database for offline wikis with multi-master replication
npm install wikidbdatabase for offline-first distributed wikis
First, we'll have a script to create documents:
`` js
var wikidb = require('wikidb');
var db = require('level')('/tmp/wiki.db');
var minimist = require('minimist');
var argv = minimist(process.argv.slice(2), {
alias: { t: [ 'tag', 'tags' ] }
});
var wdb = wikidb(db, { dir: '/tmp/wiki.blob' });
var opts = {
key: argv._[0],
prev: argv.prev,
tag: argv.tag
};
var w = wdb.createWriteStream(opts, function (err, key) {
console.log(key);
});
process.stdin.pipe(w);
`
Now we create the first document:
`
$ echo hello there | node create.js 'welcome page' --tag=welcome
2e0c45ea142f59b52d616bcf17837ff4aa07d985ec2b55d13b42c4104a67f0a2
and next we can make an updated version of the 'welcome page' by referring to
its hash in a new document:
`
$ echo 'greetings, stranger!' | node create.js 'welcome page' \
--prev=2e0c45ea142f59b52d616bcf17837ff4aa07d985ec2b55d13b42c4104a67f0a2 \
--tag=welcome --tag=greeting
e208b25dde1e33f70ab0cb047ed5ce2e87fca1e782c5598bc2ebc7465ce478d2
This script uses .recent() to pull down the recent changes:
` js
var wikidb = require('wikidb');
var db = require('level')('/tmp/wiki.db');
var wdb = wikidb(db, { dir: '/tmp/wiki.blob' });
var through = require('through2');
wdb.recent().pipe(through.obj(function (row, enc, next) {
console.log('key: ' + row.meta.key);
console.log('hash: ' + row.hash);
console.log('date: ' + new Date(row.meta.time));
console.log('-----------------------------');
var r = wdb.createReadStream(row.hash);
r.pipe(process.stdout);
r.on('end', function () { console.log(); next() });
}));
`
Running the script shows the recent changes:
`
$ node recent.js
key: welcome page
hash: e208b25dde1e33f70ab0cb047ed5ce2e87fca1e782c5598bc2ebc7465ce478d2
date: Fri Oct 24 2014 10:04:19 GMT+0100 (BST)
-----------------------------
greetings, stranger!
key: welcome page
hash: 2e0c45ea142f59b52d616bcf17837ff4aa07d985ec2b55d13b42c4104a67f0a2
date: Fri Oct 24 2014 10:04:07 GMT+0100 (BST)
-----------------------------
hello there
`
This package ships with a wikidb command:
`
wikidb create KEY {OPTIONS}
Add a new document at KEY with the content from stdin.
--prev=PREV Point at a previous hash. You can point back at multiple
hashes with more than one --prev..--tag
--tag=TAG Add a tag. You can have multiple tags with more than one
.
wikidb read HASH
Print the document contents for HASH.
wikidb get HASH
Print the metadata for HASH.
wikidb keys
List all keys.
wikidb list
List all documents.
wikidb heads KEY
Print all the heads for KEY.
wikidb recent KEY
Print all updates to KEY, most recent updates first.
wikidb history HASH
Print the entire history starting at HASH.
wikidb future HASH
Print every hash that descends from HASH.
wikidb search TERMS...
Search for documents that include TERMS.
wikidb by-tag
Search for documents by tag name.
wikidb sync
Bidirectional (push and pull) replication strategy.
stdin and stdout are used to wire up the replication protocol.
wikidb pull
Get documents from an endpoint.
stdin and stdout are used to wire up the replication protocol.
wikidb push
Push documents to an endpoint.
stdin and stdout are used to wire up the replication protocol.
`
` js`
var wikidb = require('wikidb')
Create a new wikidb instance wdb from a leveldb handle db.
Optionally, provide:
opts.dir - location to store file blobs
Create a writable stream w that will populate content at meta.key and linkmeta.prev
back to previous documents using an array of hashes in .
Optionally provide an opts.prebatch(rows, key, fn) function that will berows
called before are inserted with db.batch(). fn(err, rows) should
provide the new rows to insert.
Optionally tag a document by setting meta.tags to an array of string tags.
cb(err, hash) fires with an error or the hash of the content written into w.
Return a readable object stream r that searches for objects by opts.tag.key
The objects in the stream have and hash properties.
If opts is a string, opts will be used as the opts.tag.
If provided, cb(err, results) fires with an error or the array of results
objects.
Return a readable object stream r that produces all tags.
Use opts to bound the query (opts.lt, opts.gt, and opts.limit)cb(err, tags)
and to collect results into an array tags.
Each object in r has a single tag property with a string value.
Search for terms, an array of strings which must all be present in a full text
scan of each head document.
The result stream has the same properties as wdb.head(): 'key' and 'hash'
properties on each object.
List documents with a readable object stream r ordered by date, most recent tor
least recent. Each object output by has a hash and key property.
You can pass through opts.reverse and opts.limit to the underlying database
query.
The default ordering is most recent to least recent.
The methods in this section are inherited from forkdb and accept the same
arguments.
Replicate with the duplex stream d.
Set opts.mode to 'sync' (default), 'push', or 'pull' to use different
replication strategies.
Return a readable stream r with the content for hash.
Return the heads for key as a readable object stream that outputs objects withhash properties.
Return a readable object stream with a key property for each available key.
Return a readable object stream r for all the tails of key.
Return all the documents in the database as an object stream that outputs
objects with key and hash properties.
Return a readable object stream r that outputs all the documents that link tohash as objects with key and hash properties.
Fetch the metadata for hash as cb(err, meta).
Fetch all the history for hash as a readable object stream r until a branch
in the history when a 'branch' event fires with a new readable object stream
for each fork.
Fetch all the future for hash as a readable object stream r until a branch
in the future when a 'branch' event fires with a new readable object stream
for each fork.
With npm, do:
``
npm install wikidb
or to get the wikidb command, do:
`
npm install -g wikidb
MIT