You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.

225 lines
6.8 KiB

import os
import shutil
import glob
import itertools
from . import files
from .paper import PaperInRepo, NoDocumentFile
from . import configs
BASE_FILE = 'papers.yaml'
BIB_DIR = 'bibdata'
META_DIR = 'meta'
DOC_DIR = 'doc'
class CiteKeyCollision(Exception):
pass
class InvalidReference(Exception):
pass
class Repository(object):
def __init__(self, config, load = True):
"""Initialize the repository.
:param load: if load is True, load the repository from disk,
from path config.papers_dir.
"""
self.config = config
self.citekeys = []
if load:
self.load()
# @classmethod
# def from_directory(cls, config, papersdir=None):
# repo = cls(config)
# if papersdir is None:
# papersdir = config.papers_dir
# repo.papersdir = files.clean_path(papersdir)
# repo.load()
# return repo
def __contains__(self, citekey):
"""Allows to use 'if citekey in repo' pattern"""
return citekey in self.citekeys
def __len__(self):
return len(self.citekeys)
# load, save repo
def _init_dirs(self, autodoc = True):
"""Create, if necessary, the repository directories.
Should only be called by load or save.
"""
self.bib_dir = files.clean_path(self.config.papers_dir, BIB_DIR)
self.meta_dir = files.clean_path(self.config.papers_dir, META_DIR)
if self.config.doc_dir == 'doc':
self.doc_dir = files.clean_path(self.config.papers_dir, DOC_DIR)
else:
self.doc_dir = files.clean_path(self.config.doc_dir)
self.cfg_path = files.clean_path(self.config.papers_dir, 'papers.yaml')
for d in [self.bib_dir, self.meta_dir, self.doc_dir]:
if not os.path.exists(d):
os.makedirs(d)
def load(self):
"""Load the repository, creating dirs if necessary"""
self._init_dirs()
repo_config = files.read_yamlfile(self.cfg_path)
self.citekeys = repo_config['citekeys']
def save(self):
"""Save the repo, creating dirs if necessary"""
self._init_dirs()
repo_cfg = {'citekeys': self.citekeys}
files.write_yamlfile(self.cfg_path, repo_cfg)
# reference
def ref2citekey(self, ref):
"""Tries to get citekey from given reference.
Ref can be a citekey or a number.
"""
if ref in self.citekeys:
return ref
else:
try:
return self.citekeys[int(ref)]
except (IndexError, ValueError):
raise InvalidReference
# papers
def all_papers(self):
for key in self.citekeys:
yield self.get_paper(key)
def get_paper(self, citekey):
"""Load a paper by its citekey from disk, if necessary."""
return PaperInRepo.load(self, self._bibfile(citekey),
self._metafile(citekey))
# add, remove papers
def add_paper(self, p, overwrite = False):
if p.citekey is None: # TODO also test if citekey is valid
raise ValueError("Invalid citekey: {}.".format(p.citekey))
if not overwrite and p.citekey in self:
raise CiteKeyCollision('citekey {} already in use'.format(
p.citekey))
self.citekeys.append(p.citekey)
self.save_paper(p)
self.save()
# TODO change to logging system (17/12/2012)
print('Added: {}'.format(p.citekey))
return p
def rename_paper(self, paper, new_citekey, overwrite=False):
"""Modify the citekey of a paper, and propagate changes to disk"""
if paper.citekey not in self:
raise ValueError(
'paper {} not in repository'.format(paper.citekey))
if (not overwrite and paper.citekey != new_citekey
and new_citekey in self):
raise CiteKeyCollision('citekey {} already in use'.format(
new_citekey))
self.remove_paper(paper.citekey, remove_doc = False)
old_citekey = paper.citekey
paper.citekey = new_citekey
self.add_paper(paper, overwrite = overwrite)
self._move_doc(old_citekey, paper)
def _move_doc(self, old_citekey, paper):
"""Fragile. Make more robust"""
try:
old_docfile = self.find_document(old_citekey)
ext = os.path.splitext(old_docfile)[1]
new_docfile = os.path.join(self.doc_dir, paper.citekey + ext)
shutil.move(old_docfile, new_docfile)
paper.set_external_document(new_docfile)
except NoDocumentFile:
pass
def _bibfile(self, citekey):
return os.path.join(self.bib_dir, citekey + '.bibyaml')
def _metafile(self, citekey):
return os.path.join(self.meta_dir, citekey + '.meta')
def remove_paper(self, citekey, remove_doc = True):
paper = self.get_paper(citekey)
self.citekeys.remove(citekey)
os.remove(self._metafile(citekey))
os.remove(self._bibfile(citekey))
# Eventually remove associated document
if remove_doc:
try:
path = paper.get_document_path_in_repo()
os.remove(path)
except NoDocumentFile:
pass
self.save()
def save_paper(self, paper):
if not paper.citekey in self:
raise ValueError('Paper not in repository, first add it.')
paper.save(self._bibfile(paper.citekey),
self._metafile(paper.citekey))
def generate_citekey(self, paper, citekey=None):
"""Create a unique citekey for the given paper."""
if citekey is None:
citekey = paper.generate_citekey()
for n in itertools.count():
if not citekey + _base27(n) in self.citekeys:
return citekey + _base27(n)
def find_document(self, citekey):
found = glob.glob('{}/{}.*'.format(self.doc_dir, citekey))
if found:
return found[0]
else:
raise NoDocumentFile
def import_document(self, citekey, doc_file):
if citekey not in self.citekeys:
raise ValueError("Unknown citekey: {}.".format(citekey))
else:
if not os.path.isfile(doc_file):
raise ValueError("No file {} found.".format(doc_file))
ext = os.path.splitext(doc_file)[1]
new_doc_file = os.path.join(self.doc_dir, citekey + ext)
shutil.copy(doc_file, new_doc_file)
def get_tags(self):
tags = set()
for p in self.all_papers():
tags = tags.union(p.tags)
return tags
def _base27(n):
return _base27((n-1) // 26) + chr(97+((n-1)% 26)) if n else ''
def _base(num, b):
q, r = divmod(num - 1, len(b))
return _base(q, b) + b[r] if num else ''