##// END OF EJS Templates
More simplifications due to promises
More simplifications due to promises

File last commit:

r13353:0ca701d5
r18888:96cc9663
Show More
rwbase.py
165 lines | 5.9 KiB | text/x-python | PythonLexer
"""Base classes and utilities for readers and writers.
Authors:
* Brian Granger
"""
#-----------------------------------------------------------------------------
# Copyright (C) 2008-2011 The IPython Development Team
#
# Distributed under the terms of the BSD License. The full license is in
# the file COPYING, distributed as part of this software.
#-----------------------------------------------------------------------------
#-----------------------------------------------------------------------------
# Imports
#-----------------------------------------------------------------------------
from base64 import encodestring, decodestring
import pprint
from IPython.utils.py3compat import str_to_bytes, unicode_type, string_types
#-----------------------------------------------------------------------------
# Code
#-----------------------------------------------------------------------------
def restore_bytes(nb):
"""Restore bytes of image data from unicode-only formats.
Base64 encoding is handled elsewhere. Bytes objects in the notebook are
always b64-encoded. We DO NOT encode/decode around file formats.
"""
for ws in nb.worksheets:
for cell in ws.cells:
if cell.cell_type == 'code':
for output in cell.outputs:
if 'png' in output:
output.png = str_to_bytes(output.png, 'ascii')
if 'jpeg' in output:
output.jpeg = str_to_bytes(output.jpeg, 'ascii')
return nb
# output keys that are likely to have multiline values
_multiline_outputs = ['text', 'html', 'svg', 'latex', 'javascript', 'json']
def rejoin_lines(nb):
"""rejoin multiline text into strings
For reversing effects of ``split_lines(nb)``.
This only rejoins lines that have been split, so if text objects were not split
they will pass through unchanged.
Used when reading JSON files that may have been passed through split_lines.
"""
for ws in nb.worksheets:
for cell in ws.cells:
if cell.cell_type == 'code':
if 'input' in cell and isinstance(cell.input, list):
cell.input = u'\n'.join(cell.input)
for output in cell.outputs:
for key in _multiline_outputs:
item = output.get(key, None)
if isinstance(item, list):
output[key] = u'\n'.join(item)
else: # text cell
for key in ['source', 'rendered']:
item = cell.get(key, None)
if isinstance(item, list):
cell[key] = u'\n'.join(item)
return nb
def split_lines(nb):
"""split likely multiline text into lists of strings
For file output more friendly to line-based VCS. ``rejoin_lines(nb)`` will
reverse the effects of ``split_lines(nb)``.
Used when writing JSON files.
"""
for ws in nb.worksheets:
for cell in ws.cells:
if cell.cell_type == 'code':
if 'input' in cell and isinstance(cell.input, string_types):
cell.input = cell.input.splitlines()
for output in cell.outputs:
for key in _multiline_outputs:
item = output.get(key, None)
if isinstance(item, string_types):
output[key] = item.splitlines()
else: # text cell
for key in ['source', 'rendered']:
item = cell.get(key, None)
if isinstance(item, string_types):
cell[key] = item.splitlines()
return nb
# b64 encode/decode are never actually used, because all bytes objects in
# the notebook are already b64-encoded, and we don't need/want to double-encode
def base64_decode(nb):
"""Restore all bytes objects in the notebook from base64-encoded strings.
Note: This is never used
"""
for ws in nb.worksheets:
for cell in ws.cells:
if cell.cell_type == 'code':
for output in cell.outputs:
if 'png' in output:
if isinstance(output.png, unicode_type):
output.png = output.png.encode('ascii')
output.png = decodestring(output.png)
if 'jpeg' in output:
if isinstance(output.jpeg, unicode_type):
output.jpeg = output.jpeg.encode('ascii')
output.jpeg = decodestring(output.jpeg)
return nb
def base64_encode(nb):
"""Base64 encode all bytes objects in the notebook.
These will be b64-encoded unicode strings
Note: This is never used
"""
for ws in nb.worksheets:
for cell in ws.cells:
if cell.cell_type == 'code':
for output in cell.outputs:
if 'png' in output:
output.png = encodestring(output.png).decode('ascii')
if 'jpeg' in output:
output.jpeg = encodestring(output.jpeg).decode('ascii')
return nb
class NotebookReader(object):
"""A class for reading notebooks."""
def reads(self, s, **kwargs):
"""Read a notebook from a string."""
raise NotImplementedError("loads must be implemented in a subclass")
def read(self, fp, **kwargs):
"""Read a notebook from a file like object"""
return self.read(fp.read(), **kwargs)
class NotebookWriter(object):
"""A class for writing notebooks."""
def writes(self, nb, **kwargs):
"""Write a notebook to a string."""
raise NotImplementedError("loads must be implemented in a subclass")
def write(self, nb, fp, **kwargs):
"""Write a notebook to a file like object"""
return fp.write(self.writes(nb,**kwargs))