@eigenpal/cli 0.17.3 → 0.17.5

package/CHANGELOG.md

@@ -1,5 +1,11 @@
 # @eigenpal/cli
 
+## 0.17.5
+
+### Patch Changes
+
+- e3c0246: CLI skill reference documents the new `ai.figure-crops` workflow step for cropping and captioning figures from parsed PDFs.
+
 ## 0.16.12
 
 ### Patch Changes

package/dist/cli.js

@@ -7800,6 +7800,7 @@ var init_ids = __esm(() => {
     CLASSIFY_PAGES: "builtin/classify-pages",
     VISION: "builtin/vision",
     SEARCH_FILES: "builtin/search-files",
+    FIGURE_CROPS: "builtin/figure-crops",
     TEMPLATE: "builtin/template",
     PDF_EMBEDDER: "builtin/pdf-embedder",
     XLSX_TO_JSON: "builtin/xlsx-to-json",
@@ -53679,7 +53680,8 @@ var init_steps = __esm(() => {
     "ai.classify",
     "ai.classify-pages",
     "ai.vision",
-    "ai.search-files"
+    "ai.search-files",
+    "ai.figure-crops"
   ];
   TRANSFORM_STEP_TYPES = [
     "transform.set",
@@ -53837,7 +53839,7 @@ function getStepSchema(stepType) {
 function listStepTypes() {
   return [...STEP_TYPES];
 }
-var OptionalReasoningEffortSchema, AiParseConfigSchema, AiParseOutputSchema, AiExtractConfigSchema, AiExtractOutputSchema, VisionPageRangeSchema, AiVisionConfigSchema, AiVisionOutputSchema, AI_SEARCH_FILES_DEFAULT_MAX_ITERATIONS = 12, AI_SEARCH_FILES_MAX_ITERATIONS = 50, AI_SEARCH_FILES_DEFAULT_MAX_FILES = 50, AI_SEARCH_FILES_MAX_FILES = 200, AI_SEARCH_FILES_MAX_CITATION_EXCERPT_CHARS = 2000, AiSearchFilesConfigSchema, AiSearchFilesCitationSchema, AiSearchFilesOutputSchema, AiSplitConfigSchema, AiSplitSectionSchema, AiSplitOutputSchema, AiSegmentConfigSchema, AiSegmentOutputSchema, AiClassifyConfigSchema, AiClassifyOutputSchema, ClassifyPagesLabelSchema2, AiClassifyPagesConfigSchema, AiClassifyPagesOutputSchema, TransformSetConfigSchema, TransformSetOutputSchema, TransformRemoveConfigSchema, TransformRemoveOutputSchema, TransformCombineConfigSchema, TransformCombineOutputSchema, TransformSplitConfigSchema, TransformSplitOutputSchema, TransformMergeConfigSchema, TransformMergeOutputSchema, TransformTemplateConfigSchema, TransformTemplateOutputSchema, TransformPdfEmbedConfigSchema, TransformPdfEmbedOutputSchema, TransformXlsxToJsonConfigSchema, TransformXlsxToJsonOutputSchema, JsonToXlsxAuthoringRowsSchema, JsonToXlsxSheetConfigSchema, TransformJsonToXlsxConfigSchema, TransformJsonToXlsxOutputSchema, ARCHIVE_LIST_MAX_ENTRIES = 1e4, TransformArchiveListConfigSchema, TransformArchiveListEntrySchema, TransformArchiveListOutputSchema, TransformArchiveExtractConfigSchema, TransformArchiveExtractOutputSchema, TransformScriptConfigSchema, TransformScriptOutputSchema, ParsedDocPageSchema, TransformTextChunkerConfigSchema, TransformTextChunkerOutputSchema, RegexNormalizerSchema, RegexFormatSchema, REGEX_FLAGS, RegexFieldSchema, TransformRegexExtractConfigSchema, TransformRegexExtractOutputSchema, ActionHttpConfigSchema, ActionHttpOutputSchema, ActionInvokeWorkflowConfigSchema, ActionInvokeWorkflowOutputSchema, ActionWebsiteReaderConfigSchema, ActionWebsiteReaderOutputSchema, ACTION_EMAIL_MAX_ATTACHMENTS = 10, ACTION_EMAIL_MAX_ATTACHMENT_BYTES, ACTION_EMAIL_BODY_MAX_BYTES, emailBodyMaxBytes = (value) => new TextEncoder().encode(value).byteLength <= ACTION_EMAIL_BODY_MAX_BYTES, ActionEmailRecipientListSchema, ActionEmailToSchema, ActionEmailOptionalRecipientsSchema, ActionEmailAttachmentSchema, ActionEmailConfigSchema, ActionEmailOutputSchema, ControlIfConfigSchema, ControlIfOutputSchema, ControlSwitchConfigSchema, ControlSwitchOutputSchema, ControlForeachConfigSchema, ControlForeachOutputSchema, ControlParallelMapConfigSchema, ControlParallelMapOutputSchema, ControlParallelConfigSchema, ControlParallelOutputSchema, ControlWaitConfigSchema, ControlWaitOutputSchema, ControlHumanReviewConfigSchema, ControlHumanReviewOutputSchema, ControlFailConfigSchema, ControlFailOutputSchema, STEP_SCHEMAS;
+var OptionalReasoningEffortSchema, AiParseConfigSchema, AiParseOutputSchema, AiFigureCropsConfigSchema, AiFigureCropsFigureSchema, AiFigureCropsOutputSchema, AiExtractConfigSchema, AiExtractOutputSchema, VisionPageRangeSchema, AiVisionConfigSchema, AiVisionOutputSchema, AI_SEARCH_FILES_DEFAULT_MAX_ITERATIONS = 12, AI_SEARCH_FILES_MAX_ITERATIONS = 50, AI_SEARCH_FILES_DEFAULT_MAX_FILES = 50, AI_SEARCH_FILES_MAX_FILES = 200, AI_SEARCH_FILES_MAX_CITATION_EXCERPT_CHARS = 2000, AiSearchFilesConfigSchema, AiSearchFilesCitationSchema, AiSearchFilesOutputSchema, AiSplitConfigSchema, AiSplitSectionSchema, AiSplitOutputSchema, AiSegmentConfigSchema, AiSegmentOutputSchema, AiClassifyConfigSchema, AiClassifyOutputSchema, ClassifyPagesLabelSchema2, AiClassifyPagesConfigSchema, AiClassifyPagesOutputSchema, TransformSetConfigSchema, TransformSetOutputSchema, TransformRemoveConfigSchema, TransformRemoveOutputSchema, TransformCombineConfigSchema, TransformCombineOutputSchema, TransformSplitConfigSchema, TransformSplitOutputSchema, TransformMergeConfigSchema, TransformMergeOutputSchema, TransformTemplateConfigSchema, TransformTemplateOutputSchema, TransformPdfEmbedConfigSchema, TransformPdfEmbedOutputSchema, TransformXlsxToJsonConfigSchema, TransformXlsxToJsonOutputSchema, JsonToXlsxAuthoringRowsSchema, JsonToXlsxSheetConfigSchema, TransformJsonToXlsxConfigSchema, TransformJsonToXlsxOutputSchema, ARCHIVE_LIST_MAX_ENTRIES = 1e4, TransformArchiveListConfigSchema, TransformArchiveListEntrySchema, TransformArchiveListOutputSchema, TransformArchiveExtractConfigSchema, TransformArchiveExtractOutputSchema, TransformScriptConfigSchema, TransformScriptOutputSchema, ParsedDocPageSchema, TransformTextChunkerConfigSchema, TransformTextChunkerOutputSchema, RegexNormalizerSchema, RegexFormatSchema, REGEX_FLAGS, RegexFieldSchema, TransformRegexExtractConfigSchema, TransformRegexExtractOutputSchema, ActionHttpConfigSchema, ActionHttpOutputSchema, ActionInvokeWorkflowConfigSchema, ActionInvokeWorkflowOutputSchema, ActionWebsiteReaderConfigSchema, ActionWebsiteReaderOutputSchema, ACTION_EMAIL_MAX_ATTACHMENTS = 10, ACTION_EMAIL_MAX_ATTACHMENT_BYTES, ACTION_EMAIL_BODY_MAX_BYTES, emailBodyMaxBytes = (value) => new TextEncoder().encode(value).byteLength <= ACTION_EMAIL_BODY_MAX_BYTES, ActionEmailRecipientListSchema, ActionEmailToSchema, ActionEmailOptionalRecipientsSchema, ActionEmailAttachmentSchema, ActionEmailConfigSchema, ActionEmailOutputSchema, ControlIfConfigSchema, ControlIfOutputSchema, ControlSwitchConfigSchema, ControlSwitchOutputSchema, ControlForeachConfigSchema, ControlForeachOutputSchema, ControlParallelMapConfigSchema, ControlParallelMapOutputSchema, ControlParallelConfigSchema, ControlParallelOutputSchema, ControlWaitConfigSchema, ControlWaitOutputSchema, ControlHumanReviewConfigSchema, ControlHumanReviewOutputSchema, ControlFailConfigSchema, ControlFailOutputSchema, STEP_SCHEMAS;
 var init_step_configs = __esm(() => {
   init_zod();
   init_ai_client();
@@ -53871,6 +53873,34 @@ var init_step_configs = __esm(() => {
     figureInstructions: string2().optional().describe('Custom instruction for the figure-description pass, e.g. "Describe each figure; label a handwritten signature as `<figure>signature</figure>` and a stamp as `<figure>stamp</figure>`; for property photos note the room or exterior shown." Applied only when describeFigures runs.')
   }).superRefine(refineNativeParseModeConflicts).superRefine(refineLayoutOutputFormat);
   AiParseOutputSchema = ParseResultSchema.omit({ rawResponse: true });
+  AiFigureCropsConfigSchema = object({
+    input: string2().describe("Storage reference or template expression for the source PDF"),
+    parseOutput: unknown().describe("ai.parse output object (template {{steps.<parse>.output}}); figure elements are read from its `elements` array"),
+    model: string2().optional().describe("Vision model for captions (defaults to workspace model)"),
+    figureInstructions: string2().optional().describe("Custom caption instruction; applied per figure crop"),
+    maxFigures: number2().int().min(1).max(100).default(40).optional().describe("Cap on figures processed per document"),
+    minAreaFrac: number2().min(0).max(1).default(0).optional().describe("Skip figures below this page-area fraction (logos/decorations)"),
+    includeCrops: boolean2().default(false).optional().describe("Include base64 JPEG crops in output (heavy; default off — captions + bboxes only)"),
+    renderScale: number2().min(1).max(4).default(1).optional().describe("Scale factor for rendering PDF pages before cropping"),
+    imageQuality: number2().int().min(1).max(100).default(85).optional().describe("JPEG quality for crops"),
+    llmReasoningEffort: OptionalReasoningEffortSchema
+  });
+  AiFigureCropsFigureSchema = object({
+    id: string2(),
+    page: number2(),
+    bboxNorm: tuple([number2(), number2(), number2(), number2()]).nullable(),
+    areaFrac: number2(),
+    caption: string2().nullable(),
+    cropJpegBase64: string2().optional()
+  });
+  AiFigureCropsOutputSchema = object({
+    figures: array(AiFigureCropsFigureSchema),
+    _figureCrops: object({
+      pages: array(number2()),
+      model: string2().optional(),
+      skippedSmall: number2()
+    })
+  });
   AiExtractConfigSchema = object({
     input: string2().describe("Text content or template expression"),
     schema: object({
@@ -54596,6 +54626,15 @@ var init_step_configs = __esm(() => {
       outputSchema: AiSearchFilesOutputSchema,
       configInWith: true
     },
+    "ai.figure-crops": {
+      type: "ai.figure-crops",
+      category: "ai",
+      name: "Figure Crops",
+      description: "Crop and caption figure elements from a parsed PDF. Renders the source document, crops each kind=figure bbox from ai.parse output, and describes the crop with a vision model. Output is captions plus normalized bboxes for evidence viewers.",
+      configSchema: AiFigureCropsConfigSchema,
+      outputSchema: AiFigureCropsOutputSchema,
+      configInWith: true
+    },
     "transform.set": {
       type: "transform.set",
       category: "transform",
@@ -206278,7 +206317,7 @@ import { fileURLToPath as fileURLToPath8 } from "url";
 // package.json
 var package_default = {
   name: "@eigenpal/cli",
-  version: "0.17.3",
+  version: "0.17.5",
   description: "Eigenpal CLI — author and run document workflows from your terminal. Agent-ready.",
   keywords: [
     "eigenpal",