Beispiel #1
0
def erica_released_publications_json(callback, rebuild=None):
	j = json.load(open('/var/lib/myemsl/erica.json'))
	for entry in j['publications']:
		entry['_id'] = entry['id']
		del entry['id']
		if entry['completed'] == True and entry['publication_info']['limited_distribution'] == False:
			callback(jsonentry(entry))
	return 0
def erica_released_publications_json(callback, rebuild=None):
    j = json.load(open('/var/lib/myemsl/erica.json'))
    for entry in j['publications']:
        entry['_id'] = entry['id']
        del entry['id']
        if entry['completed'] == True and entry['publication_info'][
                'limited_distribution'] == False:
            callback(jsonentry(entry))
    return 0
Beispiel #3
0
def local_predicates_to_json(callback, rebuild=None):
	sql = """
	select lp.id, lp.desc, lp.person_id, u.first_name, u.last_name from myemsl.local_predicate as lp, eus.users as u where lp.person_id = u.person_id;
	"""
#FIXME version
	cnx = myemsldb_connect(myemsl_schema_versions=['1.3'])
	cursor = cnx.cursor()
	cursor.execute(sql)
	for row in cursoriter(cursor):
		if row[1] == None:
			print "Id %s bad atime." %(row[0])
			continue
		entry = json.loads(row[1])
		entry['_id'] = row[0]
		entry['submitter'] = {'name': row[3] + " " + row[4], 'id':row[2]}
		callback(jsonentry(entry))
	return 0
def local_predicates_to_json(callback, rebuild=None):
    sql = """
	select lp.id, lp.desc, lp.person_id, u.first_name, u.last_name from myemsl.local_predicate as lp, eus.users as u where lp.person_id = u.person_id;
	"""
    #FIXME version
    cnx = myemsldb_connect(myemsl_schema_versions=['1.3'])
    cursor = cnx.cursor()
    cursor.execute(sql)
    for row in cursoriter(cursor):
        if row[1] == None:
            print "Id %s bad atime." % (row[0])
            continue
        entry = json.loads(row[1])
        entry['_id'] = row[0]
        entry['submitter'] = {'name': row[3] + " " + row[4], 'id': row[2]}
        callback(jsonentry(entry))
    return 0
Beispiel #5
0
def proposal_info_to_json(callback, rebuild=None):
	sql = """
	select proposal_id, accepted_date, title from eus.proposals;
	"""
	cnx = myemsldb_connect(myemsl_schema_versions=['1.3'])
	cursor = cnx.cursor()
	cursor.execute(sql)
	for row in cursoriter(cursor):
		if row[1] == None:
			print "Id %s bad atime." %(row[0])
			continue
		entry = {}
		proposal_id = row[0]
		entry['title'] = row[2]
		entry['accepted_date'] = row[1]
		sql = """
select pm.person_id, u.network_id, u.first_name, u.last_name, u.email_address from eus.proposal_members as pm, eus.users as u where pm.proposal_id=%(proposal_id)s and pm.person_id = u.person_id;
		"""
		cursor2 = cnx.cursor()
		cursor2.execute(sql, {"proposal_id":proposal_id})
		person_ids = []
		network_ids = []
		person_names = []
		email_addresses = []
		for row2 in cursoriter(cursor2):
			person_id = row2[0]
			network_id = row2[1]
			first_name = row2[2]
			last_name = row2[3]
			email_address = row2[4]
			if person_id:
				person_ids.append(person_id)
			if network_id:
				network_ids.append(network_id)
			if first_name or last_name:
				if first_name and last_name:
					person_names.append("%s %s" %(first_name, last_name))
				elif last_name:
					person_names.append(last_name)
				else:
					person_names.append(first_name)
			if email_address:
				email_addresses.append(email_address)
		entry['person_ids'] = person_ids
		entry['network_ids'] = network_ids
		entry['person_names'] = person_names
		entry['email_addresses'] = email_addresses
		sql = """
select i.instrument_name, i.name_short, i.eus_display_name from eus.instruments as i, eus.proposal_instruments as pi where pi.instrument_id = i.instrument_id and pi.proposal_id=%(proposal_id)s group by i.instrument_name, i.name_short, i.eus_display_name;
		"""
		cursor2 = cnx.cursor()
		cursor2.execute(sql, {"proposal_id":proposal_id})
		instrument_names = []
		instrument_short_names = []
		for row2 in cursoriter(cursor2):
			i = []
			if row2[0]:
				i.append(row2[0])
			if row2[1]:
				i.append(row2[1])
				instrument_short_names.append(row2[1])
			if row2[2]:
				i.append(row2[2])
			if len(i) > 0:
				instrument_names.extend(i)
		if len(instrument_short_names) > 0:
			entry['instrument_names'] = instrument_short_names
		if len(instrument_names) > 0:
			entry['instrument_text'] = ' '.join(instrument_names)
		entry['_id'] = proposal_id
		callback(jsonentry(entry))
	return 0
Beispiel #6
0
def all_metadata_to_json(callback, rebuild=None):
#FIXME don't hard code this like this.
	all_icons =  dict([(i[:-len('.png')], True) for i in os.listdir('/var/www/myemsl/static/1/icons') if i.endswith('.png')])
	cnx = myemsldb_connect(myemsl_schema_versions=['1.3'])
	cursor = cnx.cursor()
	sql = """
select proposal_id, person_id from eus.proposal_members;
"""
	cursor.execute(sql)
	proposal_info = {}
	for row in cursoriter(cursor):
		p = proposal_info.get(row[0])
		if not p:
			p = {"members":[]}
			proposal_info[row[0]] = p
		p['members'].append(row[1])
	sql = """
select distinct person_id from myemsl.permission_group as pg, myemsl.permission_group_members as pgm where pg.name='MyEMSL.Downloader.omics.dms' and pg.permission_group_id = pgm.permission_group_id order by person_id;
"""
	cursor.execute(sql)
	dms_users = []
	for row in cursoriter(cursor):
		dms_users.append(row[0])
	sql = """
select distinct person_id from myemsl.permission_group as pg, myemsl.permission_group_members as pgm where pg.name='MyEMSL.NMR.Download' and pg.permission_group_id = pgm.permission_group_id order by person_id;
"""
	cursor.execute(sql)
	nmr_users = []
	for row in cursoriter(cursor):
		nmr_users.append(row[0])
	nmr_instruments = ["1002","1003","1004","1005","1142","1145","1176","34000","34075","34134","34135"]

	nmr_instruments = dict([(i, 1) for i in nmr_instruments])
	sql = """
CREATE OR REPLACE FUNCTION myemsl_group_id_sort (ANYARRAY)
RETURNS ANYARRAY LANGUAGE SQL
AS $$
SELECT ARRAY(SELECT DISTINCT unnest($1) ORDER BY 1)
$$;

CREATE TEMP TABLE "temp_group_sets" (
        "group_set_id" SERIAL PRIMARY KEY,
        "group_temp_key" text,
        "group_set_members" bigint[]
)
WITHOUT OIDS
ON COMMIT DROP;

CREATE TEMP TABLE "temp_group_flattened" (
        "group_id" bigint,
        "group_member" bigint
)
WITHOUT OIDS
ON COMMIT DROP;

CREATE TEMP TABLE "temp_group_set_members" (
        "group_set_id" int,
        "group_member" bigint
)
WITHOUT OIDS
ON COMMIT DROP;

CREATE TEMP TABLE "temp_item_to_group_set" (
        "item_id" bigint PRIMARY KEY,
        "group_set_id" int
)
WITHOUT OIDS
ON COMMIT DROP;

insert into temp_group_sets(group_temp_key, group_set_members)
select distinct group_temp_key, group_set_members from (
select array_to_string(agg, ',') as group_temp_key, agg as group_set_members from (select myemsl_group_id_sort(agg) as agg from (select array_agg(group_id) as agg from myemsl.group_items group by item_id) as g group by agg) as g
) as g;

CREATE INDEX temp_group_sets_group_temp_key_idx ON temp_group_sets USING btree (group_temp_key);

insert into temp_group_flattened(group_member, group_id)
with recursive t(parent_id, child_id) as (
select parent_id, child_id from myemsl.subgroups as s
union
select s.parent_id, t.child_id from myemsl.subgroups as s, t where t.parent_id = s.child_id
) select parent_id, child_id from t where parent_id is not NULL
union select group_id, group_id from myemsl.groups where group_id not in (select child_id from myemsl.subgroups)
union select child_id, child_id from myemsl.subgroups;

insert into temp_group_set_members(group_set_id, group_member)
select tgs.group_set_id, tgf.group_member from (select group_set_id, unnest(group_set_members) as group_id from temp_group_sets) as tgs, temp_group_flattened as tgf where tgs.group_id = tgf.group_id;

drop table temp_group_flattened;

insert into temp_item_to_group_set(item_id, group_set_id)
select item_id, group_set_id from (select item_id, myemsl_group_id_sort(array_agg(group_id)) as agg from myemsl.group_items group by item_id) as g, temp_group_sets as tgs where array_to_string(g.agg, ',') = tgs.group_temp_key;

drop table temp_group_sets;

select group_set_id, g.group_id, type, name, proposal_id from (select group_set_id, group_id, type, name from temp_group_set_members as gsm, myemsl.groups as g where gsm.group_member = g.group_id) as g left join eus.proposals as p on p.group_id = g.group_id;
"""
	cursor.execute(sql)
	group_sets = {}
#FIXME make this list use a config file somehow.
	mdp = myemsl.elasticsearch.metadata.metadata_processor()
	myemsl.elasticsearch.metadata.emsl_base.metadata(mdp)
	myemsl.elasticsearch.metadata.emsl_dms.metadata(mdp)
	print 'MD Loaded'
#FIXME unhardcode this
	extended_metadata_reference = {
		'omics.dms.datapackage_id': {'id':'gov_pnnl_emsl_dms_datapackage', 'type':int},
		'omics.dms.dataset': {'id':'gov_pnnl_emsl_dms_dataset', 'type':lambda x: mdp.emsl_dms_metadata.datasetname2datasetid.get(x)},
	}
	for row in cursoriter(cursor):
		gs = group_sets.get(row[0])
		if not gs:
			gs = {'groups':[], 'proposals':{}, 'extended_metadata':{'gov_pnnl_emsl_proposal':[], 'gov_pnnl_emsl_instrument':[]}}
			group_sets[row[0]] = gs
		if(row[4] != None):
			gs['extended_metadata']['gov_pnnl_emsl_proposal'].append(row[3])
		else:
			if row[2][:len("Instrument.")] == "Instrument.":
				gs['groups'].append([row[1], "Instrument", row[2][len("Instrument."):]])
				gs['extended_metadata']['gov_pnnl_emsl_instrument'].append(row[2][len("Instrument."):])
			elif row[2][:len("gov_pnnl_erica/irn")] == "gov_pnnl_erica/irn":
				gs['extended_metadata']['gov_pnnl_erica/irn'] = json.loads(row[3])
			else:
				gs['groups'].append([row[1], row[2], row[3]])
				e = extended_metadata_reference.get(row[2])
				if e:
					list = gs['extended_metadata'].get(e['id'])
					if not list:
						list = []
						gs['extended_metadata'][e['id']] = list
					list.append(e['type'](row[3]))
	for (gsid, gs) in group_sets.iteritems():
		users = {}
#		extended_metadata = {
#			'gov_pnnl_emsl_proposal': gs['proposals']
			#FIXME 'gov_pnnl_emsl_users': users.keys()}
#		}
#		dpid = gs.get('groups.omics.dms.datapackage_id')
#		if dpid:
#			extended_metadata['gov_pnnl_emsl_dms_datapackage'] = dpid
		extended_metadata = gs['extended_metadata']
		myemsl.elasticsearch.metadata.dedup(extended_metadata)
		dp = extended_metadata.get('gov_pnnl_emsl_dms_datapackage')
		if dp:
			for u in dms_users:
				users[u] = 1
			for id in dp[:]:
				print "Processing:", id, dp
				myemsl.elasticsearch.metadata.merge_left(extended_metadata, mdp.emsl_dms_metadata.datapackage.get(id))
		else:
			dp = extended_metadata.get('gov_pnnl_emsl_dms_dataset')
			if dp:
				for u in dms_users:
					users[u] = 1
				for id in dp[:]:
					print "Processing:", id, dp
					myemsl.elasticsearch.metadata.merge_left(extended_metadata, mdp.emsl_dms_metadata.dataset.get(id))
					extra_proposals = mdp.emsl_dms_metadata.dataset_proposals.get(id)
					if extra_proposals != None:
						gs['extended_metadata']['gov_pnnl_emsl_proposal'].extend(extra_proposals)
			else:
				inst = [i[2] for i in gs['groups'] if i[1] == 'Instrument']
				for i in inst:
					if str(i) in nmr_instruments:
						for u in nmr_users:
							users[u] = 1
		extended_metadata = mdp.resolv(extended_metadata)
		if len(extended_metadata) > 0:
#FIXME Plugin related stuff...
			for x in ['gov_pnnl_emsl_dms_pi', 'gov_pnnl_emsl_dms_technical_lead', 'gov_pnnl_emsl_dms_proj_mgr', 'gov_pnnl_emsl_dms_researcher', 'gov_pnnl_emsl_dms_dataset_operator']:
				list = extended_metadata.get(x)
				if list:
					for i in list:
						users[i['id']] = 1
			gs['extended_metadata'] = extended_metadata
		else:
			del gs['extended_metadata']
		proposals = [i['id'] for i in gs['extended_metadata']['gov_pnnl_emsl_proposal']]
#		dedup_proposals = {}
		for proposal in proposals:
#			dedup_proposals[proposal] = 1
			pi = proposal_info.get(proposal)
			if pi:
				for member in pi['members']:
					users[member] = 1
#FIXME
		gs['proposals'] = proposals
		gs['users'] = users.keys()
		print gs
	sql = """
select instrument_id, name_short from eus.instruments;
"""
	cursor.execute(sql)
	instruments = {}
	for row in cursoriter(cursor):
		instruments[row[0]] = row[1]
#FIXME remove
	import time
	count = 0
	print time.time()
	sql = """
SELECT
  name, submitter, 
  first_name, last_name,
  stime, subdir, size, 
  q.item_id,
  aged,
  group_set_id,
  hashsum as sha1,
  transaction
FROM (
  SELECT
    name, submitter,
    eus.users.first_name,
    eus.users.last_name,
    EXTRACT(EPOCH FROM stime) as stime,
    subdir, size,
    myemsl.files.item_id,
    aged,
    group_set_id,
    myemsl.transactions.transaction 
  FROM
    myemsl.files,
    myemsl.transactions,
    eus.users,
    temp_item_to_group_set
  WHERE
    myemsl.transactions.stime > NOW() - '2 days'::interval and
    myemsl.transactions.transaction = myemsl.files.transaction and
    eus.users.person_id = myemsl.transactions.submitter and
    temp_item_to_group_set.item_id = myemsl.files.item_id and (
      name != 'metadata.txt' or
      subdir != ''
    )
  ) AS q
  LEFT OUTER JOIN
  myemsl.hashsums AS h
  ON (q.item_id = h.item_id);
	"""
	cursor.execute(sql)
	for row in cursoriter(cursor):
#FIXME
		if count == 0:
			print time.time(), count
		group_set_id = row[0]
		if count % 1000 == 0:
			print time.time(), count
		count += 1
#remove till here
		if row[4] == None:
			print "Id %s bad stime." %(row[7])
			continue
		entry = {}
		item_id = row[7]
		entry['filename'] = row[0]
		idx = row[0].rfind('.')
		entry['ico'] = 'file'
		if idx != -1:
			entry['ext'] = row[0][idx + 1:]
			newicon = all_icons.get(entry['ext'])
			if newicon:
				entry['ico'] = entry['ext']
		entry['subdir'] = row[5]
		entry['size'] = row[6]
		first = row[2]
		if not first:
			first = ''
		last = row[3]
		if not last:
			last = ''
		if first != '' and last != '':
			entry['submittername'] = "%s %s" %(first, last)
		entry['first'] = first
		entry['last'] = last
		entry['submitterid'] = row[1]
		entry['trans'] = row[11]
		entry['hash'] = {'sha1':row[10]}
		aged = "false"
#FIXME make this a bool in elasticsearch later
		if row[8]:
			aged = "true"
		entry['aged'] = aged
		users = {row[1]:1}
		groups = {}
		proposals = []
		gs = group_sets.get(row[9])
		if gs:
			for g in gs['groups']:
				groups[g[1]] = g[2]
			proposals = gs['proposals']
			for user in gs['users']:
				users[user] = 1
			emd = gs.get('extended_metadata')
			if emd and len(emd) > 0:
				entry['extended_metadata'] = emd
		ds = groups.get('omics.dms.dataset')
		if ds:
			#FIXME generalize this. extended_metadata on metadata specific file entries need to be compied so modifications don't affect others.
			entry['extended_metadata'] = entry['extended_metadata'].copy()
			tsubdir = entry['subdir']
			if tsubdir.startswith(ds + '/'):
				tsubdir = tsubdir[len(ds) + 1:]
			dsid = mdp.emsl_dms_metadata.datasetname2datasetid.get(ds)
			if dsid:
				ajid = mdp.emsl_dms_metadata.datasetid_subdir2analysisjobid(mdp.emsl_dms_metadata.datasetname2datasetid[ds], tsubdir)
				if ajid == None:
					aj = {'tool': {'name':'Raw'}}
				else:
					aj = mdp.emsl_dms_metadata.analysisjob[ajid]
				entry['extended_metadata']['gov_pnnl_emsl_dms_analysisjob'] = [aj]
		if len(proposals) > 0:
			entry['proposals'] = proposals
		if len(groups) > 0:
			for k, v in groups.iteritems():
#FIXME escape
				entry["groups.%s" %(k)] = v
				if k == 'Instrument' and v.isdigit():
					tv = instruments.get(int(v))
					if tv:
						entry["instrumentname"] = tv
		entry['users'] = users.keys()
		entry['stime'] = rfc3339.rfc3339(row[4])
		entry['_id'] = item_id
		callback(jsonentry(entry))
	return 0
def proposal_info_to_json(callback, rebuild=None):
    sql = """
	select proposal_id, accepted_date, title from eus.proposals;
	"""
    cnx = myemsldb_connect(myemsl_schema_versions=['1.3'])
    cursor = cnx.cursor()
    cursor.execute(sql)
    for row in cursoriter(cursor):
        if row[1] == None:
            print "Id %s bad atime." % (row[0])
            continue
        entry = {}
        proposal_id = row[0]
        entry['title'] = row[2]
        entry['accepted_date'] = row[1]
        sql = """
select pm.person_id, u.network_id, u.first_name, u.last_name, u.email_address from eus.proposal_members as pm, eus.users as u where pm.proposal_id=%(proposal_id)s and pm.person_id = u.person_id;
		"""
        cursor2 = cnx.cursor()
        cursor2.execute(sql, {"proposal_id": proposal_id})
        person_ids = []
        network_ids = []
        person_names = []
        email_addresses = []
        for row2 in cursoriter(cursor2):
            person_id = row2[0]
            network_id = row2[1]
            first_name = row2[2]
            last_name = row2[3]
            email_address = row2[4]
            if person_id:
                person_ids.append(person_id)
            if network_id:
                network_ids.append(network_id)
            if first_name or last_name:
                if first_name and last_name:
                    person_names.append("%s %s" % (first_name, last_name))
                elif last_name:
                    person_names.append(last_name)
                else:
                    person_names.append(first_name)
            if email_address:
                email_addresses.append(email_address)
        entry['person_ids'] = person_ids
        entry['network_ids'] = network_ids
        entry['person_names'] = person_names
        entry['email_addresses'] = email_addresses
        sql = """
select i.instrument_name, i.name_short, i.eus_display_name from eus.instruments as i, eus.proposal_instruments as pi where pi.instrument_id = i.instrument_id and pi.proposal_id=%(proposal_id)s group by i.instrument_name, i.name_short, i.eus_display_name;
		"""
        cursor2 = cnx.cursor()
        cursor2.execute(sql, {"proposal_id": proposal_id})
        instrument_names = []
        instrument_short_names = []
        for row2 in cursoriter(cursor2):
            i = []
            if row2[0]:
                i.append(row2[0])
            if row2[1]:
                i.append(row2[1])
                instrument_short_names.append(row2[1])
            if row2[2]:
                i.append(row2[2])
            if len(i) > 0:
                instrument_names.extend(i)
        if len(instrument_short_names) > 0:
            entry['instrument_names'] = instrument_short_names
        if len(instrument_names) > 0:
            entry['instrument_text'] = ' '.join(instrument_names)
        entry['_id'] = proposal_id
        callback(jsonentry(entry))
    return 0
def all_metadata_to_json(callback, rebuild=None):
    #FIXME don't hard code this like this.
    all_icons = dict([(i[:-len('.png')], True)
                      for i in os.listdir('/var/www/myemsl/static/1/icons')
                      if i.endswith('.png')])
    cnx = myemsldb_connect(myemsl_schema_versions=['1.3'])
    cursor = cnx.cursor()
    sql = """
select proposal_id, person_id from eus.proposal_members;
"""
    cursor.execute(sql)
    proposal_info = {}
    for row in cursoriter(cursor):
        p = proposal_info.get(row[0])
        if not p:
            p = {"members": []}
            proposal_info[row[0]] = p
        p['members'].append(row[1])
    sql = """
select distinct person_id from myemsl.permission_group as pg, myemsl.permission_group_members as pgm where pg.name='MyEMSL.Downloader.omics.dms' and pg.permission_group_id = pgm.permission_group_id order by person_id;
"""
    cursor.execute(sql)
    dms_users = []
    for row in cursoriter(cursor):
        dms_users.append(row[0])
    sql = """
select distinct person_id from myemsl.permission_group as pg, myemsl.permission_group_members as pgm where pg.name='MyEMSL.NMR.Download' and pg.permission_group_id = pgm.permission_group_id order by person_id;
"""
    cursor.execute(sql)
    nmr_users = []
    for row in cursoriter(cursor):
        nmr_users.append(row[0])
    nmr_instruments = [
        "1002", "1003", "1004", "1005", "1142", "1145", "1176", "34000",
        "34075", "34134", "34135"
    ]

    nmr_instruments = dict([(i, 1) for i in nmr_instruments])
    sql = """
CREATE OR REPLACE FUNCTION myemsl_group_id_sort (ANYARRAY)
RETURNS ANYARRAY LANGUAGE SQL
AS $$
SELECT ARRAY(SELECT DISTINCT unnest($1) ORDER BY 1)
$$;

CREATE TEMP TABLE "temp_group_sets" (
        "group_set_id" SERIAL PRIMARY KEY,
        "group_temp_key" text,
        "group_set_members" bigint[]
)
WITHOUT OIDS
ON COMMIT DROP;

CREATE TEMP TABLE "temp_group_flattened" (
        "group_id" bigint,
        "group_member" bigint
)
WITHOUT OIDS
ON COMMIT DROP;

CREATE TEMP TABLE "temp_group_set_members" (
        "group_set_id" int,
        "group_member" bigint
)
WITHOUT OIDS
ON COMMIT DROP;

CREATE TEMP TABLE "temp_item_to_group_set" (
        "item_id" bigint PRIMARY KEY,
        "group_set_id" int
)
WITHOUT OIDS
ON COMMIT DROP;

insert into temp_group_sets(group_temp_key, group_set_members)
select distinct group_temp_key, group_set_members from (
select array_to_string(agg, ',') as group_temp_key, agg as group_set_members from (select myemsl_group_id_sort(agg) as agg from (select array_agg(group_id) as agg from myemsl.group_items group by item_id) as g group by agg) as g
) as g;

CREATE INDEX temp_group_sets_group_temp_key_idx ON temp_group_sets USING btree (group_temp_key);

insert into temp_group_flattened(group_member, group_id)
with recursive t(parent_id, child_id) as (
select parent_id, child_id from myemsl.subgroups as s
union
select s.parent_id, t.child_id from myemsl.subgroups as s, t where t.parent_id = s.child_id
) select parent_id, child_id from t where parent_id is not NULL
union select group_id, group_id from myemsl.groups where group_id not in (select child_id from myemsl.subgroups)
union select child_id, child_id from myemsl.subgroups;

insert into temp_group_set_members(group_set_id, group_member)
select tgs.group_set_id, tgf.group_member from (select group_set_id, unnest(group_set_members) as group_id from temp_group_sets) as tgs, temp_group_flattened as tgf where tgs.group_id = tgf.group_id;

drop table temp_group_flattened;

insert into temp_item_to_group_set(item_id, group_set_id)
select item_id, group_set_id from (select item_id, myemsl_group_id_sort(array_agg(group_id)) as agg from myemsl.group_items group by item_id) as g, temp_group_sets as tgs where array_to_string(g.agg, ',') = tgs.group_temp_key;

drop table temp_group_sets;

select group_set_id, g.group_id, type, name, proposal_id from (select group_set_id, group_id, type, name from temp_group_set_members as gsm, myemsl.groups as g where gsm.group_member = g.group_id) as g left join eus.proposals as p on p.group_id = g.group_id;
"""
    cursor.execute(sql)
    group_sets = {}
    #FIXME make this list use a config file somehow.
    mdp = myemsl.elasticsearch.metadata.metadata_processor()
    myemsl.elasticsearch.metadata.emsl_base.metadata(mdp)
    myemsl.elasticsearch.metadata.emsl_dms.metadata(mdp)
    print 'MD Loaded'
    #FIXME unhardcode this
    extended_metadata_reference = {
        'omics.dms.datapackage_id': {
            'id': 'gov_pnnl_emsl_dms_datapackage',
            'type': int
        },
        'omics.dms.dataset': {
            'id': 'gov_pnnl_emsl_dms_dataset',
            'type':
            lambda x: mdp.emsl_dms_metadata.datasetname2datasetid.get(x)
        },
    }
    for row in cursoriter(cursor):
        gs = group_sets.get(row[0])
        if not gs:
            gs = {
                'groups': [],
                'proposals': {},
                'extended_metadata': {
                    'gov_pnnl_emsl_proposal': [],
                    'gov_pnnl_emsl_instrument': []
                }
            }
            group_sets[row[0]] = gs
        if (row[4] != None):
            gs['extended_metadata']['gov_pnnl_emsl_proposal'].append(row[3])
        else:
            if row[2][:len("Instrument.")] == "Instrument.":
                gs['groups'].append(
                    [row[1], "Instrument", row[2][len("Instrument."):]])
                gs['extended_metadata']['gov_pnnl_emsl_instrument'].append(
                    row[2][len("Instrument."):])
            elif row[2][:len("gov_pnnl_erica/irn")] == "gov_pnnl_erica/irn":
                gs['extended_metadata']['gov_pnnl_erica/irn'] = json.loads(
                    row[3])
            else:
                gs['groups'].append([row[1], row[2], row[3]])
                e = extended_metadata_reference.get(row[2])
                if e:
                    list = gs['extended_metadata'].get(e['id'])
                    if not list:
                        list = []
                        gs['extended_metadata'][e['id']] = list
                    list.append(e['type'](row[3]))
    for (gsid, gs) in group_sets.iteritems():
        users = {}
        #		extended_metadata = {
        #			'gov_pnnl_emsl_proposal': gs['proposals']
        #FIXME 'gov_pnnl_emsl_users': users.keys()}
        #		}
        #		dpid = gs.get('groups.omics.dms.datapackage_id')
        #		if dpid:
        #			extended_metadata['gov_pnnl_emsl_dms_datapackage'] = dpid
        extended_metadata = gs['extended_metadata']
        myemsl.elasticsearch.metadata.dedup(extended_metadata)
        dp = extended_metadata.get('gov_pnnl_emsl_dms_datapackage')
        if dp:
            for u in dms_users:
                users[u] = 1
            for id in dp[:]:
                print "Processing:", id, dp
                myemsl.elasticsearch.metadata.merge_left(
                    extended_metadata,
                    mdp.emsl_dms_metadata.datapackage.get(id))
        else:
            dp = extended_metadata.get('gov_pnnl_emsl_dms_dataset')
            if dp:
                for u in dms_users:
                    users[u] = 1
                for id in dp[:]:
                    print "Processing:", id, dp
                    myemsl.elasticsearch.metadata.merge_left(
                        extended_metadata,
                        mdp.emsl_dms_metadata.dataset.get(id))
                    extra_proposals = mdp.emsl_dms_metadata.dataset_proposals.get(
                        id)
                    if extra_proposals != None:
                        gs['extended_metadata'][
                            'gov_pnnl_emsl_proposal'].extend(extra_proposals)
            else:
                inst = [i[2] for i in gs['groups'] if i[1] == 'Instrument']
                for i in inst:
                    if str(i) in nmr_instruments:
                        for u in nmr_users:
                            users[u] = 1
        extended_metadata = mdp.resolv(extended_metadata)
        if len(extended_metadata) > 0:
            #FIXME Plugin related stuff...
            for x in [
                    'gov_pnnl_emsl_dms_pi', 'gov_pnnl_emsl_dms_technical_lead',
                    'gov_pnnl_emsl_dms_proj_mgr',
                    'gov_pnnl_emsl_dms_researcher',
                    'gov_pnnl_emsl_dms_dataset_operator'
            ]:
                list = extended_metadata.get(x)
                if list:
                    for i in list:
                        users[i['id']] = 1
            gs['extended_metadata'] = extended_metadata
        else:
            del gs['extended_metadata']
        proposals = [
            i['id'] for i in gs['extended_metadata']['gov_pnnl_emsl_proposal']
        ]
        #		dedup_proposals = {}
        for proposal in proposals:
            #			dedup_proposals[proposal] = 1
            pi = proposal_info.get(proposal)
            if pi:
                for member in pi['members']:
                    users[member] = 1
#FIXME
        gs['proposals'] = proposals
        gs['users'] = users.keys()
        print gs
    sql = """
select instrument_id, name_short from eus.instruments;
"""
    cursor.execute(sql)
    instruments = {}
    for row in cursoriter(cursor):
        instruments[row[0]] = row[1]


#FIXME remove
    import time
    count = 0
    print time.time()
    sql = """
SELECT
  name, submitter, 
  first_name, last_name,
  stime, subdir, size, 
  q.item_id,
  aged,
  group_set_id,
  hashsum as sha1,
  transaction
FROM (
  SELECT
    name, submitter,
    eus.users.first_name,
    eus.users.last_name,
    EXTRACT(EPOCH FROM stime) as stime,
    subdir, size,
    myemsl.files.item_id,
    aged,
    group_set_id,
    myemsl.transactions.transaction 
  FROM
    myemsl.files,
    myemsl.transactions,
    eus.users,
    temp_item_to_group_set
  WHERE
    myemsl.transactions.stime > NOW() - '2 days'::interval and
    myemsl.transactions.transaction = myemsl.files.transaction and
    eus.users.person_id = myemsl.transactions.submitter and
    temp_item_to_group_set.item_id = myemsl.files.item_id and (
      name != 'metadata.txt' or
      subdir != ''
    )
  ) AS q
  LEFT OUTER JOIN
  myemsl.hashsums AS h
  ON (q.item_id = h.item_id);
	"""
    cursor.execute(sql)
    for row in cursoriter(cursor):
        #FIXME
        if count == 0:
            print time.time(), count
        group_set_id = row[0]
        if count % 1000 == 0:
            print time.time(), count
        count += 1
        #remove till here
        if row[4] == None:
            print "Id %s bad stime." % (row[7])
            continue
        entry = {}
        item_id = row[7]
        entry['filename'] = row[0]
        idx = row[0].rfind('.')
        entry['ico'] = 'file'
        if idx != -1:
            entry['ext'] = row[0][idx + 1:]
            newicon = all_icons.get(entry['ext'])
            if newicon:
                entry['ico'] = entry['ext']
        entry['subdir'] = row[5]
        entry['size'] = row[6]
        first = row[2]
        if not first:
            first = ''
        last = row[3]
        if not last:
            last = ''
        if first != '' and last != '':
            entry['submittername'] = "%s %s" % (first, last)
        entry['first'] = first
        entry['last'] = last
        entry['submitterid'] = row[1]
        entry['trans'] = row[11]
        entry['hash'] = {'sha1': row[10]}
        aged = "false"
        #FIXME make this a bool in elasticsearch later
        if row[8]:
            aged = "true"
        entry['aged'] = aged
        users = {row[1]: 1}
        groups = {}
        proposals = []
        gs = group_sets.get(row[9])
        if gs:
            for g in gs['groups']:
                groups[g[1]] = g[2]
            proposals = gs['proposals']
            for user in gs['users']:
                users[user] = 1
            emd = gs.get('extended_metadata')
            if emd and len(emd) > 0:
                entry['extended_metadata'] = emd
        ds = groups.get('omics.dms.dataset')
        if ds:
            #FIXME generalize this. extended_metadata on metadata specific file entries need to be compied so modifications don't affect others.
            entry['extended_metadata'] = entry['extended_metadata'].copy()
            tsubdir = entry['subdir']
            if tsubdir.startswith(ds + '/'):
                tsubdir = tsubdir[len(ds) + 1:]
            dsid = mdp.emsl_dms_metadata.datasetname2datasetid.get(ds)
            if dsid:
                ajid = mdp.emsl_dms_metadata.datasetid_subdir2analysisjobid(
                    mdp.emsl_dms_metadata.datasetname2datasetid[ds], tsubdir)
                if ajid == None:
                    aj = {'tool': {'name': 'Raw'}}
                else:
                    aj = mdp.emsl_dms_metadata.analysisjob[ajid]
                entry['extended_metadata']['gov_pnnl_emsl_dms_analysisjob'] = [
                    aj
                ]
        if len(proposals) > 0:
            entry['proposals'] = proposals
        if len(groups) > 0:
            for k, v in groups.iteritems():
                #FIXME escape
                entry["groups.%s" % (k)] = v
                if k == 'Instrument' and v.isdigit():
                    tv = instruments.get(int(v))
                    if tv:
                        entry["instrumentname"] = tv
        entry['users'] = users.keys()
        entry['stime'] = rfc3339.rfc3339(row[4])
        entry['_id'] = item_id
        callback(jsonentry(entry))
    return 0