aiex-cli 0.0.2 → 0.0.3-beta.2

package/dist/cli.mjs

@@ -1,4 +1,4 @@
-import { S as version, T as formatDoctorDiagnosticsJson, _ as createConfig, a as parseJsonSchema, b as name, c as getDefaultAIConfig, d as DEFAULT_MARKITDOWN_CONFIG, f as DEFAULT_MINERU_CONFIG, g as AIConfigSchema, h as PLACEHOLDER_TEXT, i as JsonSchemaDefinitionSchema, l as readAIConfig, m as PLACEHOLDER_SCHEMA, n as createMigrationConfig, o as toSnakeCase, p as DEFAULT_PROMPT_CONFIG, s as generateDrizzleSchema, t as collectDoctorDiagnostics, u as writeAIConfig, v as seedConfig, w as doctorDiagnosticsTableRows, x as package_default, y as description } from "./doctor-collector-DR7s0UUh.mjs";
+import { S as version, T as formatDoctorDiagnosticsJson, _ as createConfig, a as parseJsonSchema, b as name, c as getDefaultAIConfig, d as DEFAULT_MARKITDOWN_CONFIG, f as DEFAULT_MINERU_CONFIG, g as AIConfigSchema, h as PLACEHOLDER_TEXT, i as JsonSchemaDefinitionSchema, l as readAIConfig, m as PLACEHOLDER_SCHEMA, n as createMigrationConfig, o as toSnakeCase, p as DEFAULT_PROMPT_CONFIG, s as generateDrizzleSchema, t as collectDoctorDiagnostics, u as writeAIConfig, v as seedConfig, w as doctorDiagnosticsTableRows, x as package_default, y as description } from "./doctor-collector-Bnkbl48V.mjs";
 import { createRequire } from "node:module";
 import fs from "node:fs/promises";
 import os from "node:os";
@@ -13388,6 +13388,8 @@ async function savePromptSnapshot(schema, aiexDir) {
 //#region src/core/extraction-audit.ts
 const AUDIT_ID_RE = /^[\w.-]+$/;
 const STALE_AFTER_MS = 1800 * 1e3;
+const CACHE_TTL_MS = 5e3;
+const recordCache = /* @__PURE__ */ new Map();
 function auditDir(aiexDir) {
 	return path.join(aiexDir, "extracted", "_audit");
 }
@@ -13414,6 +13416,7 @@ async function createExtractionAuditRecord(aiexDir, input) {
 		spaces: 2,
 		EOL: "\n"
 	});
+	clearRecordCache(aiexDir);
 	return record;
 }
 async function updateExtractionAuditRecord(aiexDir, id, patch) {
@@ -13430,6 +13433,7 @@ async function updateExtractionAuditRecord(aiexDir, id, patch) {
 		spaces: 2,
 		EOL: "\n"
 	});
+	clearRecordCache(aiexDir);
 	return record;
 }
 async function readExtractionAuditRecord(aiexDir, id) {
@@ -13457,19 +13461,38 @@ async function markStaleIfNeeded(aiexDir, record) {
 		spaces: 2,
 		EOL: "\n"
 	});
+	clearRecordCache(aiexDir);
 	return staleRecord;
 }
+function getCachedRecords(aiexDir) {
+	const entry = recordCache.get(aiexDir);
+	if (entry && Date.now() - entry.timestamp < CACHE_TTL_MS) return entry.records;
+	return null;
+}
+function setCachedRecords(aiexDir, records) {
+	recordCache.set(aiexDir, {
+		records,
+		timestamp: Date.now()
+	});
+}
+function clearRecordCache(aiexDir) {
+	recordCache.delete(aiexDir);
+}
 async function listExtractionAuditRecords(aiexDir) {
+	const cached = getCachedRecords(aiexDir);
+	if (cached) return cached;
 	try {
 		const dir = auditDir(aiexDir);
 		const files = await fs.readdir(dir);
-		return (await Promise.all(files.filter((file) => file.endsWith(".json")).map(async (file) => {
+		const result = (await Promise.all(files.filter((file) => file.endsWith(".json")).map(async (file) => {
 			try {
 				return await markStaleIfNeeded(aiexDir, await readFile(path.join(dir, file)));
 			} catch {
 				return null;
 			}
 		}))).filter((record) => !!record).sort((a, b) => b.createdAt.localeCompare(a.createdAt));
+		setCachedRecords(aiexDir, result);
+		return result;
 	} catch {
 		return [];
 	}
@@ -13486,9 +13509,77 @@ async function deleteExtractionAuditRecord(aiexDir, id) {
 	const uploadFiles = await fs.readdir(uploadsDir).catch(() => []);
 	await Promise.all(uploadFiles.filter((file) => file.startsWith(`${id}-`)).map((file) => fs.unlink(path.join(uploadsDir, file)).catch(() => {})));
 	await fs.unlink(auditPath(aiexDir, id)).catch(() => {});
+	clearRecordCache(aiexDir);
 	return true;
 }
 
+//#endregion
+//#region src/core/file-constants.ts
+const MAX_UPLOAD_SIZE = 150 * 1024 * 1024;
+const MAX_UPLOAD_SIZE_TEXT = "150MB";
+const SUPPORTED_FILE_TYPES_TEXT = "images, PDF, text, markdown, CSV, JSON, HTML, XML, YAML";
+const MISSING_UPLOAD_FILE_TEXT = "Uploaded file is no longer available. Re-run extraction with the original file.";
+const SUPPORTED_MIME_TYPES = new Set([
+	"image/png",
+	"image/jpeg",
+	"image/gif",
+	"image/webp",
+	"image/bmp",
+	"image/svg+xml",
+	"application/pdf",
+	"text/plain",
+	"text/markdown",
+	"text/csv",
+	"application/json",
+	"text/html",
+	"text/xml",
+	"application/x-yaml",
+	"text/yaml"
+]);
+const MIME_TO_EXT = {
+	"image/png": "png",
+	"image/jpeg": "jpg",
+	"image/gif": "gif",
+	"image/webp": "webp",
+	"image/bmp": "bmp",
+	"image/svg+xml": "svg",
+	"application/pdf": "pdf",
+	"text/plain": "txt",
+	"text/markdown": "md",
+	"text/csv": "csv",
+	"application/json": "json",
+	"text/html": "html",
+	"text/xml": "xml",
+	"application/x-yaml": "yaml",
+	"text/yaml": "yaml"
+};
+function bytesToMB(bytes) {
+	return bytes / (1024 * 1024);
+}
+function getExtensionFromMime(mimeType) {
+	return MIME_TO_EXT[mimeType];
+}
+function isAllowedMimeType(mimeType) {
+	return SUPPORTED_MIME_TYPES.has(mimeType);
+}
+function unsupportedFileTypeMessage(mimeType) {
+	return `Unsupported file type "${mimeType}". Supported: ${SUPPORTED_FILE_TYPES_TEXT}.`;
+}
+function isMissingUploadFileError(error) {
+	return !!error && typeof error === "object" && error.code === "ENOENT";
+}
+var FileValidationError = class extends Error {
+	constructor(message) {
+		super(message);
+		this.name = "FileValidationError";
+	}
+};
+function validateFileUpload(file) {
+	if (file.size === 0) throw new FileValidationError("Uploaded file is empty");
+	if (file.size > MAX_UPLOAD_SIZE) throw new FileValidationError(`File size (${bytesToMB(file.size).toFixed(1)}MB) exceeds ${MAX_UPLOAD_SIZE_TEXT} limit`);
+	if (!isAllowedMimeType(file.type)) throw new FileValidationError(unsupportedFileTypeMessage(file.type));
+}
+
 //#endregion
 //#region src/core/notion-sink.ts
 const RICH_TEXT_LIMIT = 2e3;
@@ -13632,34 +13723,34 @@ function suggestFieldMap(schemaFields, databaseProperties) {
 	}
 	return fieldMap;
 }
-function hasProperties(value) {
-	return !!value && typeof value === "object" && !!value.properties && typeof value.properties === "object";
+function isDataSourceResponse(value) {
+	return !!value && typeof value === "object" && typeof value.properties === "object" && !Array.isArray(value);
 }
 function firstDataSourceId(database) {
-	return (Array.isArray(database?.data_sources) ? database.data_sources : []).find((source) => typeof source?.id === "string" && source.id.trim())?.id;
+	return database.data_sources?.find((source) => typeof source.id === "string" && source.id.trim())?.id;
 }
 async function resolveNotionDataSource(notion, inputId) {
 	const id = parseNotionDatabaseId(inputId);
 	if (!id) throw new Error("Notion database or data source URL/ID is required.");
 	try {
 		const dataSource$1 = await notion.dataSources.retrieve({ data_source_id: id });
-		if (hasProperties(dataSource$1)) return {
+		if (isDataSourceResponse(dataSource$1)) return {
 			databaseId: typeof dataSource$1.parent?.database_id === "string" ? dataSource$1.parent.database_id : id,
-			dataSourceId: dataSource$1.id ?? id,
+			dataSourceId: dataSource$1.id,
 			properties: dataSource$1.properties,
-			parent: { data_source_id: dataSource$1.id ?? id }
+			parent: { data_source_id: dataSource$1.id }
 		};
 	} catch {}
 	const database = await notion.databases.retrieve({ database_id: id });
 	const dataSourceId = firstDataSourceId(database);
 	if (!dataSourceId) throw new Error("No data source found for this Notion database. Copy the data source link from Notion, or share the source database with the integration.");
 	const dataSource = await notion.dataSources.retrieve({ data_source_id: dataSourceId });
-	if (!hasProperties(dataSource)) throw new Error("Notion data source did not return properties. Make sure it is shared with the integration and is not a linked data source.");
+	if (!isDataSourceResponse(dataSource)) throw new Error("Notion data source did not return properties. Make sure it is shared with the integration and is not a linked data source.");
 	return {
-		databaseId: database.id ?? id,
-		dataSourceId: dataSource.id ?? dataSourceId,
+		databaseId: database.id,
+		dataSourceId: dataSource.id,
 		properties: dataSource.properties,
-		parent: { data_source_id: dataSource.id ?? dataSourceId }
+		parent: { data_source_id: dataSource.id }
 	};
 }
 async function inspectNotionDatabase(input) {
@@ -13988,6 +14079,8 @@ async function listSchemas(aiexDir) {
 	}
 }
 async function readExtractFileInput(filePath, aiConfig) {
+	const stat = fs$1.statSync(filePath);
+	if (stat.size > MAX_UPLOAD_SIZE) throw new Error(`File size (${bytesToMB(stat.size).toFixed(1)}MB) exceeds ${MAX_UPLOAD_SIZE_TEXT} limit: ${filePath}`);
 	const ext = path.extname(filePath).toLowerCase().replace(".", "");
 	if (FILE_PART_EXTENSIONS.has(ext)) return {
 		text: "",
@@ -14383,6 +14476,10 @@ const retryCommand = defineCommand({
 			}
 			outro("Done!");
 		} catch (error) {
+			if (isMissingUploadFileError(error)) {
+				failCommand(MISSING_UPLOAD_FILE_TEXT);
+				return;
+			}
 			failCommand(error instanceof Error ? error.message : String(error));
 		}
 	}
@@ -14435,7 +14532,7 @@ const extractCommand = defineCommand({
 		file: {
 			type: "string",
 			alias: "f",
-			description: "File path (image/PDF) to extract from"
+			description: `File path to extract from. Supported: ${SUPPORTED_FILE_TYPES_TEXT}.`
 		},
 		model: {
 			type: "string",
@@ -15252,94 +15349,56 @@ function getFormFile(value) {
 function safeUploadName(name$1) {
 	return path.basename(name$1).replace(/[^\w.-]/g, "_") || "upload.txt";
 }
+function safeUploadNameForMime(file) {
+	const safeName = safeUploadName(file.name);
+	const ext = getExtensionFromMime(file.type);
+	if (!ext) throw new FileValidationError(unsupportedFileTypeMessage(file.type));
+	return `${path.parse(safeName).name || "upload"}.${ext}`;
+}
+function jsonResponse(body, status) {
+	return new Response(JSON.stringify(body), {
+		status,
+		headers: { "content-type": "application/json" }
+	});
+}
+async function auditFailureResponse(aiexDir, auditId, error, status) {
+	const record = await updateExtractionAuditRecord(aiexDir, auditId, {
+		status: "failed",
+		error
+	});
+	return jsonResponse({
+		success: false,
+		error: record.error,
+		auditId: record.id
+	}, status);
+}
 async function saveUploadToFile(file, uploadsDir, id) {
+	validateFileUpload(file);
 	await fs.mkdir(uploadsDir, { recursive: true });
-	const filePath = path.join(uploadsDir, `${id}-${safeUploadName(file.name)}`);
+	const filePath = path.join(uploadsDir, `${id}-${safeUploadNameForMime(file)}`);
 	const buffer = Buffer.from(await file.arrayBuffer());
 	await fs.writeFile(filePath, buffer);
 	return filePath;
 }
 async function executeAuditedExtraction(input) {
 	const aiConfig = await readAIConfig(input.aiexDir);
-	if (!aiConfig) {
-		const record$1 = await updateExtractionAuditRecord(input.aiexDir, input.auditId, {
-			status: "failed",
-			error: "AI configuration not found. Configure AI settings first."
-		});
-		return new Response(JSON.stringify({
-			success: false,
-			error: record$1.error,
-			auditId: record$1.id
-		}), {
-			status: 400,
-			headers: { "content-type": "application/json" }
-		});
-	}
-	if (!aiConfig.provider.apiKey) {
-		const record$1 = await updateExtractionAuditRecord(input.aiexDir, input.auditId, {
-			status: "failed",
-			error: "API Key not configured. Configure AI settings first."
-		});
-		return new Response(JSON.stringify({
-			success: false,
-			error: record$1.error,
-			auditId: record$1.id
-		}), {
-			status: 400,
-			headers: { "content-type": "application/json" }
-		});
-	}
-	if (!aiConfig.provider.models?.length) {
-		const record$1 = await updateExtractionAuditRecord(input.aiexDir, input.auditId, {
-			status: "failed",
-			error: "No models configured. Add at least one model in AI Settings."
-		});
-		return new Response(JSON.stringify({
-			success: false,
-			error: record$1.error,
-			auditId: record$1.id
-		}), {
-			status: 400,
-			headers: { "content-type": "application/json" }
-		});
-	}
+	if (!aiConfig) return auditFailureResponse(input.aiexDir, input.auditId, "AI configuration not found. Configure AI settings first.", 400);
+	if (!aiConfig.provider.apiKey) return auditFailureResponse(input.aiexDir, input.auditId, "API Key not configured. Configure AI settings first.", 400);
+	if (!aiConfig.provider.models?.length) return auditFailureResponse(input.aiexDir, input.auditId, "No models configured. Add at least one model in AI Settings.", 400);
 	const modelOverride = input.modelName ? aiConfig.provider.models.find((model) => model.name === input.modelName) : void 0;
-	if (input.modelName && !modelOverride) {
-		const record$1 = await updateExtractionAuditRecord(input.aiexDir, input.auditId, {
-			status: "failed",
-			error: `Model "${input.modelName}" not found in AI settings`
-		});
-		return new Response(JSON.stringify({
-			success: false,
-			error: record$1.error,
-			auditId: record$1.id
-		}), {
-			status: 400,
-			headers: { "content-type": "application/json" }
-		});
-	}
+	if (input.modelName && !modelOverride) return auditFailureResponse(input.aiexDir, input.auditId, `Model "${input.modelName}" not found in AI settings`, 400);
 	let inputText = input.text;
 	let inputFilePath = input.filePath;
-	if (input.filePath) {
+	if (input.filePath) try {
 		const source = await readExtractFileInput(input.filePath, aiConfig);
 		inputText = source.text;
 		inputFilePath = source.filePath;
+	} catch (error) {
+		if (isMissingUploadFileError(error)) return auditFailureResponse(input.aiexDir, input.auditId, MISSING_UPLOAD_FILE_TEXT, 400);
+		throw error;
 	}
 	const result = await extractSingle(input.aiexDir, input.config, aiConfig, input.schemaName, inputText, inputFilePath, modelOverride, { quiet: true });
-	if (!result.success) {
-		const record$1 = await updateExtractionAuditRecord(input.aiexDir, input.auditId, {
-			status: "failed",
-			error: result.error || "Extraction failed"
-		});
-		return new Response(JSON.stringify({
-			success: false,
-			error: record$1.error,
-			auditId: record$1.id
-		}), {
-			status: 500,
-			headers: { "content-type": "application/json" }
-		});
-	}
+	if (!result.success) return auditFailureResponse(input.aiexDir, input.auditId, result.error || "Extraction failed", 500);
 	const notionPages = [];
 	if (aiConfig.notion?.enabled && aiConfig.notion.schemas?.[input.schemaName]?.databaseId?.trim()) try {
 		if (!result.data || typeof result.data !== "object" || Array.isArray(result.data)) throw new Error("Extraction result is not an object and cannot be written to Notion.");
@@ -15353,14 +15412,11 @@ async function executeAuditedExtraction(input) {
 			tokensUsed: result.tokensUsed,
 			error: error instanceof Error ? error.message : String(error)
 		});
-		return new Response(JSON.stringify({
+		return jsonResponse({
 			success: false,
 			error: record$1.error,
 			auditId: record$1.id
-		}), {
-			status: 500,
-			headers: { "content-type": "application/json" }
-		});
+		}, 500);
 	}
 	const record = await updateExtractionAuditRecord(input.aiexDir, input.auditId, {
 		status: "succeeded",
@@ -15370,7 +15426,7 @@ async function executeAuditedExtraction(input) {
 		notionPages: notionPages.length > 0 ? notionPages : void 0,
 		tokensUsed: result.tokensUsed
 	});
-	return new Response(JSON.stringify({
+	return jsonResponse({
 		success: true,
 		outputPath: record.outputPath,
 		outputName: record.outputName,
@@ -15378,10 +15434,7 @@ async function executeAuditedExtraction(input) {
 		notionPages: record.notionPages,
 		tokensUsed: record.tokensUsed,
 		auditId: record.id
-	}), {
-		status: 200,
-		headers: { "content-type": "application/json" }
-	});
+	}, 200);
 }
 function extractRoutes(config) {
 	const app = new Hono();
@@ -15422,11 +15475,26 @@ function extractRoutes(config) {
 			});
 			let filePath;
 			if (file) {
-				filePath = await saveUploadToFile(file, uploadsDir, audit.id);
+				try {
+					filePath = await saveUploadToFile(file, uploadsDir, audit.id);
+				} catch (e) {
+					if (e instanceof FileValidationError) {
+						await updateExtractionAuditRecord(aiexDir, audit.id, {
+							status: "failed",
+							error: e.message
+						});
+						return c.json({
+							success: false,
+							error: e.message,
+							auditId: audit.id
+						}, 400);
+					}
+					throw e;
+				}
 				await updateExtractionAuditRecord(aiexDir, audit.id, { source: {
 					type: "file",
 					filePath,
-					fileName: safeUploadName(file.name)
+					fileName: path.basename(filePath)
 				} });
 			}
 			return executeAuditedExtraction({

package/dist/{doctor-collector-DR7s0UUh.mjs → doctor-collector-Bnkbl48V.mjs}

@@ -65,7 +65,7 @@ function doctorDiagnosticsTableRows(d) {
 //#endregion
 //#region package.json
 var name = "aiex-cli";
-var version = "0.0.2";
+var version = "0.0.3-beta.2";
 var description = "JSON Schema → SQLite with AI-powered data extraction";
 var package_default = {
 	name,

package/dist/index.mjs

@@ -1,3 +1,3 @@
-import { C as buildDoctorDiagnostics, T as formatDoctorDiagnosticsJson, a as parseJsonSchema, i as JsonSchemaDefinitionSchema, n as createMigrationConfig, r as generateDrizzleConfig, s as generateDrizzleSchema, t as collectDoctorDiagnostics, w as doctorDiagnosticsTableRows } from "./doctor-collector-DR7s0UUh.mjs";
+import { C as buildDoctorDiagnostics, T as formatDoctorDiagnosticsJson, a as parseJsonSchema, i as JsonSchemaDefinitionSchema, n as createMigrationConfig, r as generateDrizzleConfig, s as generateDrizzleSchema, t as collectDoctorDiagnostics, w as doctorDiagnosticsTableRows } from "./doctor-collector-Bnkbl48V.mjs";
 
 export { JsonSchemaDefinitionSchema, buildDoctorDiagnostics, collectDoctorDiagnostics, createMigrationConfig, doctorDiagnosticsTableRows, formatDoctorDiagnosticsJson, generateDrizzleConfig, generateDrizzleSchema, parseJsonSchema };

package/package.json

@@ -1,7 +1,7 @@
 {
   "name": "aiex-cli",
   "type": "module",
-  "version": "0.0.2",
+  "version": "0.0.3-beta.2",
   "description": "JSON Schema → SQLite with AI-powered data extraction",
   "author": "OSpoon <zxin088@gmail.com>",
   "license": "MIT",