def rm(q):
"Remove skid-mark associated with cached file."
cached = q.strip()
cached = re.sub('^file://', '', cached) # remove "file://" prefix
if cached.startswith(config.CACHE):
# remove cached file
os.system('rm -f %s' % cached)
# remove corresponding '.d' directory and its contents
os.system('rm -rf %s.d' % cached)
# remove file from whoosh index.
index.delete(cached)
else:
from skid.index import search
results = [dict(x) for x in search(q)]
if len(results) == 0:
# Should only happen if user hasn't done run skid-update since
# adding the paper being deleted.
print('No matches. Make sure skid is up-to-date by running `skid update`.')
elif len(results) == 1:
[hit] = results
print()
print(hit['title'])
print(colors.green % "Are you sure you'd like to delete this document [Y/n]?", end=' ')
if input().strip().lower() in ('y','yes',''):
if rm_cached(hit['cached']):
print(colors.yellow % 'Successfully deleted.')
else:
assert False, 'Multiple (%s) results found for query %r. ' \
'Refine query and try again.' \
% (len(results), q)
def main():
if len(sys.argv) <= 1:
print(', '.join(sorted(cmd.ALL)))
return
command = sys.argv.pop(1)
if command in (cmd.search, cmd.ls, cmd.similar, cmd.key):
p = ArgumentParser()
p.add_argument('query', nargs='*')
p.add_argument('--limit', type=int, default=0, #config.LIMIT,
help='query limit (use 0 for no limit)')
p.add_argument('--show', default='', help='display options')
p.add_argument('--hide', default='', help='display options')
# TODO: pager temporarily disabled because of transition to python3
p.add_argument('--pager', choices=('none', 'less', 'emacs'), default='less',
help='pager for results')
p.add_argument('--format', choices=('standard', 'org'), default='standard',
help='output format')
p.add_argument('--by', choices=('relevance', 'modified', 'added'), default='relevance',
help='Sort results by')
p.add_argument('--top', action='store_true',
help='Only show top hit.')
p.add_argument('--no-open', action='store_false',
help='do not open top hit')
p.add_argument('--note', action='store_true',
help='Open note for top hit in editor.')
args = p.parse_args()
query = ' '.join(args.query)
limit = args.limit if args.limit > 0 else None
if args.top:
args.pager = 'none'
limit = 1
if command == cmd.search:
results = index.search(query)
elif command == cmd.key:
# Supports bibtex key search, e.g. 'bottou12counterfactual'
#
# Example key
#
# 'bottou12counterfactual'
# -> 'author:bottou year:2012 title:counterfactual'
#
# - should be greedy e.g. act like '--top'
#
# - bash completion for keys should be easy to implement and useful.
#
p = bibkey(query)
if p:
# TODO: this version doesn't search for papers where author is first-author
q = ' '.join('%s:%s' % (k,v) for (k,v) in zip(['author', 'year', 'title'], p) if v)
print(q)
results = index.search(q)
else:
results = []
elif command == cmd.similar:
results = Document(query).similar(limit=limit)
elif command == cmd.ls:
results = ls(query)
else:
assert False, 'Unrecognized command %s' % command
# convert results to list and convert Whoosh.searching.Hit to skid.Document
results = list(map(todoc, results))
# sort documents according to '--by' criteria'
sortwith = {'relevance': score, 'modified': modified, 'added': added}[args.by]
if command == cmd.ls and args.by == 'relevance':
sortwith = added
results.sort(key=sortwith, reverse=True)
nresults = len(results)
# limit number of search results
results = results[:limit]
if args.format == 'org':
fmt = org
else:
fmt = display
# process display options
show = {'author', 'title', 'link', 'link:notes'} # defaults
show.update(x.strip() for x in args.show.split(','))
for x in (x.strip() for x in args.hide.split(',')):
if x in show:
show.remove(x)
with pager(args.pager):
if limit and len(results) >= limit:
if args.format == 'org':
print('# showing top %s of %s results' % (min(limit, nresults), nresults))
else:
print(colors.yellow % 'showing top %s of %s results' % (min(limit, nresults), nresults))
fmt(results, show=show)
if args.top:
assert len(results) <= 1
if not results:
print(colors.red % 'Nothing found')
return
[top] = results
# open top hit
if args.no_open:
if args.note:
# open user's note in editor
os.system('$EDITOR %s' % top.cached + '.d/notes.org')
else:
from subprocess import Popen
# open cached document
# TODO: read from config file
Popen(['xdg-open', top.cached])
elif command == cmd.add:
p = ArgumentParser()
p.add_argument('source')
p.add_argument('--name')
args = p.parse_args()
add(args.source, dest=args.name)
elif command == cmd.rm:
p = ArgumentParser()
p.add_argument('cached')
args = p.parse_args()
rm(args.cached)
elif command == cmd.update:
index.update()
elif command == cmd.authors:
authors()
elif command == cmd.tags:
tags()
elif command == cmd.drop:
print(colors.yellow % 'Dropping search index... To build a fresh one run\n$ skid update')
index.drop()
elif command == cmd.lexicon:
p = ArgumentParser()
p.add_argument('field')
args = p.parse_args()
lexicon(args.field)
elif command == cmd.title:
# doesn't require adding the document, just finds the title.
from skid.pdfhacks.pdfmill import extract_title
p = ArgumentParser()
p.add_argument('pdf')
p.add_argument('--no-extra', action='store_false', dest='extra')
args = p.parse_args()
extract_title(args.pdf, extra=args.extra)
elif command == cmd.scholar:
from skid.add import gscholar_bib
from skid.pdfhacks.pdfmill import extract_title
p = ArgumentParser()
p.add_argument('pdf')
p.add_argument('--no-extra', action='store_false', dest='extra')
args = p.parse_args()
# run google scholar search based on extracted title.
title = extract_title(args.pdf, extra=args.extra)
gscholar_bib(title=title)
else:
print(', '.join(sorted(cmd.ALL)))
def main():
if len(sys.argv) <= 1:
print ', '.join(sorted(CMDS))
return
cmd = sys.argv.pop(1)
if cmd in (SEARCH, LS, SIMILAR, KEY):
p = ArgumentParser()
p.add_argument('query', nargs='*')
p.add_argument('--limit', type=int, default=0, #config.LIMIT,
help='query limit (use 0 for no limit)')
p.add_argument('--show', default='', help='display options')
p.add_argument('--hide', default='', help='display options')
p.add_argument('--pager', choices=('none', 'less', 'emacs'), default='less',
help='pager for results')
p.add_argument('--format', choices=('standard', 'org'), default='standard',
help='output format')
p.add_argument('--by', choices=('relevance', 'modified', 'added'), default='relevance',
help='Sort results by')
p.add_argument('--top', action='store_true',
help='Only show top hit.')
p.add_argument('--no-open', action='store_false',
help='do not open top hit')
args = p.parse_args()
query = ' '.join(args.query)
limit = args.limit if args.limit > 0 else None
if args.top:
args.pager = 'none'
limit = 1
if cmd == SEARCH:
results = index.search(query)
elif cmd == KEY:
# Supports bibtex key search, e.g. 'bottou12counterfactual'
#
# Example key
#
# 'bottou12counterfactual'
# -> 'author:bottou year:2012 title:counterfactual'
#
# - should be greedy e.g. act like '--top'
#
# - bash completion for keys should be easy to implement and useful.
#
p = bibkey(query)
if p:
# TODO: this version doesn't search for papers where author is first-author
q = ' '.join('%s:%s' % (k,v) for (k,v) in zip(['author', 'year', 'title'], p) if v)
print q
results = index.search(q)
else:
results = []
elif cmd == SIMILAR:
results = Document(query).similar(limit=limit)
elif cmd == LS:
results = ls(query)
else:
assert False, 'Unrecognized command %s' % cmd
# convert results to list and convert Whoosh.searching.Hit to skid.Document
results = list(map(todoc, results))
# sort documents according to '--by' criteria'
sortwith = {'relevance': score, 'modified': modified, 'added': added}[args.by]
if cmd == LS and args.by == 'relevance':
sortwith = added
results.sort(key=sortwith, reverse=True)
nresults = len(results)
# limit number of search results
results = results[:limit]
if args.format == 'org':
fmt = org
else:
fmt = display
# process display options
show = {'author', 'title', 'link', 'link:notes'} # defaults
show.update(x.strip() for x in args.show.split(','))
for x in (x.strip() for x in args.hide.split(',')):
if x in show:
show.remove(x)
with pager(args.pager):
if limit and len(results) >= limit:
if args.format == 'org':
print '# showing top %s of %s results' % (min(limit, nresults), nresults)
else:
print yellow % 'showing top %s of %s results' % (min(limit, nresults), nresults)
fmt(results, show=show)
if args.top:
assert len(results) <= 1
if not results:
print red % 'Nothing found'
return
[top] = results
# open cached document and user notes
# os.system('gnome-open %s' % top.cached)
if args.no_open:
from subprocess import Popen
Popen(['gnome-open', top.cached])
# os.system('$EDITOR %s' % top.cached + '.d/notes.org')
elif cmd == ADD:
p = ArgumentParser()
p.add_argument('source')
args = p.parse_args()
add(args.source)
elif cmd == RM:
p = ArgumentParser()
p.add_argument('cached')
args = p.parse_args()
rm(args.cached)
elif cmd == UPDATE:
update()
elif cmd == PUSH:
push()
elif cmd == AUTHORS:
authors()
elif cmd == TAGS:
tags()
elif cmd == LEXICON:
p = ArgumentParser()
p.add_argument('field')
args = p.parse_args()
lexicon(args.field)
elif cmd == TITLE:
# doesn't require adding the document, just finds the title.
from skid.pdfhacks.pdfmill import extract_title
p = ArgumentParser()
p.add_argument('pdf')
p.add_argument('--no-extra', action='store_false', dest='extra')
args = p.parse_args()
extract_title(args.pdf, extra=args.extra)
else:
print ', '.join(sorted(CMDS))