Tabula-py

Changes in this Release

- Support 'dtype' in pandas_options with multiple_tables 219

- Download a remote file into a temporary directory 210

- Fix empty table handling when extracted table data is empty 208

- Fix a multi-threading issue with` io.BytesIO` 205

Correct duplicate column name handling with multiple_tables=True 202

This is major upgrade including backward incompatible changes. See detail in 200

- `read_pdf()` returns a list of DataFrame even if there is a table in a PDF without setting `output_format`
- `multiple_tables` becomes `True` by default
- renaming `wrapper` module to `io`
- stop using kwargs for `build_options` to avoid unexpected arguments
- add validation for `output_format` option