midnight-mcp 0.1.41 → 0.2.2

package/dist/pipeline/indexer.js

@@ -1,254 +0,0 @@
-import { githubClient } from "./github.js";
-import { parseFile } from "./parser.js";
-import { embeddingGenerator } from "./embeddings.js";
-import { vectorStore } from "../db/vectorStore.js";
-import { logger, DEFAULT_REPOSITORIES, } from "../utils/index.js";
-/**
- * Extract pragma language_version from Compact file content
- * Returns the version string or undefined if not found
- */
-function extractPragmaVersion(content) {
-    // Match patterns like: pragma language_version >= 0.14.0;
-    const pragmaMatch = content.match(/pragma\s+language_version\s*[><=]*\s*([\d.]+)/);
-    return pragmaMatch?.[1];
-}
-/**
- * Create chunks from a parsed file
- * Uses intelligent chunking based on code structure
- */
-function createChunks(file, repository, repoVersion) {
-    const chunks = [];
-    const indexedAt = new Date().toISOString();
-    // Extract pragma version for Compact files
-    const pragmaVersion = file.language === "compact"
-        ? extractPragmaVersion(file.content)
-        : undefined;
-    // Add code units as individual chunks
-    for (const unit of file.codeUnits) {
-        chunks.push({
-            text: unit.code,
-            metadata: {
-                repository,
-                filePath: file.path,
-                language: file.language,
-                chunkType: "code_unit",
-                codeUnitType: unit.type,
-                codeUnitName: unit.name,
-                startLine: unit.startLine,
-                endLine: unit.endLine,
-                isPublic: unit.isPublic,
-                repoVersion,
-                pragmaVersion,
-                indexedAt,
-            },
-        });
-    }
-    // If no code units were extracted, chunk the entire file
-    if (file.codeUnits.length === 0 && file.content.length > 0) {
-        const chunkSize = 2000; // characters
-        const overlapLines = 5; // lines to overlap
-        const lines = file.content.split("\n");
-        let currentChunk = "";
-        let startLine = 1;
-        let currentLine = 1;
-        for (const line of lines) {
-            if (currentChunk.length + line.length > chunkSize &&
-                currentChunk.length > 0) {
-                chunks.push({
-                    text: currentChunk,
-                    metadata: {
-                        repository,
-                        filePath: file.path,
-                        language: file.language,
-                        chunkType: "file_chunk",
-                        startLine,
-                        endLine: currentLine - 1,
-                        isPublic: true,
-                        repoVersion,
-                        pragmaVersion,
-                        indexedAt,
-                    },
-                });
-                // Start new chunk with overlap
-                const prevLines = currentChunk.split("\n").slice(-overlapLines);
-                currentChunk = prevLines.join("\n") + "\n";
-                startLine = Math.max(1, currentLine - overlapLines);
-            }
-            currentChunk += line + "\n";
-            currentLine++;
-        }
-        // Add remaining content
-        if (currentChunk.trim().length > 0) {
-            chunks.push({
-                text: currentChunk,
-                metadata: {
-                    repository,
-                    filePath: file.path,
-                    language: file.language,
-                    chunkType: "file_chunk",
-                    startLine,
-                    endLine: currentLine - 1,
-                    isPublic: true,
-                    repoVersion,
-                    pragmaVersion,
-                    indexedAt,
-                },
-            });
-        }
-    }
-    return chunks;
-}
-/**
- * Index a single repository
- */
-export async function indexRepository(repoConfig) {
-    const repoName = `${repoConfig.owner}/${repoConfig.repo}`;
-    logger.info(`Starting index for ${repoName}...`);
-    // Get repo version (branch name or tag)
-    const repoVersion = repoConfig.branch || "main";
-    try {
-        // Fetch all files from the repository
-        const files = await githubClient.fetchRepositoryFiles(repoConfig);
-        logger.info(`Fetched ${files.length} files from ${repoName}`);
-        let chunkCount = 0;
-        const documents = [];
-        for (const file of files) {
-            // Parse the file
-            const parsed = parseFile(file.path, file.content);
-            // Create chunks with version info
-            const chunks = createChunks(parsed, repoName, repoVersion);
-            for (const chunk of chunks) {
-                documents.push({
-                    id: `${repoName}:${file.path}:${chunk.metadata.startLine}`,
-                    content: chunk.text,
-                    metadata: {
-                        repository: chunk.metadata.repository,
-                        filePath: chunk.metadata.filePath,
-                        language: chunk.metadata.language,
-                        startLine: chunk.metadata.startLine,
-                        endLine: chunk.metadata.endLine,
-                        codeType: chunk.metadata.codeUnitType || "unknown",
-                        codeName: chunk.metadata.codeUnitName || "",
-                        isPublic: chunk.metadata.isPublic,
-                        repoVersion: chunk.metadata.repoVersion,
-                        pragmaVersion: chunk.metadata.pragmaVersion,
-                        indexedAt: chunk.metadata.indexedAt,
-                    },
-                });
-                chunkCount++;
-            }
-        }
-        // Generate embeddings and store in vector database
-        if (documents.length > 0) {
-            logger.info(`Generating embeddings for ${documents.length} chunks...`);
-            const texts = documents.map((d) => d.content);
-            const embeddings = await embeddingGenerator.generateEmbeddings(texts);
-            // Add embeddings to documents
-            for (let i = 0; i < documents.length; i++) {
-                documents[i].embedding = embeddings[i].embedding;
-            }
-            // Store in vector database
-            await vectorStore.addDocuments(documents);
-            logger.info(`Stored ${documents.length} documents in vector store`);
-        }
-        return { fileCount: files.length, chunkCount };
-    }
-    catch (error) {
-        logger.error(`Failed to index repository ${repoName}`, {
-            error: String(error),
-        });
-        throw error;
-    }
-}
-/**
- * Index all configured repositories
- */
-export async function indexAllRepositories() {
-    logger.info("Starting full index of all repositories...");
-    const stats = {
-        totalFiles: 0,
-        totalChunks: 0,
-        totalCodeUnits: 0,
-        repositoriesIndexed: [],
-        lastIndexed: new Date().toISOString(),
-    };
-    for (const repoConfig of DEFAULT_REPOSITORIES) {
-        try {
-            const { fileCount, chunkCount } = await indexRepository(repoConfig);
-            stats.totalFiles += fileCount;
-            stats.totalChunks += chunkCount;
-            stats.repositoriesIndexed.push(`${repoConfig.owner}/${repoConfig.repo}`);
-        }
-        catch (error) {
-            logger.error(`Failed to index ${repoConfig.owner}/${repoConfig.repo}`, {
-                error: String(error),
-            });
-        }
-    }
-    logger.info("Indexing complete", stats);
-    return stats;
-}
-/**
- * Incremental update - only index changed files
- */
-export async function incrementalUpdate(repoConfig, since) {
-    const repoName = `${repoConfig.owner}/${repoConfig.repo}`;
-    logger.info(`Starting incremental update for ${repoName} since ${since}...`);
-    try {
-        // Get changed files
-        const changedPaths = await githubClient.getChangedFiles(repoConfig.owner, repoConfig.repo, since);
-        // Filter by patterns
-        const filteredPaths = githubClient.filterFilesByPatterns(changedPaths, repoConfig.patterns, repoConfig.exclude);
-        logger.info(`Found ${filteredPaths.length} changed files matching patterns`);
-        let chunkCount = 0;
-        const documents = [];
-        for (const filePath of filteredPaths) {
-            // Delete existing documents for this file
-            await vectorStore.deleteByPath(repoName, filePath);
-            // Fetch new content
-            const file = await githubClient.getFileContent(repoConfig.owner, repoConfig.repo, filePath, repoConfig.branch);
-            if (file) {
-                const parsed = parseFile(file.path, file.content);
-                const repoVersion = repoConfig.branch || "main";
-                const chunks = createChunks(parsed, repoName, repoVersion);
-                for (const chunk of chunks) {
-                    documents.push({
-                        id: `${repoName}:${filePath}:${chunk.metadata.startLine}`,
-                        content: chunk.text,
-                        metadata: {
-                            repository: chunk.metadata.repository,
-                            filePath: chunk.metadata.filePath,
-                            language: chunk.metadata.language,
-                            startLine: chunk.metadata.startLine,
-                            endLine: chunk.metadata.endLine,
-                            codeType: chunk.metadata.codeUnitType || "unknown",
-                            codeName: chunk.metadata.codeUnitName || "",
-                            isPublic: chunk.metadata.isPublic,
-                            repoVersion: chunk.metadata.repoVersion,
-                            pragmaVersion: chunk.metadata.pragmaVersion,
-                            indexedAt: chunk.metadata.indexedAt,
-                        },
-                    });
-                    chunkCount++;
-                }
-            }
-        }
-        // Generate embeddings and store
-        if (documents.length > 0) {
-            const texts = documents.map((d) => d.content);
-            const embeddings = await embeddingGenerator.generateEmbeddings(texts);
-            for (let i = 0; i < documents.length; i++) {
-                documents[i].embedding = embeddings[i].embedding;
-            }
-            await vectorStore.addDocuments(documents);
-        }
-        return { fileCount: filteredPaths.length, chunkCount };
-    }
-    catch (error) {
-        logger.error(`Failed incremental update for ${repoName}`, {
-            error: String(error),
-        });
-        throw error;
-    }
-}
-//# sourceMappingURL=indexer.js.map

package/dist/pipeline/parser.d.ts

@@ -1,46 +0,0 @@
-export interface CodeUnit {
-    type: "ledger" | "circuit" | "witness" | "function" | "type" | "import" | "export" | "class" | "interface";
-    name: string;
-    code: string;
-    startLine: number;
-    endLine: number;
-    isPublic: boolean;
-    isPrivate: boolean;
-    documentation?: string;
-    parameters?: Array<{
-        name: string;
-        type: string;
-    }>;
-    returnType?: string;
-}
-export interface ParsedFile {
-    path: string;
-    language: "compact" | "typescript" | "markdown";
-    content: string;
-    codeUnits: CodeUnit[];
-    imports: string[];
-    exports: string[];
-    metadata: {
-        hasLedger: boolean;
-        hasCircuits: boolean;
-        hasWitnesses: boolean;
-        lineCount: number;
-    };
-}
-/**
- * Parse Compact smart contract files
- */
-export declare function parseCompactFile(path: string, content: string): ParsedFile;
-/**
- * Parse TypeScript files
- */
-export declare function parseTypeScriptFile(path: string, content: string): ParsedFile;
-/**
- * Parse Markdown documentation files
- */
-export declare function parseMarkdownFile(path: string, content: string): ParsedFile;
-/**
- * Parse a file based on its extension
- */
-export declare function parseFile(path: string, content: string): ParsedFile;
-//# sourceMappingURL=parser.d.ts.map