This is an example of the `fixed` paring and serialization library. In this
example we read a custom, column delimited data file into memory, run a computation
on that data and then write a result set out in a different, column delimited data file.

The basic job flow is in `main.rs`. The specifics of the computation are contained in
`alg.rs` but are not relevant to the example use of the serialization library.

## Input format

There are two types of records in the data file: person records and relation records.
The first column of the data file indicates the record type. A `P` indicates a 
person record and an `R` indicates a relation record.

### Person records

A person record is laid out as follows:

```text
P003    Charles V   1500-1558 -- example record
PAAA BBBBBBBBBB CCCCDDDD-EEEE -- coding for explanation below
```

- `P` is a literal `P` and indicates a person record
- `AAA` is a three digit, zero padded, right aligned number which serves as a
  record identifier
- `B...B` is the name of the person, right aligned
- `CCCC` is the regnal number (e.g., *V* in *Charles V*). It is left aligned and
  four columns, supporting roman numerals up to XXII.
- Life span is interpreted as two separate columns with a `-` between them.
  - `DDDD` is the year of birth
  - `EEEE` is the year of death

#### Note

In order to separate the birth and death years we need to disable **strict* mode.
Otherwise the non-whitespace character between those two columns would trigger
an error.

### Relation Records

A relation record describes a relationship between two person records.

```text
R PC 1  5   -- example record
R XX YYYZZZ -- coding for explanation below
```

- `R` is a literal `R` and indicates a relation record
- `XX` a two letter code indicating the relationship type. It has two legal values:
  - `SP` a marriage relationship
  - `PC` a parent-child relationship
- `YYY` the left member of the relationship (partner 1 or parent) as a left aligned, up to three digit number
- `ZZZ` the right member of the relationship (partner 2 or child) as a left aligned, up to three digit number

## Output Format

The output format is comparatively simple. It consists of a six column floating
point number (holding the COI), followed by a space and the person's name, left
aligned. The serialization format is defined on the `OutputRecord` struct.

## References

The data file is a custom genealogy format depicting the
[genealogy Spanish branch of the Habsburg Dynasty](https://en.wikipedia.org/wiki/Habsburg_family_tree#Ancestors_of_Charles_II_of_Spain)
and derived from that Wikipedia page. Then we run a graph algorithm on the
resulting data structure to calculate the [COI](https://en.wikipedia.org/wiki/Coefficient_of_inbreeding)
for each of the people in that genealogy. Finally, we write an output file with
that calculated data when it's value is greater than zero. Note that the value
for Charles II is lower than is somewhat reported in the literature because we
only look back seven generations. See:
[Alvarez, Ceballos, and Quinteiro (2009)](https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0005174)
for a discussion of the effect of the look back window. The COI calculation is performed
using the "tabular method" by Wright (1922) as described in [Tier (1990)](https://hal.science/hal-00893856/document).