Metadata-Version: 2.1
Name: plotagain
Version: 1.0.2
Summary: A helper tool for matplotlib. Provides a contextmanager which automatically saves data used to make plots and automatically generates a script to re-load and re-plot this data at a later stage.
Author-email: Jeremy John Henry Wilkinson <jero.wilkinson@gmail.com>
License: BSD-3-Clause
Project-URL: Homepage, https://bitbucket.org/jjhw3/plotagain/
Keywords: save data,matplotlib,plt
Classifier: Framework :: Matplotlib
Classifier: Programming Language :: Python :: 3
Requires-Python: >=3.0
Description-Content-Type: text/markdown

# PlotAgain

This package provides a simple contextmanager which automatically saves data used to make matplotlib plots and automatically generates a script to reproduce the plots at a later stage.

## Install

`pip install plotagain`

## Basic Usage

A SavePlotContext contextmanager can be used in conjunction with a *with* block as shown below. 

```python
from plotagain import SavePlotContext


with SavePlotContext("./data-save-dir", locals(), overwrite=True) as spc:
    spc.plot( ... )
    ...
    spc.title( ... )
    spc.show()
```

The resulting context manager `spc` support all matplotlib.pyplot function calls.
All calls made to `spc` are passed onto matplotlib.pyplot as-is and recorded.
The data save directory is dynamically created if it does not exist and is populated with pickle files containing the data used to make the plot.
The automatically generated `make_plot.py` script contains code which reloads all data and contains the necessary calls to matplotlib.pyplot used to create the original plot.
This script may then be modified to adjust the plot.
If the data save directory is not empty, an exception is raised unless `overwrite=True`.
Passing in `locals()` allows SavePlotContext to infer the original variable names for the objects used for re-use in `make_plot.py`.

Note that only calls directly to spc are recorded, if you use figures and axes directly, you're on your own for now.
It is suggested that plots are kept simple in the scope of the code which generates them.
Once the individual plots have been generated it is easy to move around data and edit the autogenerated scripts to merge the individual plots using subplots.

## Example Usage

```python
import numpy as np
import matplotlib.pyplot as plt
from plotagain import SavePlotContext


x_data = np.linspace(0, 2 * np.pi, 1000)
y_data = np.sin(x_data)
with SavePlotContext("./data-save-dir", locals(), overwrite=True) as spc:
    spc.plot(x_data, y_data, c='k', label='sin')
    spc.plot(x_data, np.cos(x_data), c='b', label='cos')
    spc.xlabel('xaxis')
    spc.ylabel('yaxis')
    spc.title('Title')
    spc.legend()
    spc.savefig('plot.pdf')
    spc.show()
```

This will generate the following directory structure

```
data-save-dir/
  make_plot.py
  unnamed_arg.pkl
  x_data.pkl
  y_data.pkl
```

The contents of `make_plot.py` would be

```python
import numpy as np
import matplotlib.pyplot as plt
from plotagain import load_pickle


x_data = load_pickle('x_data.pkl')
y_data = load_pickle('y_data.pkl')
unnamed_arg = load_pickle('unnamed_arg.pkl')

plt.plot(x_data, y_data, c='k', label='sin')
plt.plot(x_data, unnamed_arg, c='b', label='cos')
plt.xlabel('xaxis')
plt.ylabel('yaxis')
plt.title('Title')
plt.legend()
plt.savefig('plot.pdf')
plt.show()
```

which should perfectly reproduce the original plot.
Note that the variable names `x_data` and `y_data` are arbitrary and were inferred by the SavePlotContext instance using the locals() dict.
Good code structure will always allow SavePlotContext to infer the correct variable names.
Anonymous arguments, such as the y data of the `np.cos` plot are simply saved and named "unnamed_arg" with an integer postfix if multiple unnamed arguments are present.

## Misc

Please report any bugs or feature requests to `jero.wilkinson@gmail.com`.
