node-dirsum is a super small library that gives you the node.js equivalent of:

find <dir> -type f | xargs md5

With the added feature that this library will summarize the hashes into one master hash so you can get a single checksum for a directory treat. Note that there is no 'standard' way to do this in UNIX, but you basically have two choices:

• Pipe all the bytes of all the files into one digest
• Compute a digest of all the digests

Since (1) is hard to do if your files are large, I went with (2). Again, this may or may not be useful to you, since there's not a standard client to validate this. At a minimum, you can at least get a tree of all the files and their checksums.

Usage

var dirsum = require('../lib/dirsum');

dirsum.digest('/your/tree', 'sha1', function(err, hashes) {
  if (err) throw err;
  console.log(JSON.stringify(hashes, null, 2));
});

There you go. That's the whole API. There's the equivalent of that in the tst directory, and I copied in the openldap default files, since there's a bunch of schema files in there. Running that with 'md5' as the method gives this output:

{
  "files": {
    "ldap.conf": "a1b26e90628b18bc9cea6c9573a63b7b",
    "ldap.conf.default": "a1b26e90628b18bc9cea6c9573a63b7b",
"schema": {
  "files": {
    "apple.schema": "d74ab4c3522e0154fb409c9a6515ad8f",
    "apple_auxillary.schema": "20b1560538fe8cc73bc97b0ce8f8d7a4",
    "collective.schema": "139aefbd0541ad2f4198ea53d2689334",
    "corba.schema": "72684228a9d9691e458a9850f93231fb",
    "core.ldif": "0c9d2c5df2376192e02f11bc6edc96f7",
    "core.schema": "fdd7cfc44ca73698f877d64becdc8278",
    "cosine.ldif": "82c062506b1dc1786c7b1bdd56c53f1a",
    "duaconf.schema": "2503265c236a158ae5be7eb6d81b130a",
    "dyngroup.schema": "55b65ea42cec37ad0ba110752e1c20c2",
    "fmserver.schema": "928cb73c8f0dd0b5b18d2619a3a0dbc5",
    "inetorgperson.ldif": "71db554a56953e03669784270f69942c",
    "inetorgperson.schema": "7452ce52fcecad8aa97446fed52b1db2",
    "java.schema": "8f8f12b72432388e065e37e19eca0eec",
    "krb5-kdc.schema": "04df5507f0a3b1602efd523a42bd1f90",
    "microsoft.std.schema": "928686833526a63590d8d56fb15f3602",
    "misc.schema": "c2453f83012e97e84a4a2841c62cee88",
    "nadf.schema": "687e1d49faa7c0f4990eba6899e3ba7f",
    "netinfo.schema": "4ca7cc2b5cf43618cfbd540171f8ffd6",
    "nis.ldif": "4322c3f16f38f802bea8fdc0a3524274",
    "nis.schema": "fc3255f06a87dd7c360e88772b2d17de",
    "openldap.ldif": "f342787e81ac1101cf860dea78e7bb51",
    "openldap.schema": "bbf95ea610c9f28c4a7bdc4c867e8961",
    "ppolicy.schema": "8fbc0dda95c831a9ed84b2884c56a02b",
    "README": "dfe2414dc543da3fdf98b48ab3de1a94",
    "samba.schema": "a26fbdabbd97f076fcaa3b3774b7eba8",
    "cosine.schema": "0a85a2dbf9729ceb4bb6354c4c865135",
    "microsoft.schema": "2ced4df70c8e72f5fa1f857ed3a5c39d",
    "microsoft.ext.schema": "65dbf5a721b14de93fa011386025ab72"
  },
  "hash": "c3dce06943963c28163db6e5164844a2"
    }
  },
  "hash": "440a3933672304d72660526e4fd1463e"
}

Basically, every level in the tree is going to have two attributes in the returned Hashes object, a hash which is the hash of hashes, and files, which is a key/value pairing of filename to hash. Each file might be another object, which indicates there was a directory tree encountered.

Skipping Files/Directories Usage

There is an additional parameter to the method, an ignoreCallback function that can be provided to allow the algorithm to skip over certain files or directories. This callback has the signature: function(path, filename). An example is as follows:

var dirsum = require('../lib/dirsum');

dirsum.digest('/your/tree', 'sha1', function(err, hashes) {
  if (err) throw err;
  console.log(JSON.stringify(hashes, null, 2));
}, function(path, filename) {
  // ignore any files starting with '.'
  if (filename.indexOf('.') === 0) {
    return true;
  }
  return false;
});

NOTES

If the method parameter (i.e. the second parameter) is not passed as a string, the callback and ignoreCallback are effectively shifted left respectively. In other words the signature for the digest method looks either like this:

dirsum.digest(path : string, method : string, callback: function(err, hashes), ignoreCallback : function(path, filename));

OR

dirsum.digest(path : string, callback : function(err, hashes), ignoreCallback : function(path, filename));

Finally, the file read implementation uses streams so that large files are NOT read-entirely into memory.

Installation

npm install dirsum

License

MIT.

Bugs

See https://github.com/mcavage/node-dirsum/issues.