def importbib(db, bibfile, tags=[], overwrite=False):
errors = []
sources = Sources()
for entry in sorted(Bibtex(bibfile), key=lambda entry: entry.key):
print >> sys.stderr, entry.key
try:
docs = []
# check for doc with this bibkey
bdoc = db.doc_for_bib(entry.key)
if bdoc:
docs.append(bdoc)
# check for known sids
for source in sources.scan_bibentry(entry):
sdoc = db.doc_for_source(source.sid)
# FIXME: why can't we match docs in list?
if sdoc and sdoc.docid not in [doc.docid for doc in docs]:
docs.append(sdoc)
if len(docs) == 0:
doc = Document(db)
elif len(docs) > 0:
if len(docs) > 1:
print >> sys.stderr, " Multiple distinct docs found for entry. Using first found."
doc = docs[0]
print >> sys.stderr, " Updating id:%d..." % (doc.docid)
doc.add_bibentry(entry)
filepath = entry.get_file()
if filepath:
print >> sys.stderr, " Adding file: %s" % filepath
doc.add_file(filepath)
doc.add_tags(tags)
doc.sync()
except BibtexError as e:
print >> sys.stderr, " Error processing entry %s: %s" % (
entry.key, e)
print >> sys.stderr
errors.append(entry.key)
if errors:
print >> sys.stderr
print >> sys.stderr, "Failed to import %d" % (len(errors)),
if len(errors) == 1:
print >> sys.stderr, "entry",
else:
print >> sys.stderr, "entries",
print >> sys.stderr, "from bibtex:"
for error in errors:
print >> sys.stderr, " %s" % (error)
sys.exit(1)
else:
sys.exit(0)
class Document():
"""Represents a single Xapers document."""
def __init__(self, db, xapian_doc=None, docid=None):
# Xapers db
self.db = db
# if Xapian doc provided, initiate for that document
if xapian_doc:
self.xapian_doc = xapian_doc
self.docid = xapian_doc.get_docid()
# else, create a new empty document
# document won't be added to database until sync is called
else:
self.xapian_doc = xapian.Document()
# use specified docid if provided
if docid:
if docid in self.db:
raise DocumentError('Document already exists for id %d.' %
docid)
self.docid = docid
else:
self.docid = self.db._generate_docid()
self._add_term(self.db._find_prefix('id'), self.docid)
# specify a directory in the Xapers root for document data
self.docdir = os.path.join(self.db.root, '%010d' % self.docid)
self.bibentry = None
self._infiles = {}
def get_docid(self):
"""Return document id of document."""
return self.docid
def _make_docdir(self):
if os.path.exists(self.docdir):
if not os.path.isdir(self.docdir):
raise DocumentError(
'File exists at intended docdir location: %s' %
self.docdir)
else:
os.makedirs(self.docdir)
def _write_files(self):
for name, data in self._infiles.iteritems():
path = os.path.join(self.docdir, name)
with open(path, 'w') as f:
f.write(data)
def _write_bibfile(self):
bibpath = self.get_bibpath()
# reload bibtex only if we have new files
paths = self.get_fullpaths()
if paths:
self._load_bib()
if self.bibentry:
# we put only the first file in the bibtex
# FIXME: does jabref/mendeley spec allow for multiple files?
if paths and not self.bibentry.get_file():
self.bibentry.set_file(paths[0])
self.bibentry.to_file(bibpath)
def _write_tagfile(self):
with open(os.path.join(self.docdir, 'tags'), 'w') as f:
for tag in self.get_tags():
f.write(tag)
f.write('\n')
def _rm_docdir(self):
if os.path.exists(self.docdir) and os.path.isdir(self.docdir):
shutil.rmtree(self.docdir)
def sync(self):
"""Sync document to database."""
# FIXME: add value for modification time
# FIXME: catch db not writable errors
try:
self._make_docdir()
self._write_files()
self._write_bibfile()
self._write_tagfile()
self.db.replace_document(self.docid, self.xapian_doc)
except:
self._rm_docdir()
raise
def purge(self):
"""Purge document from database and root."""
# FIXME: catch db not writable errors
try:
self.db.delete_document(self.docid)
except xapian.DocNotFoundError:
pass
self._rm_docdir()
self.docid = None
########################################
# internal stuff
# add an individual prefix'd term for the document
def _add_term(self, prefix, value):
term = '%s%s' % (prefix, value)
self.xapian_doc.add_term(term)
# remove an individual prefix'd term for the document
def _remove_term(self, prefix, value):
term = '%s%s' % (prefix, value)
try:
self.xapian_doc.remove_term(term)
except xapian.InvalidArgumentError:
pass
# Parse 'text' and add a term to 'message' for each parsed
# word. Each term will be added both prefixed (if prefix is not
# None) and non-prefixed.
# http://xapian.org/docs/bindings/python/
# http://xapian.org/docs/quickstart.html
# http://www.flax.co.uk/blog/2009/04/02/xapian-search-architecture/
def _gen_terms(self, prefix, text):
term_gen = self.db.term_gen
term_gen.set_document(self.xapian_doc)
if prefix:
term_gen.index_text(text, 1, prefix)
term_gen.index_text(text)
# return a list of terms for prefix
def _term_iter(self, prefix=None):
term_iter = iter(self.xapian_doc)
if prefix:
plen = len(prefix)
term = term_iter.skip_to(prefix)
if not term.term.startswith(prefix):
return
yield term.term[plen:]
for term in term_iter:
if prefix:
if not term.term.startswith(prefix):
break
yield term.term[plen:]
else:
yield term.term
def term_iter(self, name=None):
"""Iterator over all terms in the document.
If a prefix is provided, will iterate over only the prefixed
terms, and the prefix will be removed from the returned terms.
"""
prefix = None
if name:
prefix = self.db._find_prefix(name)
if not prefix:
prefix = name
return self._term_iter(prefix)
# set the data object for the document
def _set_data(self, text):
self.xapian_doc.set_data(text)
def get_data(self):
"""Get data object for document."""
return self.xapian_doc.get_data()
########################################
# files
def add_file_data(self, name, data):
"""Add a file data to document.
'name' is the name of the file, 'data is the file data.
File will not copied in to docdir until sync().
"""
# FIXME: set mime type term
# parse the file data into text
text = parse_data(data)
# generate terms from the text
self._gen_terms(None, text)
# set data to be text sample
# FIXME: is this the right thing to put in the data?
summary = text[0:997] + '...'
self._set_data(summary)
# FIXME: should files be renamed to something generic (0.pdf)?
prefix = self.db._find_prefix('file')
self._add_term(prefix, name)
# add it to the cache to be written at sync()
self._infiles[name] = data
def add_file(self, infile):
"""Add a file to document.
Added file will have the same name.
File will not copied in to docdir until sync().
"""
with open(infile, 'r') as f:
data = f.read()
name = os.path.basename(infile)
self.add_file_data(name, data)
def get_files(self):
"""Return files associated with document."""
return list(self.term_iter('file'))
def get_fullpaths(self):
"""Return fullpaths of files associated with document."""
list = []
for path in self.get_files():
# FIXME: this is a hack for old path specifications that
# included the docdir
path = os.path.basename(path)
list.append(os.path.join(self.docdir, path))
return list
########################################
# SOURCES
def _purge_sources_prefix(self, source):
# purge all terms for a given source prefix
prefix = self.db._make_source_prefix(source)
for i in self._term_iter(prefix):
self._remove_term(prefix, i)
self._remove_term(self.db._find_prefix('source'), source)
def add_sid(self, sid):
"""Add source sid to document."""
source, oid = sid.split(':', 1)
source = source.lower()
# remove any existing terms for this source
self._purge_sources_prefix(source)
# add a term for the source
self._add_term(self.db._find_prefix('source'), source)
# add a term for the sid, with source as prefix
self._add_term(self.db._make_source_prefix(source), oid)
def get_sids(self):
"""Return a list of sids for document."""
sids = []
for source in self.term_iter('source'):
for oid in self._term_iter(self.db._make_source_prefix(source)):
sids.append('%s:%s' % (source, oid))
return sids
# TAGS
def add_tags(self, tags):
"""Add tags from list to document."""
prefix = self.db._find_prefix('tag')
for tag in tags:
self._add_term(prefix, tag)
def get_tags(self):
"""Return a list of tags associated with document."""
return list(self.term_iter('tag'))
def remove_tags(self, tags):
"""Remove tags from a document."""
prefix = self.db._find_prefix('tag')
for tag in tags:
self._remove_term(prefix, tag)
# TITLE
def _set_title(self, title):
pt = self.db._find_prefix('title')
for term in self._term_iter(pt):
self._remove_term(pt, term)
# FIXME: what's the clean way to get these prefixes?
for term in self._term_iter('ZS'):
self._remove_term('ZS', term)
self._gen_terms(pt, title)
# AUTHOR
def _set_authors(self, authors):
pa = self.db._find_prefix('author')
for term in self._term_iter(pa):
self._remove_term(pa, term)
# FIXME: what's the clean way to get these prefixes?
for term in self._term_iter('ZA'):
self._remove_term('ZA', term)
self._gen_terms(pa, authors)
# YEAR
def _set_year(self, year):
# FIXME: what to do if year is not an int?
try:
year = int(year)
except ValueError:
pass
prefix = self.db._find_prefix('year')
for term in self._term_iter(prefix):
self._remove_term(prefix, year)
self._add_term(prefix, year)
facet = self.db._find_facet('year')
self.xapian_doc.add_value(facet, xapian.sortable_serialise(year))
########################################
# bibtex
def get_bibpath(self):
"""Return path to document bibtex file."""
return os.path.join(self.docdir, 'bibtex')
def _set_bibkey(self, key):
prefix = self.db._find_prefix('key')
for term in self._term_iter(prefix):
self._remove_term(prefix, term)
self._add_term(prefix, key)
def _index_bibentry(self, bibentry):
authors = bibentry.get_authors()
fields = bibentry.get_fields()
if 'title' in fields:
self._set_title(fields['title'])
if 'year' in fields:
self._set_year(fields['year'])
if authors:
# authors should be a list, so we make a single text string
# FIXME: better way to do this?
self._set_authors(' '.join(authors))
# add any sources in the bibtex
for source in Sources().scan_bibentry(bibentry):
self.add_sid(source.sid)
# FIXME: index 'keywords' field as regular terms
self._set_bibkey(bibentry.key)
def add_bibentry(self, bibentry):
"""Add bibentry object."""
self.bibentry = bibentry
self._index_bibentry(self.bibentry)
def add_bibtex(self, bibtex):
"""Add bibtex to document, as string or file path."""
self.add_bibentry(Bibtex(bibtex)[0])
def _load_bib(self):
if self.bibentry:
return
bibpath = self.get_bibpath()
if os.path.exists(bibpath):
self.bibentry = Bibtex(bibpath)[0]
def get_bibtex(self):
"""Get the bib for document as a bibtex string."""
bibpath = self.get_bibpath()
if os.path.exists(bibpath):
with open(bibpath, 'r') as f:
bibtex = f.read().decode('utf-8')
return bibtex.strip()
def get_bibdata(self):
self._load_bib()
if self.bibentry:
data = self.bibentry.get_fields()
data['authors'] = self.bibentry.get_authors()
return data
def update_from_bibtex(self):
"""Update document metadata from document bibtex."""
self._load_bib()
self._index_bibentry(self.bibentry)
########################################
def get_key(self):
self._load_bib()
if not self.bibentry:
return
return self.bibentry.key
def get_title(self):
"""Get the title from document bibtex."""
self._load_bib()
if not self.bibentry:
return
fields = self.bibentry.get_fields()
if 'title' in fields:
return fields['title']
def get_year(self):
"""Get the title from document bibtex."""
self._load_bib()
if not self.bibentry:
return
fields = self.bibentry.get_fields()
if 'year' in fields:
return fields['year']
def get_urls(self):
"""Get all URLs associated with document."""
sources = Sources()
urls = []
# get urls associated with known sources
for sid in self.get_sids():
urls.append(sources[sid].url)
# get urls from bibtex
self._load_bib()
if self.bibentry:
fields = self.bibentry.get_fields()
if 'url' in fields:
urls.append(fields['url'])
if 'adsurl' in fields:
urls.append(fields['adsurl'])
return urls
def _load_bib(self):
if self.bibentry:
return
bibpath = self.get_bibpath()
if os.path.exists(bibpath):
self.bibentry = Bibtex(bibpath)[0]
def add_bibtex(self, bibtex):
"""Add bibtex to document, as string or file path."""
self.add_bibentry(Bibtex(bibtex)[0])
class Document():
"""Represents a single Xapers document."""
def __init__(self, db, doc=None, docid=None):
# Xapers db
self.db = db
self.root = self.db.root
# if Xapian doc provided, initiate for that document
if doc:
self.doc = doc
self.docid = str(doc.get_docid())
# else, create a new empty document
# document won't be added to database until sync is called
else:
self.doc = xapian.Document()
# use specified docid if provided
if docid:
if self.db[docid]:
raise DocumentError('Document already exists for id %s.' % docid)
self.docid = docid
else:
self.docid = str(self.db._generate_docid())
self._add_term(self.db._find_prefix('id'), self.docid)
# specify a directory in the Xapers root for document data
self.docdir = os.path.join(self.root, '%010d' % int(self.docid))
#
self.bibentry = None
def get_docid(self):
"""Return document id of document."""
return self.docid
def _make_docdir(self):
if os.path.exists(self.docdir):
if not os.path.isdir(self.docdir):
raise DocumentError('File exists at intended docdir location: %s' % self.docdir)
else:
os.makedirs(self.docdir)
def _write_files(self):
if '_infiles' in dir(self):
for infile, outfile in self._infiles.iteritems():
try:
shutil.copyfile(infile, outfile)
except shutil.Error:
pass
def _write_bibfile(self):
bibpath = self.get_bibpath()
# reload bibtex only if we have new files
paths = self.get_fullpaths()
if paths:
self._load_bib()
if self.bibentry:
# we put only the first file in the bibtex
# FIXME: does jabref/mendeley spec allow for multiple files?
if paths and not self.bibentry.get_file():
self.bibentry.set_file(paths[0])
self.bibentry.to_file(bibpath)
def _write_tagfile(self):
with open(os.path.join(self.docdir, 'tags'), 'w') as f:
for tag in self.get_tags():
f.write(tag)
f.write('\n')
def _rm_docdir(self):
if os.path.exists(self.docdir) and os.path.isdir(self.docdir):
shutil.rmtree(self.docdir)
def sync(self):
"""Sync document to database."""
# FIXME: add value for modification time
# FIXME: catch db not writable errors
try:
self._make_docdir()
self._write_files()
self._write_bibfile()
self._write_tagfile()
self.db.replace_document(self.docid, self.doc)
except:
self._rm_docdir()
raise
def purge(self):
"""Purge document from database and root."""
# FIXME: catch db not writable errors
try:
self.db.delete_document(self.docid)
except xapian.DocNotFoundError:
pass
self._rm_docdir()
self.docid = None
########################################
# internal stuff
# add an individual prefix'd term for the document
def _add_term(self, prefix, value):
term = '%s%s' % (prefix, value)
self.doc.add_term(term)
# remove an individual prefix'd term for the document
def _remove_term(self, prefix, value):
term = '%s%s' % (prefix, value)
try:
self.doc.remove_term(term)
except xapian.InvalidArgumentError:
pass
# Parse 'text' and add a term to 'message' for each parsed
# word. Each term will be added both prefixed (if prefix_name is
# not NULL) and also non-prefixed).
# http://xapian.org/docs/bindings/python/
# http://xapian.org/docs/quickstart.html
# http://www.flax.co.uk/blog/2009/04/02/xapian-search-architecture/
def _gen_terms(self, prefix, text):
term_gen = self.db.term_gen
term_gen.set_document(self.doc)
if prefix:
term_gen.index_text(text, 1, prefix)
term_gen.index_text(text)
# return a list of terms for prefix
# FIXME: is this the fastest way to do this?
def _get_terms(self, prefix):
list = []
for term in self.doc:
if term.term.find(prefix.encode("utf-8")) == 0:
index = len(prefix)
list.append(term.term[index:])
return list
# set the data object for the document
def _set_data(self, text):
self.doc.set_data(text)
def get_data(self):
"""Get data object for document."""
return self.doc.get_data()
########################################
# files
# index file for the document
def _index_file(self, path):
text = parse_file(path)
self._gen_terms(None, text)
summary = text[0:997].translate(None, '\n') + '...'
return summary
def _add_path(self, path):
base, full = self.db._basename_for_path(path)
prefix = self.db._find_prefix('file')
self._add_term(prefix, base)
def _get_paths(self):
return self._get_terms(self.db._find_prefix('file'))
def get_fullpaths(self):
"""Return fullpaths associated with document."""
list = []
for path in self._get_paths():
# FIXME: this is a hack for old bad path specifications and should be removed
if path.find(self.root) == 0:
index = len(self.root) + 1
path = path[index:]
path = path.lstrip('/')
# FIXME
base, full = self.db._basename_for_path(path)
list.append(full)
return list
def add_file(self, infile):
"""Add a file to document.
File will not copied in to docdir until sync()."""
# FIXME: should load entire file into {name: file} to be
# written as file>docdir/name
# FIXME: set mime type term
summary = self._index_file(infile)
# set data to be text sample
# FIXME: is this the right thing to put in the data?
self._set_data(summary)
# FIXME: should files be renamed to something generic (0.pdf)?
outfile = os.path.join(self.docdir, os.path.basename(infile))
base, full = self.db._basename_for_path(outfile)
self._add_path(base)
# add it to the cache to be written at sync()
if '_infiles' not in dir(self):
self._infiles = {}
self._infiles[infile] = outfile
########################################
# SOURCES
def _purge_sources_prefix(self, source):
# purge all terms for a given source prefix
prefix = self.db._make_source_prefix(source)
for i in self._get_terms(prefix):
self._remove_term(prefix, i)
self._remove_term(self.db._find_prefix('source'), source)
def add_sid(self, sid):
"""Add source sid to document."""
source, oid = sid.split(':', 1)
source = source.lower()
# remove any existing terms for this source
self._purge_sources_prefix(source)
# add a term for the source
self._add_term(self.db._find_prefix('source'), source)
# add a term for the sid, with source as prefix
self._add_term(self.db._make_source_prefix(source), oid)
def get_sids(self):
"""Return a list of sids for document."""
sids = []
for source in self._get_terms(self.db._find_prefix('source')):
for oid in self._get_terms(self.db._make_source_prefix(source)):
sids.append('%s:%s' % (source, oid))
return sids
# BIBTEX KEYS
def get_keys(self):
"""Return a list of bibtex citation keys associated with document."""
prefix = self.db._find_prefix('key')
return self._get_terms(prefix)
# TAGS
def add_tags(self, tags):
"""Add tags from list to document."""
prefix = self.db._find_prefix('tag')
for tag in tags:
self._add_term(prefix, tag)
def get_tags(self):
"""Return a list of tags associated with document."""
prefix = self.db._find_prefix('tag')
return self._get_terms(prefix)
def remove_tags(self, tags):
"""Remove tags from a document."""
prefix = self.db._find_prefix('tag')
for tag in tags:
self._remove_term(prefix, tag)
# TITLE
def _set_title(self, title):
pt = self.db._find_prefix('title')
for term in self._get_terms(pt):
self._remove_term(pt, term)
# FIXME: what's the clean way to get these prefixes?
for term in self._get_terms('ZS'):
self._remove_term('ZS', term)
self._gen_terms(pt, title)
# AUTHOR
def _set_authors(self, authors):
pa = self.db._find_prefix('author')
for term in self._get_terms(pa):
self._remove_term(pa, term)
# FIXME: what's the clean way to get these prefixes?
for term in self._get_terms('ZA'):
self._remove_term('ZA', term)
self._gen_terms(pa, authors)
# YEAR
def _set_year(self, year):
# FIXME: this should be a value
pass
########################################
# bibtex
def get_bibpath(self):
"""Return path to document bibtex file."""
return os.path.join(self.docdir, 'bibtex')
def _set_bibkey(self, key):
prefix = self.db._find_prefix('key')
for term in self._get_terms(prefix):
self._remove_term(prefix, term)
self._add_term(prefix, key)
def _index_bibentry(self, bibentry):
authors = bibentry.get_authors()
fields = bibentry.get_fields()
if 'title' in fields:
self._set_title(fields['title'])
if 'year' in fields:
self._set_year(fields['year'])
if authors:
# authors should be a list, so we make a single text string
# FIXME: better way to do this?
self._set_authors(' '.join(authors))
# add any sources in the bibtex
for sid in scan_bibentry_for_sources(bibentry):
self.add_sid(sid)
# FIXME: index 'keywords' field as regular terms
self._set_bibkey(bibentry.key)
def add_bibentry(self, bibentry):
"""Add bibentry object."""
self.bibentry = bibentry
self._index_bibentry(self.bibentry)
def add_bibtex(self, bibtex):
"""Add bibtex to document, as string or file path."""
self.add_bibentry(Bibtex(bibtex)[0])
def _load_bib(self):
if self.bibentry:
return
bibpath = self.get_bibpath()
if os.path.exists(bibpath):
self.bibentry = Bibtex(bibpath)[0]
def get_bibtex(self):
"""Get the bib for document as a bibtex string."""
self._load_bib()
if self.bibentry:
return self.bibentry.as_string()
else:
return None
def get_bibdata(self):
self._load_bib()
if self.bibentry:
data = self.bibentry.get_fields()
data['authors'] = self.bibentry.get_authors()
return data
else:
return None
def update_from_bibtex(self):
"""Update document metadata from document bibtex."""
self._load_bib()
self._index_bibentry(self.bibentry)
########################################
def get_title(self):
"""Get the title from document bibtex."""
self._load_bib()
if not self.bibentry:
return None
fields = self.bibentry.get_fields()
if 'title' in fields:
return fields['title']
return None
def get_urls(self):
"""Get all URLs associated with document."""
urls = []
# get urls associated with known sources
for sid in self.get_sids():
smod = get_source(sid)
urls.append(smod.gen_url())
# get urls from bibtex
self._load_bib()
if self.bibentry:
fields = self.bibentry.get_fields()
if 'url' in fields:
urls.append(fields['url'])
if 'adsurl' in fields:
urls.append(fields['adsurl'])
return urls
def _load_bib(self):
if self.bibentry:
return
bibpath = self.get_bibpath()
if os.path.exists(bibpath):
self.bibentry = Bibtex(bibpath)[0]
