// generate-tools-mdx.js
const fs = require("fs");
const path = require("path");
const axios = require("axios");

const API_URL = "http://0.0.0.0:8000/integrations/tools/";
const TOOLS_DIR = "tools";
const MAIN_FILE = path.join(TOOLS_DIR, "tools.mdx");
const MINT_JSON = "mint.json";

function sanitizeFileName(name) {
  return name.replace(/[^a-z0-9]/gi, "_");
}

function sanitizeDescription(description) {
  if (!description) return "";

  // Replace HTML line breaks with MDX line breaks
  description = description.replace(/<br\s*\/?>/gi, "\n\n");

  // Remove any unexpected HTML tags
  description = description.replace(/<\/?(description|invoke)[^>]*>/gi, "");

  // Handle code blocks and inline code more carefully
  // First, temporarily replace code blocks to protect them
  const codeBlocks = [];
  description = description.replace(/```([\s\S]*?)```/g, (match, code) => {
    codeBlocks.push(match);
    return `__CODE_BLOCK_${codeBlocks.length - 1}__`;
  });

  // Then handle inline code
  description = description.replace(/`([^`]+)`/g, (match, code) => {
    // Don't escape backticks in inline code
    return match;
  });

  // Handle any curly braces that might be interpreted as expressions
  description = description.replace(/{/g, "\\{").replace(/}/g, "\\}");

  // Escape any remaining HTML-like content
  description = description.replace(/</g, "&lt;").replace(/>/g, "&gt;");

  // Restore code blocks
  description = description.replace(/__CODE_BLOCK_(\d+)__/g, (_, index) => {
    return codeBlocks[parseInt(index)];
  });

  return description;
}

async function main() {
  const { data } = await axios.get(API_URL);

  // Group tools by provider
  const grouped = {};
  data.forEach((item) => {
    if (!grouped[item.provider]) grouped[item.provider] = [];
    grouped[item.provider].push(item);
  });

  // Ensure tools directory exists
  if (!fs.existsSync(TOOLS_DIR)) {
    fs.mkdirSync(TOOLS_DIR, { recursive: true });
  }

  // Generate a file for each provider
  const providerLinks = [];
  const providerPages = [];
  const providerNameToPage = {};
  for (const provider of Object.keys(grouped)) {
    const fileName = `${sanitizeFileName(provider).toLowerCase()}.mdx`;
    const filePath = path.join(TOOLS_DIR, fileName);

    // Use provider_description from the first tool entry and sanitize it
    const providerDescription = sanitizeDescription(
      grouped[provider][0].provider_description || ""
    );
    let mdx = `${providerDescription}\n\n`;
    grouped[provider].forEach((tool) => {
      const sanitizedToolDescription = sanitizeDescription(
        tool.tool_description
      );
      mdx += `<Accordion title=\"${tool.tool}\">\n${sanitizedToolDescription}\n</Accordion>\n\n`;
    });

    fs.writeFileSync(filePath, mdx);
    providerLinks.push(`- [${provider} Tools](./${fileName})`);
    const pagePath = `tools/${fileName.replace(/\.mdx$/, "")}`;
    providerPages.push(pagePath);
    providerNameToPage[provider] = pagePath;
  }

  // Sort providerPages alphabetically by provider name
  const sortedProviderPages = Object.keys(providerNameToPage)
    .sort((a, b) => a.localeCompare(b))
    .map((provider) => providerNameToPage[provider]);

  // Update mint.json navigation
  const mint = JSON.parse(fs.readFileSync(MINT_JSON, "utf8"));

  // Find the Tools group in navigation
  let toolsGroup = null;
  for (const navGroup of mint.navigation) {
    if (navGroup.group === "Documentation") {
      for (const subGroup of navGroup.pages) {
        if (typeof subGroup === "object" && subGroup.group === "Tools") {
          toolsGroup = subGroup;
          break;
        }
      }
    }
  }

  if (toolsGroup) {
    // Keep introduction and any static pages, then add all provider pages (no duplicates)
    const staticPages = toolsGroup.pages.filter(
      (p) => !providerPages.includes(p)
    );
    toolsGroup.pages = [...staticPages, ...sortedProviderPages];
  } else {
    console.error("Could not find Tools group in mint.json navigation.");
    process.exit(1);
  }

  fs.writeFileSync(MINT_JSON, JSON.stringify(mint, null, 2));
  console.log(`Generated provider files and updated mint.json.`);
}

main().catch((err) => {
  console.error(err);
  process.exit(1);
});


2025-04-28

2025-03-25

2025-02-27

2024-12-20

2024-12-12

2024-12-06

2024-11-26

Product Updates

Push.ai Documentation

Introduction

The Push.ai Query Agent navigates the Business Graph™ to help any user answer any business question.

Query Agent

Use Reports to share metrics in shared channels where teams can discuss and take action.

Reports

Receive a personalized feed of updates around the the parts of the business you care about.

Subscriptions

Data Processing Agreement

Data Retention Policy 

Privacy Policy

Terms of Service

Book a Demo

Open App

Website

Get Started

Support

Learn how to setup Push.ai for your organization.

Admins & Data Teams

Learn how to get started with Push.ai as a new user.

Business Users

Understand how caching and datasource queries work in Push.ai.

Caching

Metrics are the basis for measuring your business. They are times-series aggregations of data that can be defined inside of Push.ai or integrated using semantic layers such as [the dbt Semantic Layer](https://getdbt.com), [Looker](https://cloud.google.com/looker/) and [Cube](https://cube.dev).

Metrics

Dimensions are categories or attributes used to organize and analyze data within a business. Dimensions help break down data into specific segments, such as time, location, or product type, allowing a business to understand and compare different aspects of their operations.

Dimensions

If Dimensions are like groups, Dimension Values are the individual parts that make up each group. For example, Marketing Channel is a Dimension, while Paid Search, Direct Traffic and Email are Dimension Values.

Dimension Values

Artifacts are one of the core primitives of knowledge in Push.ai. They represent various types of content and provide valuable context that enhances strcutured data analysis.

Artifacts

Map data with business context using a graph-based model of your business.

How It Works?

Populate your Business Graph™ to improve the relevancy of Push.ai's AI systems.

Use Cases

Learn how Push.ai processes and analyzes unstructured data to enhance your Business Graph™.

How It Works

Unstructured data changes the future of business intelligence.

Connect with sources of KNowledge and unstrcutured data through your data warehouse, file uploads and native API integrations.

Knowledge Integrations

Receive Subscriptions and Reports where you already work in Slack.

Slack

Receive Subscriptions and Reports where you already collaborate in email.

Email

Receive Subscriptions and Reports where you already work in Microsoft Teams.

Microsoft Teams

Manage users and roles within your Push.ai organization

Users and Roles

Configure SSO authentication for your Push.ai organization

Get Started

Documentation

Changelog

Legal

Product Updates

New Feature: The Push.ai Query Agent

Updates

New Feature: Knowledge

Updates

Updates

Bug Fixes

New Feature: Outliers (experimental)

Performance Improvements

UI Upgrades

Bugfixes

Get Started

Documentation

Changelog

Legal

​New Feature: The Push.ai Query Agent

​Updates

​New Feature: Knowledge

​Updates

​Updates

​Bug Fixes

​New Feature: Outliers (experimental)

​Performance Improvements

​UI Upgrades

​Bugfixes

New Feature: The Push.ai Query Agent

Updates

New Feature: Knowledge

Updates

Updates

Bug Fixes

New Feature: Outliers (experimental)

Performance Improvements

UI Upgrades

Bugfixes