Ecosyste.ms: Awesome
An open API service indexing awesome lists of open source software.
https://github.com/zgbjgg/oox
OOX - slave isolated process for jun environments
https://github.com/zgbjgg/oox
cluster dataframe elixir erlang jun python slave
Last synced: about 1 month ago
JSON representation
OOX - slave isolated process for jun environments
- Host: GitHub
- URL: https://github.com/zgbjgg/oox
- Owner: zgbjgg
- License: mit
- Created: 2018-03-14T17:26:54.000Z (almost 7 years ago)
- Default Branch: master
- Last Pushed: 2020-02-28T16:39:45.000Z (almost 5 years ago)
- Last Synced: 2024-03-14T19:34:50.011Z (10 months ago)
- Topics: cluster, dataframe, elixir, erlang, jun, python, slave
- Language: Erlang
- Homepage:
- Size: 792 KB
- Stars: 6
- Watchers: 2
- Forks: 1
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# oox
**OOX** - slave isolated process & scheduler for jun environments
[](https://hex.pm/packages/oox)
[](https://travis-ci.org/zgbjgg/oox)
[](https://codecov.io/gh/zgbjgg/oox)
[](https://raw.githubusercontent.com/zgbjgg/oox/master/LICENSE)
[](https://hex.pm/packages/oox)
[](https://hex.pm/packages/oox)
OOX is a set of tools to execute many jun commands into an isolated slave process, taking control of the cluster (slaves) and managing
the connection to it. This tools also can be used to generate a job and check if passed or broken (similar to TDD) but applied to
jun commands and taking advantage of the slave to run all process on a separated node.
This project is under development and should not be used in production, it's not ready for that.
## Prerequisites
In order to execute the system correctly, just start the environment in a valid node (short & long names are allowed)
```shell
$ erl -pa _build/default/lib/*/ebin/ -name [email protected] -setcookie oox
```
After executing this all dependencies are set in path, just start them:
```erlang
([email protected])1> Apps = [syntax_tools, compiler, goldrush, lager, erlport, jun, oox].
[syntax_tools,compiler,goldrush,lager,erlport,jun,oox]
([email protected])2> lists:foreach(fun(App) -> application:start(App) end, Apps).
18:07:20.845 [info] Application lager started on node '[email protected]'
18:07:20.846 [info] Application erlport started on node '[email protected]'
18:07:20.848 [info] Application jun started on node '[email protected]'
18:07:20.850 [info] Application oox started on node '[email protected]'
ok
```
Don't worry about starting the apps in the environment if you use `oox` for example in a rebar.config,
this is only to show how the system works.
## Add a subscriber process
Since jobs can be executed in async mode, the response will be delivered to a process, just set it:
```erlang
([email protected])8> oox_scheduler:set_subscriber(self()).
ok
```
## Creating a new job
Now creates a new job, this will start a new slave on the host, ready to execute jun commands, it means that
slave contains an isolated jun environment, adding a job is easy:
```erlang
([email protected])9> {ok, Job} = oox_scheduler:add_job("127.0.0.1").
{ok,<0.115.0>}
11:40:12.308 [info] ensuring started for jun main dependency, state ok
11:40:13.314 [info] reached slave node '[email protected]', sending launch signal to main process <0.118.0>
11:40:13.314 [info] receiving launch signal for slave '[email protected]'
11:40:13.689 [info] sent RPC to slave node '[email protected]' with response {ok,<11388.64.0>}
11:40:13.689 [info] starting jun worker on slave node at <11388.64.0>
```
## Using commands
Define some commands in order to send & execute into slave via scheduler, the commands can be a specific overriden variables,
for example if in the command you want to use the worker, just use $worker in the args. Example:
```erlang
[[{mod,jun_pandas},
{func,read_csv},
{args,['$worker','./files/csv.txt']}],
[{mod,jun_pandas},
{func,head},
{args,['$worker','$dataframe',1,[]]}],
[{mod,jun_pandas},
{func,max},
{args,['$worker','$dataframe',no_age,[]]}]]
```
## Starting the job
Start the job to process the commands in the slave and waiting for response (remember response is delivered to the subscriber),
when job finish also the slave is stopped.
```erlang
([email protected])10> oox_scheduler:start_job(Job, Cmds).
{ok,working}
```
## Receive response of job
In the subscriber process you will receive a response based on the execution of commands, there is two posibles responses: passed or broken.
```erlang
([email protected])11> flush().
Shell got {job,<0.115.0>,{broken,{error,{'exceptions.KeyError',"Atom('fage')"}}}}
ok
```
### See also
[JUN: python pandas support for dataframes manipulation over erlang](https://github.com/zgbjgg/jun)
#### Authors
@zgbjgg Jorge Garrido