studiograph 1.1.2 → 1.2.0-beta.1

package/dist/services/sync/llm.js

@@ -0,0 +1,115 @@
+/**
+ * LLM utilities for the sync pipeline — powered by the Vercel AI SDK.
+ *
+ * - llmExtract: Structured extraction from unstructured content
+ * - llmEnrich: Enrichment with cross-source context
+ * - llmGraphMatch: Entity deduplication comparison
+ *
+ * All three functions use generateText + Output.object() for Zod-validated
+ * structured output. Provider is configurable via user config:
+ *   enrichment_model_id → model_id → 'claude-sonnet-4-5-20250929'
+ *   enrichment_model_provider → model_provider → 'anthropic'
+ */
+import { generateText, Output } from 'ai';
+import { createModel } from './model-factory.js';
+import { ExtractOutputSchema, EnrichOutputSchema, GraphMatchOutputSchema, flattenCustomFields, } from './output-schemas.js';
+import { buildExtractionPrompt, buildEnrichmentPrompt, buildGraphMatchPrompt, describeSchema } from './prompts.js';
+import { loadUserConfig, getAnthropicCredential } from '../../core/user-config.js';
+const DEFAULT_MODEL = 'claude-sonnet-4-5-20250929';
+const DEFAULT_PROVIDER = 'anthropic';
+function resolveModel(config) {
+    return {
+        provider: config.enrichment_model_provider ?? config.model_provider ?? DEFAULT_PROVIDER,
+        model: config.enrichment_model_id ?? config.model_id ?? DEFAULT_MODEL,
+    };
+}
+function resolveApiKey(config, _provider) {
+    const configKey = getAnthropicCredential(config);
+    if (configKey)
+        return configKey;
+    return undefined;
+}
+const CUSTOM_FIELDS_INSTRUCTION = '\n\nIMPORTANT: Put any extracted fields that are NOT in the base schema (entity_type, entity_id, name) into the `custom_fields` object.';
+// Cast Output to suppress TS2589 "excessively deep" type instantiation from
+// AI SDK generics × Zod. Runtime validation is still handled by Zod at parse time.
+const OutputAny = Output;
+const extractOutput = OutputAny.object({ schema: ExtractOutputSchema });
+const enrichOutput = OutputAny.object({ schema: EnrichOutputSchema });
+const graphMatchOutput = OutputAny.object({ schema: GraphMatchOutputSchema });
+/**
+ * Extract structured frontmatter from content using LLM.
+ */
+export async function llmExtract(content, entityType, schema, graphragContext) {
+    const config = loadUserConfig();
+    const { provider, model: modelId } = resolveModel(config);
+    const apiKey = resolveApiKey(config, provider);
+    const schemaDescription = describeSchema(schema);
+    const prompt = buildExtractionPrompt(content, entityType, schemaDescription, graphragContext) + CUSTOM_FIELDS_INSTRUCTION;
+    const llmModel = createModel(provider, modelId, apiKey);
+    const result = await generateText({
+        model: llmModel,
+        prompt,
+        output: extractOutput,
+        maxOutputTokens: 2048,
+        maxRetries: 2,
+    });
+    const extracted = result.output;
+    if (!extracted?.frontmatter) {
+        throw new Error('LLM extraction returned no frontmatter');
+    }
+    return flattenCustomFields(extracted.frontmatter);
+}
+/**
+ * Enrich an entity with cross-source context using LLM.
+ * Returns updated frontmatter and generated content.
+ */
+export async function llmEnrich(entityData, contextBundle, entityType, schema) {
+    const config = loadUserConfig();
+    const { provider, model: modelId } = resolveModel(config);
+    const apiKey = resolveApiKey(config, provider);
+    const schemaDescription = describeSchema(schema);
+    const prompt = buildEnrichmentPrompt(entityType, entityData, contextBundle, schemaDescription) + CUSTOM_FIELDS_INSTRUCTION;
+    const llmModel = createModel(provider, modelId, apiKey);
+    const result = await generateText({
+        model: llmModel,
+        prompt,
+        output: enrichOutput,
+        maxOutputTokens: 4096,
+        maxRetries: 2,
+    });
+    const enriched = result.output;
+    if (!enriched?.frontmatter) {
+        throw new Error('LLM enrichment returned no frontmatter');
+    }
+    return {
+        frontmatter: flattenCustomFields(enriched.frontmatter),
+        content: enriched.content ?? '',
+    };
+}
+/**
+ * Ask LLM whether two entities (staged vs existing) are the same real-world entity.
+ */
+export async function llmGraphMatch(stagedFrontmatter, existingFrontmatter, entityType) {
+    const config = loadUserConfig();
+    const { provider, model: modelId } = resolveModel(config);
+    const apiKey = resolveApiKey(config, provider);
+    const prompt = buildGraphMatchPrompt(stagedFrontmatter, existingFrontmatter, entityType);
+    const llmModel = createModel(provider, modelId, apiKey);
+    const result = await generateText({
+        model: llmModel,
+        prompt,
+        output: graphMatchOutput,
+        maxOutputTokens: 1024,
+        maxRetries: 2,
+    });
+    const match = result.output;
+    if (!match) {
+        throw new Error('LLM graph match returned no output');
+    }
+    return {
+        is_same: match.is_same,
+        confidence: match.confidence ?? 0.5,
+        reasoning: match.reasoning ?? '',
+    };
+}
+//# sourceMappingURL=llm.js.map

package/dist/services/sync/llm.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"llm.js","sourceRoot":"","sources":["../../../src/services/sync/llm.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;GAWG;AAEH,OAAO,EAAE,YAAY,EAAE,MAAM,EAAE,MAAM,IAAI,CAAC;AAC1C,OAAO,EAAE,WAAW,EAAE,MAAM,oBAAoB,CAAC;AACjD,OAAO,EACL,mBAAmB,EACnB,kBAAkB,EAClB,sBAAsB,EACtB,mBAAmB,GACpB,MAAM,qBAAqB,CAAC;AAC7B,OAAO,EAAE,qBAAqB,EAAE,qBAAqB,EAAE,qBAAqB,EAAE,cAAc,EAAE,MAAM,cAAc,CAAC;AACnH,OAAO,EAAE,cAAc,EAAE,sBAAsB,EAAE,MAAM,2BAA2B,CAAC;AAGnF,MAAM,aAAa,GAAG,4BAA4B,CAAC;AACnD,MAAM,gBAAgB,GAAG,WAAW,CAAC;AAarC,SAAS,YAAY,CAAC,MAAkB;IACtC,OAAO;QACL,QAAQ,EAAE,MAAM,CAAC,yBAAyB,IAAI,MAAM,CAAC,cAAc,IAAI,gBAAgB;QACvF,KAAK,EAAE,MAAM,CAAC,mBAAmB,IAAI,MAAM,CAAC,QAAQ,IAAI,aAAa;KACtE,CAAC;AACJ,CAAC;AAED,SAAS,aAAa,CAAC,MAAkB,EAAE,SAAiB;IAC1D,MAAM,SAAS,GAAG,sBAAsB,CAAC,MAAM,CAAC,CAAC;IACjD,IAAI,SAAS;QAAE,OAAO,SAAS,CAAC;IAChC,OAAO,SAAS,CAAC;AACnB,CAAC;AAED,MAAM,yBAAyB,GAAG,yIAAyI,CAAC;AAE5K,4EAA4E;AAC5E,mFAAmF;AACnF,MAAM,SAAS,GAAG,MAAa,CAAC;AAChC,MAAM,aAAa,GAAG,SAAS,CAAC,MAAM,CAAC,EAAE,MAAM,EAAE,mBAAmB,EAAE,CAAC,CAAC;AACxE,MAAM,YAAY,GAAG,SAAS,CAAC,MAAM,CAAC,EAAE,MAAM,EAAE,kBAAkB,EAAE,CAAC,CAAC;AACtE,MAAM,gBAAgB,GAAG,SAAS,CAAC,MAAM,CAAC,EAAE,MAAM,EAAE,sBAAsB,EAAE,CAAC,CAAC;AAE9E;;GAEG;AACH,MAAM,CAAC,KAAK,UAAU,UAAU,CAC9B,OAAe,EACf,UAAkB,EAClB,MAAW,EACX,eAAwB;IAExB,MAAM,MAAM,GAAG,cAAc,EAAE,CAAC;IAChC,MAAM,EAAE,QAAQ,EAAE,KAAK,EAAE,OAAO,EAAE,GAAG,YAAY,CAAC,MAAM,CAAC,CAAC;IAC1D,MAAM,MAAM,GAAG,aAAa,CAAC,MAAM,EAAE,QAAQ,CAAC,CAAC;IAE/C,MAAM,iBAAiB,GAAG,cAAc,CAAC,MAAM,CAAC,CAAC;IACjD,MAAM,MAAM,GAAG,qBAAqB,CAAC,OAAO,EAAE,UAAU,EAAE,iBAAiB,EAAE,eAAe,CAAC,GAAG,yBAAyB,CAAC;IAE1H,MAAM,QAAQ,GAAG,WAAW,CAAC,QAAQ,EAAE,OAAO,EAAE,MAAM,CAAC,CAAC;IACxD,MAAM,MAAM,GAAG,MAAM,YAAY,CAAC;QAChC,KAAK,EAAE,QAAQ;QACf,MAAM;QACN,MAAM,EAAE,aAAa;QACrB,eAAe,EAAE,IAAI;QACrB,UAAU,EAAE,CAAC;KACd,CAAC,CAAC;IAEH,MAAM,SAAS,GAAI,MAAc,CAAC,MAAmC,CAAC;IACtE,IAAI,CAAC,SAAS,EAAE,WAAW,EAAE,CAAC;QAC5B,MAAM,IAAI,KAAK,CAAC,wCAAwC,CAAC,CAAC;IAC5D,CAAC;IACD,OAAO,mBAAmB,CAAC,SAAS,CAAC,WAAW,CAAC,CAAC;AACpD,CAAC;AAED;;;GAGG;AACH,MAAM,CAAC,KAAK,UAAU,SAAS,CAC7B,UAA+B,EAC/B,aAAqB,EACrB,UAAkB,EAClB,MAAW;IAEX,MAAM,MAAM,GAAG,cAAc,EAAE,CAAC;IAChC,MAAM,EAAE,QAAQ,EAAE,KAAK,EAAE,OAAO,EAAE,GAAG,YAAY,CAAC,MAAM,CAAC,CAAC;IAC1D,MAAM,MAAM,GAAG,aAAa,CAAC,MAAM,EAAE,QAAQ,CAAC,CAAC;IAE/C,MAAM,iBAAiB,GAAG,cAAc,CAAC,MAAM,CAAC,CAAC;IACjD,MAAM,MAAM,GAAG,qBAAqB,CAAC,UAAU,EAAE,UAAU,EAAE,aAAa,EAAE,iBAAiB,CAAC,GAAG,yBAAyB,CAAC;IAE3H,MAAM,QAAQ,GAAG,WAAW,CAAC,QAAQ,EAAE,OAAO,EAAE,MAAM,CAAC,CAAC;IACxD,MAAM,MAAM,GAAG,MAAM,YAAY,CAAC;QAChC,KAAK,EAAE,QAAQ;QACf,MAAM;QACN,MAAM,EAAE,YAAY;QACpB,eAAe,EAAE,IAAI;QACrB,UAAU,EAAE,CAAC;KACd,CAAC,CAAC;IAEH,MAAM,QAAQ,GAAI,MAAc,CAAC,MAAkC,CAAC;IACpE,IAAI,CAAC,QAAQ,EAAE,WAAW,EAAE,CAAC;QAC3B,MAAM,IAAI,KAAK,CAAC,wCAAwC,CAAC,CAAC;IAC5D,CAAC;IACD,OAAO;QACL,WAAW,EAAE,mBAAmB,CAAC,QAAQ,CAAC,WAAW,CAAC;QACtD,OAAO,EAAE,QAAQ,CAAC,OAAO,IAAI,EAAE;KAChC,CAAC;AACJ,CAAC;AAED;;GAEG;AACH,MAAM,CAAC,KAAK,UAAU,aAAa,CACjC,iBAAsC,EACtC,mBAAwC,EACxC,UAAkB;IAElB,MAAM,MAAM,GAAG,cAAc,EAAE,CAAC;IAChC,MAAM,EAAE,QAAQ,EAAE,KAAK,EAAE,OAAO,EAAE,GAAG,YAAY,CAAC,MAAM,CAAC,CAAC;IAC1D,MAAM,MAAM,GAAG,aAAa,CAAC,MAAM,EAAE,QAAQ,CAAC,CAAC;IAE/C,MAAM,MAAM,GAAG,qBAAqB,CAAC,iBAAiB,EAAE,mBAAmB,EAAE,UAAU,CAAC,CAAC;IAEzF,MAAM,QAAQ,GAAG,WAAW,CAAC,QAAQ,EAAE,OAAO,EAAE,MAAM,CAAC,CAAC;IACxD,MAAM,MAAM,GAAG,MAAM,YAAY,CAAC;QAChC,KAAK,EAAE,QAAQ;QACf,MAAM;QACN,MAAM,EAAE,gBAAgB;QACxB,eAAe,EAAE,IAAI;QACrB,UAAU,EAAE,CAAC;KACd,CAAC,CAAC;IAEH,MAAM,KAAK,GAAI,MAAc,CAAC,MAAsC,CAAC;IACrE,IAAI,CAAC,KAAK,EAAE,CAAC;QACX,MAAM,IAAI,KAAK,CAAC,oCAAoC,CAAC,CAAC;IACxD,CAAC;IACD,OAAO;QACL,OAAO,EAAE,KAAK,CAAC,OAAO;QACtB,UAAU,EAAE,KAAK,CAAC,UAAU,IAAI,GAAG;QACnC,SAAS,EAAE,KAAK,CAAC,SAAS,IAAI,EAAE;KACjC,CAAC;AACJ,CAAC"}

package/dist/services/sync/mcp-client.d.ts

@@ -0,0 +1,59 @@
+/**
+ * SyncMCPClient
+ *
+ * Persistent MCP client for the sync pipeline. Holds a single connection
+ * for the entire sync run to avoid repeated handshake overhead.
+ *
+ * 30s timeout on all MCP calls via Promise.race.
+ */
+import type { ConnectorConfig } from '../../mcp/connector-manager.js';
+import type { DetectedEntityType } from './types.js';
+interface MCPTool {
+    name: string;
+    description?: string;
+    inputSchema?: Record<string, any>;
+}
+export declare class SyncMCPClient {
+    private client;
+    private config;
+    private tools;
+    private connected;
+    constructor(connectorConfig: ConnectorConfig);
+    /**
+     * Connect to the MCP server.
+     */
+    connect(): Promise<void>;
+    /**
+     * Close the MCP connection.
+     */
+    close(): Promise<void>;
+    /**
+     * Get available tools from the connected server.
+     */
+    getTools(): MCPTool[];
+    /**
+     * Call an MCP tool with a 30s timeout.
+     */
+    callTool(name: string, args?: Record<string, any>): Promise<string>;
+    /**
+     * Detect available entity types by inspecting tool names.
+     * Looks for list_* / get_* tool pairs.
+     */
+    detectAvailableEntityTypes(): DetectedEntityType[];
+    /**
+     * Find the list tool for a given source type.
+     * Checks exact patterns first, then falls back to a broader search
+     * for tools containing "list" and the source type (handles prefixed
+     * tool names like list_granola_documents).
+     */
+    findListTool(sourceType: string): string | null;
+    /**
+     * Find the get/detail tool for a given source type.
+     */
+    findGetTool(sourceType: string): string | null;
+    /**
+     * Detect the ID parameter name from a get tool's input schema.
+     */
+    getToolIdParam(toolName: string): string;
+}
+export {};

package/dist/services/sync/mcp-client.js

@@ -0,0 +1,285 @@
+/**
+ * SyncMCPClient
+ *
+ * Persistent MCP client for the sync pipeline. Holds a single connection
+ * for the entire sync run to avoid repeated handshake overhead.
+ *
+ * 30s timeout on all MCP calls via Promise.race.
+ */
+import { Client } from '@modelcontextprotocol/sdk/client/index.js';
+import { StreamableHTTPClientTransport } from '@modelcontextprotocol/sdk/client/streamableHttp.js';
+import { StdioClientTransport } from '@modelcontextprotocol/sdk/client/stdio.js';
+import { ConnectorManager } from '../../mcp/connector-manager.js';
+const CALL_TIMEOUT = 120_000;
+export class SyncMCPClient {
+    client;
+    config;
+    tools = [];
+    connected = false;
+    constructor(connectorConfig) {
+        this.config = connectorConfig;
+        this.client = new Client({ name: 'studiograph-sync', version: '1.0.0' });
+    }
+    /**
+     * Connect to the MCP server.
+     */
+    async connect() {
+        if (this.connected)
+            return;
+        const transport = this.config.type === 'stdio'
+            ? new StdioClientTransport({
+                command: this.config.command,
+                args: this.config.args,
+                env: this.config.env,
+                stderr: 'ignore',
+            })
+            : new StreamableHTTPClientTransport(new URL(this.config.url), {
+                requestInit: {
+                    headers: { Authorization: `Bearer ${ConnectorManager.resolveToken(this.config.token)}` },
+                },
+            });
+        await this.client.connect(transport);
+        const { tools } = await this.client.listTools();
+        this.tools = tools.map(t => ({
+            name: t.name,
+            description: t.description,
+            inputSchema: t.inputSchema,
+        }));
+        this.connected = true;
+    }
+    /**
+     * Close the MCP connection.
+     */
+    async close() {
+        if (!this.connected)
+            return;
+        try {
+            await this.client.close();
+        }
+        catch {
+            // ignore close errors
+        }
+        this.connected = false;
+    }
+    /**
+     * Get available tools from the connected server.
+     */
+    getTools() {
+        return this.tools;
+    }
+    /**
+     * Call an MCP tool with a 30s timeout.
+     */
+    async callTool(name, args) {
+        if (!this.connected) {
+            throw new Error('SyncMCPClient not connected. Call connect() first.');
+        }
+        const call = this.client.callTool({ name, arguments: args ?? {} });
+        let timer;
+        const timeout = new Promise((_, reject) => {
+            timer = setTimeout(() => reject(new Error(`MCP call to "${name}" timed out after ${CALL_TIMEOUT}ms`)), CALL_TIMEOUT);
+        });
+        const result = await Promise.race([call, timeout]).finally(() => clearTimeout(timer));
+        const text = result.content
+            .filter((c) => c.type === 'text')
+            .map((c) => c.text)
+            .join('\n');
+        let output = text || JSON.stringify(result, null, 2);
+        // Handle double-wrapped MCP responses — some servers (e.g. Pipedrive)
+        // return { content: [{ type: "text", text: "..." }] } as their text payload
+        try {
+            const parsed = JSON.parse(output);
+            if (parsed.content && Array.isArray(parsed.content)) {
+                const inner = parsed.content
+                    .filter((c) => c.type === 'text')
+                    .map((c) => c.text)
+                    .join('\n');
+                if (inner)
+                    output = inner;
+            }
+        }
+        catch {
+            // not JSON or not double-wrapped — use as-is
+        }
+        return output;
+    }
+    /**
+     * Detect available entity types by inspecting tool names.
+     * Looks for list_* / get_* tool pairs.
+     */
+    detectAvailableEntityTypes() {
+        const detected = [];
+        const toolNames = new Set(this.tools.map(t => t.name));
+        // Common patterns: list_<type>s, search_<type>s, get_<type>
+        // Also: <type>_list, <type>_search, <type>_get
+        for (const tool of this.tools) {
+            let sourceType = null;
+            let isList = false;
+            // Pattern: list_<type>s or list_<type>
+            const listMatch = tool.name.match(/^list_(.+?)s?$/);
+            if (listMatch) {
+                sourceType = listMatch[1];
+                isList = true;
+            }
+            // Pattern: search_<type>s or search_<type>
+            if (!sourceType) {
+                const searchMatch = tool.name.match(/^search_(.+?)s?$/);
+                if (searchMatch) {
+                    sourceType = searchMatch[1];
+                    isList = true;
+                }
+            }
+            // Pattern: <prefix>__list_<type>s (namespaced)
+            if (!sourceType) {
+                const nsListMatch = tool.name.match(/^[a-z_]+__list_(.+?)s?$/);
+                if (nsListMatch) {
+                    sourceType = nsListMatch[1];
+                    isList = true;
+                }
+            }
+            // Pattern: <prefix>__search_<type>s (namespaced)
+            if (!sourceType) {
+                const nsSearchMatch = tool.name.match(/^[a-z_]+__search_(.+?)s?$/);
+                if (nsSearchMatch) {
+                    sourceType = nsSearchMatch[1];
+                    isList = true;
+                }
+            }
+            // Pattern: <type>s_list or <type>s_search
+            if (!sourceType) {
+                const suffixMatch = tool.name.match(/^(.+?)s?_(?:list|search)$/);
+                if (suffixMatch) {
+                    sourceType = suffixMatch[1];
+                    isList = true;
+                }
+            }
+            if (!sourceType || !isList)
+                continue;
+            // Avoid duplicates
+            if (detected.some(d => d.source_type === sourceType))
+                continue;
+            // Look for corresponding get tool
+            const getVariants = [
+                `get_${sourceType}`,
+                `${sourceType}_get`,
+                `read_${sourceType}`,
+            ];
+            const getTool = getVariants.find(g => toolNames.has(g));
+            detected.push({
+                source_type: sourceType,
+                list_tool: tool.name,
+                get_tool: getTool,
+                suggested_entity_type: suggestEntityType(sourceType),
+                suggested_repo: suggestRepo(sourceType),
+            });
+        }
+        return detected;
+    }
+    /**
+     * Find the list tool for a given source type.
+     * Checks exact patterns first, then falls back to a broader search
+     * for tools containing "list" and the source type (handles prefixed
+     * tool names like list_granola_documents).
+     */
+    findListTool(sourceType) {
+        const candidates = [
+            `list_${sourceType}s`,
+            `list_${sourceType}`,
+            `${sourceType}s_list`,
+            `${sourceType}_list`,
+            `${sourceType}s_list_all_auto`,
+            `${sourceType}_list_all_auto`,
+            // Search-based APIs (e.g. Asana)
+            `search_${sourceType}s`,
+            `search_${sourceType}`,
+            `${sourceType}s_search`,
+            `${sourceType}_search`,
+        ];
+        for (const name of candidates) {
+            if (this.tools.some(t => t.name === name))
+                return name;
+        }
+        // Fallback: find any tool containing "list" or "search" and the source type
+        const plural = sourceType.endsWith('s') ? sourceType : sourceType + 's';
+        const match = this.tools.find(t => (t.name.includes('list') || t.name.includes('search')) &&
+            (t.name.includes(plural) || t.name.includes(sourceType)));
+        return match?.name ?? null;
+    }
+    /**
+     * Find the get/detail tool for a given source type.
+     */
+    findGetTool(sourceType) {
+        const candidates = [
+            `get_${sourceType}`,
+            `${sourceType}_get`,
+            `${sourceType}s_get`,
+            `read_${sourceType}`,
+        ];
+        for (const name of candidates) {
+            if (this.tools.some(t => t.name === name))
+                return name;
+        }
+        // Fallback: find any tool containing "get" and the source type
+        const match = this.tools.find(t => (t.name.includes('get') || t.name.includes('read')) &&
+            t.name.includes(sourceType) &&
+            !t.name.includes('list') &&
+            !t.name.includes('search'));
+        return match?.name ?? null;
+    }
+    /**
+     * Detect the ID parameter name from a get tool's input schema.
+     */
+    getToolIdParam(toolName) {
+        const tool = this.tools.find(t => t.name === toolName);
+        if (!tool?.inputSchema?.properties)
+            return 'id';
+        const required = tool.inputSchema.required ?? [];
+        // Return the first required param (usually "id")
+        for (const param of required) {
+            const prop = tool.inputSchema.properties[param];
+            if (prop && (prop.type === 'string' || prop.type === 'number')) {
+                return param;
+            }
+        }
+        return 'id';
+    }
+}
+// ── Helpers ──────────────────────────────────────────────────────────────────
+const TYPE_MAP = {
+    deal: 'project',
+    deals: 'project',
+    person: 'person',
+    persons: 'person',
+    people: 'person',
+    organization: 'client',
+    organizations: 'client',
+    company: 'client',
+    companies: 'client',
+    document: 'meeting',
+    documents: 'meeting',
+    note: 'meeting',
+    notes: 'meeting',
+    activity: 'task',
+    activities: 'task',
+    product: 'deliverable',
+    products: 'deliverable',
+};
+const REPO_MAP = {
+    project: 'projects',
+    person: 'people',
+    client: 'clients',
+    meeting: 'meetings',
+    task: 'tasks',
+    deliverable: 'deliverables',
+    deal: 'deals',
+};
+function suggestEntityType(sourceType) {
+    return TYPE_MAP[sourceType.toLowerCase()];
+}
+function suggestRepo(sourceType) {
+    const entityType = suggestEntityType(sourceType);
+    if (!entityType)
+        return undefined;
+    return REPO_MAP[entityType];
+}
+//# sourceMappingURL=mcp-client.js.map

package/dist/services/sync/mcp-client.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"mcp-client.js","sourceRoot":"","sources":["../../../src/services/sync/mcp-client.ts"],"names":[],"mappings":"AAAA;;;;;;;GAOG;AAEH,OAAO,EAAE,MAAM,EAAE,MAAM,2CAA2C,CAAC;AACnE,OAAO,EAAE,6BAA6B,EAAE,MAAM,oDAAoD,CAAC;AACnG,OAAO,EAAE,oBAAoB,EAAE,MAAM,2CAA2C,CAAC;AACjF,OAAO,EAAE,gBAAgB,EAAE,MAAM,gCAAgC,CAAC;AAIlE,MAAM,YAAY,GAAG,OAAO,CAAC;AAQ7B,MAAM,OAAO,aAAa;IAChB,MAAM,CAAS;IACf,MAAM,CAAkB;IACxB,KAAK,GAAc,EAAE,CAAC;IACtB,SAAS,GAAG,KAAK,CAAC;IAE1B,YAAY,eAAgC;QAC1C,IAAI,CAAC,MAAM,GAAG,eAAe,CAAC;QAC9B,IAAI,CAAC,MAAM,GAAG,IAAI,MAAM,CAAC,EAAE,IAAI,EAAE,kBAAkB,EAAE,OAAO,EAAE,OAAO,EAAE,CAAC,CAAC;IAC3E,CAAC;IAED;;OAEG;IACH,KAAK,CAAC,OAAO;QACX,IAAI,IAAI,CAAC,SAAS;YAAE,OAAO;QAE3B,MAAM,SAAS,GAAG,IAAI,CAAC,MAAM,CAAC,IAAI,KAAK,OAAO;YAC5C,CAAC,CAAC,IAAI,oBAAoB,CAAC;gBACvB,OAAO,EAAE,IAAI,CAAC,MAAM,CAAC,OAAQ;gBAC7B,IAAI,EAAE,IAAI,CAAC,MAAM,CAAC,IAAI;gBACtB,GAAG,EAAE,IAAI,CAAC,MAAM,CAAC,GAAG;gBACpB,MAAM,EAAE,QAAQ;aACjB,CAAC;YACJ,CAAC,CAAC,IAAI,6BAA6B,CAAC,IAAI,GAAG,CAAC,IAAI,CAAC,MAAM,CAAC,GAAI,CAAC,EAAE;gBAC3D,WAAW,EAAE;oBACX,OAAO,EAAE,EAAE,aAAa,EAAE,UAAU,gBAAgB,CAAC,YAAY,CAAC,IAAI,CAAC,MAAM,CAAC,KAAM,CAAC,EAAE,EAAE;iBAC1F;aACF,CAAC,CAAC;QAEP,MAAM,IAAI,CAAC,MAAM,CAAC,OAAO,CAAC,SAAS,CAAC,CAAC;QAErC,MAAM,EAAE,KAAK,EAAE,GAAG,MAAM,IAAI,CAAC,MAAM,CAAC,SAAS,EAAE,CAAC;QAChD,IAAI,CAAC,KAAK,GAAG,KAAK,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC;YAC3B,IAAI,EAAE,CAAC,CAAC,IAAI;YACZ,WAAW,EAAE,CAAC,CAAC,WAAW;YAC1B,WAAW,EAAE,CAAC,CAAC,WAAkC;SAClD,CAAC,CAAC,CAAC;QAEJ,IAAI,CAAC,SAAS,GAAG,IAAI,CAAC;IACxB,CAAC;IAED;;OAEG;IACH,KAAK,CAAC,KAAK;QACT,IAAI,CAAC,IAAI,CAAC,SAAS;YAAE,OAAO;QAC5B,IAAI,CAAC;YACH,MAAM,IAAI,CAAC,MAAM,CAAC,KAAK,EAAE,CAAC;QAC5B,CAAC;QAAC,MAAM,CAAC;YACP,sBAAsB;QACxB,CAAC;QACD,IAAI,CAAC,SAAS,GAAG,KAAK,CAAC;IACzB,CAAC;IAED;;OAEG;IACH,QAAQ;QACN,OAAO,IAAI,CAAC,KAAK,CAAC;IACpB,CAAC;IAED;;OAEG;IACH,KAAK,CAAC,QAAQ,CAAC,IAAY,EAAE,IAA0B;QACrD,IAAI,CAAC,IAAI,CAAC,SAAS,EAAE,CAAC;YACpB,MAAM,IAAI,KAAK,CAAC,oDAAoD,CAAC,CAAC;QACxE,CAAC;QAED,MAAM,IAAI,GAAG,IAAI,CAAC,MAAM,CAAC,QAAQ,CAAC,EAAE,IAAI,EAAE,SAAS,EAAE,IAAI,IAAI,EAAE,EAAE,CAAC,CAAC;QACnE,IAAI,KAAoC,CAAC;QACzC,MAAM,OAAO,GAAG,IAAI,OAAO,CAAQ,CAAC,CAAC,EAAE,MAAM,EAAE,EAAE;YAC/C,KAAK,GAAG,UAAU,CAAC,GAAG,EAAE,CAAC,MAAM,CAAC,IAAI,KAAK,CAAC,gBAAgB,IAAI,qBAAqB,YAAY,IAAI,CAAC,CAAC,EAAE,YAAY,CAAC,CAAC;QACvH,CAAC,CAAC,CAAC;QAEH,MAAM,MAAM,GAAG,MAAM,OAAO,CAAC,IAAI,CAAC,CAAC,IAAI,EAAE,OAAO,CAAC,CAAC,CAAC,OAAO,CAAC,GAAG,EAAE,CAAC,YAAY,CAAC,KAAM,CAAC,CAAC,CAAC;QACvF,MAAM,IAAI,GAAI,MAAM,CAAC,OAAiB;aACnC,MAAM,CAAC,CAAC,CAAM,EAAE,EAAE,CAAC,CAAC,CAAC,IAAI,KAAK,MAAM,CAAC;aACrC,GAAG,CAAC,CAAC,CAAM,EAAE,EAAE,CAAC,CAAC,CAAC,IAAI,CAAC;aACvB,IAAI,CAAC,IAAI,CAAC,CAAC;QAEd,IAAI,MAAM,GAAG,IAAI,IAAI,IAAI,CAAC,SAAS,CAAC,MAAM,EAAE,IAAI,EAAE,CAAC,CAAC,CAAC;QAErD,sEAAsE;QACtE,4EAA4E;QAC5E,IAAI,CAAC;YACH,MAAM,MAAM,GAAG,IAAI,CAAC,KAAK,CAAC,MAAM,CAAC,CAAC;YAClC,IAAI,MAAM,CAAC,OAAO,IAAI,KAAK,CAAC,OAAO,CAAC,MAAM,CAAC,OAAO,CAAC,EAAE,CAAC;gBACpD,MAAM,KAAK,GAAG,MAAM,CAAC,OAAO;qBACzB,MAAM,CAAC,CAAC,CAAM,EAAE,EAAE,CAAC,CAAC,CAAC,IAAI,KAAK,MAAM,CAAC;qBACrC,GAAG,CAAC,CAAC,CAAM,EAAE,EAAE,CAAC,CAAC,CAAC,IAAI,CAAC;qBACvB,IAAI,CAAC,IAAI,CAAC,CAAC;gBACd,IAAI,KAAK;oBAAE,MAAM,GAAG,KAAK,CAAC;YAC5B,CAAC;QACH,CAAC;QAAC,MAAM,CAAC;YACP,6CAA6C;QAC/C,CAAC;QAED,OAAO,MAAM,CAAC;IAChB,CAAC;IAED;;;OAGG;IACH,0BAA0B;QACxB,MAAM,QAAQ,GAAyB,EAAE,CAAC;QAC1C,MAAM,SAAS,GAAG,IAAI,GAAG,CAAC,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC;QAEvD,4DAA4D;QAC5D,+CAA+C;QAC/C,KAAK,MAAM,IAAI,IAAI,IAAI,CAAC,KAAK,EAAE,CAAC;YAC9B,IAAI,UAAU,GAAkB,IAAI,CAAC;YACrC,IAAI,MAAM,GAAG,KAAK,CAAC;YAEnB,uCAAuC;YACvC,MAAM,SAAS,GAAG,IAAI,CAAC,IAAI,CAAC,KAAK,CAAC,gBAAgB,CAAC,CAAC;YACpD,IAAI,SAAS,EAAE,CAAC;gBACd,UAAU,GAAG,SAAS,CAAC,CAAC,CAAC,CAAC;gBAC1B,MAAM,GAAG,IAAI,CAAC;YAChB,CAAC;YAED,2CAA2C;YAC3C,IAAI,CAAC,UAAU,EAAE,CAAC;gBAChB,MAAM,WAAW,GAAG,IAAI,CAAC,IAAI,CAAC,KAAK,CAAC,kBAAkB,CAAC,CAAC;gBACxD,IAAI,WAAW,EAAE,CAAC;oBAChB,UAAU,GAAG,WAAW,CAAC,CAAC,CAAC,CAAC;oBAC5B,MAAM,GAAG,IAAI,CAAC;gBAChB,CAAC;YACH,CAAC;YAED,+CAA+C;YAC/C,IAAI,CAAC,UAAU,EAAE,CAAC;gBAChB,MAAM,WAAW,GAAG,IAAI,CAAC,IAAI,CAAC,KAAK,CAAC,yBAAyB,CAAC,CAAC;gBAC/D,IAAI,WAAW,EAAE,CAAC;oBAChB,UAAU,GAAG,WAAW,CAAC,CAAC,CAAC,CAAC;oBAC5B,MAAM,GAAG,IAAI,CAAC;gBAChB,CAAC;YACH,CAAC;YAED,iDAAiD;YACjD,IAAI,CAAC,UAAU,EAAE,CAAC;gBAChB,MAAM,aAAa,GAAG,IAAI,CAAC,IAAI,CAAC,KAAK,CAAC,2BAA2B,CAAC,CAAC;gBACnE,IAAI,aAAa,EAAE,CAAC;oBAClB,UAAU,GAAG,aAAa,CAAC,CAAC,CAAC,CAAC;oBAC9B,MAAM,GAAG,IAAI,CAAC;gBAChB,CAAC;YACH,CAAC;YAED,0CAA0C;YAC1C,IAAI,CAAC,UAAU,EAAE,CAAC;gBAChB,MAAM,WAAW,GAAG,IAAI,CAAC,IAAI,CAAC,KAAK,CAAC,2BAA2B,CAAC,CAAC;gBACjE,IAAI,WAAW,EAAE,CAAC;oBAChB,UAAU,GAAG,WAAW,CAAC,CAAC,CAAC,CAAC;oBAC5B,MAAM,GAAG,IAAI,CAAC;gBAChB,CAAC;YACH,CAAC;YAED,IAAI,CAAC,UAAU,IAAI,CAAC,MAAM;gBAAE,SAAS;YAErC,mBAAmB;YACnB,IAAI,QAAQ,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,WAAW,KAAK,UAAU,CAAC;gBAAE,SAAS;YAE/D,kCAAkC;YAClC,MAAM,WAAW,GAAG;gBAClB,OAAO,UAAU,EAAE;gBACnB,GAAG,UAAU,MAAM;gBACnB,QAAQ,UAAU,EAAE;aACrB,CAAC;YACF,MAAM,OAAO,GAAG,WAAW,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,SAAS,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC;YAExD,QAAQ,CAAC,IAAI,CAAC;gBACZ,WAAW,EAAE,UAAU;gBACvB,SAAS,EAAE,IAAI,CAAC,IAAI;gBACpB,QAAQ,EAAE,OAAO;gBACjB,qBAAqB,EAAE,iBAAiB,CAAC,UAAU,CAAC;gBACpD,cAAc,EAAE,WAAW,CAAC,UAAU,CAAC;aACxC,CAAC,CAAC;QACL,CAAC;QAED,OAAO,QAAQ,CAAC;IAClB,CAAC;IAED;;;;;OAKG;IACH,YAAY,CAAC,UAAkB;QAC7B,MAAM,UAAU,GAAG;YACjB,QAAQ,UAAU,GAAG;YACrB,QAAQ,UAAU,EAAE;YACpB,GAAG,UAAU,QAAQ;YACrB,GAAG,UAAU,OAAO;YACpB,GAAG,UAAU,iBAAiB;YAC9B,GAAG,UAAU,gBAAgB;YAC7B,iCAAiC;YACjC,UAAU,UAAU,GAAG;YACvB,UAAU,UAAU,EAAE;YACtB,GAAG,UAAU,UAAU;YACvB,GAAG,UAAU,SAAS;SACvB,CAAC;QACF,KAAK,MAAM,IAAI,IAAI,UAAU,EAAE,CAAC;YAC9B,IAAI,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,IAAI,KAAK,IAAI,CAAC;gBAAE,OAAO,IAAI,CAAC;QACzD,CAAC;QAED,4EAA4E;QAC5E,MAAM,MAAM,GAAG,UAAU,CAAC,QAAQ,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,UAAU,CAAC,CAAC,CAAC,UAAU,GAAG,GAAG,CAAC;QACxE,MAAM,KAAK,GAAG,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,CAChC,CAAC,CAAC,CAAC,IAAI,CAAC,QAAQ,CAAC,MAAM,CAAC,IAAI,CAAC,CAAC,IAAI,CAAC,QAAQ,CAAC,QAAQ,CAAC,CAAC;YACtD,CAAC,CAAC,CAAC,IAAI,CAAC,QAAQ,CAAC,MAAM,CAAC,IAAI,CAAC,CAAC,IAAI,CAAC,QAAQ,CAAC,UAAU,CAAC,CAAC,CACzD,CAAC;QACF,OAAO,KAAK,EAAE,IAAI,IAAI,IAAI,CAAC;IAC7B,CAAC;IAED;;OAEG;IACH,WAAW,CAAC,UAAkB;QAC5B,MAAM,UAAU,GAAG;YACjB,OAAO,UAAU,EAAE;YACnB,GAAG,UAAU,MAAM;YACnB,GAAG,UAAU,OAAO;YACpB,QAAQ,UAAU,EAAE;SACrB,CAAC;QACF,KAAK,MAAM,IAAI,IAAI,UAAU,EAAE,CAAC;YAC9B,IAAI,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,IAAI,KAAK,IAAI,CAAC;gBAAE,OAAO,IAAI,CAAC;QACzD,CAAC;QAED,+DAA+D;QAC/D,MAAM,KAAK,GAAG,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,CAChC,CAAC,CAAC,CAAC,IAAI,CAAC,QAAQ,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC,IAAI,CAAC,QAAQ,CAAC,MAAM,CAAC,CAAC;YACnD,CAAC,CAAC,IAAI,CAAC,QAAQ,CAAC,UAAU,CAAC;YAC3B,CAAC,CAAC,CAAC,IAAI,CAAC,QAAQ,CAAC,MAAM,CAAC;YACxB,CAAC,CAAC,CAAC,IAAI,CAAC,QAAQ,CAAC,QAAQ,CAAC,CAC3B,CAAC;QACF,OAAO,KAAK,EAAE,IAAI,IAAI,IAAI,CAAC;IAC7B,CAAC;IAED;;OAEG;IACH,cAAc,CAAC,QAAgB;QAC7B,MAAM,IAAI,GAAG,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,IAAI,KAAK,QAAQ,CAAC,CAAC;QACvD,IAAI,CAAC,IAAI,EAAE,WAAW,EAAE,UAAU;YAAE,OAAO,IAAI,CAAC;QAEhD,MAAM,QAAQ,GAAc,IAAI,CAAC,WAAmB,CAAC,QAAQ,IAAI,EAAE,CAAC;QACpE,iDAAiD;QACjD,KAAK,MAAM,KAAK,IAAI,QAAQ,EAAE,CAAC;YAC7B,MAAM,IAAI,GAAI,IAAI,CAAC,WAAmB,CAAC,UAAU,CAAC,KAAK,CAAC,CAAC;YACzD,IAAI,IAAI,IAAI,CAAC,IAAI,CAAC,IAAI,KAAK,QAAQ,IAAI,IAAI,CAAC,IAAI,KAAK,QAAQ,CAAC,EAAE,CAAC;gBAC/D,OAAO,KAAK,CAAC;YACf,CAAC;QACH,CAAC;QACD,OAAO,IAAI,CAAC;IACd,CAAC;CACF;AAED,gFAAgF;AAEhF,MAAM,QAAQ,GAA2B;IACvC,IAAI,EAAE,SAAS;IACf,KAAK,EAAE,SAAS;IAChB,MAAM,EAAE,QAAQ;IAChB,OAAO,EAAE,QAAQ;IACjB,MAAM,EAAE,QAAQ;IAChB,YAAY,EAAE,QAAQ;IACtB,aAAa,EAAE,QAAQ;IACvB,OAAO,EAAE,QAAQ;IACjB,SAAS,EAAE,QAAQ;IACnB,QAAQ,EAAE,SAAS;IACnB,SAAS,EAAE,SAAS;IACpB,IAAI,EAAE,SAAS;IACf,KAAK,EAAE,SAAS;IAChB,QAAQ,EAAE,MAAM;IAChB,UAAU,EAAE,MAAM;IAClB,OAAO,EAAE,aAAa;IACtB,QAAQ,EAAE,aAAa;CACxB,CAAC;AAEF,MAAM,QAAQ,GAA2B;IACvC,OAAO,EAAE,UAAU;IACnB,MAAM,EAAE,QAAQ;IAChB,MAAM,EAAE,SAAS;IACjB,OAAO,EAAE,UAAU;IACnB,IAAI,EAAE,OAAO;IACb,WAAW,EAAE,cAAc;IAC3B,IAAI,EAAE,OAAO;CACd,CAAC;AAEF,SAAS,iBAAiB,CAAC,UAAkB;IAC3C,OAAO,QAAQ,CAAC,UAAU,CAAC,WAAW,EAAE,CAAC,CAAC;AAC5C,CAAC;AAED,SAAS,WAAW,CAAC,UAAkB;IACrC,MAAM,UAAU,GAAG,iBAAiB,CAAC,UAAU,CAAC,CAAC;IACjD,IAAI,CAAC,UAAU;QAAE,OAAO,SAAS,CAAC;IAClC,OAAO,QAAQ,CAAC,UAAU,CAAC,CAAC;AAC9B,CAAC"}

package/dist/services/sync/model-factory.d.ts

@@ -0,0 +1,10 @@
+/**
+ * Maps (provider, modelId, apiKey?) to an AI SDK LanguageModel.
+ *
+ * Each provider falls back to its own env var when no apiKey is given:
+ *   anthropic → ANTHROPIC_API_KEY
+ *   openai    → OPENAI_API_KEY
+ *   google    → GOOGLE_GENERATIVE_AI_API_KEY
+ */
+import type { LanguageModel } from 'ai';
+export declare function createModel(provider: string, modelId: string, apiKey?: string): LanguageModel;

package/dist/services/sync/model-factory.js

@@ -0,0 +1,24 @@
+/**
+ * Maps (provider, modelId, apiKey?) to an AI SDK LanguageModel.
+ *
+ * Each provider falls back to its own env var when no apiKey is given:
+ *   anthropic → ANTHROPIC_API_KEY
+ *   openai    → OPENAI_API_KEY
+ *   google    → GOOGLE_GENERATIVE_AI_API_KEY
+ */
+import { createAnthropic } from '@ai-sdk/anthropic';
+import { createOpenAI } from '@ai-sdk/openai';
+import { createGoogleGenerativeAI } from '@ai-sdk/google';
+export function createModel(provider, modelId, apiKey) {
+    switch (provider) {
+        case 'anthropic':
+            return createAnthropic(apiKey ? { apiKey } : {})(modelId);
+        case 'openai':
+            return createOpenAI(apiKey ? { apiKey } : {})(modelId);
+        case 'google':
+            return createGoogleGenerativeAI(apiKey ? { apiKey } : {})(modelId);
+        default:
+            throw new Error(`Unsupported provider: ${provider}. Supported: anthropic, openai, google`);
+    }
+}
+//# sourceMappingURL=model-factory.js.map

package/dist/services/sync/model-factory.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"model-factory.js","sourceRoot":"","sources":["../../../src/services/sync/model-factory.ts"],"names":[],"mappings":"AAAA;;;;;;;GAOG;AAEH,OAAO,EAAE,eAAe,EAAE,MAAM,mBAAmB,CAAC;AACpD,OAAO,EAAE,YAAY,EAAE,MAAM,gBAAgB,CAAC;AAC9C,OAAO,EAAE,wBAAwB,EAAE,MAAM,gBAAgB,CAAC;AAG1D,MAAM,UAAU,WAAW,CAAC,QAAgB,EAAE,OAAe,EAAE,MAAe;IAC5E,QAAQ,QAAQ,EAAE,CAAC;QACjB,KAAK,WAAW;YACd,OAAO,eAAe,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,MAAM,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,OAAO,CAAC,CAAC;QAC5D,KAAK,QAAQ;YACX,OAAO,YAAY,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,MAAM,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,OAAO,CAAC,CAAC;QACzD,KAAK,QAAQ;YACX,OAAO,wBAAwB,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,MAAM,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,OAAO,CAAC,CAAC;QACrE;YACE,MAAM,IAAI,KAAK,CAAC,yBAAyB,QAAQ,wCAAwC,CAAC,CAAC;IAC/F,CAAC;AACH,CAAC"}

package/dist/services/sync/name-quality.d.ts

@@ -0,0 +1,31 @@
+/**
+ * Name quality scoring for sync pipeline (STU-92)
+ *
+ * Detects username-like strings (dw8768a, dkim06) and other low-quality
+ * names that shouldn't enter the knowledge graph at full confidence.
+ *
+ * Returns a confidence value (0–1) that can be applied as a penalty
+ * to extracted records. Records aren't dropped — just flagged with
+ * lower confidence so they surface in staging review.
+ */
+/**
+ * Score the quality of an entity name.
+ *
+ * @returns confidence multiplier (0–1). 1.0 = normal, lower = suspicious.
+ */
+export declare function nameQuality(name: string, entityType: string): number;
+/**
+ * Detect username/ID-like strings.
+ *
+ * Patterns caught:
+ * - Letters + digits: dkim06, dk784, dw8768a, user123
+ * - Digits + letters: 3dsmith
+ * - Single lowercase word with no vowels: dvrbrk (rare but catches some IDs)
+ *
+ * Patterns allowed:
+ * - Multi-word names: "John Doe", "Mary Jane Watson"
+ * - Names with hyphens: "Jean-Pierre", "O'Brien"
+ * - Names with apostrophes: "O'Connor"
+ * - Single-word names without digits: "Beyonce", "Madonna"
+ */
+export declare function isLikelyUsername(name: string): boolean;

package/dist/services/sync/name-quality.js

@@ -0,0 +1,60 @@
+/**
+ * Name quality scoring for sync pipeline (STU-92)
+ *
+ * Detects username-like strings (dw8768a, dkim06) and other low-quality
+ * names that shouldn't enter the knowledge graph at full confidence.
+ *
+ * Returns a confidence value (0–1) that can be applied as a penalty
+ * to extracted records. Records aren't dropped — just flagged with
+ * lower confidence so they surface in staging review.
+ */
+/**
+ * Score the quality of an entity name.
+ *
+ * @returns confidence multiplier (0–1). 1.0 = normal, lower = suspicious.
+ */
+export function nameQuality(name, entityType) {
+    if (!name || typeof name !== 'string')
+        return 0;
+    const trimmed = name.trim();
+    if (trimmed.length === 0)
+        return 0;
+    // All entity types: all-digit strings are not names
+    if (/^\d+$/.test(trimmed))
+        return 0.1;
+    // All entity types: single char is not a name
+    if (trimmed.length === 1)
+        return 0.2;
+    // Person-specific checks
+    if (entityType === 'person') {
+        if (trimmed.length <= 2)
+            return 0.2;
+        if (isLikelyUsername(trimmed))
+            return 0.2;
+    }
+    return 1.0;
+}
+/**
+ * Detect username/ID-like strings.
+ *
+ * Patterns caught:
+ * - Letters + digits: dkim06, dk784, dw8768a, user123
+ * - Digits + letters: 3dsmith
+ * - Single lowercase word with no vowels: dvrbrk (rare but catches some IDs)
+ *
+ * Patterns allowed:
+ * - Multi-word names: "John Doe", "Mary Jane Watson"
+ * - Names with hyphens: "Jean-Pierre", "O'Brien"
+ * - Names with apostrophes: "O'Connor"
+ * - Single-word names without digits: "Beyonce", "Madonna"
+ */
+export function isLikelyUsername(name) {
+    // Multi-word names are almost never usernames
+    if (/\s/.test(name.trim()))
+        return false;
+    // Single word containing digits mixed with letters → username
+    if (/[a-z]/i.test(name) && /\d/.test(name))
+        return true;
+    return false;
+}
+//# sourceMappingURL=name-quality.js.map

package/dist/services/sync/name-quality.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"name-quality.js","sourceRoot":"","sources":["../../../src/services/sync/name-quality.ts"],"names":[],"mappings":"AAAA;;;;;;;;;GASG;AAEH;;;;GAIG;AACH,MAAM,UAAU,WAAW,CAAC,IAAY,EAAE,UAAkB;IAC1D,IAAI,CAAC,IAAI,IAAI,OAAO,IAAI,KAAK,QAAQ;QAAE,OAAO,CAAC,CAAC;IAEhD,MAAM,OAAO,GAAG,IAAI,CAAC,IAAI,EAAE,CAAC;IAC5B,IAAI,OAAO,CAAC,MAAM,KAAK,CAAC;QAAE,OAAO,CAAC,CAAC;IAEnC,oDAAoD;IACpD,IAAI,OAAO,CAAC,IAAI,CAAC,OAAO,CAAC;QAAE,OAAO,GAAG,CAAC;IAEtC,8CAA8C;IAC9C,IAAI,OAAO,CAAC,MAAM,KAAK,CAAC;QAAE,OAAO,GAAG,CAAC;IAErC,yBAAyB;IACzB,IAAI,UAAU,KAAK,QAAQ,EAAE,CAAC;QAC5B,IAAI,OAAO,CAAC,MAAM,IAAI,CAAC;YAAE,OAAO,GAAG,CAAC;QACpC,IAAI,gBAAgB,CAAC,OAAO,CAAC;YAAE,OAAO,GAAG,CAAC;IAC5C,CAAC;IAED,OAAO,GAAG,CAAC;AACb,CAAC;AAED;;;;;;;;;;;;;GAaG;AACH,MAAM,UAAU,gBAAgB,CAAC,IAAY;IAC3C,8CAA8C;IAC9C,IAAI,IAAI,CAAC,IAAI,CAAC,IAAI,CAAC,IAAI,EAAE,CAAC;QAAE,OAAO,KAAK,CAAC;IAEzC,8DAA8D;IAC9D,IAAI,QAAQ,CAAC,IAAI,CAAC,IAAI,CAAC,IAAI,IAAI,CAAC,IAAI,CAAC,IAAI,CAAC;QAAE,OAAO,IAAI,CAAC;IAExD,OAAO,KAAK,CAAC;AACf,CAAC"}