasoningEffortSchema.optional().describe("Reasoning effort for the split model. Omit to use the model default."),
-    windowTokenBudget: number2().int().positive().optional().describe("Per-window token ceiling. Defaults to env SPLIT_WINDOW_TOKEN_BUDGET or 20000. Smaller windows give sharper anchors on contract-style documents (less competing context for the LLM to mis-anchor on); 50k–100k when sections routinely exceed per-window page count.")
+    windowTokenBudget: number2().int().positive().optional().describe("Per-window token ceiling, enforced across and within parsed pages while preserving the original page index. Defaults to env SPLIT_WINDOW_TOKEN_BUDGET or 20000. Smaller windows give sharper anchors on contract-style documents; use 50k–100k only when broader context is necessary. The step rejects inputs requiring more than 64 LLM windows.")
   });
   ConfidenceLevelSchema2 = _enum(["low", "medium", "high"]).describe("LLM confidence as a coarse enum. Numeric (0..1) confidences cluster at 0.85–0.95 and gradations are noise; low/medium/high is reliable.");
   SplitSectionSchema = object({
@@ -8867,6 +8870,11 @@ var init_processor = __esm(() => {
   init_configs();
 });
 
+// ../types/src/parser/index.ts
+var init_parser2 = __esm(() => {
+  init_parser();
+});
+
 // ../types/src/execution-row.ts
 function isTerminalExecutionStatus(status) {
   return status != null && TERMINAL_STATUS_SET.has(status);
@@ -53220,7 +53228,7 @@ function describeUnsupportedType(nodeType) {
 function compileTypedScript(opts) {
   let ast;
   try {
-    ast = import_parser3.parse(opts.source, {
+    ast = import_parser4.parse(opts.source, {
       sourceType: "script",
       plugins: ["typescript"],
       ranges: true
@@ -53473,9 +53481,9 @@ function stripCommentsAndStrings(s) {
   out = out.replace(/`(\\.|[^`\\])*`/g, "``");
   return out;
 }
-var import_parser3, import_sucrase, EVALUATOR_FN_NAME = "scoreScript", EVALUATOR_PARAM_NAMES, TRANSFORM_FN_NAME = "script";
+var import_parser4, import_sucrase, EVALUATOR_FN_NAME = "scoreScript", EVALUATOR_PARAM_NAMES, TRANSFORM_FN_NAME = "script";
 var init_compile = __esm(() => {
-  import_parser3 = __toESM(require_lib(), 1);
+  import_parser4 = __toESM(require_lib(), 1);
   import_sucrase = __toESM(require_dist3(), 1);
   EVALUATOR_PARAM_NAMES = ["expected", "actual"];
 });
@@ -53897,7 +53905,7 @@ var init_step_configs = __esm(() => {
     rules: string2().optional().describe('Optional natural-language rules appended to the system prompt. E.g. "End-of-section markers like *Koniec prílohy 2* close the current section."'),
     provider: string2().optional().describe('Provider ID from eigenpal.config.yaml (e.g. "openai-gpt5.4-mini"). Falls back to the tenant default LLM provider when omitted.'),
     reasoningEffort: OptionalReasoningEffortSchema,
-    windowTokenBudget: number2().int().positive().optional().describe("Override the per-window token ceiling for this step. Defaults to env SPLIT_WINDOW_TOKEN_BUDGET or 20000. Smaller windows give sharper anchors on contract-style documents (less competing context for the LLM to mis-anchor on); bump to 50k–100k when sections routinely exceed per-window page count.")
+    windowTokenBudget: number2().int().positive().optional().describe("Override the estimated per-window token ceiling. Oversized pages are subdivided with overlap while retaining their original page index. Defaults to env SPLIT_WINDOW_TOKEN_BUDGET or 20000. Inputs requiring more than 64 LLM windows are rejected before calls begin.")
   });
   AiSplitSectionSchema = object({
     name: string2().describe("Section name from the config"),
@@ -56272,6 +56280,7 @@ var init_src = __esm(() => {
   init_core3();
   init_scoped_file_ref();
   init_processor();
+  init_parser2();
   init_execution_row();
   init_run_expand();
   init_validation();
@@ -206226,7 +206235,7 @@ import { fileURLToPath as fileURLToPath8 } from "url";
 // package.json
 var package_default = {
   name: "@eigenpal/cli",
-  version: "0.16.6",
+  version: "0.16.8",
   description: "Eigenpal CLI — author and run document workflows from your terminal. Agent-ready.",
   keywords: [
     "eigenpal",
@@ -212055,13 +212064,14 @@ function assertMissingDimensionCellRefs(path, xml, limits) {
     assertOptionalCountLimits(path, maxRow, maxCol, limits);
   }
 }
-function assertSafeSpreadsheetZip(bytes, limits = SPREADSHEET_ZIP_STRUCTURE_DEFAULTS) {
+function assertSafeSpreadsheetZip(bytes, limits = SPREADSHEET_ZIP_STRUCTURE_DEFAULTS, options = {}) {
   if (!isOfficeZipContainer(bytes)) {
     throw new Error("Unsupported spreadsheet: expected an .xls or .xlsx workbook");
   }
   const inspection = inspectSafeOfficeZip(bytes);
   const classified = classifyOfficeSpreadsheetZip(inspection);
-  if (classified.kind !== "xlsx") {
+  const allowedMacroWorkbook = classified.kind === "unsupported" && classified.format === "xlsm" && options.allowMacros === true;
+  if (classified.kind !== "xlsx" && !allowedMacroWorkbook) {
     throw new Error(classified.message);
   }
   const worksheetEntries = inspection.entries.filter((entry) => SPREADSHEET_WORKSHEET_PATH_RE.test(entry.path));
@@ -239318,30 +239328,10 @@ import { existsSync as existsSync11, readdirSync as readdirSync3, readFileSync a
 import { join as join6 } from "path";
 
 // src/lib/fs-helpers.ts
+init_src();
 import { extname as extname3, join as join5 } from "path";
-var EXT_MIME = {
-  ".pdf": "application/pdf",
-  ".png": "image/png",
-  ".jpg": "image/jpeg",
-  ".jpeg": "image/jpeg",
-  ".webp": "image/webp",
-  ".tiff": "image/tiff",
-  ".odt": "application/vnd.oasis.opendocument.text",
-  ".txt": "text/plain",
-  ".csv": "text/csv",
-  ".docx": "application/vnd.openxmlformats-officedocument.wordprocessingml.document",
-  ".xlsx": "application/vnd.openxmlformats-officedocument.spreadsheetml.sheet",
-  ".doc": "application/msword",
-  ".xls": "application/vnd.ms-excel",
-  ".rtf": "application/rtf",
-  ".msg": "application/vnd.ms-outlook",
-  ".json": "application/json",
-  ".xml": "application/xml",
-  ".html": "text/html",
-  ".zip": "application/zip"
-};
 function guessMimeType(filename) {
-  return EXT_MIME[extname3(filename).toLowerCase()] || "";
+  return mimeTypeForExtension(extname3(filename)) ?? "";
 }
 
 // src/lib/payload.ts