add option to display the subtree that holds the most tokens

for the token utility:
1. add an option to display the tree with relative tokens for each file
2. add an option to filter the tree for the top n nodes with most tokens

as of now, the `tokens` utility is a separate utility:

```
gpt-copy . | tokens
```

but to do what is described above we will need to add an option to the gpt-copy itself

```
gpt-copy . --tokens
```

possibly related to #6 