Support READ cards

[tjlaboss]

Closes #29

Where a READ card is found, recursively insert the contents of the target FILE=.... Other keywords (ECHO, NOECHO, ENCODE=..., and DECODE=...) are ignored.

---

Example:

testRead.imcnp:

Testing read card
C this is a comment
read echo file=testReadTarget.imcnp encode

1 so 0.5

mode n

testReadTarget.imcnp:

1 0 -1
read file=testReadRecursive.imcnp
c

testReadRecursive.imcnp:

2 0 +1  $ lowest level

This becomes:

Testing read card
C this is a comment
1 0 -1
2 0 +1  $ lowest level
c

1 so 0.5

mode n

And thereby:

<?xml version='1.0' encoding='utf-8'?>
<model>
  <materials>
  </materials>
  <geometry>
    <cell id="1" material="void" region="-1" universe="0"/>
    <cell id="2" material="void" region="1" universe="0"/>
    <surface coeffs="0.0 0.0 0.0 0.5" id="1" type="sphere"/>
  </geometry>
  <settings>
    <run_mode>eigenvalue</run_mode>
    <particles>100</particles>
    <batches>40</batches>
    <inactive>20</inactive>
    <source particle="neutron" strength="1.0" type="independent">
      <space type="point">
        <parameters>0.0 0.0 0.0</parameters>
      </space>
    </source>
    <output>
      <summary>true</summary>
    </output>
  </settings>
  </model>

[MicahGale]

Not going to be perfectly robust in that it may be overly generous with what counts as a read, but I think it's probably robust enough.

[paulromano]

@tjlaboss Thanks for this addition! I've started to add proper tests on this repo so it would be prudent to have a test for this as well. Can you put together a test?

[MicahGale]

@tjlaboss want any help with tests?

[tjlaboss]

Added tests. I included a new tests/inputs/ directory. It necessarily breaks from the convention of the other tests including the MCNP decks as strings because the target of read file=... must be a file path.

[MicahGale]

This looks good to me.

What's your patch coverage though?

[tjlaboss]

@paulromano, are these tests acceptable?

[paulromano]

Just made two small updates: 1) use pathlib over os.path, and 2) make the regex multiline and add ^, which makes it about 4× faster in my testing. Good to go now!