https://github.com/wxzhou/autosave
Automatically save running results of your disposable code.
https://github.com/wxzhou/autosave
science-research simulation-environment
Last synced: about 2 months ago
JSON representation
Automatically save running results of your disposable code.
- Host: GitHub
- URL: https://github.com/wxzhou/autosave
- Owner: wxzhou
- License: mit
- Created: 2019-04-16T07:39:40.000Z (about 7 years ago)
- Default Branch: master
- Last Pushed: 2019-04-17T09:13:18.000Z (about 7 years ago)
- Last Synced: 2023-12-05T08:35:57.371Z (over 2 years ago)
- Topics: science-research, simulation-environment
- Language: MATLAB
- Homepage:
- Size: 12.7 KB
- Stars: 0
- Watchers: 1
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# autosave
Automatically save running results of your disposable code.
Unlike software engineers, scientists use computers as their experiment tools and write codes for scientific computation and simulation. A tool that helps them record the computation results is thus necessary:
* The codes are often disposable and no need for branching, rebase, merging or other Git operations. So code management tools such as Git may seem too professional for them.
* The codes are often saved along with the running results, such as data/variables and figures.
* The codes are often run for multiple times with changed parameters. It is necessary to record the code/parameter and their corresponding running results.
`autosave` is such a tool.
Currently supported language: Matlab, Python.
## For Matlab
Use `begin.m` and `over.m`. Typically, `begin.m` and `over.m` are the first and last line of your code, respectively (apart from comments and subfunctions).
A simple example:
```matlab
begin on;
omega = 1;
x = linspace(-pi,pi,100);
y = sin(omega.*x);
z = cos(omega.*x);
figure;
plot(x,y);
figure;
plot(x,z);
over;
```
Save the code as `Trial.m` and run it, you will find three new files in the same directory of `Trial.m`: `Trial_run1.mat`, `Trial_run1_fig1.fig`, and `Trial_run1_fig2.fig`. Clear your workspace and load `Trial_run1.mat`, you will find not only `omega`, `x`, `y` and `z` in your code, but also a new variable `CODE_CONTENT_IN_THIS_RUN` containing the code, the running time and other information of your computer. `CODE_CONTENT_IN_THIS_RUN` is designed to have such a long and all uppercase name to avoid conflict with your own variable.
If you change the value of `omega` from 1 to 2 and run the code the second time, you will find another three files: `Trial_run2.mat`, `Trial_run2_fig1.fig`, and `Trial_run2_fig2.fig`. Also, the variable `CODE_CONTENT_IN_THIS_RUN` is in the file `Trial_run2.mat`.
The code content is saved in `CODE_CONTENT_IN_THIS_RUN` in the command `begin on`. Therefore, you are free to change your code during exeution if it takes a long time.
`over` is default to save all variables in the workspace. Delete the variables before using `over` if you don't want to save them.
## For Python
Use `autosave.py` for Python 2.x and `autosave3.py` for Python 3.x.
The python part is not as convenient as that of matlab. You have to save the figure by yourself. A simple example:
```python
import numpy as np
import matplotlib.pyplot as plt
from autosave import saveRunData
omega = 1
x = np.linspace(-10,10,100)
y = np.sin(omega*x)
run_id = saveRunData({'omega':omega, 'x':x, 'y':y})
plt.figure
plt.plot(x,y)
plt.savefig(run_id+'_fig1.png')
```
Save the code as `Trial.py` and run it, you will find two new files in the same directory of `Trial.py`: `Trial_run1.pkl` and `Trial_run1_fig1.png`. Load data in `Trial_run1.pkl` using the package `cloudpickle` or the function `load` in `autosave.py`, you will find not only the key `omega`, `x` and `y`, but also `CODE_CONTENT_IN_THIS_RUN` containing the code content of your code.
If you change the value of `omega` from 1 to 2 and run the code the second time, you will find another two files: `Trial_run2.pkl` and `Trial_run2_fig1.png`. Also, the variable `CODE_CONTENT_IN_THIS_RUN` is in the file `Trial_run2.pkl`.
The code content is saved in `CODE_CONTENT_IN_THIS_RUN` when `saveRunData` is imported. Therefore, you are free to change your code during exeution if it takes a long time.