Merge branch 'master' of github.com:pricingassistant/mrq

Author: sylvinus

README.md

@@ -49,16 +49,17 @@ Create a sample project and write a simple task :
   $ vim tasks.py
 ```
 
-
-    from mrq.task import Task
-    import urllib2
-
-    class Fetch(Task):
-        def run(self, params):
-            f = urllib2.urlopen(params.get("url"))
-            t = f.read()
-            f.close()
-            return len(t)
+```python
+from mrq.task import Task
+import urllib2
+
+class Fetch(Task):
+    def run(self, params):
+        f = urllib2.urlopen(params.get("url"))
+        t = f.read()
+        f.close()
+        return len(t)
+```
 
 You can now run it using `mrq-run` :
 ```
@@ -93,7 +94,7 @@ $ mrq-run --async --queue fetches tasks.Fetch url http://www.google.com &&
 
 Now start the dasbhoard with `mrq-dashboard &` and go check your newly created queue and job on [localhost:5555](http://localhost:5555/#jobs)
 
-Instanciate a worker with `mrq-worker` and you can follow it on the dashboard as it executes in parallel all the enqueued jobs
+Instantiate a worker with `mrq-worker` and you can follow it on the dashboard as it executes in parallel all the enqueued jobs
 
 ```
 $ mrq-worker --gevent 10 fetches
@@ -122,4 +123,4 @@ $ mrq-worker --gevent 10 fetches
 
 # More
 
-Full documentation is available on [readthedocs](http://mrq.readthedocs.org/en/latest/)
+Full documentation is available on [readthedocs](http://mrq.readthedocs.org/en/latest/)

docs/jobs.md

@@ -1,10 +1,8 @@
 # Jobs & Tasks
 
-A **Job** is an unit of work that can be added to a Queue. It will be later dequeued by a Worker.
+A **Task** is a Python class wrapping a unit of processing. A task may invoke other tasks, synchronously or asynchronously (by queuing them as proper jobs).
 
-A **Task** is a Python class wrapping your asynchronous code.
-
-Jobs have a main Task, that will be called when the Job starts. In this Job context, the main Task can then call sub-Tasks or any other Python code.
+A **Job** is an instance of the execution of a Task. It must link to a specific Task via it's path and it must have parameters. It must be queued in a Queue. A Worker then dequeues and executes it. The worker updates it with metadata about it's execution : status, traceback ...
 
 ## Job statuses
 

docs/workers.md

@@ -1,7 +1,15 @@
+## Worker
 
-## Worker statuses
+A worker is a unit of processing, that dequeues jobs and executes them.
 
-A Worker is always in one of these statuses:
+It is started with a list of queues to listen to, in a specific order.
+
+It can be started with concurrency options (multiple processes and / or multiple greenlets). We call this whole group a single 'worker' even though it is able to dequeue multiple jobs in parallel.
+
+
+## Statuses
+
+At any time, a worker is in one of these statuses:
 
 * `init`: General worker initialization
 * `wait`: Waiting for new jobs from Redis

requirements-base.txt

@@ -1,4 +1,4 @@
-argparse==1.1
+argparse>=1.1
 redis==2.10.3
 pymongo==2.7.2
 gevent==1.0.1