@bkjones Please update the pip to include https feature
https://pypi.org/project/pyrabbit/1.1.0/

Its not updated to the latest version.

Might be worth while mentioning that the Management Plugin is not enabled by default and will need to be enabled for the example code to work.

https://www.rabbitmq.com/management.html is a good URL to point people to.

Can you remove the Print in http.py on line 101. I dont want that information in my console

ie.
/api/bindings/[VHOST]/e/[SourceExchange]/e/[TargetExchange]
JSON
{"vhost":"vhostname","source":"SourceExchange","destination_type":"e","destination":"TargetExchange","routing_key":"","arguments":{}}

Method HTTPClient.do_call() (http.py, line 101) unconditionally echoes url. I use pyrabbit in an agent which catches and parses stdout. Echoed url causes error since is unexpected by the agent.

I see two possible solutions:

1. The url is a debug message and thus it shall be removed.
2. The url printing is useful somehow. Then the client class shall have option silent=False. If set to true, the url won't be echoed.

I'm using pyrabbit version 1.1.0 with RabbitMQ-3.6.0

Below is the python code I have written using pyrabbit

from pyrabbit.api import Client
cl = Client('localhost:15672', 'guest', 'guest')
print cl.is_alive()
cl.create_vhost('example_vhost')
print cl.get_vhost_names()
print cl.get_vhost_permissions('example_vhost')
cl.set_vhost_permissions('example_vhost', 'guest', '.', '.', '.*')

The output of script shows a NetworkError as shown below:

True
[u'/', u'example_vhost']
None
Traceback (most recent call last):
File "C:\Gayatri\learning\test_pyrabbit.py", line 14, in
cl.set_vhost_permissions('example_vhost', 'guest', '.', '.', '.*')
File "C:\Python27\lib\site-packages\pyrabbit-1.1.0-py2.7.egg\pyrabbit\api.py",
line 345, in set_vhost_permissions
headers=Client.json_headers)
File "C:\Python27\lib\site-packages\pyrabbit-1.1.0-py2.7.egg\pyrabbit\http.py"
, line 111, in do_call
raise NetworkError("Error: %s %s" % (type(out), out))
pyrabbit.http.NetworkError: Error: <class 'socket.error'> [Errno 10053] An estab
lished connection was aborted by the software in your host machine

Any help to resolve this issue?

Thanks in Advance,
Gayatri

I believe in docs/index.rst line

cl = Client('http://localhost:55672/api', 'guest', 'guest')

should be

cl = Client('localhost:55672', 'guest', 'guest')

While copying example code, I first got:

root@2b2e772eeff1:/code# ./sender-http.py 
http://http:/rabbitmq:15672/api/api/exchanges/myvh/amq.default/publish
Traceback (most recent call last):
  File "/usr/local/lib/python3.6/dist-packages/httplib2/__init__.py", line 1187, in _conn_request
    conn.connect()
  File "/usr/local/lib/python3.6/dist-packages/httplib2/__init__.py", line 918, in connect
    for res in socket.getaddrinfo(host, port, 0, socket.SOCK_STREAM):
  File "/usr/lib/python3.6/socket.py", line 745, in getaddrinfo
    for res in _socket.getaddrinfo(host, port, family, type, proto, flags):
socket.gaierror: [Errno -2] Name or service not known

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
  File "/usr/local/lib/python3.6/dist-packages/pyrabbit/http.py", line 106, in do_call
    headers)
  File "/usr/local/lib/python3.6/dist-packages/httplib2/__init__.py", line 1514, in request
    (response, content) = self._request(conn, authority, uri, request_uri, method, body, headers, redirections, cachekey)
  File "/usr/local/lib/python3.6/dist-packages/httplib2/__init__.py", line 1264, in _request
    (response, content) = self._conn_request(conn, request_uri, method, body, headers)
  File "/usr/local/lib/python3.6/dist-packages/httplib2/__init__.py", line 1194, in _conn_request
    raise ServerNotFoundError("Unable to find the server at %s" % conn.host)
httplib2.ServerNotFoundError: Unable to find the server at http

And string http://http:/rabbitmq:15672/api/api/exchanges/myvh/amq.default/publish has obvious errors.

Hi,
I found an issue when using the api method get_queue. It does not handle properly the symbol "+" and then, a query like get_queue("/", "a+b") fails.
My workaround is to do:
get_queue("/", re.sub(re.escape("+"), "%2B", "a+b"))
which is not needed for other special symbols.
Thanks!

c = Client(...)
c.create_queue(queue_name, '/', node='s-rabbit3', auto_delete='true', durable='true')

The above code runs fine, but if you take out any 1 or 2 of the kwargs (node/auto_delete/durable), an exception will be raised:

In [35]: c.create_queue(queue_name, '/', node='s-rabbit3', auto_delete='true')
---------------------------------------------------------------------------
HTTPError                                 Traceback (most recent call last)
<ipython-input-35-07fe5a3f45ec> in <module>()
----> 1 c.create_queue(queue_name, '/', node='s-rabbit3', auto_delete='true')

/opt/local/Library/Frameworks/Python.framework/Versions/2.6/lib/python2.6/site-packages/pyrabbit/api.pyc in create_queue(self, name, vhost, auto_delete, durable, arguments, node)
    548                                  'PUT',
    549                                  body,
--> 550                                  headers=Client.json_headers)
    551
    552     def delete_queue(self, vhost, qname):

/opt/local/Library/Frameworks/Python.framework/Versions/2.6/lib/python2.6/site-packages/pyrabbit/http.pyc in do_call(self, path, reqtype, body, headers)
    103         # 'success' HTTP status codes are 200-206
    104         if resp.status < 200 or resp.status > 206:
--> 105             raise HTTPError(content, resp.status, resp.reason, path, body)
    106         else:
    107             if content:

HTTPError: 400 - Bad Request ([u'not_boolean', None]) (queues/%2F/reputation.workers.notification_worker.NotificationWorker) ('{"node": "s-rabbit3", "auto_delete": "true", "durable": null}')

Seems like the pypi license hasn't updated from MIT to BSD-3-Clause?

The order of parameters fed into create_queue are reversed compared to all of the other functions that take similar parameters. Consistency would be nice.

I use a regular connection
Client('localhost:55672', 'guest', 'guest')

when i try to make a get_queue('/', queue-name) and the queue does not exist, it throws an error that is not handle correctly since it is not in JSON format.

When I try to set permissions and run script with set_vhost_permissions inside docker container, it is always raise BrokenPipeError. But if I run it without docker everything fine.

Other methods (like create_vhost/delete_vhost) works fine in docker container.

Hi.
I'm create work pyrabbit with shovel plugin. Can I create merge request?

How about new release?
Master branch contain a lot if fixes, but last release was in 2014 :)

Thanks for getting the new release out!

What do you think about starting a change log/release notes document with an overview of changes since the previous release? Would be a big help and much easier than trying to diff commit logs

It would be useful to expose the headers of the message when publishing to a queue so that you could specify, for example, that the message encoding is JSON.

Pyrabbit generates malformed urls to the administrative api when used in a Windows environment. Each request will receive the expected response of 'Not Found'.

The issue is caused by using os.path.join to construct urls. See pyrabbit.http.do_call.

We host our RabbitMQ server on the internet and access it via HTTPS. Currently pyrabbit can only access the RabbitMQ management API over HTTP.

If you have a binding using '#' as the routing key, invoking delete_binding with '#' will not work.

The DELETE http call is being made without properly encoding the '#' in the URL, resulting in a 405 error.

A workaround now is to manually url-encode the routing key ('%23') in the delete_binding() call

Hello,

when i try to run several RabbitMQ api requests concurrently with threads, i get strange "socket timeout" errors from pyrabbit. If I run the requests in sequence, no error.

I suppose the source cause lies in httplib2, which is probably not thread-safe.

Could you provide a 'requests' (http://docs.python-requests.org/) or urllib3 (https://urllib3.readthedocs.org/) backoffice instead ?

Regards,
Stephane

So far I've been using the pypi version of pyrabbit, but we only had http endpoints.

I have an https endpoint to reach now, and I see that a PR was merged to this effect; so I pip-installed the current git revision and was able to use https.

However I had to specify the /api/ part of the path - looks like pyrabbit doesn't include this part of the path anymore. Previously I was able to create a client by specifying a server/port like this: myserver.mydomain.com:15672. This is no longer possible, but it starts working again with myserver.mydomain.com:15672/api/.

Is this an expected change?

The HTTP endpoint returns a lot of fields not needed which increases the payload.
RabbitMQ supports this using the columns parameter (as described here):
https://rawcdn.githack.com/rabbitmq/rabbitmq-server/v3.8.16/deps/rabbitmq_management/priv/www/api/index.html

You can also restrict what information is returned per item with the columns parameter. This is a comma-separated list of subfields separated by dots.

It would be great if pyrabbit support it! ❤️

get_messages shows an error when trying it with the latest rabbitMQ, missing key 'ackmode'
When using the web browser GUI, there is a field called "Ack mode" for getting messages from a queue, upon further inspection these are the values:

<select name="ackmode">
                <option value="ack_requeue_true" selected="">Nack message requeue true</option>
                <option value="ack_requeue_false">Automatic ack</option>
                <option value="reject_requeue_true">Reject requeue true</option>
                <option value="reject_requeue_false">Reject requeue false</option>
</select>

I added it to my local pyrabbit package and it works :)

def get_messages(self, vhost, qname, count=1,
                     requeue=False, truncate=None, encoding='auto', ackmode='ack_requeue_false'):
...
base_body = {'count': count, 'requeue': requeue, 'encoding': encoding, 'ackmode': ackmode}

Python 3.3.4
Pyrabbit 1.0.1
Httlib2 0.9

The following exception is thrown with an api call such as is_alive()

  ...
  File "C:\Python33\lib\site-packages\pyrabbit\api.py", line 48, in wrapper
    if self.has_admin_rights:
  File "C:\Python33\lib\site-packages\pyrabbit\api.py", line 175, in has_admin_rights
    whoami = self.get_whoami()
  File "C:\Python33\lib\site-packages\pyrabbit\api.py", line 161, in get_whoami
    whoami = self.http.do_call(path, 'GET')
  File "C:\Python33\lib\site-packages\pyrabbit\http.py", line 108, in do_call
    content = self.decode_json_content(content)
  File "C:\Python33\lib\site-packages\pyrabbit\http.py", line 75, in decode_json_content
    py_ct = json.loads(content)
  File "C:\Python33\lib\json\__init__.py", line 316, in loads
    return _default_decoder.decode(s)
  File "C:\Python33\lib\json\decoder.py", line 351, in decode
    obj, end = self.raw_decode(s, idx=_w(s, 0).end())
TypeError: can't use a string pattern on a bytes-like object

Hey there, thanks so much for releasing/maintaining this @bkjones !

I'm having a bit of trouble using this:

Client('localhost:5672', 'guest', 'guest').get_users()

I believe that's the proper way to initialize a connection? Based on the printed url, it seems to be right.. when I go there in my browser I get AMQP back from rabbitmq server.

In [15]: Client('localhost:5672', 'guest', 'guest').get_users()
http://localhost:5672/api/whoami
---------------------------------------------------------------------------
AttributeError                            Traceback (most recent call last)
<ipython-input-15-663fd0826ea7> in <module>()
----> 1 Client('localhost:5672', 'guest', 'guest').get_users()

/Users/eric/.virtualenvs/codalab-competitions/lib/python2.7/site-packages/pyrabbit/api.pyc in wrapper(self, *args, **kwargs)
     46 
     47         """
---> 48         if self.has_admin_rights:
     49             return fun(self, *args, **kwargs)
     50         else:

/Users/eric/.virtualenvs/codalab-competitions/lib/python2.7/site-packages/pyrabbit/api.pyc in has_admin_rights(self)
    177         if self.is_admin is None:
    178             whoami = self.get_whoami()
--> 179             self.is_admin = whoami.get('tags', '') == 'administrator'
    180 
    181         return self.is_admin

AttributeError: 'NoneType' object has no attribute 'get'

It will collect statistics for each node of the cluster (or single node, if there is no cluster).
And so the same memory usage statistics (memory|binary)

add to api.py:

urls = {
            'nodes_by_name_memory': 'nodes/%s?memory=true',
            'nodes_by_name_binary': 'nodes/%s?binary=true',
            'nodes_by_name': 'nodes/%s',
#########################################


    def get_node(self, nodename, opt=None):
        """
        Get a single node, which requires nodename and optional memory or binary for memory statistic.

        :param string nodename: An individual node in the RabbitMQ cluster (even if there is no cluster).
        :param string opt: memory or binary.
            memory - to get memory statistics (it adds to the URL "?memory=true")
            binary - to get a breakdown of binary memory use  (it adds to the URL "?binary=true")
        :returns: A dictionary of queue properties.
        :rtype: dict

        """
        nodename = quote(nodename, '')
        if opt == 'memory':
            path = Client.urls['nodes_by_name_memory'] % (nodename)
        elif opt == 'binary':
            path = Client.urls['nodes_by_name_binary'] % (nodename)
        else:
            path = Client.urls['nodes_by_name'] % (nodename)

        queue = self.http.do_call(path, 'GET')
        return queue

Thank you!
Sorry for my english

I am using this library with CloudAMQP https://www.cloudamqp.com/ which gives me access to a single vhost.
I do not have administrator tags, but I have management and policymaker. Since @needs_admin_privs only checks for admin, it forbids me from calling Client.is_alive on my vhost even though I am allowed to do so (the call succeed with curl).