Class Dvi

lib/matplotlib/dviread.py:219–628

A reader for a dvi ("device-independent") file, as produced by TeX. The current implementation can only iterate through pages in order, and does not even attempt to verify the postamble. This class can be used as a context manager to close the underlying file upon exit. Pages

class Dvi:
    """
    A reader for a dvi ("device-independent") file, as produced by TeX.

    The current implementation can only iterate through pages in order,
    and does not even attempt to verify the postamble.

    This class can be used as a context manager to close the underlying
    file upon exit. Pages can be read via iteration. Here is an overly
    simple way to extract text without trying to detect whitespace::

        >>> with matplotlib.dviread.Dvi('input.dvi', 72) as dvi:
        ...     for page in dvi:
        ...         print(''.join(chr(t.glyph) for t in page.text))
    """
    # dispatch table
    _dtable = [None] * 256
    _dispatch = partial(_dispatch, _dtable)

    def __init__(self, filename, dpi):
        """
        Read the data from the file named *filename* and convert
        TeX's internal units to units of *dpi* per inch.
        *dpi* only sets the units and does not limit the resolution.
        Use None to return TeX's internal units.
        """
        _log.debug('Dvi: %s', filename)
        self.file = open(filename, 'rb')
        self.dpi = dpi
        self.fonts = {}
        self.state = _dvistate.pre
        self._missing_font = None

    def __enter__(self):
        """Context manager enter method, does nothing."""
        return self

    def __exit__(self, etype, evalue, etrace):
        """
        Context manager exit method, closes the underlying file if it is open.
        """
        self.close()

    def __iter__(self):
        """
        Iterate through the pages of the file.

        Yields
        ------
        Page
            Details of all the text and box objects on the page.
            The Page tuple contains lists of Text and Box tuples and
            the page dimensions, and the Text and Box tuples contain
            coordinates transformed into a standard Cartesian
            coordinate system at the dpi value given when initializing.
            The coordinates are floating point numbers, but otherwise
            precision is not lost and coordinate values are not clipped to
            integers.