def copyfileobj(fsrc, fdst, length=16*1024, advise_after=1024*1024):
"""
Reimplementation of shutil.copyfileobj that advises the OS to remove
parts of the source file from the OS's caches once copied to the
destination file.
Usage profile:
* You have a (potentially) large file to copy.
* You know you don't need to access the source file once copied.
* You're quite likely to access the destination file soon after.
"""
# If we can't access the the fileno then fallback to using shutil.
if not hasattr(fsrc, 'fileno'):
return shutil.copyfileobj(fsrc, fdst, length)
# Calculate the appoximate number of blocks to copy before advising the
# OS to drop pages from the cache.
advise_after_blocks = int(advise_after/length)
# Off we go ...
blocks_read = 0
while True:
data = fsrc.read(length)
if not data:
break
fdst.write(data)
blocks_read += 1
if not blocks_read % advise_after_blocks:
posix_fadvise(fsrc.fileno(), 0, length*blocks_read,
POSIX_FADV_DONTNEED)
# One final advise to flush the remaining blocks.
posix_fadvise(fsrc.fileno(), 0, 0, POSIX_FADV_DONTNEED)
def copyfileobj(fsrc, fdst, length=16 * 1024, advise_after=1024 * 1024):
"""
Reimplementation of shutil.copyfileobj that advises the OS to remove
parts of the source file from the OS's caches once copied to the
destination file.
Usage profile:
* You have a (potentially) large file to copy.
* You know you don't need to access the source file once copied.
* You're quite likely to access the destination file soon after.
"""
# If we can't access the fileno then fallback to using shutil.
if not hasattr(fsrc, 'fileno'):
return shutil.copyfileobj(fsrc, fdst, length)
# Calculate the appoximate number of blocks to copy before advising the
# OS to drop pages from the cache.
advise_after_blocks = int(advise_after / length)
# Off we go ...
blocks_read = 0
while 1:
data = fsrc.read(length)
if not data:
break
fdst.write(data)
blocks_read += 1
if not blocks_read % advise_after_blocks:
posix_fadvise(fsrc.fileno(), 0, length * blocks_read,
POSIX_FADV_DONTNEED)
# One final advise to flush the remaining blocks.
posix_fadvise(fsrc.fileno(), 0, 0, POSIX_FADV_DONTNEED)
def update_many(path, points):
"""update_many(path,points)
path is a string
points is a list of (timestamp,value) points
"""
if not points: return
points = [(int(t), float(v)) for (t, v) in points]
points.sort(key=lambda p: p[0], reverse=True) # Order points by timestamp, newest first
with open(path, 'r+b') as fh:
if CAN_FADVISE and FADVISE_RANDOM:
posix_fadvise(fh.fileno(), 0, 0, POSIX_FADV_RANDOM)
return file_update_many(fh, points)
def update(path, value, timestamp=None):
"""
update(path, value, timestamp=None)
path is a string
value is a float
timestamp is either an int or float
"""
value = float(value)
with open(path, 'r+b') as fh:
if CAN_FADVISE and FADVISE_RANDOM:
posix_fadvise(fh.fileno(), 0, 0, POSIX_FADV_RANDOM)
return file_update(fh, value, timestamp)
def main():
orig_file = sys.argv[1]
new_model = sys.argv[2]
if len(sys.argv) > 3:
start_chunk = int(sys.argv[3])
forig = open(orig_file)
prev_pos = 0
new_num = 0
for line in sys.stdin:
pos = int(line.strip())
diff = pos - prev_pos
assert diff >= 0
if new_num >= start_chunk:
filename = new_model % (new_num)
with open(new_model % (new_num), 'w') as fdest:
fadvise.posix_fadvise(forig.fileno(), 0, forig.tell(), fadvise.POSIX_FADV_DONTNEED)
fadvise.posix_fadvise(forig.fileno(), forig.tell(), 0, fadvise.POSIX_FADV_SEQUENTIAL)
fadvise.posix_fadvise(fdest.fileno(), 0, 0, fadvise.POSIX_FADV_SEQUENTIAL)
copy(forig, fdest, diff)
else:
forig.seek(diff, os.SEEK_CUR)
prev_pos = pos
new_num += 1
# Copy last chunk
with open(new_model % (new_num), 'w') as fdest:
copy(forig, fdest, 0)
def file_contents(fn, offset=None, length=None, binary=False):
"""Return the entire contents of a file, or of a specified segment. Open
the file in binary mode if specified.
>>> file_contents('/dev/null')
''
"""
mode = 'r' if not binary else 'rb'
with open(fn, mode) as f:
if offset is not None:
return file_block_at_offset(f, offset, length)
if length is not None:
s = f.read(length)
assert len(s) == length
return s
s = f.read()
# tell the OS it can discard these pages, because the aggregate size
# of the spectrum files may be huge, and we want to avoid pushing more
# important stuff out of memory
if fn.endswith('.ms2'):
posix_fadvise(f.fileno(), 0, 0, POSIX_FADV_DONTNEED)
return s
def create(path, archiveList, xFilesFactor=None, aggregationMethod=None, sparse=False, useFallocate=False):
"""create(path,archiveList,xFilesFactor=0.5,aggregationMethod='average')
path is a string
archiveList is a list of archives, each of which is of the form (secondsPerPoint,numberOfPoints)
xFilesFactor specifies the fraction of data points in a propagation interval that must have known values for a propagation to occur
aggregationMethod specifies the function to use when propagating data (see ``whisper.aggregationMethods``)
"""
# Set default params
if xFilesFactor is None:
xFilesFactor = 0.5
if aggregationMethod is None:
aggregationMethod = 'average'
# Validate archive configurations...
validateArchiveList(archiveList)
# Looks good, now we create the file and write the header
if os.path.exists(path):
raise InvalidConfiguration("File %s already exists!" % path)
with open(path, 'wb') as fh:
try:
if LOCK:
fcntl.flock(fh.fileno(), fcntl.LOCK_EX)
if CAN_FADVISE and FADVISE_RANDOM:
posix_fadvise(fh.fileno(), 0, 0, POSIX_FADV_RANDOM)
aggregationType = struct.pack(longFormat, aggregationMethodToType.get(aggregationMethod, 1))
oldest = max([secondsPerPoint * points for secondsPerPoint, points in archiveList])
maxRetention = struct.pack(longFormat, oldest)
xFilesFactor = struct.pack(floatFormat, float(xFilesFactor))
archiveCount = struct.pack(longFormat, len(archiveList))
packedMetadata = aggregationType + maxRetention + xFilesFactor + archiveCount
fh.write(packedMetadata)
headerSize = metadataSize + (archiveInfoSize * len(archiveList))
archiveOffsetPointer = headerSize
for secondsPerPoint, points in archiveList:
archiveInfo = struct.pack(archiveInfoFormat, archiveOffsetPointer, secondsPerPoint, points)
fh.write(archiveInfo)
archiveOffsetPointer += (points * pointSize)
# If configured to use fallocate and capable of fallocate use that, else
# attempt sparse if configure or zero pre-allocate if sparse isn't configured.
if CAN_FALLOCATE and useFallocate:
remaining = archiveOffsetPointer - headerSize
fallocate(fh, headerSize, remaining)
elif sparse:
fh.seek(archiveOffsetPointer - 1)
fh.write(b'\x00')
else:
remaining = archiveOffsetPointer - headerSize
chunksize = 16384
zeroes = b'\x00' * chunksize
while remaining > chunksize:
fh.write(zeroes)
remaining -= chunksize
fh.write(zeroes[:remaining])
if AUTOFLUSH:
fh.flush()
os.fsync(fh.fileno())
# Explicitly close the file to catch IOError on close()
fh.close()
except IOError:
# if we got an IOError above, the file is either empty or half created.
# Better off deleting it to avoid surprises later
os.unlink(fh.name)
raise
