added __getitem__ method

[eugene-eeo]

Allows you to slice the tables:

table.insert_multiple([{'int': k} for k in range(1, 3)])

assert table[0] == [{'int': 1}]
assert table[1:2] == [{'int': 2}]

[msiemens]

With this PR db[...] is an alias for db.all()[...]. This makes sense as len(db) is a shorthand for len(db.all()). But I'm currently thinking about two things:

1. The order of db.all(), which relies upon dict.values(), is not stable between modifications.

Keys and values are listed in an arbitrary order which is non-random, varies across Python implementations, and depends on the dictionary's history of insertions and deletions. (Python docs)

2. We could as well make db[1] make a shorthand for db.get(id=1) and slicing (db[1:3]) respectively. But that would only work as long as our IDs are numeric and somewhat sorted.

I'd propose to don't add ambiguity and stick to typing a few characters more.

[eugene-eeo]

What about implementing an index, ie db.index(*fields)? That way we can guarantee order.

[msiemens]

What would db.index(*fields) do?

[eugene-eeo]

Using tuple comparison, indexes all of the documents within a table based on the given fields. For example a very simple case would be:

table.insert_multiple([{'x': 1, 'y': 1}, {'x': 1, 'y': 2}])
idx = table.index('x', 'y')
assert idx[0] == {'x': 1, 'y': 1}
assert idx[1] == {'x': 1, 'y': 2}

[msiemens]

I've had that idea, too. Maybe we could make that an extension?

idx = tinyindex(db, 'x', 'y')
...

[eugene-eeo]

Sounds good to me!

[eugene-eeo]

Implemented said extension: https://github.com/eugene-eeo/tinyindex