def backfill_flagged_worker(rows):
"""Parallel worker for backfill_flagged_etags()
Expects a list of records to work on and commit as a group
"""
storage = IDigBioStorage()
with apidbpool.connection(autocommit=False) as conn:
cur = conn.cursor()
for row in rows:
try:
table = make_temp_table_name(row['prefix'])
b = storage.get_bucket(
row['ceph_bucket']
) # Two phase here because validate=False in storage class so etag is not populated
row["etag"] = b.get_key(row["ceph_name"]).etag[1:-1]
cur.execute(
"""UPDATE {0}
SET ceph_etag=%(etag)s
WHERE ceph_bucket=%(ceph_bucket)s AND ceph_name=%(ceph_name)s
""".format(table), row)
except:
logger.error(
"Failed to update etag for {0}:{1} in {2} {3}".format(
row["ceph_bucket"], row["ceph_name"], row["prefix"],
traceback.format_exc()))
conn.commit()
return 0
def get_objects_from_ceph():
local_cur.execute("SELECT etag FROM objects")
existing_objects = set()
for r in local_cur:
existing_objects.add(r[0])
print len(existing_objects)
s = IDigBioStorage()
buckets = ["datasets","images"]
count = 0
rowcount = 0
lrc = 0
for b_k in buckets:
b = s.get_bucket("idigbio-" + b_k + "-prod")
for k in b.list():
if k.name not in existing_objects:
try:
ks = k.get_contents_as_string(headers={'Range' : 'bytes=0-100'})
detected_mime = magic.from_buffer(ks, mime=True)
local_cur.execute("INSERT INTO objects (bucket,etag,detected_mime) SELECT %(bucket)s,%(etag)s,%(dm)s WHERE NOT EXISTS (SELECT 1 FROM objects WHERE etag=%(etag)s)", {"bucket": b_k, "etag": k.name, "dm": detected_mime})
existing_objects.add(k.name)
rowcount += local_cur.rowcount
except:
print "Ceph Error", b_k, k.name
count += 1
if rowcount != lrc and rowcount % 10000 == 0:
print count, rowcount
local_pg.commit()
lrc = rowcount
print count, rowcount
local_pg.commit()
def bucket_list_worker(work):
logger.debug("Listing and inserting prefix {0} from bucket {1}".format(
work["prefix"], work["bucket"]))
storage = IDigBioStorage()
# Read through bucket inserting into temp table
with apidbpool.connection(
autocommit=False
) as conn: # use a single connection from the pool to commit groups of statements
cur = conn.cursor()
# inserted = 1
#logger.info("Importing bucket listing for {0}.".format(bucket))
b = storage.get_bucket(work["bucket"])
for f in b.list(prefix=work["prefix"]):
# see backfill_new_etags() for why no etag here
cur.execute(("INSERT INTO {0} "
"(ceph_bucket, ceph_name, ceph_date, ceph_bytes) "
"VALUES (%s, %s, %s, %s)").format(
make_temp_table_name(work["prefix"])),
(work["bucket"], f.name, f.last_modified, f.size))
# inserted += 1
# if (inserted % 10000) == 0:
# logger.info("Committing {0}".format(inserted))
# conn.commit()
conn.commit()
return 1
def get_objects_from_ceph():
local_cur.execute("SELECT etag FROM objects")
existing_objects = set()
for r in local_cur:
existing_objects.add(r[0])
print len(existing_objects)
s = IDigBioStorage()
buckets = ["datasets","images"]
count = 0
rowcount = 0
lrc = 0
for b_k in buckets:
b = s.get_bucket("idigbio-" + b_k + "-prod")
for k in b.list():
if k.name not in existing_objects:
try:
ks = k.get_contents_as_string(headers={'Range' : 'bytes=0-100'})
detected_mime = magic.from_buffer(ks, mime=True)
local_cur.execute("INSERT INTO objects (bucket,etag,detected_mime) SELECT %(bucket)s,%(etag)s,%(dm)s WHERE NOT EXISTS (SELECT 1 FROM objects WHERE etag=%(etag)s)", {"bucket": b_k, "etag": k.name, "dm": detected_mime})
existing_objects.add(k.name)
rowcount += local_cur.rowcount
except:
print "Ceph Error", b_k, k.name
count += 1
if rowcount != lrc and rowcount % 10000 == 0:
print count, rowcount
local_pg.commit()
lrc = rowcount
print count, rowcount
local_pg.commit()
def set_deriv_from_ceph():
s = IDigBioStorage()
b = s.get_bucket("idigbio-images-prod-thumbnail")
count = 0
for k in b.list():
local_cur.execute("UPDATE objects SET derivatives=true WHERE etag=%s", (k.name.split(".")[0],))
count += 1
if count % 10000 == 0:
print count
local_pg.commit()
print count
local_pg.commit()
def set_deriv_from_ceph():
s = IDigBioStorage()
b = s.get_bucket("idigbio-images-prod-thumbnail")
count = 0
for k in b.list():
local_cur.execute("UPDATE objects SET derivatives=true WHERE etag=%s", (k.name.split(".")[0],))
count += 1
if count % 10000 == 0:
print count
local_pg.commit()
print count
local_pg.commit()
def get_objects_from_ceph():
import magic
existing_objects = set(r[0] for r in apidbpool.fetchiter(
"SELECT etag FROM objects", cursor_factory=cursor))
logger.info("Found %d objects", len(existing_objects))
s = IDigBioStorage()
buckets = ["datasets", "images"]
count = 0
rowcount = 0
lrc = 0
with apidbpool.connection() as conn:
with apidbpool.cursor() as cur:
for b_k in buckets:
b = s.get_bucket("idigbio-" + b_k + "-prod")
for k in b.list():
if k.name not in existing_objects:
try:
ks = k.get_contents_as_string(
headers={'Range': 'bytes=0-100'})
detected_mime = magic.from_buffer(ks, mime=True)
cur.execute(
"""INSERT INTO objects (bucket,etag,detected_mime)
SELECT %(bucket)s,%(etag)s,%(dm)s
WHERE NOT EXISTS(
SELECT 1 FROM objects WHERE etag=%(etag)s)""",
{
"bucket": b_k,
"etag": k.name,
"dm": detected_mime
})
existing_objects.add(k.name)
rowcount += cur.rowcount
except:
logger.exception(
"Ceph Error; bucket:%s keyname:%s", b_k,
k.name)
count += 1
if rowcount != lrc and rowcount % 10000 == 0:
logger.info("Count: %8d, rowcount: %8d", count,
rowcount)
conn.commit()
lrc = rowcount
conn.commit()
logger.info("Count: %8d, rowcount: %8d (Finished %s)", count,
rowcount, b_k)
def main():
index_file_name = "index.txt"
query = {
"size": 0,
"aggs": {
"rs": {
"terms": {
"field": "recordset",
"size": 1000
},
"aggs": {
"ic":{
"terms": {
"field": "institutioncode",
"size": 1000,
},
"aggs": {
"cc": {
"terms":{
"field": "collectioncode",
"size": 1000,
}
}
}
}
}
}
}
}
r = requests.post("http://search.idigbio.org/idigbio/records/_search",
data=json.dumps(query),
headers={"Content-Type": "application/json"})
r.raise_for_status()
ro = r.json()
recordsets = {}
for rs_b in ro["aggregations"]["rs"]["buckets"]:
rsid = rs_b["key"]
ic = ""
cc = ""
if len(rs_b["ic"]["buckets"]) == 0:
ic = ""
cc = ""
elif len(rs_b["ic"]["buckets"]) == 1 or (
float(rs_b["ic"]["buckets"][0]["doc_count"]) / float(rs_b["doc_count"]) > 0.9
):
ic_b = rs_b["ic"]["buckets"][0]
ic = get_true_ic(ic_b["key"])
if len(ic_b["cc"]["buckets"]) == 0:
cc = ""
elif len(ic_b["cc"]["buckets"]) == 1:
cc = ic_b["cc"]["buckets"][0]["key"]
else:
cc = "MULTIPLE"
else:
# print(rs_b)
ic = "MULTIPLE"
cc = "MULTIPLE"
recordsets[rsid] = {
"institutioncode": ic,
"collectioncode": cc
}
s = IDigBioStorage()
b = s.get_bucket("idigbio-static-downloads")
headers = ["zipfile","emlfile","etag","modified","recordset_id", "institutioncode", "collectioncode"]
files = {}
for k in b.list():
# Skip the index itself
if k == index_file_name:
continue
# Skip files older than 8 days
lm_d = dateutil.parser.parse(k.last_modified).date()
if lm_d < (datetime.datetime.now() - datetime.timedelta(7)).date():
continue
fkey = k.name.split(".")[0]
if fkey not in files:
files[fkey] = {k:"" for k in headers}
if k.name.endswith(".eml"):
files[fkey]["emlfile"] = k.name
elif k.name.endswith(".zip"):
files[fkey]["zipfile"] = k.name
files[fkey]["modified"] = k.last_modified
files[fkey]["etag"] = k.etag
if is_uuid(fkey):
files[fkey]["recordset_id"] = fkey
if fkey in recordsets:
files[fkey]["institutioncode"] = recordsets[fkey]["institutioncode"]
files[fkey]["collectioncode"] = recordsets[fkey]["collectioncode"]
else:
files[fkey]["institutioncode"] = ""
files[fkey]["collectioncode"] = ""
fil = StringIO()
cw = csv.writer(fil,delimiter="\t")
cw.writerow(headers)
for k in files:
if files[k]["zipfile"] != "":
cw.writerow([files[k][h].replace("\"","") for h in headers])
fil.seek(0)
ik = b.get_key(index_file_name,validate=False)
ik.content_type = 'text/tsv'
ik.set_contents_from_file(fil)
ik.make_public()
