misanthropicbit / bibpy Goto Github PK

Support for Python 2.7 will be dropped 1 January, 2020. The plan is to only support Python 3.4+ thereafter.

Such a codec would translate LaTeX symbol notation to Python unicode and vice versa. Observe:

>>> from bibpy.codec import decode, encode
>>> decode('Kurt G{\"o}del')
'Kurt Gödel'
>>> encode(_)
'Kurt G{\"o}del'

A unicode to LaTeX mapping exists here.

The lexers/parsers throw away information about line number positions. This is problematic for tools like bibformat which should maintain the same relative positions of comments and entries in files as those given in the input. This is sensible for comments that are given for a specific entry.

However, it is more complicated than that. What if a comment appears two entries that both change position after being processed by bibformat? We cannot sort by line number as this would undo the reordering. Instead we could store a relative position, e.g. a comment with position 1 should be placed after the first entry has been output. Since we cannot know if a comment belongs to a certain entry, the least we can do is keep its position in the file.

Retrieving the value of a missing field with default value fails.

import bibpy

ref = """
@misc{reddit_2022,
title  = {R/reinforcementlearning. Are exploration and credit assignment independent? What's your opinion?},
url    = {https://www.reddit.com/r/reinforcementlearning/comments/vobvly/are_exploration_and_credit_assignment_independent/},
ournal = {reddit},
year   = {2022},
month  = {Jun}
}
"""

references = bibpy.read_string(ref)
for ref in references:
    author = ref.get("author", default="")
    assert author == "", "author is not an empty string, but {}".format(str(author))