Supporting infrastructure to run scientific experiments without a scientific workflow management system.
Project description
Copyright (c) 2014 Universidade Federal Fluminense (UFF). Copyright (c) 2014 Polytechnic Institute of New York University. All rights reserved.
The noWorkflow project aims at allowing scientists to benefit from provenance data analysis even when they don’t use a workflow system. Also, the goal is to allow them to avoid using naming conventions to store files originated in previous executions. Currently, when this is not done, the result and intermediate files are overwritten by every new execution of the pipeline.
noWorkflow was developed in Python and it currently is able to capture provenance of Python scripts using Software Engineering techniques such as abstract syntax tree (AST) analysis, reflection, and profiling, to collect provenance without the need of a version control system or any other environment.
Installing and using noWorkflow is simple and easy. Please check our installation and basic usage guidelines below.
Team
The noWorkflow team is composed by researchers from Universidade Federal Fluminense (UFF) in Brazil and New York University (NYU), in the USA.
Vanessa Braganholo (UFF)
Fernando Chirigati (NYU)
Juliana Freire (NYU)
David Koop (NYU)
Leonardo Murta (UFF)
João Felipe Pimentel (UFF)
Publications
[MURTA, L. G. P.; BRAGANHOLO, V.; CHIRIGATI, F. S.; KOOP, D.; FREIRE, J.; noWorkflow: Capturing and Analyzing Provenance of Scripts. In: International Provenance and Annotation Workshop (IPAW), 2014, Cologne, Germany.] (https://github.com/gems-uff/noworkflow/raw/master/docs/ipaw2014.pdf)
Quick Installation
To install noWorkflow, you should follow these basic instructions:
If you have pip, just run:
$ pip install noworkflow[all]
This installs noWorkflow, PyPosAST, Flask, IPython Notebook and PySWIP. The only requirement for running noWorkflow is PyPosAST. The other libraries are only used for provenance analysis.
If you only want to install noWorkflow and PyPosAST, please do:
$ pip install noworkflow
If you do not have pip, but already have Git (to clone our repository) and Python:
$ git clone git@github.com:JoaoFelipe/pyposast.git
$ cd pyposast
$ ./setup.py install
$ cd ..
$ git clone git@github.com:gems-uff/noworkflow.git
$ cd noworkflow/capture
$ ./setup.py install
This installs noWorkflow on your system.
Basic Usage
noWorkflow is transparent in the sense that it requires neither changes to the script, nor any laborious configuration. Run
now --help
to learn the usage options.
To run noWorkflow with a script called simulation.py with input data data1.dat and data2.dat, you should run
now run -v simulation.py data1.dat data2.dat
The -v option turns the verbose mode on, so that noWorkflow gives you feedback on the steps taken by the tool. The output, in this case, is similar to what follows.
$ now run -v simulation.py data1.dat data2.dat
[now] removing noWorkflow boilerplate
[now] setting up local provenance store
[now] collecting definition provenance
[now] registering user-defined functions
[now] collecting deployment provenance
[now] registering environment attributes
[now] searching for module dependencies
[now] registering provenance from 703 modules
[now] collecting execution provenance
[now] executing the script
[now] the execution of trial 1 finished successfully
Each new run produces a different trial that will be stored with a sequential identification number in the relational database.
Verifying the module dependencies is a time consuming step, and scientists can bypass this step by using the -b flag if they know that no library or source code has changed. The current trial then inherits the module dependencies of the previous one.
It is possible to collect more information than what is collected by default, such as variable usages and dependencias. To perform a dynamic program slicing and capture those information, just run
now run -e Tracer simulation.py data1.dat data2.dat
To list all trials, just run
now list
Assuming we run the experiment again and then run , the output would be as follows.
$ now list
[now] trials available in the provenance store:
Trial 1: simulation.py data1.dat data2.dat
with code hash aa49daae4ae8084af3602db436e895f08f14aba8
ran from 2014-03-04 13:10:34.595995 to 2014-03-04 13:11:33.793083
Trial 2: simulation.py data1.dat data2.dat
with code hash aa49daae4ae8084af3602db436e895f08f14aba8
ran from 2014-03-04 17:59:02.917920 to 2014-03-04 18:00:10.383637
To look at details of an specific trial, use
now show
This command has several options, such as -m to show module dependencies; -d to show function definitions; -e to show the environment context; -a to show function activations; and -f to show file accesses.
Running
now show -a 1
would show details of trial 1. Notice that the function name is preceded by the line number where the call was activated.
$ now show -a 1
[now] trial information:
Id: 1
Inherited Id: None
Script: simulation.py
Code hash: aa49daae4ae8084af3602db436e895f08f14aba8
Start: 2014-03-04 13:10:34.595995
Finish: 2014-03-04 13:11:33.793083
[now] this trial has the following function activation graph:
42: run_simulation (2014-03-04 13:11:30.969055 -
2014-03-04 13:11:32.978796)
Arguments: data_b = 'data2.dat', data_a = 'data1.dat'
Globals: wait = 2
Return value: [['0.0', '0.6'], ['1.0', '0.0'], ['1.0', '0.0'],
...
To restore files used by trial 1, run
$ now checkout -l -i 1
By default, the checkout command only restores the script used for the trial (“simulation.py”), even when it has imports and read files as input. Use the option “-l” to restore imported modules and the option “-i” to restore input files. The checkout command track the evolution history. By default, subsequent trials are based on the previous Trial (e.g. Trial 2 is based on Trial 1). When you checkout a Trial, the next Trial will be based on the checked out Trial (e.g. Trial 3 based on Trial 1).
The remaining options of noWorkflow are diff, export and vis. The diff option compares two trials, and the export option exports provenance data of a given trial to Prolog facts, so inference queries can be run over the database.
The vis option starts a visualization tool that allows interactive analysis:
$ now vis -b
The visualization tool shows the evolotion history, the trial information, an activation graph. It is also possible to compare different trials in the visualization tool.
The visualization tool requires Flask to be installed. To install Flask, you can run
$ pip install flask
IPython Interface
Another way to run, visualize, and query trials is to use IPython notebook. To install IPython notebook, you can run
$ pip install ipython[all]
Then, to run ipython notebook, go to the project directory and execute:
$ ipython notebook
It will start a local webserver where you can create notebooks and run python code.
Before loading anything related to noworkflow on a notebook, you must initialize it:
In [1]: %load_ext noworkflow
...: import noworkflow.now.ipython as nip
It is equivalent to:
In [1]: %load_ext noworkflow
...: nip = %now_ip
After that, you can either run a new trial or load an existing object (History,Trial,Diff).
There are two ways to run a new trial: 1- Load an external file
In [1]: arg = 6
In [2]: trial = %now script1.py $arg
...: trial
Out [2]: <Trial 5> # Loads the trial object represented as a graph
2- Load the code inside a cell
In [3]: %%now --name script2 --out=out_var $arg
...: import sys
...: l = range(sys.argv[1])
...: c = sum(l)
...: print(c)
6
Out [3]: <Trial 6> # Loads the trial object represented as a graph
In [4]: out_var
Out [4]: "6\n"
Both modes supports all the now run parameters It is worth noticing that the noworkflow cannot access ipython variables and ipython cannot access noworkflow variables. To workaround this limitation, it is necessary to pass arguments to the program and parse the output. Pickle may be a good solution for serializing complex objects
Loading existing trials, histories and diffs:
In [5]: trial = nip.Trial(2) # Loads trial with Id = 2
...: trial # Shows trial graph
Out [5]: <Trial 2>
In [6]: history = nip.History() # Loads history
...: history # Shows history graph
Out [6]: <History>
In [7]: diff = nip.Diff(1, 2) # Loads diff between trial 1 and 2
...: diff # Shows diff graph
Out [7]: <Diff 1 2>
There are attributes on those objects to change the graph visualization, width, height and filter values. Please, check the documentation by running the following code on ipython notebook
In [8]: trial?
In [9]: history?
It is also possible to run prolog queries on IPython notebook. To do so, you will need to install SWI-Prolog with shared libraries and the pyswip module.
You can install pyswip module with the command:
$ pip install pyswip-alt
Check how to install SWI-Prolog with shared libraries at https://github.com/yuce/pyswip/blob/master/INSTALL
You can install pyswip
To query a specific trial, you can do:
In [10]: result = trial.query("activation(_, 550, X, _, _, _)")
...: next(result) # The result is a generator
Out [10]: {'X': 'range'}
To check the existing rules, please do:
In [11]: trial.prolog_rules()
Out [11]: [...]
Finally, it is possible to run the CLI commands inside ipython notebook:
In [12]: !now export ${trial.id}
Out [12]: %
...: % FACT: activation(trial_id, id, name, start, finish, caller_activation_id).
...: %
...: ...
Running
Included Software
Parts of the following software were used by noWorkflow directly or in an adapted form:
The Python Debugger Copyright (c) 2001-2013 Python Software Foundation. All Rights Reserved.
Acknowledgements
We would like to thank JetBrains for providing us a license for PyCharm. We also want to thank CNPq, FAPERJ, and the National Science Foundation (CNS-1229185, CNS-1153503, IIS-1142013) for partially supporting this work.
License Terms
Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the “Software”), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions:
The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software.
THE SOFTWARE IS PROVIDED “AS IS”, WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.