Nano Records

A module for interacting with CouchDB through nano with Nodejs.

Conflicts are avoided with automated retries, errors are sanitised, designs are persisted on an as-needed basis, and so on. Intended to make life simpler by abstracting away some general database busywork and hopefully makes it easy to get up and running.

npm

Notice

If you are using CouchDB2.0/Cloudant this library nano-records has been superceded by Couch Recliner.

Upgrading from 1.0.9 to 1.1.0

• db.doc.forcedWrite is now db.doc.write
• db.doc.write has been removed
• db.doc.forcedUpdate is now db.doc.updateOrWrite
• db.designs.show is now db.show.catalog
• db.designs.view is now db.view.catalog

Install

npm install nano --save
npm install nano-records --save

var nano = require('nano')("localhost");
var NanoRecords = require('nano-records');
var dbName = "my-database";
var db = new NanoRecords(nano, dbName);

// typescript
import NanoRecords = require('nano-records');

Each NanoRecords instance represents one database.

Provide NanoRecords with a running instance of nano and a chosen database name. This should be all you need to get started.

Documents

db.doc.create(body, (err, doc) => {
  if (err)
    return;
  // doc is a NanoRecords document
  console.log(doc.body);
});

Easiest way to persist a new record, returns a NanoRecords document, the _id attribute is generated by CouchDB.

doc.read((err) => {
  if (!err)
    console.log('success!');
});
db.doc.read(id, (err, doc) => {
  if (err)
    return;
  // doc is a NanoRecords document
  console.log(doc.body);
});

Retrieve the latest version of the body of a document.

doc.write(body, (err) => {
  if (!err)
    console.log('success!');
});
db.doc.write(id, body, (err, doc) => {
  if (err)
    return;
  // doc is a NanoRecords document
  console.log(doc.body);
});

Overwrite an existing document or create a new document if one doesn't exist.

doc.update(body, (err) => {
  if (!err)
    console.log('success!');
});
db.doc.update(id, body, (err, doc) => {
  if (err)
    return;
  // doc is a NanoRecords document
  console.log(doc.body);
});

Deeply extend an existing document.

db.doc.updateOrWrite(id, body, (err, doc) => {
  if (err)
    return;
  // doc is a NanoRecords document
  console.log(doc.body);
});

Deeply extend an existing document or create a document if one doesn't exist.

doc.destroy((err) => {
  if (!err)
    console.log('success!');
});
db.doc.destroy(id, (err) => {
  if (!err)
    console.log('success!');
});

Remove a document.

doc.head((err, rev, result) => {
  if (err)
    return;
  // rev is the most up to date document revision
  console.log(result);
});
db.doc.head(id, (err, rev, result) => {
  if (err)
    return;
  // rev is the most up to date document revision
  console.log(result);
});

Retrieve header information.

doc.getId() == doc.body['_id']; // true
doc.getRev() == doc.body['_rev']; // true

doc.getBody(); // {}

Access the document body's _id and _rev properties. Another way of writing doc.body['_id'] and doc.body['_rev'].

Create a deep clone of the full document body deepExtend({}, doc.body)1.

Attachments

doc.attachment.read(name, (err, data) => {
  if (err)
    return;
  console.log(data.length);
});
db.doc.attachment.read(id, name, (err, data) => {
  if (err)
    return;
  console.log(data.length);
});

Read an attachment.

doc.attachment.write(name, data, mimeType, (err) => {
  if (!err)
    console.log('success!');
});
db.doc.attachment.write(id, name, data, mimeType, (err) => {
  if (!err)
    console.log('success!');
});

Overwrite an existing attachment with the same name or create a new attachment if one doesn't exist.

doc.attachment.destroy(name, (err) => {
  if (!err)
    console.log('success!');
});
db.doc.attachment.destroy(id, name, (err) => {
  if (!err)
    console.log('success!');
});

Remove an attachment.

doc.attachment.list();
doc.attachment.exists(name);

Return attachment names as an array and whether a specific attachment exists.

Attachment streams

doc.attachment.createReadStream(name, (err) => {
  if (!err)
    console.log('success!');
}).pipe(fs.createWriteStream('./my-file.txt'));
db.doc.attachment.createReadStream(id, name, (err) => {
  if (!err)
    console.log('success!');
}).pipe(fs.createWriteStream('./my-file.txt'));

Read the attachment as a stream.

var reader = fs.createReadStream('./my-file.txt');
reader.pipe(doc.attachment.createWriteStream(name, mimeType, (err) => {
  if (!err)
    console.log('success!');
}));

Write an attachment using a stream. It's important to note that streams cannot be retried, if there is an error you will have to pipe a new stream manually.

Designs

// > ./designs.json
{
  "foo": {
    "views": {
      "comments": {
        "map": "function (doc) { ... };",
        "reduce": "function (keys, values, rereduce) { ... };"
      }
    },
    "shows": {
      "post": "function (doc, req) { ... };"
    }
  },
  "bar": {
    "language": "javascript",
    "views": {}
  }
}

var designs = require('./designs.json');
var db = new NanoRecords(nano, dbName, designs);

When creating your NanoRecords instance optionally provide it a catalog of designs to use. You can learn more about Views and Design Documents on the CouchDB website.

db.show.catalog(id, design, name, (err, result) => {
  if (err)
    return;
  console.log(result);
});

Persist a show using the provided design (ie. "foo") and name (ie. "post") if it's not already there, then return the result for the doc.

db.view.catalog(design, name, params, (err, list) => {
  if (err)
    return;
  // list is a NanoRecords list
  console.log(list.values());
});

Persist a view similar to db.show.catalog and return the result. An empty list object is returned if there is an error.

View helpers

db.view.all(keys, params, (err, list) => {
  if (err)
    return;
  // list is a NanoRecords list
  console.log(list.values());
});

Will generate a view for you using the provided keys, which returns a list of documents.

Useful for simple search functions, for example keys may be "user_id" then you can provide { key: "myuserid" } to params and find relevant results. The keys parameter may also be an array of values, or nested values. It's best not to provide parameters to this function which are dynamic in nature, as a new view is persisted to the database for each set of keys provided.

Complex views are still best constructed manually. Read more about view helpers on: CouchDB design documents using view helpers in nano-records.

db.view.only(keys, values, params, (err, list) => {
  if (err)
    return;
  // list is a NanoRecords list
  console.log(list.values());
});

Will generate a view similar to view.all which only returns a specific set of values from each document. For example ["created_at", "title", "author.name"] would return { created_at: "mydate", title: "mytitle", author: { name: "myauthorname" } } as each result.

This is more efficient than performing a full document lookup.

Lists

list.total; // 11
list.offset; // 0
list.rows; // raw result
list.ids(); // [ids...]
list.keys(); // [keys...]
list.values(); // [values...]
list.docs(); // [NanoRecords documents...]
list.doc(index); // NanoRecords document

The list.docs method may not give you complete document objects depending on the values that were returned by the view. However running doc.read will fetch the full document from the database, similarly all normal NanoRecords document functions should work as you expect.

A common way to get just the first document is list.doc(0).

Db

db.create((err) => {
  if (!err)
    console.log('success!');
});

You should never need to but it is there if you want to use it.

db.destroy("_DESTROY_", (err) => {
  if (!err)
    console.log('success!');
});

Destroy the database, first parameter must be _DESTROY_.

db.reset("_RESET_", (err) => {
  if (!err)
    console.log('success!');
});

Destroy and then recreate the database, first parameter must be _RESET_.

Errors

err.scope; // source of the error
err.name; // error code
err.message; // more information
err.raw; // the full error returned from nano

// common errors
// ==
// not_found: Not found.
// missing_id: Id parameter required.
// conflict: There was a conflict.
// malformed_script: Problem with one of your designs.
// no_db_file: Database missing.

When an error is returned it has the above format. Generally you should never see no_db_file or conflict so maybe these are not so common errors.

A conflict would only happen if the max number of retries was reached on a request, possibly you have too much activity on a single document.

You might see no_db_file if your CouchDB has security locked down.

Contribute

If you like what you see please feel encouraged to get involved report problems and submit pull requests! As of the time of this writing the project is new with one maintainer.

New repository under development: couch-recliner