extra/ucsc_encode_submission/encode_find.py

   1 #!/usr/bin/env python
   2
   3 from BeautifulSoup import BeautifulSoup
   4 from datetime import datetime
   5 import httplib2
   6 from operator import attrgetter
   7 from optparse import OptionParser, OptionGroup
   8 # python keyring
   9 import keyring
  10 import logging
  11 import os
  12 import re
  13 # redland rdf lib
  14 import RDF
  15 import sys
  16 import urllib
  17 import urlparse
  18
  19 from htsworkflow.util import api
  20 from htsworkflow.util.rdfhelp import \
  21      dublinCoreNS, \
  22      get_model, \
  23      get_serializer, \
  24      sparql_query, \
  25      submitOntology, \
  26      libraryOntology, \
  27      load_into_model, \
  28      rdfNS, \
  29      rdfsNS, \
  30      xsdNS
  31
  32 # URL mappings
  33 libraryNS = RDF.NS("http://jumpgate.caltech.edu/library/")
  34
  35
  36 from htsworkflow.submission.ucsc import submission_view_url, UCSCEncodePipeline
  37 download_ddf = urlparse.urljoin(UCSCEncodePipeline, "download_ddf#", allow_fragments=True)
  38 ddfNS = RDF.NS(download_ddf)
  39
  40 DBDIR = os.path.expanduser("~diane/proj/submission")
  41
  42 logger = logging.getLogger("encode_find")
  43
  44 LOGIN_URL = 'http://encodesubmit.ucsc.edu/account/login'
  45 USER_URL = 'http://encodesubmit.ucsc.edu/pipeline/show_user'
  46
  47 USERNAME = 'detrout'
  48 CHARSET = 'utf-8'
  49
  50 def main(cmdline=None):
  51     parser = make_parser()
  52     opts, args = parser.parse_args(cmdline)
  53
  54     if opts.verbose:
  55         logging.basicConfig(level=logging.INFO)
  56
  57     htsw_authdata = api.make_auth_from_opts(opts, parser)
  58     htswapi = api.HtswApi(opts.host, htsw_authdata)
  59
  60     cookie = None
  61     model = get_model(opts.load_model, DBDIR)
  62
  63     if opts.load_rdf is not None:
  64         ns_uri = submitOntology[''].uri
  65         load_into_model(model, opts.rdf_parser_name, opts.load_rdf, ns_uri)
  66
  67     if opts.update:
  68         cookie = login(cookie=cookie)
  69         load_my_submissions(model, cookie=cookie)
  70         load_encode_libraries(model, htswapi)
  71
  72     if opts.sparql is not None:
  73         sparql_query(model, opts.sparql)
  74
  75     if opts.find_submission_with_no_library:
  76         missing = find_submissions_with_no_library(model)
  77
  78     if opts.print_rdf:
  79         serializer = get_serializer(name=opts.rdf_parser_name)
  80         print serializer.serialize_model_to_string(model)
  81
  82
  83 def make_parser():
  84     parser = OptionParser()
  85     commands = OptionGroup(parser, "Commands")
  86     commands.add_option('--load-model', default=None,
  87       help="Load model database")
  88     commands.add_option('--load-rdf', default=None,
  89       help="load rdf statements into model")
  90     commands.add_option('--print-rdf', action="store_true", default=False,
  91       help="print ending model state")
  92     commands.add_option('--update', action="store_true", default=False,
  93       help="Query remote data sources and update our database")
  94     #commands.add_option('--update-ucsc-status', default=None,
  95     #  help="download status from ucsc, requires filename for extra rules")
  96     #commands.add_option('--update-ddfs', action="store_true", default=False,
  97     #  help="download ddf information for known submission")
  98     #commands.add_option('--update-library', default=None,
  99     #  help="download library info from htsw, requires filename for extra rules")
 100     parser.add_option_group(commands)
 101
 102     queries = OptionGroup(parser, "Queries")
 103     queries.add_option('--sparql', default=None,
 104       help="execute arbitrary sparql query")
 105     queries.add_option('--find-submission-with-no-library', default=False,
 106       action="store_true",
 107       help="find submissions with no library ID")
 108     parser.add_option_group(queries)
 109
 110     options = OptionGroup(parser, "Options")
 111     options.add_option("--rdf-parser-name", default="turtle",
 112       help="set rdf file parser type")
 113     options.add_option("-v", "--verbose", action="store_true", default=False)
 114     parser.add_option_group(options)
 115
 116     api.add_auth_options(parser)
 117
 118     return parser
 119
 120 def load_my_submissions(model, cookie=None):
 121     if cookie is None:
 122         cookie = login()
 123
 124     soup = get_url_as_soup(USER_URL, 'GET', cookie)
 125     p = soup.find('table', attrs={'id':'projects'})
 126     tr = p.findNext('tr')
 127     # first record is header
 128     tr = tr.findNext()
 129     TypeN = rdfsNS['type']
 130     NameN = submitOntology['name']
 131     SpeciesN = submitOntology['species']
 132     LibraryURN = submitOntology['library_urn']
 133
 134     while tr is not None:
 135         td = tr.findAll('td')
 136         if td is not None and len(td) > 1:
 137             subUrnText = td[0].contents[0].contents[0].encode(CHARSET)
 138             subUrn = RDF.Uri(submission_view_url(subUrnText))
 139
 140             add_stmt(model, subUrn, TypeN, submitOntology['Submission'])
 141
 142             name = get_contents(td[4])
 143             add_stmt(model, subUrn, NameN, name)
 144
 145             species = get_contents(td[2])
 146             if species is not None:
 147                 add_stmt(model, subUrn, SpeciesN, species)
 148
 149             library_id = get_library_id(name)
 150             if library_id is not None:
 151                 add_submission_to_library_urn(model,
 152                                               subUrn,
 153                                               LibraryURN,
 154                                               library_id)
 155
 156             add_submission_creation_date(model, subUrn, cookie)
 157
 158             # grab changing atttributes
 159             status = get_contents(td[6]).strip()
 160             last_mod_datetime = get_date_contents(td[8])
 161             last_mod = last_mod_datetime.isoformat()
 162
 163             update_submission_detail(model, subUrn, status, last_mod, cookie=cookie)
 164
 165             logging.info("Processed {0}".format( subUrn))
 166
 167         tr = tr.findNext('tr')
 168
 169
 170 def add_submission_to_library_urn(model, submissionUrn, predicate, library_id):
 171     """Add a link from a UCSC submission to woldlab library if needed
 172     """
 173     libraryUrn = libraryNS[library_id]
 174     query = RDF.Statement(submissionUrn, predicate, libraryUrn)
 175     if not model.contains_statement(query):
 176         link = RDF.Statement(submissionUrn, predicate, libraryNS[library_id])
 177         logger.info("Adding Sub -> Lib link: {0}".format(link))
 178         model.add_statement(link)
 179     else:
 180         logger.debug("Found: {0}".format(str(query)))
 181
 182
 183 def find_submissions_with_no_library(model):
 184     missing_lib_query = RDF.SPARQLQuery("""
 185 PREFIX submissionOntology:<{submissionOntology}>
 186
 187 SELECT
 188  ?subid ?name
 189 WHERE {{
 190   ?subid submissionOntology:name ?name
 191   OPTIONAL {{ ?subid submissionOntology:library_urn ?libid }}
 192   FILTER  (!bound(?libid))
 193 }}""".format(submissionOntology=submitOntology[''].uri)
 194 )
 195
 196     results = missing_lib_query.execute(model)
 197     for row in results:
 198         subid = row['subid']
 199         name = row['name']
 200         print "# {0}".format(name)
 201         print "<{0}>".format(subid.uri)
 202         print "  encodeSubmit:library_urn <http://jumpgate.caltech.edu/library/> ."
 203         print ""
 204
 205
 206 def add_submission_creation_date(model, subUrn, cookie):
 207     # in theory the submission page might have more information on it.
 208     creationDateN = libraryOntology['date']
 209     dateTimeType = xsdNS['dateTime']
 210     query = RDF.Statement(subUrn, creationDateN, None)
 211     creation_dates = list(model.find_statements(query))
 212     if len(creation_dates) == 0:
 213         logger.info("Getting creation date for: {0}".format(str(subUrn)))
 214         soup = get_url_as_soup(str(subUrn.uri), 'GET', cookie)
 215         created_label = soup.find(text="Created: ")
 216         if created_label:
 217             created_date = get_date_contents(created_label.next)
 218             created_date_node = RDF.Node(literal=created_date.isoformat(),
 219                                          datatype=dateTimeType.uri)
 220             add_stmt(model, subUrn, creationDateN, created_date_node)
 221     else:
 222         logger.debug("Found creation date for: {0}".format(str(subUrn)))
 223
 224 def update_submission_detail(model, subUrn, status, recent_update, cookie):
 225     HasStatusN = submitOntology['has_status']
 226     StatusN = submitOntology['status']
 227     LastModifyN = submitOntology['last_modify_date']
 228
 229     status_nodes_query = RDF.Statement(subUrn, HasStatusN, None)
 230     status_nodes = list(model.find_statements(status_nodes_query))
 231
 232     if len(status_nodes) == 0:
 233         # has no status node, add one
 234         logging.info("Adding status node to {0}".format(subUrn))
 235         status_blank = RDF.Node()
 236         add_stmt(model, subUrn, HasStatusN, status_blank)
 237         add_stmt(model, status_blank, rdfsNS['type'], StatusN)
 238         add_stmt(model, status_blank, StatusN, status)
 239         add_stmt(model, status_blank, LastModifyN, recent_update)
 240         update_ddf(model, subUrn, status_blank, cookie=cookie)
 241     else:
 242         logging.info("Found {0} status blanks".format(len(status_nodes)))
 243         for status_statement in status_nodes:
 244             status_blank = status_statement.object
 245             last_modified_query = RDF.Statement(status_blank, LastModifyN, None)
 246             last_mod_nodes = model.find_statements(last_modified_query)
 247             for last_mod_statement in last_mod_nodes:
 248                 last_mod_date = str(last_mod_statement.object)
 249                 if recent_update == str(last_mod_date):
 250                     update_ddf(model, subUrn, status_blank, cookie=cookie)
 251                     break
 252
 253
 254
 255 def update_ddf(model, subUrn, statusNode, cookie):
 256     TypeN = rdfsNS['type']
 257
 258     download_ddf_url = str(subUrn).replace('show', 'download_ddf')
 259     ddfUrn = RDF.Uri(download_ddf_url)
 260
 261     status_is_ddf = RDF.Statement(statusNode, TypeN, ddfNS['ddf'])
 262     if not model.contains_statement(status_is_ddf):
 263         logging.info('Adding ddf to {0}, {1}'.format(subUrn, statusNode))
 264         ddf_text = get_url_as_text(download_ddf_url, 'GET', cookie)
 265         add_ddf_statements(model, statusNode, ddf_text)
 266         model.add_statement(status_is_ddf)
 267
 268
 269 def add_ddf_statements(model, statusNode, ddf_string):
 270     """Convert a ddf text file into RDF Statements
 271     """
 272     ddf_lines = ddf_string.split('\n')
 273     # first line is header
 274     header = ddf_lines[0].split()
 275     attributes = [ ddfNS[x] for x in header ]
 276     statements = []
 277
 278     for ddf_line in ddf_lines[1:]:
 279         ddf_line = ddf_line.strip()
 280         if len(ddf_line) == 0:
 281             continue
 282         if ddf_line.startswith("#"):
 283             continue
 284
 285         ddf_record = ddf_line.split('\t')
 286         files = ddf_record[0].split(',')
 287         file_attributes = ddf_record[1:]
 288
 289         for f in files:
 290             fileNode = RDF.Node()
 291             add_stmt(model, statusNode, submitOntology['has_file'], fileNode)
 292             add_stmt(model, fileNode, rdfsNS['type'], ddfNS['file'])
 293             add_stmt(model, fileNode, ddfNS['filename'], f)
 294
 295             for predicate, object in zip( attributes[1:], file_attributes):
 296                 add_stmt(model, fileNode, predicate, object)
 297
 298
 299 def load_encode_libraries(model, htswapi):
 300     """Get libraries associated with encode.
 301     """
 302     encodeUrl = os.path.join(htswapi.root_url + "/library/?affiliations__id__exact=44")
 303     rdfaParser = RDF.Parser(name='rdfa')
 304     print encodeUrl
 305     rdfaParser.parse_into_model(model, encodeUrl)
 306     query = RDF.Statement(None, libraryOntology['library_id'], None)
 307     libraries = model.find_statements(query)
 308     for statement in libraries:
 309         libraryUrn = statement.subject
 310         load_library_detail(model, libraryUrn)
 311
 312
 313 def load_library_detail(model, libraryUrn):
 314     """Grab detail information from library page
 315     """
 316     rdfaParser = RDF.Parser(name='rdfa')
 317     query = RDF.Statement(libraryUrn, libraryOntology['date'], None)
 318     results = list(model.find_statements(query))
 319     if len(results) == 0:
 320         logger.info("Loading {0}".format(str(libraryUrn)))
 321         rdfaParser.parse_into_model(model, libraryUrn.uri)
 322     elif len(results) == 1:
 323         pass # Assuming that a loaded dataset has one record
 324     else:
 325         logging.warning("Many dates for {0}".format(libraryUrn))
 326
 327 def get_library_id(name):
 328     """Guess library ID from library name
 329     """
 330     match = re.search(r"[ -](?P<id>([\d]{5})|(SL[\d]{4}))", name)
 331     library_id = None
 332     if match is not None:
 333         library_id = match.group('id')
 334     return library_id
 335
 336
 337 def get_contents(element):
 338     """Return contents or none.
 339     """
 340     if len(element.contents) == 0:
 341         return None
 342
 343     a = element.find('a')
 344     if a is not None:
 345         return a.contents[0].encode(CHARSET)
 346
 347     return element.contents[0].encode(CHARSET)
 348
 349
 350 def get_date_contents(element):
 351     data = get_contents(element)
 352     if data:
 353         return datetime.strptime(data, "%Y-%m-%d %H:%M")
 354     else:
 355         return None
 356
 357
 358 def load_into_model(model, parser_name, filename):
 359     if not os.path.exists(filename):
 360         raise IOError("Can't find {0}".format(filename))
 361
 362     data = open(filename, 'r').read()
 363     rdf_parser = RDF.Parser(name=parser_name)
 364     rdf_parser.parse_string_into_model(model, data, ns_uri)
 365
 366 def add_stmt(model, subject, predicate, object):
 367     """Convienence create RDF Statement and add to a model
 368     """
 369     return model.add_statement(
 370         RDF.Statement(subject, predicate, object)
 371     )
 372
 373 def login(cookie=None):
 374     """Login if we don't have a cookie
 375     """
 376     if cookie is not None:
 377         return cookie
 378
 379     keys = keyring.get_keyring()
 380     password = keys.get_password(LOGIN_URL, USERNAME)
 381     credentials = {'login': USERNAME,
 382                    'password': password}
 383     headers = {'Content-type': 'application/x-www-form-urlencoded'}
 384     http = httplib2.Http()
 385     response, content = http.request(LOGIN_URL,
 386                                      'POST',
 387                                      headers=headers,
 388                                      body=urllib.urlencode(credentials))
 389     logging.debug("Login to {0}, status {1}".format(LOGIN_URL,
 390                                                     response['status']))
 391
 392     cookie = response.get('set-cookie', None)
 393     if cookie is None:
 394         raise RuntimeError("Wasn't able to log into: {0}".format(LOGIN_URL))
 395     return cookie
 396
 397
 398 def get_url_as_soup(url, method, cookie=None):
 399     http = httplib2.Http()
 400     headers = {}
 401     if cookie is not None:
 402         headers['Cookie'] = cookie
 403     response, content = http.request(url, method, headers=headers)
 404     if response['status'] == '200':
 405         soup = BeautifulSoup(content,
 406                              fromEncoding="utf-8", # should read from header
 407                              convertEntities=BeautifulSoup.HTML_ENTITIES
 408                              )
 409         return soup
 410     else:
 411         msg = "error accessing {0}, status {1}"
 412         msg = msg.format(url, response['status'])
 413         e = httplib2.HttpLib2ErrorWithResponse(msg, response, content)
 414
 415 def get_url_as_text(url, method, cookie=None):
 416     http = httplib2.Http()
 417     headers = {}
 418     if cookie is not None:
 419         headers['Cookie'] = cookie
 420     response, content = http.request(url, method, headers=headers)
 421     if response['status'] == '200':
 422         return content
 423     else:
 424         msg = "error accessing {0}, status {1}"
 425         msg = msg.format(url, response['status'])
 426         e = httplib2.HttpLib2ErrorWithResponse(msg, response, content)
 427
 428 ################
 429 #  old stuff
 430 SUBMISSIONS_LACKING_LIBID = [
 431     ('1x75-Directional-HeLa-Rep1',    '11208'),
 432     ('1x75-Directional-HeLa-Rep2',    '11207'),
 433     ('1x75-Directional-HepG2-Rep1',   '11210'),
 434     ('1x75-Directional-HepG2-Rep2',   '11209'),
 435     ('1x75-Directional-H1-hESC-Rep1', '10947'),
 436     ('1x75-Directional-H1-hESC-Rep2', '11009'),
 437     ('1x75-Directional-HUVEC-Rep1',   '11206'),
 438     ('1x75-Directional-HUVEC-Rep2',   '11205'),
 439     ('1x75-Directional-K562-Rep1',    '11008'),
 440     ('1x75-Directional-K562-Rep2',    '11007'),
 441     ('1x75-Directional-NHEK-Rep1',    '11204'),
 442     ('1x75-Directional-GM12878-Rep1', '11011'),
 443     ('1x75-Directional-GM12878-Rep2', '11010'),
 444     ]
 445
 446
 447
 448 def select_by_library_id(submission_list):
 449     subl = [ (x.library_id, x) for x in submission_list if x.library_id ]
 450     libraries = {}
 451     for lib_id, subobj in subl:
 452         libraries.setdefault(lib_id, []).append(subobj)
 453
 454     for submission in libraries.values():
 455         submission.sort(key=attrgetter('date'), reverse=True)
 456
 457     return libraries
 458
 459 def library_to_freeze(selected_libraries):
 460     freezes = ['2010-Jan', '2010-Jul', '2011-Jan']
 461     lib_ids = sorted(selected_libraries.keys())
 462     report = ['<html><table border="1">']
 463     report = ["""<html>
 464 <head>
 465 <style type="text/css">
 466  td {border-width:0 0 1px 1px; border-style:solid;}
 467 </style>
 468 </head>
 469 <body>
 470 <table>
 471 """]
 472     report.append('<thead>')
 473     report.append('<tr><td>Library ID</td><td>Name</td>')
 474     for f in freezes:
 475         report.append('<td>{0}</td>'.format(f))
 476     report.append('</tr>')
 477     report.append('</thead>')
 478     report.append('<tbody>')
 479     for lib_id in lib_ids:
 480         report.append('<tr>')
 481         lib_url = libraryNS[lib_id].uri
 482         report.append('<td><a href="{0}">{1}</a></td>'.format(lib_url, lib_id))
 483         submissions = selected_libraries[lib_id]
 484         report.append('<td>{0}</td>'.format(submissions[0].name))
 485         batched = {}
 486         for sub in submissions:
 487             date = date_to_freeze(sub.date)
 488             batched.setdefault(date, []).append(sub)
 489         print lib_id, batched
 490         for d in freezes:
 491             report.append('<td>')
 492             for s in batched.get(d, []):
 493                 show_url = submission_view_url(s.subid)
 494                 subid = '<a href="{0}">{1}</a>'.format(show_url, s.subid)
 495                 report.append("{0}:{1}".format(subid, s.status))
 496             report.append('</td>')
 497         else:
 498             report.append('<td></td>')
 499         report.append("</tr>")
 500     report.append('</tbody>')
 501     report.append("</table></html>")
 502     return "\n".join(report)
 503
 504
 505 def date_to_freeze(d):
 506     freezes = [ (datetime(2010, 1, 30), '2010-Jan'),
 507                 (datetime(2010, 7, 30), '2010-Jul'),
 508                 (datetime(2011, 1, 30), '2011-Jan'),
 509                 ]
 510     for end, name in freezes:
 511         if d < end:
 512             return name
 513     else:
 514         return None
 515
 516 if __name__ == "__main__":
 517     main()
 518