thingdom / node-neo4j Goto Github PK

Is it possible to get the node id when a new node is created?

Hi I am considering to use node-neo4j for a project however during my test i faced a problem.
When I execute the code below

var node1 = db.getNodeById(1, print1);
console.log("node : " + node1);        

what i see is node1 is undefined. I have added logs into GraphDatabase getNodeById and I
figured that getNodeById returns before getting data from the server. I am not very familiar with
asychronuous functions in javascript but i believe this is the case here.

I'm running nodejs with neo4j in windows.

When i try to execute an app with the basic:

var neo4j = require('neo4j');

I get the error:

Object expected
Line: 2
Char : 1
Code: 800A138F

On Windows 7
[email protected]

Hi,
I tried exactly the steps from "Development" section from Readme. When i start the tests it fails:

neo4j-node/neo4j$ npm test

> [email protected] test /home/thomas/Projekte/Laboratory/neo4j-node/neo4j
> node test

UNCAUGHT EXCEPTION: The syntax for bound nodes has changed in v1.5 of Neo4j. Now, it is START a=node(<nodeId>), or START a=node:idxName(key='value').
"start n=(21) return n"
         ^
Error: The syntax for bound nodes has changed in v1.5 of Neo4j. Now, it is START a=node(<nodeId>), or START a=node:idxName(key='value').
"start n=(21) return n"
         ^
    at /home/thomas/Projekte/Laboratory/neo4j-node/neo4j/lib/util_.coffee:14:10
    at GraphDatabase_prototype_query__10 (/home/thomas/Projekte/Laboratory/neo4j-node/neo4j/lib/GraphDatabase_.coffee:180:11)
    at main (/home/thomas/Projekte/Laboratory/neo4j-node/neo4j/test/cypher_.coffee_.js:65:10)

UNCAUGHT EXCEPTION: 405
Error: 405
    at ___ (/home/thomas/Projekte/Laboratory/neo4j-node/neo4j/lib/Node_.coffee:272:11)
    at Request.___ [as callback] (/home/thomas/Projekte/Laboratory/neo4j-node/neo4j/lib/Node_.coffee:1:887)
    at Request.<anonymous> (/home/thomas/Projekte/Laboratory/neo4j-node/neo4j/node_modules/request/main.js:314:21)
    at Request.emit (events.js:67:17)
    at IncomingMessage.<anonymous> (/home/thomas/Projekte/Laboratory/neo4j-node/neo4j/node_modules/request/main.js:295:19)
    at IncomingMessage.emit (events.js:88:20)
    at HTTPParser.onMessageComplete (http.js:137:23)
    at Socket.ondata (http.js:1125:24)
    at TCP.onread (net.js:334:27)
    at main (/home/thomas/Projekte/Laboratory/neo4j-node/neo4j/test/crud_.coffee_.js:124:0)

It seems that it doesn't work with the latest Neo4J 1.5?

Hi,

Is there a way to add a node to multiple indexes in one go? At the moment i'm still doing this... which isnt really adding multiple indexs at one time

node.save(function (err) {
    if (err) return callback(err);
    node.index("users", "username", data.username, function (err) {
        if (err) return callback(err);
        node.index("node", "type", "user", function (err) {
            if (err) return callback(err);
            callback(null, user);
        }); 
    });       
});

Whats the proper way to do it ? thanks.

With the latest Neo4j server from brew I get this:

Error: [object Object]
    at exports.adjustError (node_modules/neo4j/lib/util.js:67:15)
    at __$GraphDatabase_prototype_query__14 (node_modules/neo4j/lib/GraphDatabase.js:276:26)
    at __tryCatch (node_modules/neo4j/node_modules/streamline/lib/callbacks/runtime.js:141:4)
    at ___ (node_modules/neo4j/lib/GraphDatabase.js:274:118)
    at ___ (node_modules/neo4j/lib/GraphDatabase.js:255:28)
    at Request.___ [as _callback] (node_modules/neo4j/node_modules/streamline/lib/callbacks/runtime.js:85:13)
    at Request.init.self.callback (node_modules/request/main.js:120:22)
    at Request.EventEmitter.emit (events.js:96:17)
    at Request.<anonymous> (node_modules/request/main.js:555:16)
    at Request.EventEmitter.emit (events.js:93:17)

I don't know what provoked it, sorry.

Executing the following line takes more than 3 seconds to execute on my server. This is too slow to be acceptable.

var neo4j = require('neo4j');

This huge delay is caused by streamline. It reparses the whole library everytime you require neo4j.

How to fix:

cd node_modules/neo4j/lib
../node_modules/coffee-script/bin/coffee -c *.coffee
../node_modules/streamline/bin/node-streamline -c *.js
# Manually replace all require('blabla_') by require('blabla')

It would be really great if you could provide the compiled files in the NPM install :)

Thanks

I want to be able to create indexes for relationships same as for nodes.

Just started looking into how it was done for nodes and I'm trying to figure out how if services.node_index just becomes services.relationship_index?

Will experiment tomorrow.

Well, here i'm again.

Maybe you gona hate me so much this week ^^

Thats a strange thing and i think it's not a node-neo4j issue, just my issue... but i don't get the point and i don't know where else to ask.

I'm trying to make a more dynamic whay to create,delete nodes and add relations to the system so i do this:
https://github.com/naxhh/playground/blob/master/neo4j/app.js

The problem is, when i create a new relationship i get 3 or 4 new properties

And i don't know why.

When i use the browser cms and i create a relationship it adds a relationship as normal

See thread:

https://groups.google.com/forum/?fromgroups=#!topic/neo4j/0q70qE1HR6Y

[Title edited by @aseemk. Original title: "Node.getRelationshipNodes(): Don't hard-code maxDepth and add additional settings".]

Right now this method is hard-coded with maxDepth=1. This method would be a lot more useful if maxDepth could be passed in as a parameter. A few more options a la Neography's NodeTraverser would also be very helpful

Nice library. But would be great to have API docs and better test cases.

Let's use a nice test framework like Mocha for testing. Much more user-friendly.

any chance of expanding the query functionality to include params?

db.query(query, params, callback)

The API supports this:
{"query": "start x = node(23) match x -[r]-> n return type(r), n.name?, n.age?","params": {}}

It appears from the CoffeeScript code that node-neo4j can throw Exceptions under certain conditions. Under other conditions, an error is returned in the callback. If so then it seems that I have to surround calls to node-neo4j with try/catch as well as handle any errors passed to me in my callback. Is this true?

hey, can you help me please ?. I'm running a websocket server named server.js
var ws = require(server + '../ws');

var server = ws.createServer();

but my Microsoft Javascript show me the following error
Error : "objet attendu"
Code : 800A138F

getIndexedNodes() method doesnt seem to work with automated indexing. Any idea on how this can be accomplished?

Hello,

When deleting a node, the index is not cleared automatically. We found during our testing that Neo4J will not delete the node from the index and if we attempt to create (during testing) a node with the same ID, it'll give an error. While in testing it's not important, in production the index will continue to grow even if we delete nodes. so we might have 1 million node in the database but 2 million in the index (assuming we deleted a million of course).

Would it be possible to add 2 functions;
1- One to remove a node/relationship from Index.
2- Another to delete an index completely.

Any and All help would be much appreciated.

Regards.

UNCAUGHT EXCEPTION: Unexpected token ILLEGAL
SyntaxError: Unexpected token ILLEGAL
at Object.parse (native)
at Node_prototype_save__1 (/Users/foo/Temp/neo4j-enterprise-1.4.1/node_modules/neo4j/lib/Node_.coffee:54:78)
at Node_prototype_save__1 (undefined_.js:43:19)

I'm getting error message above when I try the example hello world you have in README. I'm running Node 0.4.11. I don't know coffeescript so when I tried to debug the code I couldn't do much :/

The new neo4j streaming support provides some performance enhancements. However, clients of the node-neo4j API still have to wait for the entire result set to be collected, possibly converted to Node or Relationship objects, and then finally returned.

With neo4j streaming, a streaming JSON API such as Clarinet (https://github.com/dscape/clarinet), and an EventEmitter-based implementation, node-neo4j could return each result in a result set as it is received in something like a stream.on('result', obj) API.

This would obviously be a big win especially for large result sets.

Neo4j apparently supports streaming JSON back instead of buffering it all up in memory first. This will probably improve perf considerably across the board for most apps.

Need to investigate: How does this work? Are there any downsides? Should this be a transparent change? Or should it require an API call/parameter?

From Neo4j 1.6's changelog:

o Added request timeout, controlled with org.neo4j.server.webserver.limit.executiontime (disabled by default). Request header (max-execution-time) can shorten it if specified.

Might be nice to add support into our library to specify this on a request basis. We probably won't make use of it yet though, so chime in if you think you might. (And pull requests always welcome. =D)

Hi guys, I using your library on heroku and have spotted authentication issue. Requests to the remote neo4h server authenticated using basic http authentication, but server response doesn't contain auth part in urls, so, all requests other then _getRoot becomes unauthorized. I've solved it using monkey-patching request library, but, maybe you can fix it better in future releases. Here is my solution:

at the beginning of lib/GraphDatabase_.coffee

status = require 'http-status'
request = require 'request'

applyPatch = (method, auth) ->
    return if applyPatch.patched[method]
    applyPatch.patched[method] = true
    __m = request[method]
    request[method] = ->
        args = [].slice.call(arguments)
        url = args[0]
        console.log(args)
        if url && url.match && !url.match(/https?:\/\/[^\/]*@/)
            args[0] = url.replace(/http:\/\//, 'http://' + auth + '@')
        if url && url.url && url.url.match && !url.url.match(/https?:\/\/[^\/]*@/)
            args[0].url = url.url.replace(/http:\/\//, 'http://' + auth + '@')
        __m.apply(request, args)

applyPatch.patched = {}

util = require './util_'
adjustError = util.adjustError

Relationship = require './Relationship_'
Node = require './Node_'

module.exports = class GraphDatabase
    constructor: (url) ->

        @url = url
        @auth = require('url').parse(url).auth

        applyPatch('get', @auth)
        applyPatch('post', @auth)

I know, this is a quick and dirty solution, just wanted to demonstrate what happens.

I can't seem to get basic functionality working with this library. I was starting off getting a reference node to add new data too but something as trivial as that doesn't seem to work. I'm sure I'm just missing something basic but there's no error messages and nothing to go on.

var neo4j = require('neo4j');
var db = new neo4j.GraphDatabase('http://localhost:7474');

var node = db.getNodeById(0);
console.log(node.self);

Outputs

undefined

The neo4j service is up and running. The following GET works just fine from the command line:

GET http://localhost:7474/db/data/node/0

From the changelog:

o Queries can now be pre-pended with which parser version to use

This seems like it might be useful, in case future Neo4j versions introduce breaking changes (which is likely). I don't see how to specify this in the REST API though. Need to investigate, but low-pri for us.

Connecting to the local db works just fine but trying to connect to the hosted db provided through heroku keeps throwing HTTP 500 errors. My understanding is that it's an authentication issue.

var db = new neo4j.GraphDatabase(process.env.NEO4J_URL || 'http://localhost:7474');

Why does this work for the template app but not ours? Any suggestions?

i could not find a way to get bulk or batch operations working with node-neo4j. i am looking on creating a dozen nodes and relating them to an existing node all in one query.

Any ideas?

Hi

(not an issue)
node-neo4j cool social analytics is in GitMoon node-neo4j, you might want to check it out.

Thanks,
Yaron

Hi All,

I am using the node-neo4j package on a project that I am hosting on Heroku. Upon pushing the project, it fails on the post-install step of node-neo4j. However, the package installs just fine on my local machine, so I am not sure if there is an actual issue here. Alternatively, I am thinking that there could be an issue with the buildpack, or with the specific versions of Node.js (v0.8.2) and NPM (v1.1.39) I am using on Heroku.

I will continue to investigate the problem, but if anyone has seen this before, or knows what might be going on, that would be most helpful.

Please find the Heroku output below.

Best,
-Eric

> [email protected] postinstall /tmp/build_20zj9urr4tkew/node_modules/neo4j
> npm run build

sh: npm: not found
npm http GET https://registry.npmjs.org/crc/0.2.0
npm http GET https://registry.npmjs.org/mime/1.2.6
npm ERR! [email protected] postinstall: `npm run build`
npm ERR! `sh "-c" "npm run build"` failed with 127
npm ERR! 
npm ERR! Failed at the [email protected] postinstall script.
npm ERR! This is most likely a problem with the neo4j package,
npm ERR! not with npm itself.
npm ERR! Tell the author that this fails on your system:
npm ERR!     npm run build
npm ERR! You can get their info via:
npm ERR!     npm owner ls neo4j
npm ERR! There is likely additional logging output above.

npm ERR! System Linux 2.6.32-343-ec2
npm ERR! command "/tmp/node-node-oFAf/bin/node" "/tmp/node-npm-vcxz/cli.js" "install" "--production"
npm ERR! cwd /tmp/build_20zj9urr4tkew
npm ERR! node -v v0.8.2
npm ERR! npm -v 1.1.39
npm ERR! code ELIFECYCLE
npm ERR! message [email protected] postinstall: `npm run build`
npm ERR! message `sh "-c" "npm run build"` failed with 127

I'm using neo4j from an express app. What's the best practice for create vs. reuse connections? I've noticed that a new connection always sends some handshake request to the server. So often creation is not good. But is it ok to reuse one connection object throughout the whole app sessions?

hi !
i'm trying to get the nodes properties for the nodes contained in the array return by Node.path().
but path.nodes[i].data is always empty...
did i miss something ?
cheers,
guillaume

1.6 adds this:

o Exposes get-or-create-uniquely via REST in ex. POST /index/node/?unique {...}

We should consider adding support for specifying this in our library too. Low-pri for us because we prefer to use the auto-index, so we never manually index stuff currently.

Hi, I like node-neo4j very much.

I need to implement someting like this
" ... relation = db.createRelationship(currentNodeId, id, "VSEBUJE", saveRelation) ..."

but in source code is stated:

"...
# Relationships
createRelationship: (startNode, endNode, type, _) ->
# TODO: Implement
..."

Can you give me some advise how can I implement this metod.

Thank you , Jadran

Hi

I now use neo4j outside of the firewall and use an http proxy to connect outsude.

Is there a built-in way to handle this? for now I have overriden this in util.js which works fine:

request = require("request").defaults({'proxy':'http://my-proxy:8080/'})

Thanks,
Yaron

When making a gremlin based request the return object i full of strings, not other objects with usable data.

Server url
http://localhost:7474
Kernel version
Neo4j - Graph Database Kernel 1.8.1

[email protected] test /home/paul/workspace/helpdeskgenius/node_modules/neo4j

_coffee test

passed CRUD tests

/home/paul/workspace/helpdeskgenius/node_modules/neo4j/node_modules/streamline/lib/callbacks/runtime.js:134
throw err;
^
AssertionError: "user8" == "user7"
at ___ (/home/paul/workspace/helpdeskgenius/node_modules/neo4j/test/cypher._coffee:196:4338)
at ___ (/home/paul/workspace/helpdeskgenius/node_modules/neo4j/node_modules/streamline/lib/callbacks/runtime.js:85:13)
at GraphDatabase_prototype_query__13 (/home/paul/workspace/helpdeskgenius/node_modules/neo4j/lib/GraphDatabase.js:241:56)
at main (/home/paul/workspace/helpdeskgenius/node_modules/neo4j/test/cypher._coffee:119:12)

npm ERR! Test failed. See above for more details.
npm ERR! not ok code 0

http://docs.neo4j.org/chunked/stable/rest-api-cypher.html#rest-api-send-queries-with-parameters

See neo4j/community#494 for an example why.

When googling for "neo4j nodejs", it does not show this project in page 1.
But googling for "neo4j node js" does.

As it says. =)

http://about.travis-ci.org/docs/

Hello, I'm having trouble executing concurrent, (mutating?) requests. Below is a gist that should give you an idea of where the problem arises:

https://gist.github.com/3036561

Of course, this example is contrived, but in a webserver context (especially with node), it is not impossible to imagine such an event occurring naturally. I don't believe this is an issue with the node-neo4j library, but the error returned isn't very helpful so it's hard to tell... any info you have on this would be very much appreciated!

I'm really confused with this lib.

Maybe is due of the lack of doc..

I was checking the codes and i don't come with the answer.

I have 3 nodes

node1, node2, node3

And 2 relationships
node1->node2
node2->node3

I want to get the best path from node1 to node3

But i don't know how to create a path and how to use it....

Any help?

I see a lot of nice methods in neo4j for finding best paths but seems to don't be implemented here?

My actual code can be reviewed in https://github.com/naxhh/playground/blob/master/neo4j/app.js

Hi, Is there a method for us to nest objects within properties of other nodes?
Otherwise is there an efficient way of converting a json with several nested objects, into many nodes and relationship ?

{
Name: bob,
School: SMU,

.
.
.
(Nested properties and objects)
}

The documentation implies that the callback passed to Node.save() receives an error argument and a results argument. However, no arguments are sent on a successful request. Even though the Node instance itself is updated on a successful request, it might be helpful to pass that Node to the callback in the results argument. Failing that, the documentation should be updated to make it clear what the behavior is. I spun my wheels for a while thinking that my Node.save() calls were failing because the request argument was 'undefined.'

Hi

I'm using neo4j 1.8 and node-neo4j.

I need to delete 100 relationships outgoing from node X to different nodes. When I do this synchronously there is no problem (e.g. I wait for one delete to finish before calling the next one).
When I delete them in an async manner (many rest requests sent concurrently) I intermittently get a few failures. Even when I limit the concurrent deletes to 3. I see this coming back in the http response once per each error:

HTTP/1.1 500 Transaction(15744)[STATUS_ACTIVE,Resources=1] can't wait on resource RWLock[Relationship[8253]] since => Transaction(15744)[STATUS_ACTIVE,Resources=1] <-[:HELD_BY]- RWLock[Node[1841]] <-[:WAITING_FOR]- Transaction(15744)[STATUS_ACTIVE,Resources=1] <-[:HELD_BY]- RWLock[Relationship[8253]]
Content-Length: 0
Server: Jetty(6.1.25)

I believe neo4j uses keep-alive to reuse the same connection but this does not seem relevant. I have verified my code is (logically) correct, e.g. I do not delete an already deleted link and etc. Also I delete using "relationship.del _, true".

Any idea?

Thanks,
Yaron

Is is possible to have the Node.outgoing() method to support passing in a sorting option?

I'm currently using a raw query to get an ordered relationship...

I cannot get the simple example code working since the code fails with exception:

TypeError: Cannot read property 'url' of undefined
    at exports.wrapRequest.modifyArgs (/home/jhh/git/nko3/oulu/api/node_modules/neo4j/lib/util.js:27:17)
    at Object.exports.wrapRequest.wrapper.(anonymous function) [as get] (/home/jhh/git/nko3/oulu/api/node_modules/neo4j/lib/util.js:46:37)
    at ___ (/home/jhh/git/nko3/oulu/api/node_modules/neo4j/lib/GraphDatabase.js:59:40)
    at ___ (/home/jhh/git/nko3/oulu/api/node_modules/neo4j/node_modules/streamline/lib/callbacks/runtime.js:85:13)
    at GraphDatabase_prototype__getRoot__1 (/home/jhh/git/nko3/oulu/api/node_modules/neo4j/lib/GraphDatabase.js:48:56)
    at GraphDatabase_prototype_getServices__2 (/home/jhh/git/nko3/oulu/api/node_modules/neo4j/lib/GraphDatabase.js:59:19)

My test code is like this:

var GraphDatabase = require('neo4j').GraphDatabase;
var _db = new GraphDatabase('http://foo:[email protected]:7526/db/data/');
// ...
    var n = _db.createNode(req.params);
    n.save(function (err, n) {
        if (err) {
            return next('Error saving new node to database');
        } else {
            res.send(201, {'node':n});
            return next();
        }
    });

What am I doing wrong? Or is this a bug?

I'm looking into using neo4j's built in dijkstra algorithm, it looked like there's already some support in node-neo4j for shortest path but dijkstra support would be great.

First off, thanks for the code!

Getting an issue on heroku where I'm able to enter and retrieve new users just fine but the webpage still responds that there's no possible followers.

"There's no one else left for (name) to follow!"

I've been poking around with this in users.js:

// GET /users/:id
exports.show = function (req, res, next) {
    User.get(req.params.id, function (err, user) {
        if (err) return next(err);
        // TODO also fetch and show followers?
        user.getFollowingAndOthers(function (err, following, others) {
            if (err) return next(err);
            res.render('user', {
                user: user,
                following: following,
                others: others
            });
        });
    });
};

and this in users.js

User.prototype.getFollowingAndOthers = function (callback) {
    // query all users and whether we follow each one or not:
    var query = [
        'START user=node({userId}), other=node:INDEX_NAME(INDEX_KEY="INDEX_VAL")',
        'MATCH (user) -[rel?:FOLLOWS_REL]-> (other)',
        'RETURN other'//, COUNT(rel)'  // COUNT(rel) is a hack for 1 or 0
    ].join('\n')
        .replace('INDEX_NAME', INDEX_NAME)
        .replace('INDEX_KEY', INDEX_KEY)
        .replace('INDEX_VAL', INDEX_VAL)
        .replace('FOLLOWS_REL', FOLLOWS_REL);

    var params = {
        userId: this.id,
    };

But can't figure out why it doesn't show other users as possible followers.

Is this the same code as is running in the web example? There's other data in my neo4j instance, could that possibly be messing with it?

Grateful for any help,

Thanks!

Edited to format the code! –@aseemk