Adding documentation

[NicolasMICAUX]

We should keep trace of technical choice made, and it would be nice to have some explanations of the functioning of the lib.
Probably in the wiki?

[NicolasMICAUX]

A few loose pieces of information:
The code is compatible with python 2.7, and python3.5+ (so 99% of python users). It should stay as such as much as possible.

The main algorithm in _search_object is a BFS.
We keep the information about which property/method of the object we are currently searching in with a Path of Node.

A Node has a name (useful to represent it in the search results), a depth and a pointer to the "object"=python entity it represents.
A Path is not a list of Node, it only stores the last node, and the names of all the nodes encountered. It is useless to store all the nodes a priori.

The BFS function calls a function is_in that is meant to check if a query is in an "object", whatever this object type is (number, string, standard python object). For the moment, it is really basic, but could be more complex in the future.
is_in returns False if no match, but if there is a match, it returns a SearchMatch object, useful for printing the match result afterwards.

A SearchMatch basically rpresents the place where the match has been found in the object. This enables a meaningful search result print.

A SearchResult is a "clean" object, that is used to print search results, but also and above all that could be returned bu searchin if someone wants raw results, not just print.

[NicolasMICAUX]

This wiki page should overall talk about the "High-level overview of the design and architecture of Search'In"