veryfront 0.1.73 → 0.1.75

package/esm/cli/commands/knowledge/command-help.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"command-help.d.ts","sourceRoot":"","sources":["../../../../src/cli/commands/knowledge/command-help.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,WAAW,EAAE,MAAM,qBAAqB,CAAC;AAEvD,eAAO,MAAM,aAAa,EAAE,WAqC3B,CAAC"}
+{"version":3,"file":"command-help.d.ts","sourceRoot":"","sources":["../../../../src/cli/commands/knowledge/command-help.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,WAAW,EAAE,MAAM,qBAAqB,CAAC;AAEvD,eAAO,MAAM,aAAa,EAAE,WAuC3B,CAAC"}

package/esm/cli/commands/knowledge/command-help.js

@@ -1,7 +1,7 @@
 export const knowledgeHelp = {
     name: "knowledge",
     description: "Ingest documents into the project knowledge base",
-    usage: "veryfront knowledge ingest <source> [options]",
+    usage: "veryfront knowledge ingest <source...> [options]",
     options: [
         {
             flag: "--project, -p <slug>",
@@ -26,6 +26,7 @@ export const knowledgeHelp = {
     ],
     examples: [
         "veryfront knowledge ingest uploads/contracts/q1.pdf --json",
+        "veryfront knowledge ingest uploads/contracts/a.pdf uploads/contracts/b.pdf uploads/contracts/c.pdf --json",
         "veryfront knowledge ingest /workspace/uploads/q1.pdf --json",
         "veryfront knowledge ingest --path uploads/ --all --json",
     ],
@@ -34,5 +35,6 @@ export const knowledgeHelp = {
         "`uploads/...` means a remote project upload; use `./uploads/...` or `/workspace/uploads/...` to force a local file",
         "`ingest` orchestrates upload resolution, parsing, and project file writes",
         "Requires python3; non-text formats also require the supported parser packages unless you run inside the Veryfront sandbox",
+        "The Veryfront sandbox image includes `kreuzberg`, and knowledge ingest falls back to the built-in parser when `kreuzberg` is unavailable or extraction fails",
     ],
 };

package/esm/cli/commands/knowledge/command.d.ts

@@ -1,6 +1,7 @@
 import { z } from "zod";
 import { type ApiClient } from "../../shared/config.js";
 import type { ParsedArgs } from "../../shared/types.js";
+import { type Logger } from "../../../src/utils/index.js";
 export interface KnowledgeParserResult {
     success: true;
     source_path: string;
@@ -41,10 +42,10 @@ type DownloadResult = {
     localPath: string;
     bytes?: number;
 };
-declare const KnowledgeIngestArgsSchema: z.ZodObject<{
+declare const KnowledgeIngestArgsSchema: z.ZodEffects<z.ZodObject<{
     projectSlug: z.ZodOptional<z.ZodString>;
     projectDir: z.ZodOptional<z.ZodString>;
-    source: z.ZodOptional<z.ZodString>;
+    sources: z.ZodDefault<z.ZodArray<z.ZodString, "many">>;
     path: z.ZodOptional<z.ZodString>;
     all: z.ZodDefault<z.ZodBoolean>;
     recursive: z.ZodDefault<z.ZodBoolean>;
@@ -58,6 +59,7 @@ declare const KnowledgeIngestArgsSchema: z.ZodObject<{
     json: boolean;
     recursive: boolean;
     all: boolean;
+    sources: string[];
     quiet: boolean;
     knowledgePath: string;
     path?: string | undefined;
@@ -66,7 +68,6 @@ declare const KnowledgeIngestArgsSchema: z.ZodObject<{
     outputDir?: string | undefined;
     projectDir?: string | undefined;
     slug?: string | undefined;
-    source?: string | undefined;
 }, {
     path?: string | undefined;
     projectSlug?: string | undefined;
@@ -77,7 +78,33 @@ declare const KnowledgeIngestArgsSchema: z.ZodObject<{
     projectDir?: string | undefined;
     slug?: string | undefined;
     all?: boolean | undefined;
-    source?: string | undefined;
+    sources?: string[] | undefined;
+    quiet?: boolean | undefined;
+    knowledgePath?: string | undefined;
+}>, {
+    json: boolean;
+    recursive: boolean;
+    all: boolean;
+    sources: string[];
+    quiet: boolean;
+    knowledgePath: string;
+    path?: string | undefined;
+    projectSlug?: string | undefined;
+    description?: string | undefined;
+    outputDir?: string | undefined;
+    projectDir?: string | undefined;
+    slug?: string | undefined;
+}, {
+    path?: string | undefined;
+    projectSlug?: string | undefined;
+    description?: string | undefined;
+    outputDir?: string | undefined;
+    json?: boolean | undefined;
+    recursive?: boolean | undefined;
+    projectDir?: string | undefined;
+    slug?: string | undefined;
+    all?: boolean | undefined;
+    sources?: string[] | undefined;
     quiet?: boolean | undefined;
     knowledgePath?: string | undefined;
 }>;
@@ -102,8 +129,9 @@ export declare function runKnowledgeParser(input: {
     description?: string;
     slug?: string;
     sourceReference?: string;
+    env?: Record<string, string>;
 }): Promise<KnowledgeParserResult>;
-export declare function collectKnowledgeSources(options: Pick<KnowledgeIngestOptions, "source" | "path" | "all" | "recursive">, deps: {
+export declare function collectKnowledgeSources(options: Pick<KnowledgeIngestOptions, "sources" | "path" | "all" | "recursive">, deps: {
     client: ApiClient;
     projectSlug: string;
     downloadUploads: (uploadPaths: string[]) => Promise<DownloadResult[]>;
@@ -116,6 +144,7 @@ export declare function ingestResolvedSources(sources: KnowledgeSource[], option
     uploadKnowledgeFile: (remotePath: string, localPath: string) => Promise<{
         path: string;
     }>;
+    eventLogger?: Logger | null;
 }): Promise<KnowledgeIngestFileResult[]>;
 export declare function knowledgeCommand(args: ParsedArgs): Promise<void>;
 export {};

package/esm/cli/commands/knowledge/command.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"command.d.ts","sourceRoot":"","sources":["../../../../src/cli/commands/knowledge/command.ts"],"names":[],"mappings":"AACA,OAAO,EAAE,CAAC,EAAE,MAAM,KAAK,CAAC;AAKxB,OAAO,EAAE,KAAK,SAAS,EAA0C,MAAM,wBAAwB,CAAC;AAChG,OAAO,KAAK,EAAE,UAAU,EAAE,MAAM,uBAAuB,CAAC;AAqBxD,MAAM,WAAW,qBAAqB;IACpC,OAAO,EAAE,IAAI,CAAC;IACd,WAAW,EAAE,MAAM,CAAC;IACpB,eAAe,EAAE,MAAM,CAAC;IACxB,WAAW,EAAE,MAAM,CAAC;IACpB,IAAI,EAAE,MAAM,CAAC;IACb,mBAAmB,EAAE,MAAM,CAAC;IAC5B,sBAAsB,EAAE,MAAM,CAAC;IAC/B,WAAW,EAAE,MAAM,CAAC;IACpB,KAAK,EAAE,MAAM,CAAC;IACd,OAAO,EAAE,MAAM,CAAC;IAChB,KAAK,EAAE,MAAM,CAAC,MAAM,EAAE,OAAO,CAAC,CAAC;IAC/B,QAAQ,EAAE,MAAM,EAAE,CAAC;CACpB;AAED,MAAM,WAAW,yBAAyB;IACxC,MAAM,EAAE,MAAM,CAAC;IACf,eAAe,EAAE,MAAM,CAAC;IACxB,UAAU,EAAE,MAAM,CAAC;IACnB,UAAU,EAAE,MAAM,CAAC;IACnB,IAAI,EAAE,MAAM,CAAC;IACb,UAAU,EAAE,MAAM,CAAC;IACnB,OAAO,EAAE,MAAM,CAAC;IAChB,KAAK,EAAE,MAAM,CAAC,MAAM,EAAE,OAAO,CAAC,CAAC;IAC/B,QAAQ,EAAE,MAAM,EAAE,CAAC;CACpB;AAED,KAAK,eAAe,GAChB;IAAE,IAAI,EAAE,OAAO,CAAC;IAAC,KAAK,EAAE,MAAM,CAAC;IAAC,SAAS,EAAE,MAAM,CAAA;CAAE,GACnD;IAAE,IAAI,EAAE,QAAQ,CAAC;IAAC,KAAK,EAAE,MAAM,CAAC;IAAC,UAAU,EAAE,MAAM,CAAC;IAAC,SAAS,EAAE,MAAM,CAAA;CAAE,CAAC;AAE7E,KAAK,cAAc,GAAG;IAAE,UAAU,EAAE,MAAM,CAAC;IAAC,SAAS,EAAE,MAAM,CAAC;IAAC,KAAK,CAAC,EAAE,MAAM,CAAA;CAAE,CAAC;AAEhF,QAAA,MAAM,yBAAyB;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;EAa7B,CAAC;AAEH,MAAM,MAAM,sBAAsB,GAAG,CAAC,CAAC,KAAK,CAAC,OAAO,yBAAyB,CAAC,CAAC;AA+B/E,wBAAgB,wBAAwB,CACtC,IAAI,EAAE,UAAU,GACf,CAAC,CAAC,mBAAmB,CAAC,OAAO,EAAE,sBAAsB,CAAC,CAexD;AAED,wBAAgB,2BAA2B,CAAC,SAAS,EAAE,MAAM,GAAG,MAAM,CAMrE;AAED,wBAAgB,0BAA0B,CAAC,SAAS,EAAE,MAAM,GAAG,MAAM,CAGpE;AAED,wBAAgB,2BAA2B,CAAC,UAAU,EAAE,MAAM,GAAG,MAAM,CAKtE;AAED,wBAAgB,iBAAiB,CAAC,KAAK,EAAE,MAAM,GAAG,OAAO,CAGxD;AAoBD,wBAAgB,iCAAiC,CAAC,SAAS,EAAE,MAAM,GAAG,MAAM,CAE3E;AAgED,wBAAgB,yBAAyB,CACvC,UAAU,EAAE,MAAM,EAClB,SAAS,EAAE,MAAM,EACjB,aAAa,EAAE,MAAM,GACpB,MAAM,CAQR;AAED,wBAAgB,2BAA2B,CAAC,KAAK,EAAE;IACjD,MAAM,EAAE,MAAM,CAAC;IACf,eAAe,EAAE,MAAM,CAAC;IACxB,UAAU,EAAE,MAAM,CAAC;IACnB,UAAU,EAAE,MAAM,CAAC;IACnB,MAAM,EAAE,IAAI,CAAC,qBAAqB,EAAE,MAAM,GAAG,OAAO,GAAG,UAAU,GAAG,aAAa,GAAG,SAAS,CAAC,CAAC;CAChG,GAAG,yBAAyB,CAY5B;AAED,wBAAsB,kBAAkB,CAAC,KAAK,EAAE;IAC9C,QAAQ,EAAE,MAAM,CAAC;IACjB,SAAS,EAAE,MAAM,CAAC;IAClB,WAAW,CAAC,EAAE,MAAM,CAAC;IACrB,IAAI,CAAC,EAAE,MAAM,CAAC;IACd,eAAe,CAAC,EAAE,MAAM,CAAC;CAC1B,GAAG,OAAO,CAAC,qBAAqB,CAAC,CA6CjC;AAED,wBAAsB,uBAAuB,CAC3C,OAAO,EAAE,IAAI,CAAC,sBAAsB,EAAE,QAAQ,GAAG,MAAM,GAAG,KAAK,GAAG,WAAW,CAAC,EAC9E,IAAI,EAAE;IACJ,MAAM,EAAE,SAAS,CAAC;IAClB,WAAW,EAAE,MAAM,CAAC;IACpB,eAAe,EAAE,CAAC,WAAW,EAAE,MAAM,EAAE,KAAK,OAAO,CAAC,cAAc,EAAE,CAAC,CAAC;CACvE,GACA,OAAO,CAAC,eAAe,EAAE,CAAC,CAmE5B;AAED,wBAAsB,qBAAqB,CACzC,OAAO,EAAE,eAAe,EAAE,EAC1B,OAAO,EAAE,sBAAsB,EAC/B,IAAI,EAAE;IACJ,MAAM,EAAE,SAAS,CAAC;IAClB,WAAW,EAAE,MAAM,CAAC;IACpB,SAAS,EAAE,MAAM,CAAC;IAClB,SAAS,EAAE,OAAO,kBAAkB,CAAC;IACrC,mBAAmB,EAAE,CAAC,UAAU,EAAE,MAAM,EAAE,SAAS,EAAE,MAAM,KAAK,OAAO,CAAC;QAAE,IAAI,EAAE,MAAM,CAAA;KAAE,CAAC,CAAC;CAC3F,GACA,OAAO,CAAC,yBAAyB,EAAE,CAAC,CA8BtC;AAED,wBAAsB,gBAAgB,CAAC,IAAI,EAAE,UAAU,GAAG,OAAO,CAAC,IAAI,CAAC,CAwEtE"}
+{"version":3,"file":"command.d.ts","sourceRoot":"","sources":["../../../../src/cli/commands/knowledge/command.ts"],"names":[],"mappings":"AACA,OAAO,EAAE,CAAC,EAAE,MAAM,KAAK,CAAC;AAKxB,OAAO,EAAE,KAAK,SAAS,EAA0C,MAAM,wBAAwB,CAAC;AAChG,OAAO,KAAK,EAAE,UAAU,EAAE,MAAM,uBAAuB,CAAC;AAIxD,OAAO,EAAuB,KAAK,MAAM,EAAgB,MAAM,6BAA6B,CAAC;AAkB7F,MAAM,WAAW,qBAAqB;IACpC,OAAO,EAAE,IAAI,CAAC;IACd,WAAW,EAAE,MAAM,CAAC;IACpB,eAAe,EAAE,MAAM,CAAC;IACxB,WAAW,EAAE,MAAM,CAAC;IACpB,IAAI,EAAE,MAAM,CAAC;IACb,mBAAmB,EAAE,MAAM,CAAC;IAC5B,sBAAsB,EAAE,MAAM,CAAC;IAC/B,WAAW,EAAE,MAAM,CAAC;IACpB,KAAK,EAAE,MAAM,CAAC;IACd,OAAO,EAAE,MAAM,CAAC;IAChB,KAAK,EAAE,MAAM,CAAC,MAAM,EAAE,OAAO,CAAC,CAAC;IAC/B,QAAQ,EAAE,MAAM,EAAE,CAAC;CACpB;AAED,MAAM,WAAW,yBAAyB;IACxC,MAAM,EAAE,MAAM,CAAC;IACf,eAAe,EAAE,MAAM,CAAC;IACxB,UAAU,EAAE,MAAM,CAAC;IACnB,UAAU,EAAE,MAAM,CAAC;IACnB,IAAI,EAAE,MAAM,CAAC;IACb,UAAU,EAAE,MAAM,CAAC;IACnB,OAAO,EAAE,MAAM,CAAC;IAChB,KAAK,EAAE,MAAM,CAAC,MAAM,EAAE,OAAO,CAAC,CAAC;IAC/B,QAAQ,EAAE,MAAM,EAAE,CAAC;CACpB;AAED,KAAK,eAAe,GAChB;IAAE,IAAI,EAAE,OAAO,CAAC;IAAC,KAAK,EAAE,MAAM,CAAC;IAAC,SAAS,EAAE,MAAM,CAAA;CAAE,GACnD;IAAE,IAAI,EAAE,QAAQ,CAAC;IAAC,KAAK,EAAE,MAAM,CAAC;IAAC,UAAU,EAAE,MAAM,CAAC;IAAC,SAAS,EAAE,MAAM,CAAA;CAAE,CAAC;AAE7E,KAAK,cAAc,GAAG;IAAE,UAAU,EAAE,MAAM,CAAC;IAAC,SAAS,EAAE,MAAM,CAAC;IAAC,KAAK,CAAC,EAAE,MAAM,CAAA;CAAE,CAAC;AAIhF,QAAA,MAAM,yBAAyB;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;EAmD7B,CAAC;AAEH,MAAM,MAAM,sBAAsB,GAAG,CAAC,CAAC,KAAK,CAAC,OAAO,yBAAyB,CAAC,CAAC;AAoD/E,wBAAgB,wBAAwB,CACtC,IAAI,EAAE,UAAU,GACf,CAAC,CAAC,mBAAmB,CAAC,OAAO,EAAE,sBAAsB,CAAC,CAexD;AAED,wBAAgB,2BAA2B,CAAC,SAAS,EAAE,MAAM,GAAG,MAAM,CAMrE;AAED,wBAAgB,0BAA0B,CAAC,SAAS,EAAE,MAAM,GAAG,MAAM,CAGpE;AAED,wBAAgB,2BAA2B,CAAC,UAAU,EAAE,MAAM,GAAG,MAAM,CAKtE;AAED,wBAAgB,iBAAiB,CAAC,KAAK,EAAE,MAAM,GAAG,OAAO,CAGxD;AAoBD,wBAAgB,iCAAiC,CAAC,SAAS,EAAE,MAAM,GAAG,MAAM,CAE3E;AAgED,wBAAgB,yBAAyB,CACvC,UAAU,EAAE,MAAM,EAClB,SAAS,EAAE,MAAM,EACjB,aAAa,EAAE,MAAM,GACpB,MAAM,CAQR;AAED,wBAAgB,2BAA2B,CAAC,KAAK,EAAE;IACjD,MAAM,EAAE,MAAM,CAAC;IACf,eAAe,EAAE,MAAM,CAAC;IACxB,UAAU,EAAE,MAAM,CAAC;IACnB,UAAU,EAAE,MAAM,CAAC;IACnB,MAAM,EAAE,IAAI,CAAC,qBAAqB,EAAE,MAAM,GAAG,OAAO,GAAG,UAAU,GAAG,aAAa,GAAG,SAAS,CAAC,CAAC;CAChG,GAAG,yBAAyB,CAY5B;AAED,wBAAsB,kBAAkB,CAAC,KAAK,EAAE;IAC9C,QAAQ,EAAE,MAAM,CAAC;IACjB,SAAS,EAAE,MAAM,CAAC;IAClB,WAAW,CAAC,EAAE,MAAM,CAAC;IACrB,IAAI,CAAC,EAAE,MAAM,CAAC;IACd,eAAe,CAAC,EAAE,MAAM,CAAC;IACzB,GAAG,CAAC,EAAE,MAAM,CAAC,MAAM,EAAE,MAAM,CAAC,CAAC;CAC9B,GAAG,OAAO,CAAC,qBAAqB,CAAC,CA8CjC;AAED,wBAAsB,uBAAuB,CAC3C,OAAO,EAAE,IAAI,CAAC,sBAAsB,EAAE,SAAS,GAAG,MAAM,GAAG,KAAK,GAAG,WAAW,CAAC,EAC/E,IAAI,EAAE;IACJ,MAAM,EAAE,SAAS,CAAC;IAClB,WAAW,EAAE,MAAM,CAAC;IACpB,eAAe,EAAE,CAAC,WAAW,EAAE,MAAM,EAAE,KAAK,OAAO,CAAC,cAAc,EAAE,CAAC,CAAC;CACvE,GACA,OAAO,CAAC,eAAe,EAAE,CAAC,CA0G5B;AAED,wBAAsB,qBAAqB,CACzC,OAAO,EAAE,eAAe,EAAE,EAC1B,OAAO,EAAE,sBAAsB,EAC/B,IAAI,EAAE;IACJ,MAAM,EAAE,SAAS,CAAC;IAClB,WAAW,EAAE,MAAM,CAAC;IACpB,SAAS,EAAE,MAAM,CAAC;IAClB,SAAS,EAAE,OAAO,kBAAkB,CAAC;IACrC,mBAAmB,EAAE,CAAC,UAAU,EAAE,MAAM,EAAE,SAAS,EAAE,MAAM,KAAK,OAAO,CAAC;QAAE,IAAI,EAAE,MAAM,CAAA;KAAE,CAAC,CAAC;IAC1F,WAAW,CAAC,EAAE,MAAM,GAAG,IAAI,CAAC;CAC7B,GACA,OAAO,CAAC,yBAAyB,EAAE,CAAC,CA6DtC;AAED,wBAAsB,gBAAgB,CAAC,IAAI,EAAE,UAAU,GAAG,OAAO,CAAC,IAAI,CAAC,CA+FtE"}

package/esm/cli/commands/knowledge/command.js

@@ -1,6 +1,6 @@
 import * as dntShim from "../../../_dnt.shims.js";
 import { z } from "zod";
-import { createFileSystem } from "../../../src/platform/index.js";
+import { createFileSystem, getEnv } from "../../../src/platform/index.js";
 import { basename, extname, join, normalize, relative } from "../../../src/platform/compat/path/index.js";
 import { withSpan } from "../../../src/observability/tracing/otlp-setup.js";
 import { cliLogger } from "../../utils/index.js";
@@ -8,6 +8,7 @@ import { createApiClient, resolveConfigWithAuth } from "../../shared/config.js";
 import { downloadUploadToFile, listAllUploads } from "../uploads/command.js";
 import { putRemoteFileFromLocal } from "../files/command.js";
 import { knowledgeIngestPythonSource } from "./parser-source.js";
+import { createJobUserLogger, serverLogger } from "../../../src/utils/index.js";
 const SUPPORTED_EXTENSIONS = new Set([
     ".pdf",
     ".csv",
@@ -23,10 +24,11 @@ const SUPPORTED_EXTENSIONS = new Set([
     ".md",
     ".mdx",
 ]);
+const knowledgeJobLogger = serverLogger.component("knowledge-ingest");
 const KnowledgeIngestArgsSchema = z.object({
     projectSlug: z.string().optional(),
     projectDir: z.string().optional(),
-    source: z.string().optional(),
+    sources: z.array(z.string()).default([]),
     path: z.string().optional(),
     all: z.boolean().default(false),
     recursive: z.boolean().default(false),
@@ -36,6 +38,39 @@ const KnowledgeIngestArgsSchema = z.object({
     slug: z.string().optional(),
     json: z.boolean().default(false),
     quiet: z.boolean().default(false),
+}).superRefine((value, ctx) => {
+    const hasExplicitSources = value.sources.length > 0;
+    const hasPath = typeof value.path === "string" && value.path.length > 0;
+    if (hasExplicitSources && (hasPath || value.all)) {
+        ctx.addIssue({
+            code: z.ZodIssueCode.custom,
+            message: "Use either explicit source paths or --path with --all, not both.",
+        });
+    }
+    if (!hasExplicitSources && !hasPath && !value.all) {
+        ctx.addIssue({
+            code: z.ZodIssueCode.custom,
+            message: "Provide one or more source paths or use --path with --all.",
+        });
+    }
+    if (hasPath && !value.all) {
+        ctx.addIssue({
+            code: z.ZodIssueCode.custom,
+            message: "--path requires --all.",
+        });
+    }
+    if (!hasPath && value.all) {
+        ctx.addIssue({
+            code: z.ZodIssueCode.custom,
+            message: "--all requires --path.",
+        });
+    }
+    if (value.slug && value.sources.length !== 1) {
+        ctx.addIssue({
+            code: z.ZodIssueCode.custom,
+            message: "--slug can only be used with a single explicit source.",
+        });
+    }
 });
 function getStringArg(args, ...keys) {
     for (const key of keys) {
@@ -51,12 +86,29 @@ function getBooleanArg(args, ...keys) {
 function printJson(value) {
     console.log(JSON.stringify(value, null, 2));
 }
+function createKnowledgeIngestEventLogger() {
+    const projectId = getEnv("TENANT_PROJECT_ID");
+    const jobId = getEnv("JOB_ID");
+    if (!projectId || !jobId) {
+        return null;
+    }
+    return createJobUserLogger(knowledgeJobLogger, {
+        projectId,
+        jobId,
+        batchId: getEnv("JOB_BATCH_ID") ?? undefined,
+        jobTarget: getEnv("JOB_TARGET") ?? undefined,
+        task: "knowledge-ingest",
+    });
+}
+function buildKnowledgeSourceName(source) {
+    return basename(source.kind === "upload" ? source.uploadPath : source.localPath);
+}
 function showKnowledgeUsage() {
     console.log(`
 Veryfront Knowledge
 
 Usage:
-  veryfront knowledge ingest <source> [options]
+  veryfront knowledge ingest <source...> [options]
   veryfront knowledge ingest --path <prefix-or-dir> --all [options]
 
 Subcommands:
@@ -67,7 +119,7 @@ export function parseKnowledgeIngestArgs(args) {
     return KnowledgeIngestArgsSchema.safeParse({
         projectSlug: getStringArg(args, "project", "p", "project-slug"),
         projectDir: getStringArg(args, "project-dir", "dir", "d"),
-        source: typeof args._[2] === "string" ? args._[2] : undefined,
+        sources: args._.slice(2).filter((value) => typeof value === "string"),
         path: getStringArg(args, "path"),
         all: getBooleanArg(args, "all"),
         recursive: getBooleanArg(args, "recursive"),
@@ -218,6 +270,7 @@ export async function runKnowledgeParser(input) {
         try {
             result = await new dntShim.Deno.Command("python3", {
                 args: [scriptPath, "--input-json", inputJsonPath, "--output-json", outputJsonPath],
+                ...(input.env ? { env: input.env } : {}),
                 stdout: "piped",
                 stderr: "piped",
             }).output();
@@ -241,27 +294,56 @@ export async function runKnowledgeParser(input) {
 }
 export async function collectKnowledgeSources(options, deps) {
     const fs = createFileSystem();
-    if (options.source) {
-        if (!isProjectUploadReference(options.source) && await fs.exists(options.source)) {
-            const localFiles = await collectLocalFiles(options.source, options.recursive);
-            if (!localFiles.length)
-                throw new Error(`No supported files found at ${options.source}`);
-            return localFiles.map((localPath) => ({ kind: "local", input: options.source, localPath }));
+    if (options.sources.length > 0) {
+        const explicitSources = [];
+        const uploadTargets = [];
+        for (const input of options.sources) {
+            if (!isProjectUploadReference(input) && await fs.exists(input)) {
+                const localFiles = await collectLocalFiles(input, options.recursive);
+                if (!localFiles.length)
+                    throw new Error(`No supported files found at ${input}`);
+                explicitSources.push({
+                    kind: "local",
+                    sources: localFiles.map((localPath) => ({ kind: "local", input, localPath })),
+                });
+                continue;
+            }
+            if (isLikelyLocalPath(input)) {
+                throw new Error(`Local file not found: ${input}`);
+            }
+            const uploadPath = normalizeProjectUploadPath(input);
+            explicitSources.push({ kind: "upload", input, uploadPath });
+            uploadTargets.push(uploadPath);
         }
-        if (isLikelyLocalPath(options.source)) {
-            throw new Error(`Local file not found: ${options.source}`);
+        const downloads = uploadTargets.length > 0 ? await deps.downloadUploads(uploadTargets) : [];
+        const downloadsByPath = new Map();
+        for (const download of downloads) {
+            const existing = downloadsByPath.get(download.uploadPath) ?? [];
+            existing.push(download);
+            downloadsByPath.set(download.uploadPath, existing);
         }
-        const uploadPath = normalizeProjectUploadPath(options.source);
-        const downloads = await deps.downloadUploads([uploadPath]);
-        return downloads.map((download) => ({
-            kind: "upload",
-            input: options.source,
-            uploadPath: download.uploadPath,
-            localPath: download.localPath,
-        }));
+        const resolvedSources = [];
+        for (const source of explicitSources) {
+            if (source.kind === "local") {
+                resolvedSources.push(...source.sources);
+                continue;
+            }
+            const matchingDownloads = downloadsByPath.get(source.uploadPath);
+            const download = matchingDownloads?.shift();
+            if (!download) {
+                throw new Error(`Upload not found: ${formatKnowledgeUploadSource(source.uploadPath)}`);
+            }
+            resolvedSources.push({
+                kind: "upload",
+                input: source.input,
+                uploadPath: download.uploadPath,
+                localPath: download.localPath,
+            });
+        }
+        return resolvedSources;
     }
     if (!options.path || !options.all) {
-        throw new Error("Provide a source path or use --path with --all.");
+        throw new Error("Provide one or more source paths or use --path with --all.");
     }
     if (!isProjectUploadReference(options.path) && await fs.exists(options.path)) {
         const localFiles = await collectLocalFiles(options.path, options.recursive);
@@ -298,9 +380,18 @@ export async function collectKnowledgeSources(options, deps) {
     }));
 }
 export async function ingestResolvedSources(sources, options, deps) {
-    const slugs = options.slug && sources.length === 1 ? [options.slug] : ensureUniqueSlugs(sources);
+    if (options.slug && sources.length !== 1) {
+        throw new Error("--slug can only be used with a single explicit source.");
+    }
+    const slugs = options.slug ? [options.slug] : ensureUniqueSlugs(sources);
     const results = [];
     for (const [index, source] of sources.entries()) {
+        deps.eventLogger?.info("Processing knowledge source", {
+            phase: "file_processing",
+            progress_current: index + 1,
+            progress_total: sources.length,
+            source_name: buildKnowledgeSourceName(source),
+        });
         const parser = await deps.runParser({
             filePath: source.localPath,
             outputDir: deps.outputDir,
@@ -310,6 +401,23 @@ export async function ingestResolvedSources(sources, options, deps) {
         });
         const remotePath = deriveKnowledgeRemotePath(parser.sandbox_output_path, deps.outputDir, options.knowledgePath);
         const uploaded = await deps.uploadKnowledgeFile(remotePath, parser.sandbox_output_path);
+        deps.eventLogger?.info("Knowledge source ingested", {
+            phase: "file_completed",
+            progress_current: index + 1,
+            progress_total: sources.length,
+            source_name: buildKnowledgeSourceName(source),
+            remote_path: uploaded.path,
+            warning_count: parser.warnings.length,
+        });
+        if (parser.warnings.length > 0) {
+            deps.eventLogger?.warn("Knowledge source emitted warnings", {
+                phase: "file_warning",
+                progress_current: index + 1,
+                progress_total: sources.length,
+                source_name: buildKnowledgeSourceName(source),
+                warning_count: parser.warnings.length,
+            });
+        }
         results.push(createKnowledgeIngestResult({
             source: buildSourceReference(source),
             localSourcePath: source.localPath,
@@ -341,19 +449,34 @@ export async function knowledgeCommand(args) {
                 const outputDir = options.outputDir ?? await defaultOutputRoot();
                 const shouldCleanupOutputDir = options.outputDir === undefined;
                 const downloadOutputDir = resolveKnowledgeDownloadOutputDir(outputDir);
+                const eventLogger = createKnowledgeIngestEventLogger();
                 try {
+                    eventLogger?.info("Starting knowledge ingest", {
+                        phase: "started",
+                        mode: options.path ? "path_prefix" : "explicit_sources",
+                    });
                     const sources = await collectKnowledgeSources(options, {
                         client,
                         projectSlug: config.projectSlug,
                         downloadUploads: (uploadPaths) => Promise.all(uploadPaths.map((uploadPath) => downloadUploadToFile(client, config.projectSlug, uploadPath, downloadOutputDir))),
                     });
+                    eventLogger?.info("Resolved knowledge sources", {
+                        phase: "sources_resolved",
+                        progress_total: sources.length,
+                    });
                     const results = await ingestResolvedSources(sources, options, {
                         client,
                         projectSlug: config.projectSlug,
                         outputDir,
                         runParser: runKnowledgeParser,
+                        eventLogger,
                         uploadKnowledgeFile: (remotePath, localPath) => putRemoteFileFromLocal(client, config.projectSlug, remotePath, localPath),
                     });
+                    eventLogger?.info("Completed knowledge ingest", {
+                        phase: "completed",
+                        progress_current: results.length,
+                        progress_total: results.length,
+                    });
                     if (options.json) {
                         printJson(results);
                         return;
@@ -365,6 +488,12 @@ export async function knowledgeCommand(args) {
                         }
                     }
                 }
+                catch (error) {
+                    eventLogger?.error("Knowledge ingest failed", {
+                        phase: "failed",
+                    });
+                    throw error;
+                }
                 finally {
                     if (shouldCleanupOutputDir) {
                         await Promise.all([

package/esm/cli/commands/knowledge/parser-source.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"parser-source.d.ts","sourceRoot":"","sources":["../../../../src/cli/commands/knowledge/parser-source.ts"],"names":[],"mappings":"AAAA,eAAO,MAAM,2BAA2B,QA8ZvC,CAAC"}
+{"version":3,"file":"parser-source.d.ts","sourceRoot":"","sources":["../../../../src/cli/commands/knowledge/parser-source.ts"],"names":[],"mappings":"AAAA,eAAO,MAAM,2BAA2B,QAugBvC,CAAC"}

package/esm/cli/commands/knowledge/parser-source.js

@@ -3,6 +3,7 @@ import argparse
 import csv
 import json
 import re
+import subprocess
 from datetime import date
 from pathlib import Path
 from typing import Any, Optional
@@ -71,6 +72,107 @@ def build_frontmatter(source: str, source_type: str, description: str) -> str:
     ])
 
 
+def metadata_int(metadata: dict[str, Any], *keys: str) -> Optional[int]:
+    for key in keys:
+        value = metadata.get(key)
+        if isinstance(value, int) and not isinstance(value, bool):
+            return value
+    return None
+
+
+def metadata_string_list(metadata: dict[str, Any], *keys: str) -> Optional[list[str]]:
+    for key in keys:
+        value = metadata.get(key)
+        if isinstance(value, list) and all(isinstance(item, str) for item in value):
+            return value
+    return None
+
+
+def build_kreuzberg_stats(source_type: str, content: str, metadata: dict[str, Any]):
+    stats: dict[str, Any] = {
+        "characters": len(content),
+        "lines": len(content.splitlines()) if content else 0,
+        "engine": "kreuzberg",
+    }
+
+    if isinstance(metadata.get("mime_type"), str):
+        stats["mime_type"] = metadata["mime_type"]
+
+    if source_type == "pdf":
+        stats["pages"] = metadata_int(metadata, "page_count") or 0
+        stats["tables"] = metadata_int(metadata, "table_count") or 0
+    elif source_type in {"xlsx", "xls"}:
+        stats["sheets"] = metadata_int(metadata, "sheet_count") or 0
+        stats["rows"] = metadata_int(metadata, "row_count") or 0
+        stats["sheet_names"] = metadata_string_list(metadata, "sheet_names") or []
+    elif source_type == "docx":
+        stats["paragraphs"] = metadata_int(metadata, "paragraph_count") or 0
+        stats["tables"] = metadata_int(metadata, "table_count") or 0
+    elif source_type == "pptx":
+        stats["slides"] = metadata_int(metadata, "slide_count", "page_count") or 0
+        stats["tables"] = metadata_int(metadata, "table_count") or 0
+    elif source_type == "html":
+        stats["tables"] = metadata_int(metadata, "table_count") or 0
+
+    return stats
+
+
+def parse_with_kreuzberg(path: str, source_type: str):
+    warnings: list[str] = []
+    completed = subprocess.run(
+        [
+            "kreuzberg",
+            "extract",
+            path,
+            "--format",
+            "json",
+            "--output-format",
+            "markdown",
+        ],
+        capture_output=True,
+        text=True,
+        check=False,
+    )
+
+    if completed.returncode != 0:
+        detail = completed.stderr.strip() or completed.stdout.strip() or f"exit code {completed.returncode}"
+        raise RuntimeError(f"kreuzberg extract failed: {detail}")
+
+    try:
+        payload = json.loads(completed.stdout)
+    except json.JSONDecodeError as error:
+        raise RuntimeError(f"kreuzberg extract returned invalid JSON: {error}") from error
+
+    content = payload.get("content", "")
+    if not isinstance(content, str):
+        raise RuntimeError("kreuzberg extract did not return string content")
+
+    metadata = payload.get("metadata") if isinstance(payload.get("metadata"), dict) else {}
+    normalized_content = clean_text(content)
+    stats = build_kreuzberg_stats(source_type, normalized_content, metadata)
+
+    return normalized_content or "_No extractable text found in document._", stats, warnings
+
+
+def prefer_kreuzberg(source_type: str, fallback_parser):
+    def parser(path: str):
+        try:
+            return parse_with_kreuzberg(path, source_type)
+        except FileNotFoundError as error:
+            if getattr(error, "filename", "") == "kreuzberg":
+                return fallback_parser(path)
+            raise
+        except RuntimeError as error:
+            content, stats, warnings = fallback_parser(path)
+            warnings.append(
+                "kreuzberg extraction failed; fell back to the built-in parser: "
+                + str(error)
+            )
+            return content, stats, warnings
+
+    return parser
+
+
 def parse_csv_like(path: str, delimiter: str = ","):
     warnings: list[str] = []
     with open(path, newline="", encoding="utf-8-sig") as file:
@@ -305,18 +407,19 @@ def parse_json(path: str):
 def select_parser(path: Path):
     ext = path.suffix.lower()
     if ext == ".pdf":
-        return "pdf", parse_pdf
+        return "pdf", prefer_kreuzberg("pdf", parse_pdf)
     if ext in {".csv", ".tsv"}:
         delimiter = "\t" if ext == ".tsv" else ","
         return ext.lstrip("."), lambda file_path: parse_csv_like(file_path, delimiter)
     if ext in {".xlsx", ".xls"}:
-        return ext.lstrip("."), parse_excel
+        source_type = ext.lstrip(".")
+        return source_type, prefer_kreuzberg(source_type, parse_excel)
     if ext == ".docx":
-        return "docx", parse_docx
+        return "docx", prefer_kreuzberg("docx", parse_docx)
     if ext == ".pptx":
-        return "pptx", parse_pptx
+        return "pptx", prefer_kreuzberg("pptx", parse_pptx)
     if ext in {".html", ".htm"}:
-        return "html", parse_html
+        return "html", prefer_kreuzberg("html", parse_html)
     if ext in {".txt", ".md", ".mdx"}:
         return ext.lstrip("."), parse_text
     if ext == ".json":
@@ -325,6 +428,8 @@ def select_parser(path: Path):
 
 
 def build_summary(source_type: str, stats: dict[str, Any]) -> str:
+    if stats.get("engine") == "kreuzberg":
+        return f"Converted {source_type.upper()} to markdown ({stats.get('characters', 0)} chars)."
     if source_type in {"csv", "tsv"}:
         return f"Parsed {stats.get('rows', 0)} rows across {stats.get('columns', 0)} columns."
     if source_type in {"xlsx", "xls"}:

package/esm/deno.d.ts

@@ -24,6 +24,7 @@ declare namespace _default {
         "./resource": string;
         "./mcp": string;
         "./middleware": string;
+        "./utils": string;
         "./oauth": string;
         "./provider": string;
         "./fs": string;
@@ -61,6 +62,7 @@ declare namespace _default {
         "veryfront/workflow/claude-code": string;
         "veryfront/workflow/claude-code/react": string;
         "veryfront/workflow/discovery": string;
+        "veryfront/utils": string;
         "veryfront/utils/box": string;
         "veryfront/utils/case-utils": string;
         "veryfront/utils/constants/server": string;

package/esm/deno.js

@@ -1,6 +1,6 @@
 export default {
     "name": "veryfront",
-    "version": "0.1.73",
+    "version": "0.1.75",
     "license": "Apache-2.0",
     "nodeModulesDir": "auto",
     "exclude": [
@@ -34,6 +34,7 @@ export default {
         "./resource": "./src/resource/index.ts",
         "./mcp": "./src/mcp/index.ts",
         "./middleware": "./src/middleware/index.ts",
+        "./utils": "./src/utils/index.ts",
         "./oauth": "./src/oauth/index.ts",
         "./provider": "./src/provider/index.ts",
         "./fs": "./src/fs/index.ts",
@@ -71,6 +72,7 @@ export default {
         "veryfront/workflow/claude-code": "./src/workflow/claude-code/index.ts",
         "veryfront/workflow/claude-code/react": "./src/workflow/claude-code/react/index.ts",
         "veryfront/workflow/discovery": "./src/workflow/discovery/index.ts",
+        "veryfront/utils": "./src/utils/index.ts",
         "veryfront/utils/box": "./src/utils/box.ts",
         "veryfront/utils/case-utils": "./src/utils/case-utils.ts",
         "veryfront/utils/constants/server": "./src/utils/constants/server.ts",

package/esm/src/data/data-fetcher.d.ts

@@ -1,11 +1,21 @@
 import type { DataContext, DataResult, PageWithData, StaticPathsResult } from "./types.js";
+/**
+ * Options for isolated data fetching. Passed through to ServerDataFetcher
+ * when worker isolation is enabled.
+ */
+export interface FetchDataOptions {
+    /** Absolute path to the module containing getServerData */
+    modulePath?: string;
+    /** Project directory for worker scoping */
+    projectDir?: string;
+}
 export declare class DataFetcher {
     private cacheManager;
     private serverFetcher;
     private staticFetcher;
     private pathsFetcher;
     constructor(_adapter?: unknown);
-    fetchData(pageModule: PageWithData, context: DataContext, mode?: "development" | "production"): Promise<DataResult>;
+    fetchData(pageModule: PageWithData, context: DataContext, mode?: "development" | "production", options?: FetchDataOptions): Promise<DataResult>;
     getStaticPaths(pageModule: PageWithData): Promise<StaticPathsResult | null>;
     clearCache(pattern?: string): void;
 }

package/esm/src/data/data-fetcher.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"data-fetcher.d.ts","sourceRoot":"","sources":["../../../src/src/data/data-fetcher.ts"],"names":[],"mappings":"AAMA,OAAO,KAAK,EAAE,WAAW,EAAE,UAAU,EAAE,YAAY,EAAE,iBAAiB,EAAE,MAAM,YAAY,CAAC;AAE3F,qBAAa,WAAW;IACtB,OAAO,CAAC,YAAY,CAAe;IACnC,OAAO,CAAC,aAAa,CAAoB;IACzC,OAAO,CAAC,aAAa,CAAoB;IACzC,OAAO,CAAC,YAAY,CAAqB;gBAE7B,QAAQ,CAAC,EAAE,OAAO;IAO9B,SAAS,CACP,UAAU,EAAE,YAAY,EACxB,OAAO,EAAE,WAAW,EACpB,IAAI,GAAE,aAAa,GAAG,YAA4B,GACjD,OAAO,CAAC,UAAU,CAAC;IA0BtB,cAAc,CAAC,UAAU,EAAE,YAAY,GAAG,OAAO,CAAC,iBAAiB,GAAG,IAAI,CAAC;IAI3E,UAAU,CAAC,OAAO,CAAC,EAAE,MAAM,GAAG,IAAI;CAQnC"}
+{"version":3,"file":"data-fetcher.d.ts","sourceRoot":"","sources":["../../../src/src/data/data-fetcher.ts"],"names":[],"mappings":"AAMA,OAAO,KAAK,EAAE,WAAW,EAAE,UAAU,EAAE,YAAY,EAAE,iBAAiB,EAAE,MAAM,YAAY,CAAC;AAE3F;;;GAGG;AACH,MAAM,WAAW,gBAAgB;IAC/B,2DAA2D;IAC3D,UAAU,CAAC,EAAE,MAAM,CAAC;IACpB,2CAA2C;IAC3C,UAAU,CAAC,EAAE,MAAM,CAAC;CACrB;AAED,qBAAa,WAAW;IACtB,OAAO,CAAC,YAAY,CAAe;IACnC,OAAO,CAAC,aAAa,CAAoB;IACzC,OAAO,CAAC,aAAa,CAAoB;IACzC,OAAO,CAAC,YAAY,CAAqB;gBAE7B,QAAQ,CAAC,EAAE,OAAO;IAO9B,SAAS,CACP,UAAU,EAAE,YAAY,EACxB,OAAO,EAAE,WAAW,EACpB,IAAI,GAAE,aAAa,GAAG,YAA4B,EAClD,OAAO,CAAC,EAAE,gBAAgB,GACzB,OAAO,CAAC,UAAU,CAAC;IA8BtB,cAAc,CAAC,UAAU,EAAE,YAAY,GAAG,OAAO,CAAC,iBAAiB,GAAG,IAAI,CAAC;IAI3E,UAAU,CAAC,OAAO,CAAC,EAAE,MAAM,GAAG,IAAI;CAQnC"}