Merge pull request #203 from jlaura/master

%% Cell type:code id: tags:

``` python

import pandas as pd

```

%% Cell type:code id: tags:

``` python

df = pd.DataFrame([[1, 2], [3, 4]], columns=list('AB'))

df

```

%% Output

       A  B

    0  1  2

    1  3  4

%% Cell type:code id: tags:

``` python

# This is super slow lots of times

```

%% Cell type:code id: tags:

``` python

df.loc[2] = [4,5]

df

```

%% Output

       A  B

    0  1  2

    1  3  4

    2  4  5

%% Cell type:code id: tags:

``` python

# Build a new dataframe and then bulk append to the old one.

```

%% Cell type:code id: tags:

``` python

newrows = [[1,2] for i in range(5)] # This could be a dict of lists as well.

df2 = pd.DataFrame(newrows, columns=list('AB'))

df2

```

%% Output

       A  B

    0  1  2

    1  1  2

    2  1  2

    3  1  2

    4  1  2

%% Cell type:code id: tags:

``` python

df.append(df2)

```

%% Output

       A  B

    0  1  2

    1  3  4

    2  4  5

    0  1  2

    1  1  2

    2  1  2

    3  1  2

    4  1  2

matplotlib

numpy

scipy

nbsphinx

   What is AutoCNet? <introduction>

   Installing AutoCNet <installation>

   Getting Started with AutoCNet <tutorials/index>

%% Cell type:markdown id: tags:

# Creating the CandidateGraph Object

%% Cell type:code id: tags:

``` python

import os

import sys

sys.path.insert(0, os.path.abspath('/data/autocnet'))

from autocnet import CandidateGraph

%pylab inline

figsize(12,4)

```

%% Output

    Populating the interactive namespace from numpy and matplotlib

    /home/jlaura/anaconda3/envs/autocnet/lib/python3.5/site-packages/IPython/core/magics/pylab.py:161: UserWarning: pylab import has clobbered these variables: ['shape']

    `%matplotlib` prevents importing * from pylab and numpy

      "\n`%matplotlib` prevents importing * from pylab and numpy"

%% Cell type:markdown id: tags:

## Which Apollo Pan images?

The first question to ask is which images to use.  Without good geospatial information, using 5, ~2GB each, raw images is not practical for a number of reasons (memory constraints, lack of a priori information to constrain the search space, ambiguity in overlap requiring human intervention).  The JP2000 option is attractive (and viable), but image sizes of >6GB are also concerning from a memory perspective.  Therefore, the highest resolution `png` images were selected.  These were then cropped in [GIMP](https://www.gimp.org) to the extent of the valid data.  The scans are not perfectly square resulting in some black border remaining at some corners.  This should not be an issue.

%% Cell type:code id: tags:

``` python

a = 'AS15-P-0111_CENTER_LRG_CROPPED.png'

b = 'AS15-P-0112_CENTER_LRG_CROPPED.png'

adj = {a:[b],

       b:[a]}

cg = CandidateGraph.from_adjacency(adj)

```

%% Cell type:code id: tags:

``` python

cg.plot(labels=True)

```

%% Output

    <matplotlib.axes._subplots.AxesSubplot at 0x7fe16f80df60>

%% Cell type:markdown id: tags:

## Graph

About the graph

## Nodes

About the nodes

## Edges

About the edges

%% Cell type:code id: tags:

``` python

```