@aigne/doc-smith 0.2.8 → 0.2.9

package/CHANGELOG.md

@@ -1,5 +1,12 @@
 # Changelog
 
+## [0.2.9](https://github.com/AIGNE-io/aigne-doc-smith/compare/v0.2.8...v0.2.9) (2025-08-13)
+
+
+### Bug Fixes
+
+* polish ignore check ([#25](https://github.com/AIGNE-io/aigne-doc-smith/issues/25)) ([90bc866](https://github.com/AIGNE-io/aigne-doc-smith/commit/90bc866513fef7b47047b1016e07bf38881c101c))
+
 ## [0.2.8](https://github.com/AIGNE-io/aigne-doc-smith/compare/v0.2.7...v0.2.8) (2025-08-13)
 
 

package/agents/find-item-by-path.mjs

@@ -8,15 +8,8 @@ function getActionText(isTranslate, baseText) {
 }
 
 export default async function findItemByPath(
-  {
-    "doc-path": docPath,
-    structurePlanResult,
-    boardId,
-    docsDir,
-    isTranslate,
-    feedback,
-  },
-  options
+  { "doc-path": docPath, structurePlanResult, boardId, docsDir, isTranslate, feedback },
+  options,
 ) {
   let foundItem = null;
   let selectedFileContent = null;
@@ -30,9 +23,7 @@ export default async function findItemByPath(
       // Filter for main language .md files (exclude _sidebar.md and language-specific files)
       const mainLanguageFiles = files.filter(
         (file) =>
-          file.endsWith(".md") &&
-          file !== "_sidebar.md" &&
-          !file.match(/\.\w+(-\w+)?\.md$/) // Exclude language-specific files like .en.md, .zh-CN.md, etc.
+          file.endsWith(".md") && file !== "_sidebar.md" && !file.match(/\.\w+(-\w+)?\.md$/), // Exclude language-specific files like .en.md, .zh-CN.md, etc.
       );
 
       if (mainLanguageFiles.length === 0) {
@@ -52,7 +43,7 @@ export default async function findItemByPath(
 
           const searchTerm = input.trim().toLowerCase();
           const filteredFiles = mainLanguageFiles.filter((file) =>
-            file.toLowerCase().includes(searchTerm)
+            file.toLowerCase().includes(searchTerm),
           );
 
           return filteredFiles.map((file) => ({
@@ -71,10 +62,7 @@ export default async function findItemByPath(
         const selectedFilePath = join(docsDir, selectedFile);
         selectedFileContent = await readFile(selectedFilePath, "utf-8");
       } catch (readError) {
-        console.warn(
-          `⚠️  Could not read content from ${selectedFile}:`,
-          readError.message
-        );
+        console.warn(`⚠️  Could not read content from ${selectedFile}:`, readError.message);
         selectedFileContent = null;
       }
 
@@ -87,9 +75,7 @@ export default async function findItemByPath(
 
       // First try without boardId prefix
       foundItemByFile = structurePlanResult.find((item) => {
-        const itemFlattenedPath = item.path
-          .replace(/^\//, "")
-          .replace(/\//g, "-");
+        const itemFlattenedPath = item.path.replace(/^\//, "").replace(/\//g, "-");
         return itemFlattenedPath === flatName;
       });
       if (!foundItemByFile) {
@@ -102,8 +88,8 @@ export default async function findItemByPath(
       throw new Error(
         getActionText(
           isTranslate,
-          "Please provide a doc-path parameter to specify which document to {action}"
-        )
+          "Please provide a doc-path parameter to specify which document to {action}",
+        ),
       );
     }
   }
@@ -121,18 +107,14 @@ export default async function findItemByPath(
       // Find item by comparing flattened paths
       foundItem = structurePlanResult.find((item) => {
         // Convert item.path to flattened format (replace / with -)
-        const itemFlattenedPath = item.path
-          .replace(/^\//, "")
-          .replace(/\//g, "-");
+        const itemFlattenedPath = item.path.replace(/^\//, "").replace(/\//g, "-");
         return itemFlattenedPath === flattenedPath;
       });
     }
   }
 
   if (!foundItem) {
-    throw new Error(
-      `Item with path "${docPath}" not found in structurePlanResult`
-    );
+    throw new Error(`Item with path "${docPath}" not found in structurePlanResult`);
   }
 
   // Prompt for feedback if not provided
@@ -140,7 +122,7 @@ export default async function findItemByPath(
   if (!userFeedback) {
     const feedbackMessage = getActionText(
       isTranslate,
-      "Please provide feedback for the {action} (press Enter to skip):"
+      "Please provide feedback for the {action} (press Enter to skip):",
     );
 
     userFeedback = await options.prompts.input({

package/agents/input-generator.mjs

@@ -1,16 +1,8 @@
 import { mkdir, readFile, writeFile } from "node:fs/promises";
 import { dirname, join } from "node:path";
 import chalk from "chalk";
-import {
-  DOCUMENT_STYLES,
-  SUPPORTED_LANGUAGES,
-  TARGET_AUDIENCES,
-} from "../utils/constants.mjs";
-import {
-  getAvailablePaths,
-  getProjectInfo,
-  validatePath,
-} from "../utils/utils.mjs";
+import { DOCUMENT_STYLES, SUPPORTED_LANGUAGES, TARGET_AUDIENCES } from "../utils/constants.mjs";
+import { getAvailablePaths, getProjectInfo, validatePath } from "../utils/utils.mjs";
 
 // UI constants
 const _PRESS_ENTER_TO_FINISH = "Press Enter to finish";
@@ -23,12 +15,8 @@ const _PRESS_ENTER_TO_FINISH = "Press Enter to finish";
  * @returns {Promise<Object>}
  */
 export default async function init(
-  {
-    outputPath = ".aigne/doc-smith",
-    fileName = "config.yaml",
-    skipIfExists = false,
-  },
-  options
+  { outputPath = ".aigne/doc-smith", fileName = "config.yaml", skipIfExists = false },
+  options,
 ) {
   if (skipIfExists) {
     const filePath = join(outputPath, fileName);
@@ -104,7 +92,7 @@ export default async function init(
   // 4. Translation languages
   // Filter out the primary language from available choices
   const availableTranslationLanguages = SUPPORTED_LANGUAGES.filter(
-    (lang) => lang.code !== primaryLanguageChoice
+    (lang) => lang.code !== primaryLanguageChoice,
   );
 
   const translateLanguageChoices = await options.prompts.checkbox({
@@ -154,11 +142,7 @@ export default async function init(
     });
 
     // Check if user chose to exit
-    if (
-      !selectedPath ||
-      selectedPath.trim() === "" ||
-      selectedPath === "Press Enter to finish"
-    ) {
+    if (!selectedPath || selectedPath.trim() === "" || selectedPath === "Press Enter to finish") {
       break;
     }
 
@@ -207,13 +191,9 @@ export default async function init(
     console.log(chalk.cyan("---"));
     console.log(chalk.cyan(yamlContent));
     console.log(chalk.cyan("---"));
+    console.log("💡 You can edit the configuration file anytime to modify settings.\n");
     console.log(
-      "💡 You can edit the configuration file anytime to modify settings.\n"
-    );
-    console.log(
-      `🚀 Run ${chalk.cyan(
-        "'aigne doc generate'"
-      )} to start documentation generation!\n`
+      `🚀 Run ${chalk.cyan("'aigne doc generate'")} to start documentation generation!\n`,
     );
 
     return {};
@@ -284,5 +264,4 @@ function generateYAML(input) {
   return yaml;
 }
 
-init.description =
-  "Generate a configuration file for the documentation generation process";
+init.description = "Generate a configuration file for the documentation generation process";

package/agents/load-sources.mjs

@@ -1,87 +1,9 @@
 import { access, readFile, stat } from "node:fs/promises";
 import path from "node:path";
-import { glob } from "glob";
 import { DEFAULT_EXCLUDE_PATTERNS, DEFAULT_INCLUDE_PATTERNS } from "../utils/constants.mjs";
+import { getFilesWithGlob, loadGitignore } from "../utils/file-utils.mjs";
 import { getCurrentGitHead, getModifiedFilesBetweenCommits } from "../utils/utils.mjs";
 
-/**
- * Load .gitignore patterns from a directory
- * @param {string} dir - Directory path
- * @returns {object|null} Ignore instance or null if no .gitignore found
- */
-async function loadGitignore(dir) {
-  const gitignorePath = path.join(dir, ".gitignore");
-  try {
-    await access(gitignorePath);
-    const gitignoreContent = await readFile(gitignorePath, "utf8");
-    // Create ignore patterns from .gitignore content
-    const ignorePatterns = gitignoreContent
-      .split("\n")
-      .map((line) => line.trim())
-      .filter((line) => line && !line.startsWith("#"))
-      .map((line) => line.replace(/^\//, "")); // Remove leading slash
-
-    return ignorePatterns.length > 0 ? ignorePatterns : null;
-  } catch {
-    // .gitignore file doesn't exist
-    return null;
-  }
-}
-
-/**
- * Get files using glob patterns
- * @param {string} dir - Directory to search
- * @param {string[]} includePatterns - Include patterns
- * @param {string[]} excludePatterns - Exclude patterns
- * @param {string[]} gitignorePatterns - .gitignore patterns
- * @returns {Promise<string[]>} Array of file paths
- */
-async function getFilesWithGlob(dir, includePatterns, excludePatterns, gitignorePatterns) {
-  // Prepare all ignore patterns
-  const allIgnorePatterns = [];
-
-  if (excludePatterns) {
-    allIgnorePatterns.push(...excludePatterns);
-  }
-
-  if (gitignorePatterns) {
-    allIgnorePatterns.push(...gitignorePatterns);
-  }
-
-  // Add default exclusions if not already present
-  const defaultExclusions = ["node_modules/**", "test/**", "temp/**"];
-  for (const exclusion of defaultExclusions) {
-    if (!allIgnorePatterns.includes(exclusion)) {
-      allIgnorePatterns.push(exclusion);
-    }
-  }
-
-  // Convert patterns to be relative to the directory
-  const patterns = includePatterns.map((pattern) => {
-    // If pattern doesn't start with / or **, make it relative to dir
-    if (!pattern.startsWith("/") && !pattern.startsWith("**")) {
-      return `**/${pattern}`; // Use ** to search recursively
-    }
-    return pattern;
-  });
-
-  try {
-    const files = await glob(patterns, {
-      cwd: dir,
-      ignore: allIgnorePatterns.length > 0 ? allIgnorePatterns : undefined,
-      absolute: true,
-      nodir: true, // Only return files, not directories
-      dot: false, // Don't include dot files by default
-      gitignore: true, // Enable .gitignore support
-    });
-
-    return files;
-  } catch (error) {
-    console.warn(`Warning: Error during glob search in ${dir}: ${error.message}`);
-    return [];
-  }
-}
-
 export default async function loadSources({
   sources = [],
   sourcesPath = [],
@@ -242,6 +164,21 @@ export default async function loadSources({
     }
   }
 
+  // Count words and lines in allSources
+  let totalWords = 0;
+  let totalLines = 0;
+
+  for (const source of Object.values(allSources)) {
+    if (typeof source === "string") {
+      // Count English words (simple regex for words containing a-zA-Z)
+      const words = source.match(/[a-zA-Z]+/g) || [];
+      totalWords += words.length;
+
+      // Count lines
+      totalLines += source.split("\n").length;
+    }
+  }
+
   return {
     datasourcesList: sourceFiles,
     datasources: allSources,
@@ -249,6 +186,8 @@ export default async function loadSources({
     originalStructurePlan,
     files,
     modifiedFiles,
+    totalWords,
+    totalLines,
   };
 }
 

package/agents/publish-docs.mjs

@@ -18,11 +18,7 @@ const DEFAULT_APP_URL = "https://docsmith.aigne.io";
  * @returns {Promise<string>} - The access token
  */
 async function getAccessToken(appUrl) {
-  const DOC_SMITH_ENV_FILE = join(
-    homedir(),
-    ".aigne",
-    "doc-smith-connected.yaml"
-  );
+  const DOC_SMITH_ENV_FILE = join(homedir(), ".aigne", "doc-smith-connected.yaml");
   const { hostname } = new URL(appUrl);
 
   let accessToken = process.env.DOC_DISCUSS_KIT_ACCESS_TOKEN;
@@ -47,10 +43,7 @@ async function getAccessToken(appUrl) {
   // If still no access token, prompt user to authorize
   if (!accessToken) {
     const DISCUSS_KIT_URL = appUrl;
-    const connectUrl = joinURL(
-      new URL(DISCUSS_KIT_URL).origin,
-      WELLKNOWN_SERVICE_PATH_PREFIX
-    );
+    const connectUrl = joinURL(new URL(DISCUSS_KIT_URL).origin, WELLKNOWN_SERVICE_PATH_PREFIX);
 
     try {
       const result = await createConnect({
@@ -59,8 +52,7 @@ async function getAccessToken(appUrl) {
         source: `AIGNE DocSmith connect to Discuss Kit`,
         closeOnSuccess: true,
         appName: "AIGNE DocSmith",
-        appLogo:
-          "https://www.aigne.io/image-bin/uploads/a7910a71364ee15a27e86f869ad59009.svg",
+        appLogo: "https://www.aigne.io/image-bin/uploads/a7910a71364ee15a27e86f869ad59009.svg",
         openPage: (pageUrl) => open(pageUrl),
       });
 
@@ -85,12 +77,12 @@ async function getAccessToken(appUrl) {
             DOC_DISCUSS_KIT_ACCESS_TOKEN: accessToken,
             DOC_DISCUSS_KIT_URL: DISCUSS_KIT_URL,
           },
-        })
+        }),
       );
     } catch (error) {
       console.error("Failed to get access token:", error);
       throw new Error(
-        "Failed to obtain access token. Please check your network connection and try again later."
+        "Failed to obtain access token. Please check your network connection and try again later.",
       );
     }
   }
@@ -100,7 +92,7 @@ async function getAccessToken(appUrl) {
 
 export default async function publishDocs(
   { docsDir, appUrl, boardId, projectName, projectDesc, projectLogo },
-  options
+  options,
 ) {
   // Check if DOC_DISCUSS_KIT_URL is set in environment variables
   const envAppUrl = process.env.DOC_DISCUSS_KIT_URL;
@@ -184,7 +176,7 @@ export default async function publishDocs(
         await saveValueToConfig(
           "boardId",
           newBoardId,
-          "⚠️ Warning: boardId is auto-generated by system, please do not edit manually"
+          "⚠️ Warning: boardId is auto-generated by system, please do not edit manually",
         );
       }
     }

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@aigne/doc-smith",
-  "version": "0.2.8",
+  "version": "0.2.9",
   "description": "",
   "publishConfig": {
     "access": "public"

package/utils/file-utils.mjs

@@ -0,0 +1,205 @@
+import { access, readFile } from "node:fs/promises";
+import { execSync } from "node:child_process";
+import path from "node:path";
+import { glob } from "glob";
+
+/**
+ * Check if a directory is inside a git repository using git command
+ * @param {string} dir - Directory path to check
+ * @returns {boolean} True if inside a git repository
+ */
+function isInGitRepository(dir) {
+  try {
+    execSync("git rev-parse --is-inside-work-tree", {
+      cwd: dir,
+      stdio: "pipe",
+      encoding: "utf8",
+    });
+    return true;
+  } catch {
+    return false;
+  }
+}
+
+/**
+ * Find git repository root directory using git command
+ * @param {string} startDir - Starting directory path
+ * @returns {string|null} Git repository root path or null if not found
+ */
+function findGitRoot(startDir) {
+  try {
+    const gitRoot = execSync("git rev-parse --show-toplevel", {
+      cwd: startDir,
+      stdio: "pipe",
+      encoding: "utf8",
+    }).trim();
+    return gitRoot;
+  } catch {
+    return null;
+  }
+}
+
+/**
+ * Convert gitignore patterns to glob-compatible patterns
+ * @param {string} pattern - A single gitignore pattern
+ * @returns {string[]} Array of glob patterns that match gitignore behavior
+ */
+function gitignoreToGlobPatterns(pattern) {
+  const patterns = [];
+
+  // Remove leading slash (already handled by gitignore parsing)
+  const cleanPattern = pattern.replace(/^\//, "");
+
+  // If pattern doesn't contain wildcards and doesn't end with /
+  // it could match both files and directories
+  if (!cleanPattern.includes("*") && !cleanPattern.includes("?") && !cleanPattern.endsWith("/")) {
+    // Add patterns to match both file and directory
+    patterns.push(cleanPattern); // Exact match
+    patterns.push(`${cleanPattern}/**`); // Directory contents
+    patterns.push(`**/${cleanPattern}`); // Nested exact match
+    patterns.push(`**/${cleanPattern}/**`); // Nested directory contents
+  } else if (cleanPattern.endsWith("/")) {
+    // Directory-only pattern
+    const dirPattern = cleanPattern.slice(0, -1);
+    patterns.push(`${dirPattern}/**`);
+    patterns.push(`**/${dirPattern}/**`);
+  } else {
+    // Pattern with wildcards or specific file
+    patterns.push(cleanPattern);
+    if (!cleanPattern.startsWith("**/")) {
+      patterns.push(`**/${cleanPattern}`);
+    }
+  }
+
+  return patterns;
+}
+
+/**
+ * Parse .gitignore content into patterns
+ * @param {string} content - .gitignore file content
+ * @returns {string[]} Array of ignore patterns converted to glob format
+ */
+function parseGitignoreContent(content) {
+  const lines = content
+    .split("\n")
+    .map((line) => line.trim())
+    .filter((line) => line && !line.startsWith("#"))
+    .map((line) => line.replace(/^\//, "")); // Remove leading slash
+
+  // Convert each gitignore pattern to glob patterns
+  const allPatterns = [];
+  for (const line of lines) {
+    allPatterns.push(...gitignoreToGlobPatterns(line));
+  }
+
+  return [...new Set(allPatterns)]; // Remove duplicates
+}
+
+/**
+ * Load .gitignore patterns from multiple directories (current + all parent directories up to git root)
+ * @param {string} dir - Directory path (will search up to find all .gitignore files)
+ * @returns {string[]|null} Array of merged ignore patterns or null if no .gitignore found
+ */
+export async function loadGitignore(dir) {
+  // First, check if we're in a git repository
+  const inGitRepo = isInGitRepository(dir);
+  if (!inGitRepo) {
+    // Not in a git repository, just check the current directory
+    const gitignorePath = path.join(dir, ".gitignore");
+    try {
+      await access(gitignorePath);
+      const gitignoreContent = await readFile(gitignorePath, "utf8");
+      const ignorePatterns = parseGitignoreContent(gitignoreContent);
+      return ignorePatterns.length > 0 ? ignorePatterns : null;
+    } catch {
+      return null;
+    }
+  }
+
+  // We're in a git repository, collect all .gitignore files from current dir to git root
+  const gitRoot = findGitRoot(dir);
+  if (!gitRoot) {
+    return null;
+  }
+
+  const allPatterns = [];
+  let currentDir = path.resolve(dir);
+
+  // Collect .gitignore patterns from current directory up to git root
+  while (currentDir.startsWith(gitRoot)) {
+    const gitignorePath = path.join(currentDir, ".gitignore");
+    try {
+      await access(gitignorePath);
+      const gitignoreContent = await readFile(gitignorePath, "utf8");
+      const patterns = parseGitignoreContent(gitignoreContent);
+
+      // Add patterns with context of which directory they came from
+      // Patterns from deeper directories take precedence
+      allPatterns.unshift(...patterns);
+    } catch {
+      // .gitignore doesn't exist in this directory, continue
+    }
+
+    // Move up one directory
+    if (currentDir === gitRoot) {
+      break;
+    }
+    currentDir = path.dirname(currentDir);
+  }
+
+  return allPatterns.length > 0 ? [...new Set(allPatterns)] : null;
+}
+
+/**
+ * Get files using glob patterns
+ * @param {string} dir - Directory to search
+ * @param {string[]} includePatterns - Include patterns
+ * @param {string[]} excludePatterns - Exclude patterns
+ * @param {string[]} gitignorePatterns - .gitignore patterns
+ * @returns {Promise<string[]>} Array of file paths
+ */
+export async function getFilesWithGlob(dir, includePatterns, excludePatterns, gitignorePatterns) {
+  // Prepare all ignore patterns
+  const allIgnorePatterns = [];
+
+  if (excludePatterns) {
+    allIgnorePatterns.push(...excludePatterns);
+  }
+
+  if (gitignorePatterns) {
+    allIgnorePatterns.push(...gitignorePatterns);
+  }
+
+  // Add default exclusions if not already present
+  const defaultExclusions = ["node_modules/**", "test/**", "temp/**"];
+  for (const exclusion of defaultExclusions) {
+    if (!allIgnorePatterns.includes(exclusion)) {
+      allIgnorePatterns.push(exclusion);
+    }
+  }
+
+  // Convert patterns to be relative to the directory
+  const patterns = includePatterns.map((pattern) => {
+    // If pattern doesn't start with / or **, make it relative to dir
+    if (!pattern.startsWith("/") && !pattern.startsWith("**")) {
+      return `**/${pattern}`; // Use ** to search recursively
+    }
+    return pattern;
+  });
+
+  try {
+    const files = await glob(patterns, {
+      cwd: dir,
+      ignore: allIgnorePatterns.length > 0 ? allIgnorePatterns : undefined,
+      absolute: true,
+      nodir: true, // Only return files, not directories
+      dot: false, // Don't include dot files by default
+      gitignore: true, // Enable .gitignore support
+    });
+
+    return files;
+  } catch (error) {
+    console.warn(`Warning: Error during glob search in ${dir}: ${error.message}`);
+    return [];
+  }
+}