I met bug when my csv header began with comma

### situation
When I upload a file like this,the first header is '' since the line begins with comma.(See the up left empty cell)

Then DataPusher raise error as `The data was invalid (for example: a numeric value is out of range or was inserted into a text field). `
### debug
The directly reason of the error is `headers_dicts` in DataPusher:
I log the headers_dicts then it shows:
```
[
{
'type': u'numeric',
'id': u'name'
},
{
'type': u'text',
'id': u'total'
},
{
'type': u'numeric',
'id': u'male_share'
},
{
'type': u'numeric',
'id': u'female_share'
},
{
'type': u'numeric',
'id': u'gap'
}
]
```
Obviously there is some shifting thing cuz the `name` should be `text` and the `total` should be `numeric`

I read the code :
1. `offset, headers = messytables.headers_guess(row_set.sample)`
2. `types = messytables.type_guess(row_set.sample, types=TYPES, strict=False)`
The first empty header would be stored as u'',It's totally fine.
Here len(types)==len(headers)
3. `headers = [header.strip() for header in headers if header.strip()]`
Then this line dropped the u'',but didn't drop correspond type, that is ,len(types)== len(headers) -1

4. `headers_dicts = [dict(id=field[0], type=TYPE_MAPPING[str(field[1])]) for field in zip(headers, types)]`
Finally in this line we get a wrong headers_dicts, `zip`func would not raise error even if parameters have different length.