fables 0.0.1

(F)ile T(ables)

Fables - (F)ile T(ables)

For users:

Parse all of the tabular data in the input file:

import fables

for parse_result in fables.parse('myfile.zip'):
    for table in parse_result.tables:
        print(table.name)
        print(table.df.head())
    for error in parse_result.errors:
        print(error.message)

Inspect the contents of the input file:

node = fables.detect('myfile.zip')
print(node.name)
print(node.mimetype)

for child in node.children:
    print(child.name)
    print(child.mimetype)

Note if you've already discovered the input tree from detect(), you can pass it into parse():

parse_results = parse(tree=node)

Handling encrypted zip, xlsx, and xls files:

node = fables.detect('encrypted.xlsx')
assert node.encrypted
node.add_password('encrypted.xlsx', 'fables')
assert not node.encrypted

You can also supply a passwords dictionary (filename -> password) into detect and parse:

node = fables.detect(
    'encrypted.zip',
    passwords={
        'encrypted.zip': 'fables',
        # an encrypted file inside the zip
        'encrypted.xlsx': 'foobles',
    }
)
# and/or parse
parse_results = fables.parse(
    'sub_dir',
    passwords={
        'sub_dir/encrypted.xlsx': 'fables',
        'sub_dir/encrypted.xls': 'foobles',
    },
)

Installation

The python library python-magic requires additional system dependencies. There are installation instructions there, but here are recommended routes to try:

• on OSX: brew install libmagic.

• on Windows: this pip install python-magic-bin will install a built version using ctypes to access the libmagic file type identification library.

Then pip install -r requirements.txt should do the trick.

For contributors:

Tests

• all tests: pytest
  • coverage: pytest --cov=fables tests
• integration: pytest tests/integration
  • coverage: pytest --cov=fables tests/integration
• unit: pytest tests/unit
  • coverage: pytest --cov=fables tests/unit

Note all the coverage statistics are for statements.

Type checking with mypy

• mypy fables

Linting

• We enforce flake8:
  • flake8 .

Run test, type checking, and linter all at once

• nox