aiex-cli 0.0.2-beta.9 → 0.0.3-beta.1

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-Cb9X2mxU.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-CjFTz8p4.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 [];
 	}
@@ -13486,6 +13509,7 @@ 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;
 }
 
@@ -13632,34 +13656,34 @@ 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) {
@@ -13987,7 +14011,12 @@ async function listSchemas(aiexDir) {
 		return [];
 	}
 }
+function getFileSizeMB(filePath) {
+	return fs$1.statSync(filePath).size / (1024 * 1024);
+}
 async function readExtractFileInput(filePath, aiConfig) {
+	const sizeMB = getFileSizeMB(filePath);
+	if (sizeMB > 150) throw new Error(`File size (${sizeMB.toFixed(1)}MB) exceeds 150MB limit: ${filePath}`);
 	const ext = path.extname(filePath).toLowerCase().replace(".", "");
 	if (FILE_PART_EXTENSIONS.has(ext)) return {
 		text: "",
@@ -14964,6 +14993,7 @@ function aiRoutes(config) {
 //#region src/server/routes/data.ts
 const FILE_REGEX = /\.json$/;
 const EXTRACTION_TIMESTAMP_RE = /-\d{4}-\d{2}-\d{2}T/;
+const INTERNAL_ROWID_COLUMN = "__aiex_rowid";
 const TIMESTAMP_CLEANUP = /(\d{2})-(\d{2})-(\d{2})/;
 const TIMESTAMP_TZ = /(\d{3})Z/;
 const tableParamSchema = z.object({ name: z.string().regex(/^[a-z][a-z0-9_]*$/) });
@@ -14980,6 +15010,31 @@ function invalidParamResponse$1(message) {
 		if (!result.success) return c.json({ error: message }, 400);
 	};
 }
+function getAuditNotionStatus(record) {
+	if (record.notionPages?.length) return "synced";
+	if (record.status === "failed") return "failed";
+	return "not_synced";
+}
+async function getRowExtractionActions(aiexDir, tableName) {
+	const actions = /* @__PURE__ */ new Map();
+	const auditRecords = await listExtractionAuditRecords(aiexDir);
+	for (const record of auditRecords) {
+		if (!record.outputName) continue;
+		for (const inserted of record.tablesInserted ?? []) {
+			if (inserted.table !== tableName) continue;
+			const key = String(inserted.rowId);
+			if (actions.has(key)) continue;
+			const notionPages = record.notionPages?.length ? record.notionPages : void 0;
+			actions.set(key, {
+				extractionName: record.outputName,
+				notionStatus: getAuditNotionStatus(record),
+				notionPages,
+				notionError: !notionPages && record.status === "failed" ? record.error : void 0
+			});
+		}
+	}
+	return actions;
+}
 function schemaNameFromExtractionFile(name$1) {
 	const stem = name$1.replace(FILE_REGEX, "");
 	const match = stem.match(EXTRACTION_TIMESTAMP_RE);
@@ -14997,22 +15052,27 @@ function dataRoutes(config) {
 		try {
 			await fs.mkdir(extractedDir, { recursive: true });
 			const jsonFiles = (await fs.readdir(extractedDir)).filter((f) => f.endsWith(".json") && !f.endsWith(".prompt.md"));
+			const auditRecords = await listExtractionAuditRecords(aiexDir);
+			const auditByOutputName = new Map(auditRecords.map((record) => [record.outputName, record]));
 			const records = [];
 			for (const file of jsonFiles) {
-				const name$1 = file.replace(FILE_REGEX, "");
-				const idx = name$1.lastIndexOf("-");
-				if (idx === -1) continue;
-				const schemaName = name$1.slice(0, idx);
-				const timestamp = name$1.slice(idx + 1).replace(/-/g, (d, i) => i === 4 || i === 7 ? "-" : d).replace(TIMESTAMP_CLEANUP, (_, h, m, s) => `${h}:${m}:${s}`).replace(TIMESTAMP_TZ, ".$1Z");
+				const schemaName = schemaNameFromExtractionFile(file);
+				if (!schemaName) continue;
+				const timestamp = file.replace(FILE_REGEX, "").slice(schemaName.length + 1).replace(/-/g, (d, i) => i === 4 || i === 7 ? "-" : d).replace(TIMESTAMP_CLEANUP, (_, h, m, s) => `${h}:${m}:${s}`).replace(TIMESTAMP_TZ, ".$1Z");
 				const filePath = path.join(extractedDir, file);
 				try {
 					const stat = await fs.stat(filePath);
+					const audit = auditByOutputName.get(file);
+					const notionPages = audit?.notionPages?.length ? audit.notionPages : void 0;
 					records.push({
 						name: file,
 						schemaName,
 						timestamp,
 						fileSize: stat.size,
-						modifiedAt: stat.mtime.toISOString()
+						modifiedAt: stat.mtime.toISOString(),
+						notionStatus: notionPages ? "synced" : audit?.status === "failed" ? "failed" : "not_synced",
+						notionPages,
+						notionError: !notionPages && audit?.status === "failed" ? audit.error : void 0
 					});
 				} catch {
 					continue;
@@ -15099,16 +15159,24 @@ function dataRoutes(config) {
 			const offset = (page - 1) * pageSize;
 			const totalPages = Math.max(1, Math.ceil(total / pageSize));
 			const result = await sql`
-          select *
+          select rowid as ${sql.raw(INTERNAL_ROWID_COLUMN)}, *
           from ${sql.table(tableName)}
           ${searchCondition}
           ${orderBy}
           limit ${pageSize}
           offset ${offset}
         `.execute(db);
+			const actionsByRowId = await getRowExtractionActions(aiexDir, tableName);
+			const rowActions = Object.fromEntries(result.rows.map((row, index) => {
+				const rowId = row[INTERNAL_ROWID_COLUMN];
+				const action = rowId === null || rowId === void 0 ? void 0 : actionsByRowId.get(String(rowId));
+				return action ? [String(index), action] : null;
+			}).filter((entry) => !!entry));
+			const rows = result.rows.map(({ [INTERNAL_ROWID_COLUMN]: _rowid, ...row }) => row);
 			return c.json({
 				columns,
-				rows: result.rows,
+				rows,
+				rowActions,
 				total,
 				page,
 				pageSize,
@@ -15162,9 +15230,19 @@ function dataRoutes(config) {
 				databaseId: page.databaseId,
 				pageId: page.pageId
 			}];
-			const record = (await listExtractionAuditRecords(aiexDir)).find((record$1) => record$1.outputName === name$1);
+			let record = (await listExtractionAuditRecords(aiexDir)).find((record$1) => record$1.outputName === name$1);
+			if (!record) record = await createExtractionAuditRecord(aiexDir, {
+				schemaName,
+				source: {
+					type: "file",
+					filePath,
+					fileName: name$1
+				}
+			});
 			if (record) await updateExtractionAuditRecord(aiexDir, record.id, {
 				status: "succeeded",
+				outputPath: filePath,
+				outputName: name$1,
 				notionPages,
 				error: void 0
 			});
@@ -15173,15 +15251,79 @@ function dataRoutes(config) {
 				notionPages
 			});
 		} catch (error) {
+			const message = error instanceof Error ? error.message : String(error);
+			const record = (await listExtractionAuditRecords(aiexDir)).find((record$1) => record$1.outputName === name$1);
+			if (record) await updateExtractionAuditRecord(aiexDir, record.id, {
+				status: "failed",
+				outputPath: filePath,
+				outputName: name$1,
+				error: message
+			});
 			return c.json({
 				success: false,
-				error: error instanceof Error ? error.message : String(error)
+				error: message
 			}, 500);
 		}
 	});
 	return app;
 }
 
+//#endregion
+//#region src/core/file-constants.ts
+const MAX_UPLOAD_SIZE = 150 * 1024 * 1024;
+const MAX_UPLOAD_SIZE_TEXT = "150MB";
+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 getExtensionFromMime(mimeType) {
+	return MIME_TO_EXT[mimeType];
+}
+function isAllowedMimeType(mimeType) {
+	return SUPPORTED_MIME_TYPES.has(mimeType);
+}
+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 (${(file.size / 1024 / 1024).toFixed(1)}MB) exceeds ${MAX_UPLOAD_SIZE_TEXT} limit`);
+	if (!isAllowedMimeType(file.type)) throw new FileValidationError(`Unsupported file type "${file.type}". Supported types: ${[...SUPPORTED_MIME_TYPES].join(", ")}`);
+}
+
 //#endregion
 //#region src/server/routes/extract.ts
 function getFormString(value) {
@@ -15195,9 +15337,16 @@ function getFormFile(value) {
 function safeUploadName(name$1) {
 	return path.basename(name$1).replace(/[^\w.-]/g, "_") || "upload.txt";
 }
+function safeUploadNameForMime(file) {
+	const safeName = safeUploadName(file.name);
+	const ext = getExtensionFromMime(file.type);
+	if (!ext) throw new FileValidationError(`Unsupported file type "${file.type}"`);
+	return `${path.parse(safeName).name || "upload"}.${ext}`;
+}
 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;
@@ -15365,11 +15514,25 @@ 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
+						}, 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-Cb9X2mxU.mjs → doctor-collector-CjFTz8p4.mjs}

@@ -65,7 +65,7 @@ function doctorDiagnosticsTableRows(d) {
 //#endregion
 //#region package.json
 var name = "aiex-cli";
-var version = "0.0.2-beta.9";
+var version = "0.0.3-beta.1";
 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-Cb9X2mxU.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-CjFTz8p4.mjs";
 
 export { JsonSchemaDefinitionSchema, buildDoctorDiagnostics, collectDoctorDiagnostics, createMigrationConfig, doctorDiagnosticsTableRows, formatDoctorDiagnosticsJson, generateDrizzleConfig, generateDrizzleSchema, parseJsonSchema };