@clue-ai/cli 0.0.8 → 0.0.10

package/src/contracts.mjs

@@ -2,163 +2,168 @@ import { createRequire } from "node:module";
 
 const require = createRequire(import.meta.url);
 const schemaPackage = require("./public-schema.cjs");
+
 const {
-  clueInitToolRequestSchema,
-  clueInitToolReportSchema,
-  semanticSnapshotRequestSchema,
-  semanticSnapshotResponseSchema,
+	clueInitToolRequestSchema,
+	clueInitToolReportSchema,
+	semanticSnapshotRequestSchema,
+	semanticSnapshotResponseSchema,
 } = schemaPackage;
 
 const formatSchemaError = (result, field) => {
-  const issues = result.error.issues.map((issue) => {
-    const path = issue.path.length ? issue.path.join(".") : field;
-    return `${path}: ${issue.message}`;
-  });
-  return `${field} is invalid: ${issues.join("; ")}`;
+	const issues = result.error.issues.map((issue) => {
+		const path = issue.path.length ? issue.path.join(".") : field;
+		return `${path}: ${issue.message}`;
+	});
+	return `${field} is invalid: ${issues.join("; ")}`;
 };
 
 const parseWithSchema = (schema, input, field) => {
-  const result = schema.safeParse(input);
-  if (!result.success) {
-    throw new Error(formatSchemaError(result, field));
-  }
-  return result.data;
+	const result = schema.safeParse(input);
+	if (!result.success) {
+		throw new Error(formatSchemaError(result, field));
+	}
+	return result.data;
 };
 
 const nonEmpty = (value, field) => {
-  if (typeof value !== "string" || value.trim() === "") {
-    throw new Error(`${field} is required`);
-  }
-  return value.trim();
+	if (typeof value !== "string" || value.trim() === "") {
+		throw new Error(`${field} is required`);
+	}
+	return value.trim();
 };
 
 const optionalNonEmpty = (value) =>
-  typeof value === "string" && value.trim() ? value.trim() : null;
+	typeof value === "string" && value.trim() ? value.trim() : null;
 
 const stringArray = (value, field, { min = 0 } = {}) => {
-  if (!Array.isArray(value)) {
-    throw new Error(`${field} must be an array`);
-  }
-  const result = value.map((entry) => nonEmpty(entry, field));
-  if (result.length < min) {
-    throw new Error(`${field} must include at least ${min} item(s)`);
-  }
-  return [...new Set(result)];
+	if (!Array.isArray(value)) {
+		throw new Error(`${field} must be an array`);
+	}
+	const result = value.map((entry) => nonEmpty(entry, field));
+	if (result.length < min) {
+		throw new Error(`${field} must include at least ${min} item(s)`);
+	}
+	return [...new Set(result)];
 };
 
 export const validateInitRequest = (input) => {
-  const parsed = parseWithSchema(
-    clueInitToolRequestSchema,
-    input,
-    "init request",
-  );
-  return {
-    ...parsed,
-    allowed_source_paths: [...new Set(parsed.allowed_source_paths)],
-    excluded_source_paths: [...new Set(parsed.excluded_source_paths ?? [])],
-  };
+	const parsed = parseWithSchema(
+		clueInitToolRequestSchema,
+		input,
+		"init request",
+	);
+	return {
+		...parsed,
+		allowed_source_paths: [...new Set(parsed.allowed_source_paths)],
+		excluded_source_paths: [...new Set(parsed.excluded_source_paths ?? [])],
+	};
 };
 
 export const buildInitReport = ({
-  request,
-  lifecycleInsertions,
-  requiredVariables = [],
-  warnings = [],
+	request,
+	lifecycleInsertions,
+	requiredVariables = [],
+	warnings = [],
 }) =>
-  parseWithSchema(
-    clueInitToolReportSchema,
-    {
-      target_tool: request.target_tool,
-      ci_workflow_path: request.ci_workflow_path,
-      ci_workflow_added: true,
-      required_secrets: ["CLUE_API_KEY", "CLUE_AI_PROVIDER_API_KEY"],
-      required_variables: requiredVariables,
-      lifecycle_insertions: lifecycleInsertions,
-      semantic_generation_timing: "after_merge_ci",
-      semantic_preview_generated: false,
-      client_repo_generated_files_committed: false,
-      clue_track_inserted: false,
-      clue_dom_tag_inserted: false,
-      allowed_source_paths: request.allowed_source_paths,
-      excluded_source_paths: request.excluded_source_paths,
-      warnings,
-    },
-    "init report",
-  );
+	parseWithSchema(
+		clueInitToolReportSchema,
+		{
+			target_tool: request.target_tool,
+			ci_workflow_path: request.ci_workflow_path,
+			ci_workflow_added: true,
+			required_secrets: ["CLUE_API_KEY", "CLUE_AI_PROVIDER_API_KEY"],
+			required_variables: requiredVariables,
+			lifecycle_insertions: lifecycleInsertions,
+			semantic_generation_timing: "after_merge_ci",
+			semantic_preview_generated: false,
+			client_repo_generated_files_committed: false,
+			frontend_api_key_exposed: false,
+			browser_token_endpoint_required: true,
+			browser_token_provider_required: true,
+			browser_token_endpoint_path: "/api/v1/ingest/browser-tokens",
+			clue_track_inserted: false,
+			clue_dom_tag_inserted: false,
+			allowed_source_paths: request.allowed_source_paths,
+			excluded_source_paths: request.excluded_source_paths,
+			warnings,
+		},
+		"init report",
+	);
 
 export const validateSemanticCiRequest = (input) => ({
-  project_key: nonEmpty(input.project_key, "project_key"),
-  environment: nonEmpty(input.environment, "environment"),
-  service_key: nonEmpty(input.service_key, "service_key"),
-  repository: {
-    provider: nonEmpty(input.repository?.provider, "repository.provider"),
-    repository_id: nonEmpty(
-      input.repository?.repository_id,
-      "repository.repository_id",
-    ),
-    ...(optionalNonEmpty(input.repository?.owner)
-      ? { owner: optionalNonEmpty(input.repository.owner) }
-      : {}),
-    ...(optionalNonEmpty(input.repository?.name)
-      ? { name: optionalNonEmpty(input.repository.name) }
-      : {}),
-    ...(optionalNonEmpty(input.repository?.default_branch)
-      ? { default_branch: optionalNonEmpty(input.repository.default_branch) }
-      : {}),
-    merge_commit: nonEmpty(
-      input.repository?.merge_commit,
-      "repository.merge_commit",
-    ),
-    ...(optionalNonEmpty(input.repository?.merged_at)
-      ? { merged_at: optionalNonEmpty(input.repository.merged_at) }
-      : {}),
-    ...(input.repository?.pull_request_number === undefined ||
-    input.repository?.pull_request_number === null
-      ? {}
-      : { pull_request_number: input.repository.pull_request_number }),
-    ...(optionalNonEmpty(input.repository?.workflow_run_id)
-      ? { workflow_run_id: optionalNonEmpty(input.repository.workflow_run_id) }
-      : {}),
-  },
-  service: {
-    service_key: nonEmpty(
-      input.service?.service_key ?? input.service_key,
-      "service.service_key",
-    ),
-    root_path: input.service?.root_path ?? null,
-    framework: input.service?.framework ?? "fastapi",
-    language: input.service?.language ?? "python",
-  },
-  allowed_source_paths: stringArray(
-    input.allowed_source_paths,
-    "allowed_source_paths",
-    { min: 1 },
-  ),
-  excluded_source_paths: stringArray(
-    input.excluded_source_paths ?? [],
-    "excluded_source_paths",
-  ),
-  clue_api_base_url: nonEmpty(input.clue_api_base_url, "clue_api_base_url"),
-  ai_provider_base_url: null,
-  ai_provider: optionalNonEmpty(input.ai_provider),
-  ai_model: optionalNonEmpty(input.ai_model),
+	project_key: nonEmpty(input.project_key, "project_key"),
+	environment: nonEmpty(input.environment, "environment"),
+	service_key: nonEmpty(input.service_key, "service_key"),
+	repository: {
+		provider: nonEmpty(input.repository?.provider, "repository.provider"),
+		repository_id: nonEmpty(
+			input.repository?.repository_id,
+			"repository.repository_id",
+		),
+		...(optionalNonEmpty(input.repository?.owner)
+			? { owner: optionalNonEmpty(input.repository.owner) }
+			: {}),
+		...(optionalNonEmpty(input.repository?.name)
+			? { name: optionalNonEmpty(input.repository.name) }
+			: {}),
+		...(optionalNonEmpty(input.repository?.default_branch)
+			? { default_branch: optionalNonEmpty(input.repository.default_branch) }
+			: {}),
+		merge_commit: nonEmpty(
+			input.repository?.merge_commit,
+			"repository.merge_commit",
+		),
+		...(optionalNonEmpty(input.repository?.merged_at)
+			? { merged_at: optionalNonEmpty(input.repository.merged_at) }
+			: {}),
+		...(input.repository?.pull_request_number === undefined ||
+		input.repository?.pull_request_number === null
+			? {}
+			: { pull_request_number: input.repository.pull_request_number }),
+		...(optionalNonEmpty(input.repository?.workflow_run_id)
+			? { workflow_run_id: optionalNonEmpty(input.repository.workflow_run_id) }
+			: {}),
+	},
+	service: {
+		service_key: nonEmpty(
+			input.service?.service_key ?? input.service_key,
+			"service.service_key",
+		),
+		root_path: input.service?.root_path ?? null,
+		framework: input.service?.framework ?? "fastapi",
+		language: input.service?.language ?? "python",
+	},
+	allowed_source_paths: stringArray(
+		input.allowed_source_paths,
+		"allowed_source_paths",
+		{ min: 1 },
+	),
+	excluded_source_paths: stringArray(
+		input.excluded_source_paths ?? [],
+		"excluded_source_paths",
+	),
+	clue_api_base_url: nonEmpty(input.clue_api_base_url, "clue_api_base_url"),
+	ai_provider_base_url: null,
+	ai_provider: optionalNonEmpty(input.ai_provider),
+	ai_model: optionalNonEmpty(input.ai_model),
 });
 
 export const buildOperationSourceKey = (method, pathTemplate) =>
-  `route.${method.toUpperCase()}.${pathTemplate}`;
+	`route.${method.toUpperCase()}.${pathTemplate}`;
 
 export const validateSemanticSnapshotRequest = (input) => {
-  return parseWithSchema(
-    semanticSnapshotRequestSchema,
-    input,
-    "semantic snapshot",
-  );
+	return parseWithSchema(
+		semanticSnapshotRequestSchema,
+		input,
+		"semantic snapshot",
+	);
 };
 
 export const validateSemanticSnapshotResponse = (input) => {
-  return parseWithSchema(
-    semanticSnapshotResponseSchema,
-    input,
-    "semantic snapshot response",
-  );
+	return parseWithSchema(
+		semanticSnapshotResponseSchema,
+		input,
+		"semantic snapshot response",
+	);
 };

package/src/init-tool.mjs

@@ -1,6 +1,7 @@
 import { mkdir, writeFile } from "node:fs/promises";
 import { dirname, join } from "node:path";
 import { buildInitReport, validateInitRequest } from "./contracts.mjs";
+import { SEMANTIC_GEN_WORKFLOW_COMMAND } from "./cli-invocation.mjs";
 import {
   applyLifecyclePlan,
   planLifecycleInsertions,
@@ -158,7 +159,7 @@ jobs:
           CLUE_SEMANTIC_REQUEST_JSON: |
 ${indentMultiline(workflowRequestPayload(request), 12)}
         run: |
-          npx @clue-ai/cli semantic-gen --request-env CLUE_SEMANTIC_REQUEST_JSON --repo .
+          ${SEMANTIC_GEN_WORKFLOW_COMMAND}
 `;
 
 export const writeSemanticWorkflow = async ({ repoRoot, request }) => {

package/src/lifecycle-guard.mjs

@@ -1,134 +1,197 @@
 const LIFECYCLE_CALL_PATTERN =
-	/\b(ClueInit|ClueIdentify|ClueSetAccount|ClueLogout)\s*\(/g;
-const SAFE_HELPER_PATTERN =
-	/\b(?:safeClue|safe_clue|safe_clue_call|safeClueCall|withClueGuard|with_clue_guard)\s*\(/g;
+	/\b(ClueInit|ClueIdentify|ClueSetAccount|ClueLogout|clue_init_fastapi|clue_init_django)\s*\(/g;
 
-const findMatchingDelimiter = (text, openIndex, open, close) => {
-	let depth = 0;
-	for (let index = openIndex; index < text.length; index += 1) {
-		const character = text[index];
-		if (character === open) depth += 1;
-		if (character === close) {
-			depth -= 1;
-			if (depth === 0) return index;
+export const stripSourceNoise = (text, { stripStrings = false } = {}) => {
+	let output = "";
+	let index = 0;
+	while (index < text.length) {
+		const char = text[index];
+		const next = text[index + 1];
+		if (char === "/" && next === "/") {
+			while (index < text.length && text[index] !== "\n") index += 1;
+			continue;
+		}
+		if (char === "#") {
+			while (index < text.length && text[index] !== "\n") index += 1;
+			continue;
 		}
+		if (char === "/" && next === "*") {
+			index += 2;
+			while (
+				index < text.length &&
+				!(text[index] === "*" && text[index + 1] === "/")
+			) {
+				if (text[index] === "\n") output += "\n";
+				index += 1;
+			}
+			index += index < text.length ? 2 : 0;
+			continue;
+		}
+		if (char === "'" || char === '"' || char === "`") {
+			const quote = char;
+			const triple =
+				quote !== "`" && text.slice(index, index + 3) === quote.repeat(3);
+			const endToken = triple ? quote.repeat(3) : quote;
+			if (!stripStrings) {
+				output += triple ? endToken : quote;
+			} else {
+				output += quote === "`" ? "``" : `${quote}${quote}`;
+			}
+			index += triple ? 3 : 1;
+			while (index < text.length) {
+				if (!triple && text[index] === "\\") {
+					if (!stripStrings) output += text.slice(index, index + 2);
+					index += 2;
+					continue;
+				}
+				if (text.slice(index, index + endToken.length) === endToken) {
+					if (!stripStrings) output += endToken;
+					index += endToken.length;
+					break;
+				}
+				if (text[index] === "\n") output += "\n";
+				if (!stripStrings && text[index] !== "\n") output += text[index];
+				index += 1;
+			}
+			continue;
+		}
+		output += char;
+		index += 1;
 	}
-	return -1;
+	return output;
 };
 
-const lineNumberForIndex = (text, index) =>
-	text.slice(0, index).split("\n").length;
+const isIdentifierCharacter = (character) =>
+	typeof character === "string" && /[A-Za-z0-9_$]/.test(character);
 
-const isAwaitedOnCallLine = (text, callIndex) => {
-	const lineStart = text.lastIndexOf("\n", callIndex - 1) + 1;
-	return /\bawait\b/.test(text.slice(lineStart, callIndex));
-};
+const startsKeyword = (text, index, keyword) =>
+	text.startsWith(keyword, index) &&
+	!isIdentifierCharacter(text[index - 1]) &&
+	!isIdentifierCharacter(text[index + keyword.length]);
 
-const isInsideSafeHelperCall = (text, callIndex) => {
-	for (const match of text.matchAll(SAFE_HELPER_PATTERN)) {
-		const helperIndex = match.index ?? 0;
-		if (helperIndex > callIndex) return false;
-		const openParenIndex = text.indexOf("(", helperIndex);
-		const closeParenIndex = findMatchingDelimiter(
-			text,
-			openParenIndex,
-			"(",
-			")",
-		);
-		if (openParenIndex < callIndex && callIndex < closeParenIndex) return true;
+const advanceQuotedString = (text, startIndex, { preserve = false } = {}) => {
+	const quote = text[startIndex];
+	const triple = quote !== "`" && text.slice(startIndex, startIndex + 3) === quote.repeat(3);
+	const endToken = triple ? quote.repeat(3) : quote;
+	let index = startIndex + (triple ? 3 : 1);
+	let value = preserve ? (triple ? endToken : quote) : "";
+	while (index < text.length) {
+		if (!triple && text[index] === "\\") {
+			if (preserve) value += text.slice(index, index + 2);
+			index += 2;
+			continue;
+		}
+		if (text.slice(index, index + endToken.length) === endToken) {
+			if (preserve) value += endToken;
+			index += endToken.length;
+			break;
+		}
+		if (preserve) value += text[index];
+		index += 1;
 	}
-	return false;
+	return { index, value };
 };
 
-const isInsideJsTryBlock = (text, callIndex) => {
-	for (const match of text.matchAll(/\btry\s*{/g)) {
-		const tryIndex = match.index ?? 0;
-		if (tryIndex > callIndex) return false;
-		const openBraceIndex = text.indexOf("{", tryIndex);
-		const closeBraceIndex = findMatchingDelimiter(
-			text,
-			openBraceIndex,
-			"{",
-			"}",
-		);
-		if (openBraceIndex < callIndex && callIndex < closeBraceIndex) return true;
+export const extractExecutableModuleStatements = (text) => {
+	const statements = [];
+	let index = 0;
+	while (index < text.length) {
+		const char = text[index];
+		const next = text[index + 1];
+		if (char === "/" && next === "/") {
+			while (index < text.length && text[index] !== "\n") index += 1;
+			continue;
+		}
+		if (char === "#") {
+			while (index < text.length && text[index] !== "\n") index += 1;
+			continue;
+		}
+		if (char === "/" && next === "*") {
+			index += 2;
+			while (
+				index < text.length &&
+				!(text[index] === "*" && text[index + 1] === "/")
+			) {
+				index += 1;
+			}
+			index += index < text.length ? 2 : 0;
+			continue;
+		}
+		if (char === "'" || char === '"' || char === "`") {
+			index = advanceQuotedString(text, index).index;
+			continue;
+		}
+		if (!startsKeyword(text, index, "import") && !startsKeyword(text, index, "export")) {
+			index += 1;
+			continue;
+		}
+		let statement = "";
+		let depth = 0;
+		while (index < text.length) {
+			const current = text[index];
+			const following = text[index + 1];
+			if (current === "/" && following === "/") {
+				while (index < text.length && text[index] !== "\n") index += 1;
+				continue;
+			}
+			if (current === "/" && following === "*") {
+				index += 2;
+				while (
+					index < text.length &&
+					!(text[index] === "*" && text[index + 1] === "/")
+				) {
+					index += 1;
+				}
+				index += index < text.length ? 2 : 0;
+				continue;
+			}
+			if (current === "'" || current === '"' || current === "`") {
+				const quoted = advanceQuotedString(text, index, { preserve: true });
+				statement += quoted.value;
+				index = quoted.index;
+				continue;
+			}
+			statement += current;
+			if (current === "{" || current === "(" || current === "[") depth += 1;
+			if (current === "}" || current === ")" || current === "]") {
+				depth = Math.max(0, depth - 1);
+			}
+			index += 1;
+			if (current === ";" && depth === 0) break;
+			if (current === "\n" && depth === 0) break;
+		}
+		const trimmed = statement.trim();
+		if (trimmed) statements.push(trimmed);
 	}
-	return false;
+	return statements;
 };
 
-const buildLines = (text) => {
-	const lines = [];
-	let start = 0;
-	for (const line of text.split("\n")) {
-		const indent = line.match(/^\s*/)?.[0].length ?? 0;
-		lines.push({
-			start,
-			end: start + line.length,
-			indent,
-			text: line,
-		});
-		start += line.length + 1;
-	}
-	return lines;
-};
+const lineNumberForIndex = (text, index) =>
+	text.slice(0, index).split("\n").length;
 
-const isInsidePythonTryBlock = (text, callIndex) => {
-	const lines = buildLines(text);
-	const callLineIndex = lines.findIndex(
-		(line) => line.start <= callIndex && callIndex <= line.end,
-	);
-	if (callLineIndex < 0) return false;
-	const callLine = lines[callLineIndex];
-	for (let index = callLineIndex - 1; index >= 0; index -= 1) {
-		const candidate = lines[index];
-		if (!candidate.text.trim()) continue;
-		if (!/^\s*try\s*:\s*(?:#.*)?$/.test(candidate.text)) continue;
-		if (candidate.indent >= callLine.indent) continue;
-		const escaped = lines
-			.slice(index + 1, callLineIndex)
-			.some(
-				(line) =>
-					line.text.trim() &&
-					line.indent <= candidate.indent &&
-					!/^\s*(?:except|finally|else)\b/.test(line.text),
-			);
-		if (!escaped) return true;
-	}
-	return false;
+const isAwaitedOnCallLine = (text, callIndex) => {
+	const lineStart = text.lastIndexOf("\n", callIndex - 1) + 1;
+	return /\bawait\b/.test(text.slice(lineStart, callIndex));
 };
 
-const hasCatchHandlerOnCall = (text, callIndex) => {
-	const openParenIndex = text.indexOf("(", callIndex);
-	const closeParenIndex = findMatchingDelimiter(text, openParenIndex, "(", ")");
-	if (closeParenIndex < 0) return false;
-	return /^\s*\.catch\s*\(/.test(text.slice(closeParenIndex + 1));
+const canonicalLifecycleApiName = (apiName) => {
+	if (apiName === "clue_init_fastapi" || apiName === "clue_init_django") {
+		return "ClueInit";
+	}
+	return apiName;
 };
 
-const isGuardedLifecycleCall = (text, callIndex) =>
-	hasCatchHandlerOnCall(text, callIndex) ||
-	isInsideSafeHelperCall(text, callIndex) ||
-	isInsideJsTryBlock(text, callIndex) ||
-	isInsidePythonTryBlock(text, callIndex);
-
 export const findLifecycleGuardViolations = (text) => {
 	const violations = [];
 	for (const match of text.matchAll(LIFECYCLE_CALL_PATTERN)) {
 		const callIndex = match.index ?? 0;
-		const apiName = match[1];
+		const apiName = canonicalLifecycleApiName(match[1]);
 		if (isAwaitedOnCallLine(text, callIndex)) {
 			violations.push({
 				api_name: apiName,
 				line: lineNumberForIndex(text, callIndex),
 				reason: "awaited_lifecycle_call",
 			});
-			continue;
-		}
-		if (!isGuardedLifecycleCall(text, callIndex)) {
-			violations.push({
-				api_name: apiName,
-				line: lineNumberForIndex(text, callIndex),
-				reason: "unguarded_lifecycle_call",
-			});
 		}
 	}
 	return violations;
@@ -136,6 +199,8 @@ export const findLifecycleGuardViolations = (text) => {
 
 export const findLifecycleCallApiNames = (text) => [
 	...new Set(
-		[...text.matchAll(LIFECYCLE_CALL_PATTERN)].map((match) => match[1]),
+		[...text.matchAll(LIFECYCLE_CALL_PATTERN)].map((match) =>
+			canonicalLifecycleApiName(match[1]),
+		),
 	),
 ];