def myopen(path, *args) :
path = get_absolute_path(path)
if not file_exists(path) and file_exists(path + '.gz') :
path = path + '.gz'
if path[-3:] == '.gz' :
import gzip
return gzip.open(path, *args)
elif path[-4:] == '.bgz' :
import bgzf # Make sure sys.path has appropriate dir
return bgzf.open(path, *args)
else :
return open(path, *args)
def read_header(self):
self.fh = bgzf.open(self.filepath)
magic = struct.unpack('i', self.fh.read(4))[0]
l_text = struct.unpack('i', self.fh.read(4))[0]
self.header_text = self.fh.read(l_text)
n_ref = struct.unpack('i', self.fh.read(4))[0]
for tid in range(n_ref):
l_name = struct.unpack('i', self.fh.read(4))[0]
name = self.fh.read(l_name)
name = name[:-1]
l_ref = struct.unpack('i', self.fh.read(4))[0]
self.reference_table[name] = AlignmentReference(tid, name, l_ref)
self.reference_names.append(name)
def loadbgzip(filepath, estimated=1362493):
entries = []
with bgzf.open(filepath, "rb") as fd:
pbar = tqdm(total=estimated)
while True:
pbar.update(1)
sizeBuffer = fd.read(8 * 1)
if (len(sizeBuffer) == 8 * 1):
dataSize = struct.unpack("<Q", sizeBuffer)[0]
data = fd.read(dataSize)
entry = ujson.loads(data.decode("utf8"))
entries.append(entry)
else:
break
return entries
def processTitleYearJournal(bgzPapersFilePath,papersDataTypes):
estimatedCount = 233745561;
index2magID = []
index2Title = []
index2DOI = []
index2Year = []
index2OnlineDate = []
index2JournalID = []
count = 0;
with bgzf.open(bgzPapersFilePath,"rb") as infd:
pbar = tqdm(total=estimatedCount);
while True:
pbar.update(1);
lengthData = infd.read(8);
if(len(lengthData)==8):
length = struct.unpack("<q",lengthData)[0];
else:
break;
data = infd.read(length);
paperIndex = struct.unpack("<q",data[0:8])[0];
currentPointer = 8;
# paperID = struct.unpack("<q",data[8:16])[0];
dataDict = {};
for (typeIndex,(typeName,typeType)) in enumerate(papersDataTypes):
if(typeType==int):
entryValue, = struct.unpack("<q",data[currentPointer:currentPointer+8]);
currentPointer+=8;
dataDict[typeName] = entryValue;
else:
entryLength, = struct.unpack("<q",data[currentPointer:currentPointer+8]);
currentPointer+=8;
entryData = data[currentPointer:currentPointer+entryLength].decode("utf8");
currentPointer+=entryLength;
dataDict[typeName] = entryData;
index2Title.append(dataDict["PaperTitle"])
index2Year.append(dataDict["Year"])
index2magID.append(dataDict["PaperId"])
index2DOI.append(dataDict["Doi"])
index2OnlineDate.append(dataDict["OnlineDate"])
index2JournalID.append(dataDict["JournalId"])
count+=1;
# if(count>100000):
# break;
return index2Title,index2magID,index2Year,index2OnlineDate,index2JournalID,index2DOI;
# setOfJournalsWithPublications = set()
HCACitationsPerYear = {}
HBMapCitationsPerYear = {}
HCAHuBMapCitationsPerYear = {}
HGPCitationsPerYear = {}
projectCitations = {
"HCA":(HCAProjectMAGIndices,HCACitationsPerYear),
"HuBMAP":(HBMapProjectMAGIndices,HBMapCitationsPerYear),
"HuBMAP+HCA":(HBMapProjectMAGIndices.union(HCAProjectMAGIndices),HCAHuBMapCitationsPerYear),
"HGP":(HGPProjectMAGIndices,HGPCitationsPerYear)
}
MAGID2CitationYear = {}
with bgzf.open(bgzReferencesFromToPath,"rb") as infd:
pbar = tqdm(total=estimatedCount)
count = 0
while True:
pbar.update(1)
lengthData = infd.read(8)
if(len(lengthData)==8):
length = struct.unpack("<q",lengthData)[0]
else:
break
data = infd.read(length)
edgesCount = length/8-1
fmt = "<%dq" % (edgesCount+1)
edgeData = list(struct.unpack(fmt,data))
fromIndex = edgeData[0]
fromID = index2magID[fromIndex]
def savebgzip(filepath, entries):
with bgzf.open(filepath, "wb") as fd:
for entry in tqdm(entries):
data = ujson.dumps(entry).encode("utf8")
fd.write(struct.pack("<Q", len(data)))
fd.write(data)
np.savetxt(PJ(dataPath, "MAGJournalPaperCounts.txt.gz"), journalPaperCounts,
"%d")
np.savetxt(PJ(dataPath, "MAGJournalISSN.txt.gz"), journalISSN, "%s")
print("Total With ISSN: %.0f%%" % (totalISSNCount / totalBigCount * 100))
print("Final journals Count: %d" % len(index2JournalID))
print("\n Reading and Saving Paper 2 Journal Information")
estimatedCount = 242387326
# Used just to display status
paper2journalYear = {}
paper2journalYearBGZPath = PJ(dataPath, "MAGPaper2JournalYear.bgz")
if (not rebuildAll and os.path.exists(paper2journalYearBGZPath)):
print("Reading from BGZ file...")
pbar = tqdm(total=estimatedCount)
with bgzf.open(paper2journalYearBGZPath, "rb") as fd:
while True:
pbar.update(1)
data = fd.read(8 * 3)
if (len(data) == 8 * 3):
paperID, journal, year = struct.unpack("<QQQ", data)
paper2journalYear[paperID] = (journal, year)
else:
break
else:
print("Reading from RAW file...")
with open(PJ(MAGPath, "Papers.txt"), "rt") as fd:
pbar = tqdm(total=estimatedCount)
for line in fd:
pbar.update(1)
entries = line.split("\t")
