Consider situation:
I have an Ubuntu server with installed Python, tensorflow and other libs.
My code is python script, that load several models, some of them pretrained vectors .bin, some files from server folders, etc.
When i run script in terminal it launch interactive session, where i input some text and script output me back (like chatbot). During answer it call my Ai models (Tensorflow, keras).
Question: how do i access this running session from other python script? I mean i want use it as a function: to send text and receive answer back.
And of course i need to run this terminal session in background for long time.
I read this and similar answers, but not sure is that right solution (seems not a full):
In Linux, how to prevent a background process from being stopped after closing SSH client
What i am asking, commonly is done by REST server with API that expose and then this api is called from a external code. But there is no API wotking: Tensorflow throw errors when run via Flask (was not able to fix).
If you want your script stays up after closing ssh session, add & disown at the end of your execution command and it will run in background.
Related
I try to run all the rules by followed commands:
touch scripts/*.py
snakemake --cores <YOUR NUMBER>
The problem is my local internet connection is unstable, I could submit the command through putty to the linux computation platform, while it seems that there're always outputs returns back to the putty interface. So when my local internet connection is interrupted, the code running is also interrupted.
Is there any methods that I could let the codes just run on the linux backends? Then outputs could be written in the log file at last.
This could be a very basic question.
This is a common problem (not just for snakemake), and there are several options, at least the following:
use a program that can persists across multiple connection: popular options are screen, tmux. The workflow would look like this: log on to the server, launch screen or tmux, once inside the program launch the code you would like to run, log off, next time you login to the server, you can reconnect to the previous session and observe computations that were done in the meantime. I recommend tmux, see this tmux tutorial.
use nohup, this launches the computation in the background and it will continue running on the server if you disconnect:
nohup snakemake --cores <YOUR NUMBER>
Note that with this option, if you want to see the progress of computation, you will need to watch the appropriate .log inside the .snakemake folder.
I'm having some problem making a python file run everytime the AWS server boots.
I am trying to run a python file to start a web server on Amazon Webservice EC2 server.
But I am limited to edit systemd folder and other folders such as init.d
Is there anything wrong?
Sorry I don't really understand EC2's OS, it seems a lot of methods are not working on it.
What I usually do via ssh to start my server is:
python hello.py
Can anyone tell me how to run this file automatically every time system reboots?
It depends on your linux OS but you are on the right track (init.d). This is exactly where you'd want to run arbitrary shell scripts on start up.
Here is a great HOWTO and explanation:
https://www.tldp.org/HOWTO/HighQuality-Apps-HOWTO/boot.html
and another stack overflow specific to running a python script:
Run Python script at startup in Ubuntu
if you want to share you linux OS I can be more specific.
EDIT: This may help, looks like they have some sort of launch wizard:
https://docs.aws.amazon.com/AWSEC2/latest/UserGuide/user-data.html
When you launch an instance in Amazon EC2, you have the option of
passing user data to the instance that can be used to perform common
automated configuration tasks and even run scripts after the instance
starts. You can pass two types of user data to Amazon EC2: shell
scripts and cloud-init directives. You can also pass this data into
the launch wizard as plain text, as a file (this is useful for
launching instances using the command line tools), or as
base64-encoded text (for API calls).
I have coded a Python Script for Twitter Automation using Tweepy. Now, when i run on my own Linux Machine as python file.py The file runs successfully and it keeps on running because i have specified repeated Tasks inside the Script and I also don't want to stop the script either. But as it is on my Local Machine, the script might get stopped when my Internet Connection is off or at Night. So i couldn't keep running the Script Whole day on my PC..
So is there any way or website or Method where i could deploy my Script and make it Execute forever there ? I have heard about CRON JOBS before in Cpanel which can Help repeated Tasks but here in my case i want to keep running my Script on the Machine till i don't close the script .
Are their any such solutions. Because most of twitter bots i see are running forever, meaning their Script is getting executed somewhere 24x7 . This is what i want to know, How is that Task possible?
As mentioned by Jon and Vincent, it's better to run the code from a cloud service. But either way, I think what you're looking for is what to put into the terminal to run the code even after you close the terminal. This is what worked for me:
nohup python code.py &
You can add a systemd .service file, which can have the added benefit of:
logging (compressed logs at a central place, or over network to a log server)
disallowing access to /tmp and /home-directories
restarting the service if it fails
starting the service at boot
setting capabilities (ref setcap/getcap), disallowing file access if the process only needs network access, for instance
So I am trying to set up a Continuous integration environment using Jenkins.
One of the build step requires a series of mouse actions/movements to accomplish a task in Excel. I have already written a python script using the ctypes library to do this.
The script works perfectly fine if I run it either through Jenkins or on the server itself when I am actively logged in to the server using remote desktop connection, but as soon as I minimize/close the connection and then run the script from Jenkins, it seems the mouse events never get executed. Is there something I can add to the script to make this work? Thanks for any help you can provide.
I have been interested in finding an alternative to the UI in SAS for quite some time now. We license SAS on our server instead of our desktops, so furthermore we have to launch a remote desktop application to execute code.
I was able to use a Telnet connection instead to remotely connect to the server, and batch execute SAS programs. Then I was interested in whether a python script could be made to connect remotely, and batch execute code, and this script could be executed in jEdit as a BeanShell script.
So far, I have Python code which successfully opens and closes the Telnet connection. It can do basic shell functions like call "dir". However, when I pass the exact same line that I use to execute SAS from command prompt on the remote server with a telnet connection in Python, nothing happens.
Is it possible the server is preventing me from executing code from a script? I use a "read_until" statement for the prompt before running any code.
Here's a few ideas...
The issue you are having above may be related to Local Security Policy settings in Windows (if it is running on a windows server). I'm far from an expert on that stuff but I remember older SAS/Intranet installations required some rumaging around in there to get them working.
As an alternative to the approach you are trying above you could also setup a SAS session on the server that listens for incoming socket requests as per this article:
http://analytics.ncsu.edu/sesug/2000/p-1003.pdf
And finally... Not sure if this helps or not by I remotely execute SAS jobs using PSEXEC. A description of how I set it all up can be found here:
http://www.runsubmit.com/questions/260/hide-sas-batch-jobs-winxp
Good luck
This paper outlines how you can use a Python script to connect to a Unix server using SSH, copy the SAS program written locally onto the server, batch submit it, and download the results back to your local machine, all using a BeanShell macro script for jEdit.