コード例 #1
0
 def test_bool_filtered_query2(self):
     ''' Test building and running a filtered boolean query. '''
     query_bool = BoolQuery()
     query_bool.should(RangeQuery("start", lte=20000)) \
               .should(Query.term("seqid", 2)) \
               .must(Query.term("seqid", 1))
     query_string = Query.query_string("rs768019142", fields=["id", "seqid"])
     query = ElasticQuery.filtered_bool(query_string, query_bool, sources=["id", "seqid", "start"])
     elastic = Search(query, idx=ElasticSettings.idx('DEFAULT'))
     self.assertTrue(elastic.search().hits_total == 1, "Elastic filtered query retrieved marker (rs768019142)")
コード例 #2
0
    def test_bool_nested_filter(self):
        ''' Test combined Bool filter '''
        query_bool_nest = BoolQuery()
        query_bool_nest.must(Query.match("id", "rs768019142").query_wrap()) \
                       .must(Query.term("seqid", 1))

        query_bool = BoolQuery()
        query_bool.should(query_bool_nest) \
                  .should(Query.term("seqid", 2))
        query = ElasticQuery.filtered_bool(Query.match_all(), query_bool, sources=["id", "seqid", "start"])
        elastic = Search(query, idx=ElasticSettings.idx('DEFAULT'))
        self.assertTrue(elastic.search().hits_total >= 1, "Nested bool filter query")
コード例 #3
0
    def test_bool_filtered_query4(self):
        ''' Test building and running a filtered boolean query.
        Note: ElasticQuery used to wrap match in a query object. '''
        query_bool = BoolQuery()
        query_bool.should(RangeQuery("start", lte=20000)) \
                  .should(Query.term("seqid", 2)) \
                  .must(Query.match("id", "rs768019142").query_wrap()) \
                  .must(Query.term("seqid", 1))

        query = ElasticQuery.filtered_bool(Query.match_all(), query_bool, sources=["id", "seqid", "start"])
        elastic = Search(query, idx=ElasticSettings.idx('DEFAULT'))
        self.assertTrue(elastic.search().hits_total == 1, "Elastic filtered query retrieved marker (rs768019142)")
コード例 #4
0
ファイル: views.py プロジェクト: tottlefields/pydgin
def _get_query_filters(q_dict, user):
    ''' Build query bool filter. If biotypes are specified add them to the filter and
    allow for other non-gene types.
    @type  q_dict: dict
    @param q_dict: request dictionary.
    '''
    if not q_dict.getlist("biotypes"):
        return None

    query_bool = BoolQuery()
    if q_dict.getlist("biotypes"):
        query_bool.should(Query.terms("biotype", q_dict.getlist("biotypes")))
        type_filter = [Query.query_type_for_filter(ElasticSettings.search_props(c.upper(), user)['idx_type'])
                       for c in q_dict.getlist("categories") if c != "gene"]
        if len(type_filter) > 0:
            query_bool.should(type_filter)
    return Filter(query_bool)
コード例 #5
0
    def test_elastic_group_name(self):
        '''
        Testing the workflow defined in: https://killin.cimr.cam.ac.uk/nextgensite/2015/08/05/region-authorization/
        Testing various elastic queries

        idx doc:
         "_source":{"attr": {"region_id": "803", "group_name": "[\"DIL\"]", "Name": "4q27"},
         "seqid": "chr4", "source": "immunobase", "type": "region",
         "score": ".", "strand": ".", "phase": ".", "start": 122061159, "end": 122684373}
        idx_query:
        Private(in given group) OR Public
        -d '{"query":{"filtered":{"filter":{"bool": {
                                            "should": [
                                                        {"terms": {"group_name":["dil"]}},
                                                        { "missing": { "field": "group_name"   }}
                                                      ]
                                                    }}}}}'
        Private(in given group):
        -d '{"query":{"filtered":{"filter":{"terms":{"group_name":["dil"]}}}}}'
        Public:
        -d {'query': {'filtered': {'filter': {'missing': {'field': 'group_name'}},
-                         'query': {'term': {'match_all': '{}'}}}}}
        '''
        # get the groups for the given user
        response = self.client.post('/accounts/login/', {
            'username': '******',
            'password': '******'
        })
        self.assertTrue(response.status_code, "200")

        logged_in_user = User.objects.get(
            id=self.client.session['_auth_user_id'])
        if logged_in_user and logged_in_user.is_authenticated():
            user_groups = get_user_groups(logged_in_user)
            self.assertTrue('READ' in user_groups,
                            "user present in READ group")
            # make sure the user is not yet in DIL group
            self.assertFalse('DIL' in user_groups,
                             "user not present in DIL group")

        group_names = get_user_groups(logged_in_user)
        if 'READ' in group_names: group_names.remove('READ')  # @IgnorePep8
        group_names = [x.lower() for x in group_names]
        self.assertTrue(len(group_names) == 0, "No group present")

        # Match all query, as there is no group we do a match all
        query = ElasticQuery(Query.match_all())
        expected_query_string = {"query": {"match_all": {}}}
        self.assertJSONEqual(json.dumps(query.query),
                             json.dumps(expected_query_string),
                             "Query string matched")

        Search.index_refresh(self.index_name)
        elastic = Search(query, idx=self.index_name)
        docs = elastic.search().docs
        self.assertTrue(
            len(docs) == 12,
            "Elastic string query retrieved all public regions")

        # Filtered query for group names, add the user to DIL group and get the query string
        self.dil_group = Group.objects.create(name='DIL')
        logged_in_user.groups.add(self.dil_group)
        group_names = get_user_groups(logged_in_user)
        if 'READ' in group_names: group_names.remove('READ')  # @IgnorePep8
        group_names = [x.lower() for x in group_names]
        self.assertTrue(len(group_names) > 0, "More than 1 group present")
        self.assertTrue("dil" in group_names, "DIL group present")

        # retrieves all docs with missing field group_name - 11 docs
        terms_filter = TermsFilter.get_missing_terms_filter(
            "field", "attr.group_name")
        query = ElasticQuery.filtered(Query.match_all(), terms_filter)
        elastic = Search(query, idx=self.index_name)
        docs = elastic.search().docs
        self.assertTrue(
            len(docs) == 11,
            "Elastic string query retrieved all public regions")

        # build filtered boolean query to bring all public docs + private docs 11+1 = 12 docs
        query_bool = BoolQuery()
        query_bool.should(Query.missing_terms("field", "group_name")) \
                  .should(Query.terms("group_name", group_names).query_wrap())

        query = ElasticQuery.filtered_bool(Query.match_all(), query_bool)
        elastic = Search(query, idx=self.index_name)
        docs = elastic.search().docs
        self.assertTrue(
            len(docs) == 12,
            "Elastic string query retrieved both public + private regions")

        terms_filter = TermsFilter.get_terms_filter("attr.group_name",
                                                    group_names)
        query = ElasticQuery.filtered(Query.match_all(), terms_filter)
        elastic = Search(query, idx=self.index_name)
        docs = elastic.search().docs
        self.assertTrue(
            len(docs) == 1,
            "Elastic string query retrieved one private regions")
        self.assertEqual(docs[0].attr['Name'], "4q27", "type matched region")
        self.assertEqual(docs[0].attr['region_id'], "803",
                         "type matched region")
        self.assertEqual(docs[0].attr['group_name'], "[\"DIL\"]",
                         "type matched region")
コード例 #6
0
    def test_elastic_group_name(self):
        '''
        Testing the workflow defined in: https://killin.cimr.cam.ac.uk/nextgensite/2015/08/05/region-authorization/
        Testing various elastic queries

        idx doc:
         "_source":{"attr": {"region_id": "803", "group_name": "[\"DIL\"]", "Name": "4q27"},
         "seqid": "chr4", "source": "immunobase", "type": "region",
         "score": ".", "strand": ".", "phase": ".", "start": 122061159, "end": 122684373}
        idx_query:
        Private(in given group) OR Public
        -d '{"query":{"filtered":{"filter":{"bool": {
                                            "should": [
                                                        {"terms": {"group_name":["dil"]}},
                                                        { "missing": { "field": "group_name"   }}
                                                      ]
                                                    }}}}}'
        Private(in given group):
        -d '{"query":{"filtered":{"filter":{"terms":{"group_name":["dil"]}}}}}'
        Public:
        -d {'query': {'filtered': {'filter': {'missing': {'field': 'group_name'}},
-                         'query': {'term': {'match_all': '{}'}}}}}
        '''
        # get the groups for the given user
        response = self.client.post('/accounts/login/', {'username': '******', 'password': '******'})
        self.assertTrue(response.status_code, "200")

        logged_in_user = User.objects.get(id=self.client.session['_auth_user_id'])
        if logged_in_user and logged_in_user.is_authenticated():
            user_groups = get_user_groups(logged_in_user)
            self.assertTrue('READ' in user_groups, "user present in READ group")
            # make sure the user is not yet in DIL group
            self.assertFalse('DIL' in user_groups, "user not present in DIL group")

        group_names = get_user_groups(logged_in_user)
        if 'READ' in group_names : group_names.remove('READ')  # @IgnorePep8
        group_names = [x.lower() for x in group_names]
        self.assertTrue(len(group_names) == 0, "No group present")

        # Match all query, as there is no group we do a match all
        query = ElasticQuery(Query.match_all())
        expected_query_string = {"query": {"match_all": {}}}
        self.assertJSONEqual(json.dumps(query.query), json.dumps(expected_query_string), "Query string matched")

        Search.index_refresh(self.index_name)
        elastic = Search(query, idx=self.index_name)
        docs = elastic.search().docs
        self.assertTrue(len(docs) == 12, "Elastic string query retrieved all public regions")

        # Filtered query for group names, add the user to DIL group and get the query string
        self.dil_group = Group.objects.create(name='DIL')
        logged_in_user.groups.add(self.dil_group)
        group_names = get_user_groups(logged_in_user)
        if 'READ' in group_names : group_names.remove('READ')  # @IgnorePep8
        group_names = [x.lower() for x in group_names]
        self.assertTrue(len(group_names) > 0, "More than 1 group present")
        self.assertTrue("dil" in group_names, "DIL group present")

        # retrieves all docs with missing field group_name - 11 docs
        terms_filter = TermsFilter.get_missing_terms_filter("field", "attr.group_name")
        query = ElasticQuery.filtered(Query.match_all(), terms_filter)
        elastic = Search(query, idx=self.index_name)
        docs = elastic.search().docs
        self.assertTrue(len(docs) == 11, "Elastic string query retrieved all public regions")

        # build filtered boolean query to bring all public docs + private docs 11+1 = 12 docs
        query_bool = BoolQuery()
        query_bool.should(Query.missing_terms("field", "group_name")) \
                  .should(Query.terms("group_name", group_names).query_wrap())

        query = ElasticQuery.filtered_bool(Query.match_all(), query_bool)
        elastic = Search(query, idx=self.index_name)
        docs = elastic.search().docs
        self.assertTrue(len(docs) == 12, "Elastic string query retrieved both public + private regions")

        terms_filter = TermsFilter.get_terms_filter("attr.group_name", group_names)
        query = ElasticQuery.filtered(Query.match_all(), terms_filter)
        elastic = Search(query, idx=self.index_name)
        docs = elastic.search().docs
        self.assertTrue(len(docs) == 1, "Elastic string query retrieved one private regions")
        self.assertEqual(docs[0].attr['Name'], "4q27", "type matched region")
        self.assertEqual(docs[0].attr['region_id'], "803", "type matched region")
        self.assertEqual(docs[0].attr['group_name'], "[\"DIL\"]", "type matched region")
コード例 #7
0
ファイル: views.py プロジェクト: D-I-L/django-chicp
def chicpeaSearch(request, url):
    queryDict = request.GET
    user = request.user
    targetIdx = queryDict.get("targetIdx")
    blueprint = {}
    hic = []
    addList = []
    searchType = 'gene'
    searchTerm = queryDict.get("searchTerm").upper()
    searchTerm = searchTerm.replace(",", "")
    searchTerm = searchTerm.replace("..", "-")
    searchTerm = searchTerm.replace(" ", "") # Chris suggestion to prevent issue with spaces in queries
    snpTrack = queryDict.get("snp_track")

    (idx_keys_auth, idx_type_keys_auth) = get_authenticated_idx_and_idx_types(
                                            user=user, idx_keys=None, idx_type_keys=None)

    if snpTrack:
        mo = re.match(r"(.*)-(.*)", snpTrack)
        (group, track) = mo.group(1, 2)  # @UnusedVariable
        if group != 'ud' and 'CP_STATS_'+group.upper()+'.'+snpTrack.upper() not in idx_type_keys_auth:
            snpTrack = None

    if targetIdx not in utils.tissues:
        for target in getattr(chicp_settings, 'CP_TARGET'):
            if 'CP_TARGET_'+target not in idx_keys_auth:
                if targetIdx == target:
                    retJSON = {'error': 'Sorry, you do not have permission to view this dataset.'}
                    return JsonResponse(retJSON)
                continue
            elasticJSON = Search(idx=ElasticSettings.idx('CP_TARGET_'+target)).get_mapping(mapping_type="gene_target")
            tissueList = list(elasticJSON[ElasticSettings.idx('CP_TARGET_'+target)]
                              ['mappings']['gene_target']['_meta']['tissue_type'].keys())
            utils.tissues['CP_TARGET_'+target] = tissueList

    if queryDict.get("region") or re.match(r"(.*):(\d+)-(\d+)", searchTerm):
        searchType = 'region'
        region = searchTerm
        if queryDict.get("region"):
            region = queryDict.get("region")
        else:
            searchTerm = ""
        mo = re.match(r"(.*):(\d+)-(\d+)", region)
        (chrom, segmin, segmax) = mo.group(1, 2, 3)
        chrom = chrom.replace('chr', "")
        chrom = chrom.replace('CHR', "")
    if re.search("^rs[0-9]+", searchTerm.lower()):
        searchTerm = searchTerm.lower()
        addList.append(_find_snp_position(snpTrack, searchTerm))
        if addList[0].get("error"):
            return JsonResponse({'error': addList[0]['error']})
        position = addList[0]['end']
        if searchType != 'region':
            searchType = 'snp'

    logger.warn("### "+searchType+" - "+searchTerm+' ###')

    if searchType == 'region':
        query_bool = BoolQuery()
        filter_bool = BoolQuery()
        if searchTerm and len(addList) == 0 and re.match(r"(.*):(\d+)-(\d+)",
                                                         queryDict.get("searchTerm").replace(",", "")) == None:
            query_bool.must([Query.query_string(searchTerm, fields=["name", "ensg"]),
                             Query.term("baitChr", chrom),
                             Query.term("oeChr", chrom),
                             RangeQuery("dist", gte=-2e6, lte=2e6)])
        else:
            query_bool.must([Query.term("baitChr", chrom),
                             Query.term("oeChr", chrom),
                             RangeQuery("dist", gte=-2e6, lte=2e6)])

        query_bool = _add_tissue_filter(query_bool, targetIdx)

        if len(addList) > 0:
            filter_bool.should([BoolQuery(must_arr=[RangeQuery("baitStart", lte=position),
                                                    RangeQuery("baitEnd", gte=position)]),
                                BoolQuery(must_arr=[RangeQuery("oeStart", lte=position),
                                                    RangeQuery("oeEnd", gte=position)])])
        else:
            filter_bool.should([BoolQuery(must_arr=[RangeQuery("baitStart", gte=segmin, lte=segmax),
                                                    RangeQuery("baitEnd", gte=segmin, lte=segmax)]),
                                BoolQuery(must_arr=[RangeQuery("oeStart", gte=segmin, lte=segmax),
                                                    RangeQuery("oeEnd", gte=segmin, lte=segmax)])])

        query = ElasticQuery.filtered_bool(query_bool, filter_bool,
                                           sources=utils.hicFields + utils.tissues['CP_TARGET_'+targetIdx])
        (hic, v1, v2) = _build_hic_query(query, targetIdx, segmin, segmax)  # @UnusedVariable

        if "error" in hic:
            return JsonResponse(hic)
        if len(hic) == 0:
            retJSON = {'error': queryDict.get("searchTerm")+' does not overlap any bait/target regions in this dataset.'}
            return JsonResponse(retJSON)

    elif searchType == 'snp':
        if len(addList) > 0:
            chrom = addList[0]['chr']

            query_bool = BoolQuery()
            query_bool.must([Query.term("baitChr", chrom),
                             Query.term("oeChr", chrom),
                             RangeQuery("dist", gte=-2e6, lte=2e6)])
            query_bool = _add_tissue_filter(query_bool, targetIdx)

            filter_bool = BoolQuery()
            filter_bool.should([BoolQuery(must_arr=[RangeQuery("baitStart", lte=position),
                                                    RangeQuery("baitEnd", gte=position)]),
                                BoolQuery(must_arr=[RangeQuery("oeStart", lte=position),
                                                    RangeQuery("oeEnd", gte=position)])])

            query = ElasticQuery.filtered_bool(query_bool, filter_bool,
                                               sources=utils.hicFields + utils.tissues['CP_TARGET_'+targetIdx])
            hic, segmin, segmax = _build_hic_query(query, targetIdx)

            if "error" in hic:
                return JsonResponse(hic)
            if len(hic) == 0:
                retJSON = {'error': 'Marker '+searchTerm+' does not overlap any bait/target regions in this dataset.'}
                return JsonResponse(retJSON)
    else:
        # geneQuery = ElasticQuery.query_string(searchTerm, fields=["gene_name"])
        geneQuery = ElasticQuery.filtered(Query.match_all(), Filter(Query.match("gene_name", searchTerm).query_wrap()))
        resultObj = Search(idx=getattr(chicp_settings, 'CP_GENE_IDX') + '/genes/',
                           search_query=geneQuery, size=0, qsort=Sort('seqid:asc,start')).search()
        if resultObj.hits_total > 1:
            geneResults = []
            resultObj2 = Search(idx=getattr(chicp_settings, 'CP_GENE_IDX') + '/genes/', search_query=geneQuery,
                                size=(resultObj.hits_total+1), qsort=Sort('seqid:asc,start')).search()

            docs = resultObj2.docs
            gene_ids = [getattr(doc, 'attr')['gene_id'][1:-1] for doc in docs]

            query = ElasticQuery.filtered(Query.match_all(), TermsFilter.get_terms_filter('ensg', gene_ids))
            agg = Agg('ensg_agg', "terms", {"field": "ensg", "size": 0})
            res = Search(idx=ElasticSettings.idx('CP_TARGET_'+targetIdx), search_query=query, aggs=Aggs(agg),
                         size=0).search()

            ensg_count = res.aggs['ensg_agg'].get_buckets()
            gene_ids = [g['key'] for g in ensg_count]

            for d in resultObj2.docs:
                if getattr(d, "attr")["gene_id"].replace('\"', '') in gene_ids:
                    geneResults.append({
                        'gene_name': getattr(d, "attr")["gene_name"].replace('\"', ''),
                        'gene_id': getattr(d, "attr")["gene_id"].replace('\"', ''),
                        'location': "chr" + getattr(d, "seqid") + ":" +
                        locale.format_string("%d", getattr(d, "start"), grouping=True) + ".." +
                        locale.format_string("%d", getattr(d, "end"), grouping=True),
                    })

            if len(geneResults) == 0:
                retJSON = {'error': 'Gene name '+searchTerm+' not found in this dataset.'}
                return JsonResponse(retJSON)
            elif len(geneResults) > 1:
                retJSON = {
                    'error': 'Gene name <strong>'+searchTerm+'</strong> returns too many hits, please select your prefered result from the list below.',
                    'results': geneResults,
                    'cols': ['HGNC Symbol', 'Ensembl Gene ID', 'Location']
                }
                return JsonResponse(retJSON)

        query_bool = BoolQuery()
        query_bool.must([RangeQuery("dist", gte=-2e6, lte=2e6)])
        query_bool = _add_tissue_filter(query_bool, targetIdx)
        query = ElasticQuery.filtered_bool(Query.query_string(searchTerm, fields=["name", "ensg", "oeName"]),
                                           query_bool, sources=utils.hicFields + utils.tissues['CP_TARGET_'+targetIdx])

        (hic, segmin, segmax) = _build_hic_query(query, targetIdx)

        if "error" in hic:
            return JsonResponse(hic)
        if len(hic) == 0:
            retJSON = {'error': 'Gene name '+searchTerm+' not found in this dataset.'}
            return JsonResponse(retJSON)
        chrom = hic[0]['baitChr']

    try:
        chrom
    except NameError:
        retJSON = {'error': 'No chromosome defined for search'}
        return JsonResponse(retJSON)

    # get genes based on this segment
    genes = _build_gene_query(chrom, segmin, segmax)
    (snps, snpMeta) = _build_snp_query(snpTrack, chrom, segmin, segmax)
    frags = _build_frags_query(getattr(chicp_settings, 'DEFAULT_FRAG'), chrom, segmin, segmax)

    addList = utils.makeRelative(int(segmin), int(segmax), ['start', 'end'], addList)

    retJSON = {"hic": hic,
               "frags": frags,
               "meta": {"ostart": int(segmin),
                        "oend": int(segmax),
                        "rstart": 1,
                        "rend": int(segmax) - int(segmin),
                        "rchr": str(chrom),
                        "tissues": utils.tissues['CP_TARGET_'+targetIdx]},
               "snps": snps,
               "snp_meta": snpMeta,
               "genes": genes,
               "region": str(chrom) + ":" + str(segmin) + "-" + str(segmax),
               "blueprint": blueprint,
               "extra": addList
               }

    response = JsonResponse(retJSON)
    return response
コード例 #8
0
ファイル: views.py プロジェクト: tcarver/django-chicp
def chicpeaSearch(request, url):
    queryDict = request.GET
    targetIdx = queryDict.get("targetIdx")
    blueprint = {}
    hic = []
    addList = []
    searchType = 'gene'
    searchTerm = queryDict.get("searchTerm").upper()

    if targetIdx not in utils.tissues:
        for idx in getattr(chicp_settings, 'TARGET_IDXS'):
            elasticJSON = Search(idx=idx).get_mapping(mapping_type="gene_target")
            tissueList = list(elasticJSON[idx]['mappings']['gene_target']['_meta']['tissue_type'].keys())
            utils.tissues[idx] = tissueList

    if queryDict.get("region") or re.match(r"(.*):(\d+)-(\d+)", queryDict.get("searchTerm")):
        searchType = 'region'
        region = queryDict.get("searchTerm")
        if queryDict.get("region"):
            region = queryDict.get("region")
        else:
            searchTerm = ""
        mo = re.match(r"(.*):(\d+)-(\d+)", region)
        (chrom, segmin, segmax) = mo.group(1, 2, 3)
        chrom = chrom.replace('chr', "")
    if re.search("^rs[0-9]+", queryDict.get("searchTerm").lower()):
        searchTerm = queryDict.get("searchTerm").lower()
        addList.append(_find_snp_position(queryDict.get("snp_track"), searchTerm))
        if addList[0].get("error"):
            return JsonResponse({'error': addList[0]['error']})
        position = addList[0]['end']
        if searchType != 'region':
            searchType = 'snp'

    logger.warn("### "+searchType+" - "+searchTerm+' ###')

    if searchType == 'region':
        query_bool = BoolQuery()
        filter_bool = BoolQuery()
        if searchTerm and len(addList) == 0 and re.match(r"(.*):(\d+)-(\d+)", queryDict.get("searchTerm")) == None:
            query_bool.must([Query.query_string(searchTerm, fields=["name", "ensg"]),
                             Query.term("baitChr", chrom),
                             Query.term("oeChr", chrom),
                             RangeQuery("dist", gte=-2e6, lte=2e6)])
        else:
            query_bool.must([Query.term("baitChr", chrom),
                             Query.term("oeChr", chrom),
                             RangeQuery("dist", gte=-2e6, lte=2e6)])

        query_bool = _add_tissue_filter(query_bool, targetIdx)

        if len(addList) > 0:
            filter_bool.should([BoolQuery(must_arr=[RangeQuery("baitStart", lte=position),
                                                    RangeQuery("baitEnd", gte=position)]),
                                BoolQuery(must_arr=[RangeQuery("oeStart", lte=position),
                                                    RangeQuery("oeEnd", gte=position)])])
        else:
            filter_bool.should([BoolQuery(must_arr=[RangeQuery("baitStart", gte=segmin, lte=segmax),
                                                    RangeQuery("baitEnd", gte=segmin, lte=segmax)]),
                                BoolQuery(must_arr=[RangeQuery("oeStart", gte=segmin, lte=segmax),
                                                    RangeQuery("oeEnd", gte=segmin, lte=segmax)])])

        query = ElasticQuery.filtered_bool(query_bool, filter_bool, sources=utils.hicFields + utils.tissues[targetIdx])
        (hic, v1, v2) = _build_hic_query(query, targetIdx, segmin, segmax)
        # print(hic)

        if len(hic) == 0:
            retJSON = {'error': queryDict.get("searchTerm")+' does not overlap any bait/target regions in this dataset.'}
            return JsonResponse(retJSON)

    elif searchType == 'snp':
        if len(addList) > 0:
            chrom = addList[0]['chr']

            query_bool = BoolQuery()
            query_bool.must([Query.term("baitChr", chrom),
                             Query.term("oeChr", chrom),
                             RangeQuery("dist", gte=-2e6, lte=2e6)])
            query_bool = _add_tissue_filter(query_bool, targetIdx)

            filter_bool = BoolQuery()
            filter_bool.should([BoolQuery(must_arr=[RangeQuery("baitStart", lte=position),
                                                    RangeQuery("baitEnd", gte=position)]),
                                BoolQuery(must_arr=[RangeQuery("oeStart", lte=position),
                                                    RangeQuery("oeEnd", gte=position)])])

            query = ElasticQuery.filtered_bool(query_bool, filter_bool,
                                               sources=utils.hicFields + utils.tissues[targetIdx])
            hic, segmin, segmax = _build_hic_query(query, targetIdx)

            if len(hic) == 0:
                retJSON = {'error': 'Marker '+searchTerm+' does not overlap any bait/target regions in this dataset.'}
                return JsonResponse(retJSON)
    else:
        query_bool = BoolQuery()
        query_bool.must([RangeQuery("dist", gte=-2e6, lte=2e6)])
        query_bool = _add_tissue_filter(query_bool, targetIdx)
        query = ElasticQuery.filtered_bool(Query.query_string(searchTerm, fields=["name", "ensg", "oeName"]),
                                           query_bool, sources=utils.hicFields + utils.tissues[targetIdx])

        hic, segmin, segmax = _build_hic_query(query, targetIdx)

        if len(hic) == 0:
            retJSON = {'error': 'Gene name '+searchTerm+' not found in this dataset.'}
            return JsonResponse(retJSON)
        chrom = hic[0]['baitChr']

    try:
        chrom
    except NameError:
        retJSON = {'error': 'No chromosome defined for search'}
        return JsonResponse(retJSON)

    # get genes based on this segment
    genes = _build_gene_query(chrom, segmin, segmax)
    (snps, snpMeta) = _build_snp_query(queryDict.get("snp_track"), chrom, segmin, segmax)
    frags = _build_frags_query(getattr(chicp_settings, 'DEFAULT_FRAG'), chrom, segmin, segmax)

    addList = utils.makeRelative(int(segmin), int(segmax), ['start', 'end'], addList)

    retJSON = {"hic": hic,
               "frags": frags,
               "meta": {"ostart": int(segmin),
                        "oend": int(segmax),
                        "rstart": 1,
                        "rend": int(segmax) - int(segmin),
                        "rchr": str(chrom),
                        "tissues": utils.tissues[targetIdx]},
               "snps": snps,
               "snp_meta": snpMeta,
               "genes": genes,
               "region": str(chrom) + ":" + str(segmin) + "-" + str(segmax),
               "blueprint": blueprint,
               "extra": addList
               }

    response = JsonResponse(retJSON)
    return response
コード例 #9
0
def chicpeaSearch(request, url):
    queryDict = request.GET
    user = request.user
    targetIdx = queryDict.get("targetIdx")
    blueprint = {}
    hic = []
    addList = []
    searchType = 'gene'
    searchTerm = queryDict.get("searchTerm").upper()
    searchTerm = searchTerm.replace(",", "")
    searchTerm = searchTerm.replace("..", "-")
    snpTrack = queryDict.get("snp_track")

    (idx_keys_auth, idx_type_keys_auth) = get_authenticated_idx_and_idx_types(
                                            user=user, idx_keys=None, idx_type_keys=None)

    if snpTrack:
        mo = re.match(r"(.*)-(.*)", snpTrack)
        (group, track) = mo.group(1, 2)  # @UnusedVariable
        if group != 'ud' and 'CP_STATS_'+group.upper()+'.'+snpTrack.upper() not in idx_type_keys_auth:
            snpTrack = None

    if targetIdx not in utils.tissues:
        for target in getattr(chicp_settings, 'CP_TARGET'):
            if 'CP_TARGET_'+target not in idx_keys_auth:
                if targetIdx == target:
                    retJSON = {'error': 'Sorry, you do not have permission to view this dataset.'}
                    return JsonResponse(retJSON)
                continue
            elasticJSON = Search(idx=ElasticSettings.idx('CP_TARGET_'+target)).get_mapping(mapping_type="gene_target")
            tissueList = list(elasticJSON[ElasticSettings.idx('CP_TARGET_'+target)]
                              ['mappings']['gene_target']['_meta']['tissue_type'].keys())
            utils.tissues['CP_TARGET_'+target] = tissueList

    if queryDict.get("region") or re.match(r"(.*):(\d+)-(\d+)", searchTerm):
        searchType = 'region'
        region = searchTerm
        if queryDict.get("region"):
            region = queryDict.get("region")
        else:
            searchTerm = ""
        mo = re.match(r"(.*):(\d+)-(\d+)", region)
        (chrom, segmin, segmax) = mo.group(1, 2, 3)
        chrom = chrom.replace('chr', "")
        chrom = chrom.replace('CHR', "")
    if re.search("^rs[0-9]+", searchTerm.lower()):
        searchTerm = searchTerm.lower()
        addList.append(_find_snp_position(snpTrack, searchTerm))
        if addList[0].get("error"):
            return JsonResponse({'error': addList[0]['error']})
        position = addList[0]['end']
        if searchType != 'region':
            searchType = 'snp'

    logger.warn("### "+searchType+" - "+searchTerm+' ###')

    if searchType == 'region':
        query_bool = BoolQuery()
        filter_bool = BoolQuery()
        if searchTerm and len(addList) == 0 and re.match(r"(.*):(\d+)-(\d+)",
                                                         queryDict.get("searchTerm").replace(",", "")) == None:
            query_bool.must([Query.query_string(searchTerm, fields=["name", "ensg"]),
                             Query.term("baitChr", chrom),
                             Query.term("oeChr", chrom),
                             RangeQuery("dist", gte=-2e6, lte=2e6)])
        else:
            query_bool.must([Query.term("baitChr", chrom),
                             Query.term("oeChr", chrom),
                             RangeQuery("dist", gte=-2e6, lte=2e6)])

        query_bool = _add_tissue_filter(query_bool, targetIdx)

        if len(addList) > 0:
            filter_bool.should([BoolQuery(must_arr=[RangeQuery("baitStart", lte=position),
                                                    RangeQuery("baitEnd", gte=position)]),
                                BoolQuery(must_arr=[RangeQuery("oeStart", lte=position),
                                                    RangeQuery("oeEnd", gte=position)])])
        else:
            filter_bool.should([BoolQuery(must_arr=[RangeQuery("baitStart", gte=segmin, lte=segmax),
                                                    RangeQuery("baitEnd", gte=segmin, lte=segmax)]),
                                BoolQuery(must_arr=[RangeQuery("oeStart", gte=segmin, lte=segmax),
                                                    RangeQuery("oeEnd", gte=segmin, lte=segmax)])])

        query = ElasticQuery.filtered_bool(query_bool, filter_bool,
                                           sources=utils.hicFields + utils.tissues['CP_TARGET_'+targetIdx])
        (hic, v1, v2) = _build_hic_query(query, targetIdx, segmin, segmax)  # @UnusedVariable

        if "error" in hic:
            return JsonResponse(hic)
        if len(hic) == 0:
            retJSON = {'error': queryDict.get("searchTerm")+' does not overlap any bait/target regions in this dataset.'}
            return JsonResponse(retJSON)

    elif searchType == 'snp':
        if len(addList) > 0:
            chrom = addList[0]['chr']

            query_bool = BoolQuery()
            query_bool.must([Query.term("baitChr", chrom),
                             Query.term("oeChr", chrom),
                             RangeQuery("dist", gte=-2e6, lte=2e6)])
            query_bool = _add_tissue_filter(query_bool, targetIdx)

            filter_bool = BoolQuery()
            filter_bool.should([BoolQuery(must_arr=[RangeQuery("baitStart", lte=position),
                                                    RangeQuery("baitEnd", gte=position)]),
                                BoolQuery(must_arr=[RangeQuery("oeStart", lte=position),
                                                    RangeQuery("oeEnd", gte=position)])])

            query = ElasticQuery.filtered_bool(query_bool, filter_bool,
                                               sources=utils.hicFields + utils.tissues['CP_TARGET_'+targetIdx])
            hic, segmin, segmax = _build_hic_query(query, targetIdx)

            if "error" in hic:
                return JsonResponse(hic)
            if len(hic) == 0:
                retJSON = {'error': 'Marker '+searchTerm+' does not overlap any bait/target regions in this dataset.'}
                return JsonResponse(retJSON)
    else:
        # geneQuery = ElasticQuery.query_string(searchTerm, fields=["gene_name"])
        geneQuery = ElasticQuery.filtered(Query.match_all(), Filter(Query.match("gene_name", searchTerm).query_wrap()))
        resultObj = Search(idx=getattr(chicp_settings, 'CP_GENE_IDX') + '/genes/',
                           search_query=geneQuery, size=0, qsort=Sort('seqid:asc,start')).search()
        if resultObj.hits_total > 1:
            geneResults = []
            resultObj2 = Search(idx=getattr(chicp_settings, 'CP_GENE_IDX') + '/genes/', search_query=geneQuery,
                                size=(resultObj.hits_total+1), qsort=Sort('seqid:asc,start')).search()

            docs = resultObj2.docs
            gene_ids = [getattr(doc, 'attr')['gene_id'][1:-1] for doc in docs]

            query = ElasticQuery.filtered(Query.match_all(), TermsFilter.get_terms_filter('ensg', gene_ids))
            agg = Agg('ensg_agg', "terms", {"field": "ensg", "size": 0})
            res = Search(idx=ElasticSettings.idx('CP_TARGET_'+targetIdx), search_query=query, aggs=Aggs(agg),
                         size=0).search()

            ensg_count = res.aggs['ensg_agg'].get_buckets()
            gene_ids = [g['key'] for g in ensg_count]

            for d in resultObj2.docs:
                if getattr(d, "attr")["gene_id"].replace('\"', '') in gene_ids:
                    geneResults.append({
                        'gene_name': getattr(d, "attr")["gene_name"].replace('\"', ''),
                        'gene_id': getattr(d, "attr")["gene_id"].replace('\"', ''),
                        'location': "chr" + getattr(d, "seqid") + ":" +
                        locale.format_string("%d", getattr(d, "start"), grouping=True) + ".." +
                        locale.format_string("%d", getattr(d, "end"), grouping=True),
                    })

            if len(geneResults) == 0:
                retJSON = {'error': 'Gene name '+searchTerm+' not found in this dataset.'}
                return JsonResponse(retJSON)
            elif len(geneResults) > 1:
                retJSON = {
                    'error': 'Gene name <strong>'+searchTerm+'</strong> returns too many hits, please select your prefered result from the list below.',
                    'results': geneResults,
                    'cols': ['HGNC Symbol', 'Ensembl Gene ID', 'Location']
                }
                return JsonResponse(retJSON)

        query_bool = BoolQuery()
        query_bool.must([RangeQuery("dist", gte=-2e6, lte=2e6)])
        query_bool = _add_tissue_filter(query_bool, targetIdx)
        query = ElasticQuery.filtered_bool(Query.query_string(searchTerm, fields=["name", "ensg", "oeName"]),
                                           query_bool, sources=utils.hicFields + utils.tissues['CP_TARGET_'+targetIdx])

        (hic, segmin, segmax) = _build_hic_query(query, targetIdx)

        if "error" in hic:
            return JsonResponse(hic)
        if len(hic) == 0:
            retJSON = {'error': 'Gene name '+searchTerm+' not found in this dataset.'}
            return JsonResponse(retJSON)
        chrom = hic[0]['baitChr']

    try:
        chrom
    except NameError:
        retJSON = {'error': 'No chromosome defined for search'}
        return JsonResponse(retJSON)

    # get genes based on this segment
    genes = _build_gene_query(chrom, segmin, segmax)
    (snps, snpMeta) = _build_snp_query(snpTrack, chrom, segmin, segmax)
    frags = _build_frags_query(getattr(chicp_settings, 'DEFAULT_FRAG'), chrom, segmin, segmax)

    addList = utils.makeRelative(int(segmin), int(segmax), ['start', 'end'], addList)

    retJSON = {"hic": hic,
               "frags": frags,
               "meta": {"ostart": int(segmin),
                        "oend": int(segmax),
                        "rstart": 1,
                        "rend": int(segmax) - int(segmin),
                        "rchr": str(chrom),
                        "tissues": utils.tissues['CP_TARGET_'+targetIdx]},
               "snps": snps,
               "snp_meta": snpMeta,
               "genes": genes,
               "region": str(chrom) + ":" + str(segmin) + "-" + str(segmax),
               "blueprint": blueprint,
               "extra": addList
               }

    response = JsonResponse(retJSON)
    return response