LLM Prompt Creation Scripts

mongo2prompt.js

// Connect to the database
db = db.getSiblingDB("cdtserver");

var ignore = ["collection-metadata", "Roulette", "fs.", "test"];

var collectionNames = db.getCollectionNames();
var descriptions = {};
if (collectionNames.includes("collection-metadata")) {
  db.getCollection("collection-metadata")
    .find()
    .forEach(function (doc) {
      descriptions[doc["collectionName"]] = doc["description"];
    });
}

function stringify(obj) {
  function isPrimitiveArray(value) {
    if (!Array.isArray(value)) return false;
    for (var i = 0; i < value.length; i++) {
      if (typeof value[i] === "object") return false;
    }
    return true;
  }

  function toHexString(bytes) {
    return Array.from(Object.values(bytes), (byte) =>
      ("0" + (Number(byte) & 0xff).toString(16)).slice(-2)
    ).join("");
  }

  function isBinary(obj) {
    return obj && typeof obj === "object" && typeof obj.buffer == "object";
  }

  function replacer(key, value) {
    if (isBinary(value)) {
      return `Binary('${toHexString(value.buffer)}',${value.sub_type})`;
    } else if (isPrimitiveArray(value)) {
      return "[" + value.join(", ") + "]";
    }
    return value;
  }

  return JSON.stringify(obj, replacer, 2);
}

function limitArrayEntries(obj, maxEntries = 3) {
  function recurse(value) {
    if (Array.isArray(value)) {
      if (value.length > maxEntries) {
        value.length = maxEntries;
      }
      return value.map(recurse);
    } else if (value && typeof value === "object") {
      return Object.keys(value).reduce((acc, key) => {
        acc[key] = recurse(value[key]);
        return acc;
      }, {});
    }
    return value;
  }
  return recurse(obj);
}

function formatDocument(doc, indent = 2, maxDepth = 10) {
  var jsonString = stringify(doc, null, indent);
  var lines = jsonString.split("\n");
  var finalLines = [];
  var skipDepth = null;
  var maxIndent = maxDepth * indent;
  var indentRx = new RegExp("^ {" + maxIndent + ",}");

  for (let i = 0; i < lines.length; i++) {
    let line = lines[i];
    let nextLine = lines[i + 1] || "";
    let currentDepth = line.match(/^ */)[0].length;
    if (indentRx.test(line) && skipDepth === null) {
      skipDepth = currentDepth;
    }
    if (
      skipDepth !== null &&
      (!indentRx.test(nextLine) || currentDepth < skipDepth)
    ) {
      if (indentRx.test(line)) {
        let previousLine = finalLines.pop();
        previousLine = previousLine.replace(/: {/, ": {...},");
        finalLines.push(previousLine);
      }
      skipDepth = null;
    } else if (skipDepth === null) {
      finalLines.push(line);
    }
  }
  return finalLines.join("\n");
}

// Get all collection names in the database
collectionNames.forEach(function (collectionName) {
  //  if (collectionName !== "CDT.Models.RPR.Pool") return; // TEST

  if (ignore.find((kw) => collectionName.includes(kw)))
    // Skip any we know we don't need
    return;

  // Print collection name
  print("----------");
  print('Collection: "' + collectionName + '"\n');
  print(`Description: ${descriptions[collectionName] || "n/a"}\n`);

  // Fetch the example document using $natural sort to get the most recently inserted document
  var exampleDocument = db[collectionName]
    .find()
    .sort({ $natural: -1 })
    .limit(1)
    .toArray();
  if (exampleDocument && exampleDocument.length > 0) {
    print("Example Document:");
    print(formatDocument(limitArrayEntries(exampleDocument[0], 10), 2, 12));
  } else {
    print("No documents found in this collection.");
  }

  // Fetch and format indexes for the collection
  var indexes = db[collectionName].getIndexes();
  indexes.forEach(function (index) {
    if (index.name === "_id_") return;
    var keyFields = Object.keys(index.key).join(", ");
    var unique = index.unique ? "Unique" : "";
    print(`${unique}Index: "${index.name}" on ${keyFields}`);
  });
});