def access_precomputed(
store_path: str,
key: str,
mode: str,
array_type=None,
dtype=None,
num_channels=None,
shape=None,
resolution=None,
encoding=None,
chunks=None,
jpeg_quality=None,
voxel_offset=None,
scale_index=None,
) -> TensorStoreArray:
driver = "neuroglancer_precomputed"
kvstore_driver, _store_path = fsspec.core.split_protocol(store_path)
if kvstore_driver == None:
kvstore_driver = "file"
kvstore_path = "/"
# remove the leading slash after making the absolute path
_store_path = os.path.abspath(_store_path)[1:]
else:
kvstore_path = _store_path.split(os.path.sep)[0]
if kvstore_driver not in KVSTORE_DRIVERS:
raise ValueError(
f"File system protocol {kvstore_driver} is not supported by tensorstore."
)
info_path = os.path.join(store_path, "info")
if mode == 'r':
with fsspec.open(info_path) as fh:
json_data = json.loads(fh.read())
precomputed_metadata = parse_info(json_data)
scale_matches = [
scale.key == key for scale in precomputed_metadata.scales
]
if not any(scale_matches):
raise ValueError(
'Could not find key: {key} in info file at {info_path}')
else:
scale_index = scale_matches.index(True)
scale_meta = precomputed_metadata.scales[scale_index]
else:
scale_meta = ScaleMetadata(
size=shape,
resolution=resolution,
encoding=encoding,
chunk_size=chunks,
key=key,
voxel_offset=voxel_offset,
jpeg_quality=jpeg_quality,
)
precomputed_metadata = PrecomputedMetadata(type=array_type,
data_type=dtype,
num_channels=num_channels,
scales=[scale_meta])
if mode == "r":
read = True
# So cool that tensorstore errors when these are set to False for reading...
write = None
create = None
delete_existing = None
elif mode == 'a':
read = True
write = True
create = True
delete_existing = False
elif mode == "rw":
read = True
write = True
create = True
delete_existing = True
elif mode == "w":
read = False
write = True
create = True
delete_existing = True
elif mode == "w-":
read = False
write = True
create = True
delete_existing = False
else:
raise ValueError('Mode must be "r", "rw", "a", "w", or "w-"')
tsa = TensorStoreArray(driver=driver,
path=_store_path,
kvstore_path=kvstore_path,
kvstore_driver=kvstore_driver,
encoding=scale_meta.encoding,
scale_index=scale_index,
key=scale_meta.key,
num_channels=precomputed_metadata.num_channels,
volume_type=precomputed_metadata.type,
dtype=precomputed_metadata.data_type,
resolution=scale_meta.resolution,
size=scale_meta.size,
chunk_size=scale_meta.chunk_size,
jpeg_quality=jpeg_quality)
return tsa.open(read=read,
write=write,
create=create,
delete_existing=delete_existing).result()
def input_opener(fname, **kwargs):
logger.info(f"Opening input '{fname}'")
with fsspec.open(fname, **kwargs) as f:
yield f
def read_text_from_href(self, href: str, *args: Any, **kwargs: Any) -> str:
with fsspec.open(href, "r") as f:
return f.read()
def read_bytes(
vineyard_socket: str,
path: str,
storage_options: Dict,
read_options: Dict,
proc_num: int,
proc_index: int,
):
"""Read bytes from external storage and produce a ByteStream,
which will later be assembled into a ParallelStream.
Args:
vineyard_socket (str): Ipc socket
path (str): External storage path to write to
storage_options (dict): Configurations of external storage
read_options (dict): Additional options that could control the behavior of read
proc_num (int): Total amount of process
proc_index (int): The sequence of this process
Raises:
ValueError: If the stream is invalid.
"""
client = vineyard.connect(vineyard_socket)
builder = ByteStreamBuilder(client)
serialization_mode = read_options.pop('serialization_mode', False)
if serialization_mode:
parsed = urlparse(path)
try:
fs = fsspec.filesystem(parsed.scheme)
except ValueError as e:
report_status("error", str(e))
raise
meta_file = f"{path}_{proc_index}.meta"
blob_file = f"{path}_{proc_index}"
if not fs.exists(meta_file) or not fs.exists(blob_file):
report_status("error", f"Some serialization file cannot be found. Expected: {meta_file} and {blob_file}")
raise FileNotFoundError('{}, {}'.format(meta_file, blob_file))
# Used for read bytes of serialized graph
meta_file = fsspec.open(meta_file, mode="rb", **storage_options)
with meta_file as f:
meta = f.read().decode('utf-8')
meta = json.loads(meta)
lengths = meta.pop("lengths")
for k, v in meta.items():
builder[k] = v
stream = builder.seal(client)
client.persist(stream)
ret = {"type": "return", "content": repr(stream.id)}
print(json.dumps(ret), flush=True)
writer = stream.open_writer(client)
of = fsspec.open(blob_file, mode="rb", **storage_options)
with of as f:
try:
total_size = f.size()
except TypeError:
total_size = f.size
assert total_size == sum(lengths), "Target file is corrupted"
for length in lengths:
buf = f.read(length)
chunk = writer.next(length)
buf_writer = pa.FixedSizeBufferWriter(chunk)
buf_writer.write(buf)
buf_writer.close()
writer.finish()
else:
# Used when reading tables from external storage.
# Usually for load a property graph
header_row = read_options.get("header_row", False)
for k, v in read_options.items():
if k in ("header_row", "include_all_columns"):
builder[k] = "1" if v else "0"
elif k == "delimiter":
builder[k] = bytes(v, "utf-8").decode("unicode_escape")
else:
builder[k] = v
offset = 0
chunk_size = 1024 * 1024 * 4
try:
of = fsspec.open(path, mode="rb", **storage_options)
except Exception as e:
report_status("error", str(e))
raise
with of as f:
header_line = read_block(f, 0, 1, b'\n')
builder["header_line"] = header_line.decode("unicode_escape")
if header_row:
offset = len(header_line)
stream = builder.seal(client)
client.persist(stream)
ret = {"type": "return", "content": repr(stream.id)}
print(json.dumps(ret), flush=True)
writer = stream.open_writer(client)
try:
total_size = f.size()
except TypeError:
total_size = f.size
part_size = (total_size - offset) // proc_num
begin = part_size * proc_index + offset
end = min(begin + part_size, total_size)
if proc_index == 0:
begin -= int(header_row)
while begin < end:
buf = read_block(f, begin, min(chunk_size, end - begin), delimiter=b"\n")
size = len(buf)
if not size:
break
begin += size - 1
chunk = writer.next(size)
buf_writer = pa.FixedSizeBufferWriter(chunk)
buf_writer.write(buf)
buf_writer.close()
writer.finish()
def from_json(cls, fname: Union[str, Path], open_kwargs: dict = {}):
with fsspec.open(str(fname), mode='rt', **open_kwargs) as fh:
jblob = json.loads(fh.read())
return cls(**jblob)
def save_speaker_mapping(out_path, speaker_mapping):
"""Saves speaker mapping if not yet present."""
if out_path is not None:
speakers_json_path = _set_file_path(out_path)
with fsspec.open(speakers_json_path, "w") as f:
json.dump(speaker_mapping, f, indent=4)
def _save_json(json_file_path: str, data: dict) -> None:
with fsspec.open(json_file_path, "w") as f:
json.dump(data, f, indent=4)
def inspect(self, dataset, columns_dict, output_file):
"""
Parameters
-----------
path: str, list of str, or <dask.dataframe|cudf|pd>.DataFrame
Dataset path (or list of paths), or a DataFrame. If string,
should specify a specific file or directory path. If this is a
directory path, the directory structure must be flat (nested
directories are not yet supported).
dataset_format: str
Dataset format (i.e parquet or csv)
columns_dict: dict
Dictionary indicating the different columns type
output_file: str
Filename to write the output statistics
"""
# Get dataset columns
cats = columns_dict["cats"]
conts = columns_dict["conts"]
labels = columns_dict["labels"]
# Create Dataset, Workflow, and get Stats
stats = DataStats()
features = ColumnSelector(cats + conts + labels) >> stats
workflow = Workflow(features, client=self.client)
workflow.fit(dataset)
# get statistics from the datastats op
output = stats.output
# Dictionary to store collected information
data = {}
# Store num_rows
data["num_rows"] = dataset.num_rows
# Store cols
for col_type in ["conts", "cats", "labels"]:
data[col_type] = {}
for col in columns_dict[col_type]:
data[col_type][col] = {}
data[col_type][col]["dtype"] = output[col]["dtype"]
if col_type != "conts":
data[col_type][col]["cardinality"] = output[col][
"cardinality"]
if col_type == "cats":
data[col_type][col]["min_entry_size"] = output[col]["min"]
data[col_type][col]["max_entry_size"] = output[col]["max"]
data[col_type][col]["avg_entry_size"] = output[col]["mean"]
elif col_type == "conts":
data[col_type][col]["min_val"] = output[col]["min"]
data[col_type][col]["max_val"] = output[col]["max"]
data[col_type][col]["mean"] = output[col]["mean"]
data[col_type][col]["std"] = output[col]["std"]
data[col_type][col]["per_nan"] = output[col]["per_nan"]
# Write json file
with fsspec.open(output_file, "w") as outfile:
json.dump(data, outfile, cls=NpEncoder)
def test_automkdir_readonly(tmpdir):
dir = os.path.join(str(tmpdir), "d")
with pytest.raises(FileNotFoundError):
of = fsspec.open(os.path.join(dir, "dfile"), "r")
with of:
pass
def open_file(url, *args, **kwargs):
of = fsspec.open(url, *args, **kwargs)
with of as f:
yield f
N_EXAMPLES = 2
if __name__ == "__main__":
for i, config in enumerate(Oscar.BUILDER_CONFIGS):
print(f"Loading config '{config.name}' ({i + 1}/{len(Oscar.BUILDER_CONFIGS)})")
# Get data url
checksum_filename = _BASE_CHECKSUM_FILE_NAME.format(language=config.language)
checksum_url = config.base_data_url + checksum_filename
checksum_file_content = requests.get(checksum_url).text.splitlines()
data_filename = checksum_file_content[0].split("\t")[0]
data_url = config.base_data_url + data_filename
# Get a few examples
with fs.open(data_url, "rt", compression="gzip") as f:
current_examples = 0
dummy_content = []
for line in f:
dummy_content.append(line)
current_examples += len(line.strip()) == 0
if current_examples == N_EXAMPLES:
break
dummy_content = "".join(dummy_content).rstrip()
# Write dummy files
dummy_data_dir = Path(__file__).resolve().parent / "dummy" / config.name / str(config.version) / "dummy_data"
dummy_data_dir.mkdir(parents=True, exist_ok=True)
(dummy_data_dir / checksum_filename).open("w").write(data_filename + "\t insert_hash_here")
with fs.open(str(dummy_data_dir / data_filename), "wt", compression="gzip") as f:
f.write(dummy_content)
def to_image(
mols: Union[List[Chem.rdchem.Mol], Chem.rdchem.Mol],
legends: Union[List[Union[str, None]], str, None] = None,
n_cols: int = 4,
use_svg: bool = False,
mol_size: Union[Tuple[int, int], int] = (200, 200),
highlight_atom: List[List[int]] = None,
highlight_bond: List[List[int]] = None,
outfile: str = None,
max_mols: int = 32,
copy: bool = False,
indices: bool = False,
):
"""Generate an image out of a molecule or a list of molecule.
Args:
mols: one or a list of molecules.
legends: a string or a list of string as legend for every molecules.
n_cols: number of molecules per column.
use_svg: whether to ouput an SVG (or a PNG).
mol_size: a int or a tuple of int defining the size per molecule.
highlight_atom: atom to highlight.
highlight_bond: bonds to highlight.
outfile: path where to save the image (local or remote path).
max_mols: the maximum number of molecules to display.
copy: whether to copy the molecules or not.
indices: Whether to draw the atom indices.
"""
if isinstance(mol_size, int):
mol_size = (mol_size, mol_size)
if isinstance(mols, Chem.rdchem.Mol):
mols = [mols]
if isinstance(legends, str):
legends = [legends]
if copy:
mols = [dm.copy_mol(mol) for mol in mols]
if max_mols is not None:
mols = mols[:max_mols]
if legends is not None:
legends = legends[:max_mols]
if indices is True:
[dm.atom_indices_to_mol(mol) for mol in mols]
_highlight_atom = highlight_atom
if highlight_atom is not None and isinstance(highlight_atom[0], int):
_highlight_atom = [highlight_atom]
_highlight_bond = highlight_bond
if highlight_bond is not None and isinstance(highlight_bond[0], int):
_highlight_bond = [highlight_bond]
# Don't make the image bigger than it
if len(mols) < n_cols:
n_cols = len(mols)
image = Draw.MolsToGridImage(
mols,
legends=legends,
molsPerRow=n_cols,
useSVG=use_svg,
subImgSize=mol_size,
highlightAtomLists=_highlight_atom,
highlightBondLists=_highlight_bond,
)
if outfile is not None:
with fsspec.open(outfile, "wb") as f:
if use_svg:
if isinstance(image, str):
# in a terminal process
f.write(image.encode())
else:
# in a jupyter kernel process
f.write(image.data.encode()) # type: ignore
else:
if isinstance(image,
PIL.PngImagePlugin.PngImageFile): # type: ignore
# in a terminal process
image.save(f)
else:
# in a jupyter kernel process
f.write(image.data) # type: ignore
return image
def makedirs(url, exist_ok=False):
fs, path = get_fs_and_path(url)
fs.makedirs(path, exist_ok=exist_ok)
if not path_exists(url):
with fsspec.open(url, mode="wb"):
pass
def read_file(path: str, **kwargs):
"""Support fetching files from arbitrary filesystems
"""
with fsspec.open(path, **kwargs) as f:
content = f.read()
return content
def _create_checkpoint(checkpoint_stream, global_step: int,
filename: str):
with tune.checkpoint_dir(step=global_step) as checkpoint_dir:
file_path = os.path.join(checkpoint_dir, filename)
with fsspec.open(file_path, "wb") as f:
f.write(checkpoint_stream)
def upload(entry, parquet_dir, url):
of = fsspec.open(url)
if of.fs.exists(url):
of.fs.delete(url, recursive=True)
entry.fs.upload(parquet_dir, url, recursive=True)
return True
],
'pressure_level': [
'250', '500', '700',
'850', '925', '1000',
],
'year': date.year,
'month': date.month,
'day': date.day,
'time': [
'00:00', '12:00',
],
'area' : [75.,185, 15., 320.] #N,W,S,E
}
r1 = c1.retrieve(name, request, None)
with fsspec.open(r1.location) as f1:
iso_vars = xr.open_dataset(f1, engine='scipy')
# print(iso_vars)
print('getting single-level data')
## also need MSLP & PW
## get the data from ECMWF API
c2 = cdsapi.Client()
name = 'reanalysis-era5-single-levels'
request = {
'product_type': 'reanalysis',
'format': 'netcdf',
'variable': ['mean_sea_level_pressure','total_column_water_vapour'],
'year': date.year,
'month': date.month,
def _fsspec_safe_open(fname, **kwargs):
# workaround for inconsistent behavior of fsspec.open
# https://github.com/intake/filesystem_spec/issues/579
with fsspec.open(fname, **kwargs) as fp:
with fp as fp2:
yield fp2
def _load_json(json_file_path: str) -> Dict:
with fsspec.open(json_file_path, "r") as f:
return json.load(f)
def _get_url_size(fname):
with fsspec.open(fname, mode="rb") as of:
size = of.size
return size
# tests for pillow-8.1.2-py38ha0e1e83_1 (this is a generated file);
print('===== testing package: pillow-8.1.2-py38ha0e1e83_1 =====')
print('running run_test.py')
# --- run_test.py (begin) ---
import fsspec
from PIL import Image
# Test JPEG2k
with fsspec.open("https://www.fnordware.com/j2k/relax.jp2") as f:
image = Image.open(f)
image.load()
# --- run_test.py (end) ---
print('===== pillow-8.1.2-py38ha0e1e83_1 OK =====')
print("import: 'PIL'")
import PIL
print("import: 'PIL.Image'")
import PIL.Image
print("import: 'PIL.ImageCms'")
import PIL.ImageCms
def __init__(self, filename: str, hdf5group: str = None, hdf5file_mode: str = 'r',
store: Union[MutableMapping, str, Path] = None, store_path: str = None,
store_mode: str = 'a', LRU: bool = False, LRU_max_size: int = 2**30,
max_chunksize=2*2**20):
"""
Args:
filename: str or File-like object, file name string or File-like object to be read by zarr
hdf5group: str, hdf5 group in hdf5 file to be read by zarr
along with its children. default is the root group.
hdf5file_mode str, subset of h5py file access modes, filename must exist
'r' readonly, default 'r'
'r+' read and write
store: collections.abc.MutableMapping or str, zarr store.
if string path is passed, zarr.DirectoryStore
is created at the given path, if None, zarr.MemoryStore is used
store_mode: store data access mode, default 'a'
'r' readonly, compatible zarr hierarchy should
already exist in the passed store
'r+' read and write, return error if file does not exist,
for updating zarr hierarchy
'w' create store, remove data if it exists
'w-' or 'x' create store, fail if exists
'a' read and write, create if it does not exist, default 'r'
store_path: string, path in zarr store
LRU: bool, if store is not already zarr.LRUStoreCache, add
a zarr.LRUStoreCache store layer on top of currently used store
LRU_max_size: int, maximum zarr.LRUStoreCache cache size, only used
if store is zarr.LRUStoreCache, or LRU argument is True
max_chunksize: maximum chunk size to use when creating zarr hierarchy, this is useful if
only a small slice of data needs to be read
"""
# Verify arguments
if hdf5file_mode not in ('r', 'r+'):
raise ValueError("hdf5file_mode must be 'r' or 'r+'")
self.hdf5file_mode = hdf5file_mode
# Verify arguments
if not isinstance(LRU, bool):
raise TypeError(f"Expected bool for LRU, received {type(LRU)}")
self.LRU = LRU
if not isinstance(LRU_max_size, int):
raise TypeError(f"Expected int for LRU_max_size, received {type(LRU_max_size)}")
self.LRU_max_size = LRU_max_size
if not isinstance(max_chunksize, int):
raise TypeError(f"Expected int for max_chunksize, received {type(max_chunksize)}")
self.max_chunksize = max_chunksize
# store, store_path, and store_mode are passed through to zarr
self.store_path = store_path
self.store_mode = store_mode
if store is not None and LRU is True and not isinstance(store, zarr.LRUStoreCache):
self.store = zarr.LRUStoreCache(store, max_size=self.LRU_max_size)
else:
self.store = store
# create dictionary mapping hdf5 filter numbers to compatible zarr codec
self._hdf5_regfilters_subset = {}
self._fill_regfilters()
# dictionary to hold addresses of hdf5 objects in file
self._address_dict = {}
# create zarr format hierarchy for datasets and attributes compatible with hdf5 file,
# dataset contents are not copied, unless it contains variable-length strings
self.zgroup = zarr.open_group(self.store, mode=self.store_mode, path=self.store_path)
if self.store is None:
self.store = self.zgroup.store
# FileChunkStore requires uri
if isinstance(filename, str):
self.uri = filename
else:
try:
self.uri = getattr(filename, 'path', None)
if self.uri is None:
self.uri = filename.name
except:
self.uri = ''
# Access hdf5 file and create zarr hierarchy
if hdf5group is not None and not isinstance(hdf5group, str):
raise TypeError(f"Expected str for hdf5group, recieved {type(hdf5group)}")
self.hdf5group = hdf5group
self.filename = filename
if self.store_mode != 'r':
self.file = h5py.File(self.filename, mode=self.hdf5file_mode)
self.group = self.file[self.hdf5group] if self.hdf5group is not None else self.file
self.create_zarr_hierarchy(self.group, self.zgroup)
self.file.close()
if isinstance(self.filename, str):
self.chunkstore_file = fsspec.open(self.filename, mode='rb')
self.chunk_store = FileChunkStore(self.store, chunk_source=self.chunkstore_file.open())
else:
self.chunk_store = FileChunkStore(self.store, chunk_source=self.filename)
if LRU is True and not isinstance(self.chunk_store, zarr.LRUStoreCache):
self.chunk_store = zarr.LRUStoreCache(self.chunk_store, max_size=self.LRU_max_size)
# open zarr group
store_mode_cons = 'r' if self.store_mode == 'r' else 'r+'
self.zgroup = zarr.open_group(self.store, mode=store_mode_cons, path=self.store_path, chunk_store=self.chunk_store)
def open_gzip(path: PathType, storage_options: Optional[Dict[str, str]]) -> IO[Any]:
url = str(path)
storage_options = storage_options or {}
openfile: IO[Any] = fsspec.open(url, compression="gzip", **storage_options)
return openfile
def generate_wb_timeseries(shapes, config_dict):
"""
This is where the code processing is actually done. This code takes in a
polygon, and the and a config dict which contains: shapefile's crs, output
directory, id_field, time_span, and include_uncertainty which says whether
to include all data as well as an invalid pixel count which can be used
for measuring uncertainty performs a polygon drill into the wofs_albers
product. The resulting xarray, which contains the water classified pixels
for that polygon over every available timestep, is used to calculate the
percentage of the water body that is wet at each time step. The outputs
are written to a csv file named using the polygon UID, which is a geohash
of the polygon's centre coords.
Inputs:
shapes - polygon to be interrogated
config_dict - many config settings including crs, id_field, time_span,
shapefile
Outputs:
Nothing is returned from the function, but a csv file is written out to
disk
"""
output_dir = config_dict['output_dir']
crs = config_dict['crs']
id_field = config_dict['id_field']
time_span = config_dict['time_span']
include_uncertainty = config_dict['include_uncertainty']
wofls = config_dict['wofls']
assert wofls
# Some query parameters will be different for different WOfL products.
output_res = get_resolution(wofls)
dataset_maturity = get_dataset_maturity(wofls)
if include_uncertainty:
unknown_percent_threshold = 100
else:
unknown_percent_threshold = 10
with Datacube(app='Polygon drill') as dc:
first_geometry = shapes['geometry']
str_poly_name = shapes['properties'][id_field]
try:
fpath = os.path.join(output_dir,
f'{str_poly_name[0:4]}/{str_poly_name}.csv')
except TypeError:
str_poly_name = str(int(str_poly_name)).zfill(6)
fpath = os.path.join(output_dir,
f'{str_poly_name[0:4]}/{str_poly_name}.csv')
geom = geometry.Geometry(first_geometry, crs=crs)
current_year = datetime.now().year
if time_span == 'ALL':
if shapely_geom.shape(first_geometry).envelope.area > 2000000:
years = range(1986, current_year + 1, 5)
time_periods = [(str(year), str(year + 4)) for year in years]
else:
time_periods = [('1986', str(current_year))]
elif time_span == 'APPEND':
start_date = get_last_date(fpath)
if start_date is None:
logger.debug(f'There is no csv for {str_poly_name}')
return 1
time_periods = [(start_date, str(current_year))]
elif time_span == 'CUSTOM':
time_periods = [(config_dict['start_dt'], config_dict['end_date'])]
valid_capacity_pc = []
valid_capacity_ct = []
invalid_capacity_ct = []
date_list = []
for time in time_periods:
wb_capacity_pc = []
wb_capacity_ct = []
wb_invalid_ct = []
dry_observed = []
invalid_observations = []
# Set up the query, and load in all of the WOFS layers
query = {
'geopolygon': geom,
'time': time,
'output_crs': crs,
'resolution': output_res,
'resampling': 'nearest'
}
if dataset_maturity:
query['dataset_maturity'] = dataset_maturity
logger.debug('Query: {}'.format(
{k: v
for k, v in query.items() if k != 'geopolygon'}))
wofl = dc.load(product=wofls,
group_by='solar_day',
fuse_func=wofls_fuser,
**query)
if len(wofl.attrs) == 0:
logger.debug(
f'There is no new data for {str_poly_name} in {time}')
# TODO(MatthewJA): Confirm (with Ness?) that changing this
# return to a continue doesn't break things.
continue
# Make a mask based on the polygon (to remove extra data
# outside of the polygon)
mask = rasterio.features.geometry_mask(
[geom.to_crs(wofl.geobox.crs) for geoms in [geom]],
out_shape=wofl.geobox.shape,
transform=wofl.geobox.affine,
all_touched=False,
invert=True)
# mask the data to the shape of the polygon
# the geometry width and height must both be larger than one pixel
# to mask.
if (geom.boundingbox.width > 25.3
and geom.boundingbox.height > 25.3):
wofl_masked = wofl.water.where(mask)
else:
wofl_masked = wofl.water
# Work out how full the waterbody is at every time step
for ix, times in enumerate(wofl.time):
# Grab the data for our timestep
all_the_bit_flags = wofl_masked.isel(time=ix)
# Find all the wet/dry pixels for that timestep
lsa_wet = all_the_bit_flags.where(
all_the_bit_flags == 136).count().item()
lsa_dry = all_the_bit_flags.where(
all_the_bit_flags == 8).count().item()
sea_wet = all_the_bit_flags.where(
all_the_bit_flags == 132).count().item()
sea_dry = all_the_bit_flags.where(
all_the_bit_flags == 4).count().item()
sea_lsa_wet = all_the_bit_flags.where(
all_the_bit_flags == 140).count().item()
sea_lsa_dry = all_the_bit_flags.where(
all_the_bit_flags == 12).count().item()
wet_pixels = (all_the_bit_flags.where(
all_the_bit_flags == 128).count().item() + lsa_wet +
sea_wet + sea_lsa_wet)
dry_pixels = (all_the_bit_flags.where(
all_the_bit_flags == 0).count().item() + lsa_dry +
sea_dry + sea_lsa_dry)
# Count the number of masked observations
masked_all = all_the_bit_flags.count().item()
# Turn our counts into percents
try:
water_percent = round((wet_pixels / masked_all * 100), 1)
dry_percent = round((dry_pixels / masked_all * 100), 1)
missing_pixels = masked_all - (wet_pixels + dry_pixels)
unknown_percent = missing_pixels / masked_all * 100
except ZeroDivisionError:
water_percent = 0.0
dry_percent = 0.0
unknown_percent = 100.0
missing_pixels = masked_all
logger.debug(f'{str_poly_name} has divide by zero error')
# Append the percentages to a list for each timestep
# Filter out timesteps with < 90% valid observations. Add
# empty values for timesteps with < 90% valid. if you set
# 'UNCERTAINTY = True' in your config file then you will
# only filter out timesteps with 100% invalid pixels.
# You will also record the number invalid pixels per timestep.
if unknown_percent < unknown_percent_threshold:
wb_capacity_pc.append(water_percent)
invalid_observations.append(unknown_percent)
wb_invalid_ct.append(missing_pixels)
dry_observed.append(dry_percent)
wb_capacity_ct.append(wet_pixels)
else:
wb_capacity_pc.append('')
invalid_observations.append('')
wb_invalid_ct.append('')
dry_observed.append('')
wb_capacity_ct.append('')
valid_obs = wofl.time.dropna(dim='time')
valid_obs = valid_obs.to_dataframe()
if 'spatial_ref' in valid_obs.columns:
valid_obs = valid_obs.drop(columns=['spatial_ref'])
valid_capacity_pc += wb_capacity_pc
valid_capacity_ct += wb_capacity_ct
invalid_capacity_ct += wb_invalid_ct
date_list += valid_obs.to_csv(
None,
header=False,
index=False,
date_format="%Y-%m-%dT%H:%M:%SZ").split('\n')
date_list.pop()
if date_list:
if include_uncertainty:
rows = zip(date_list, valid_capacity_pc, valid_capacity_ct,
invalid_capacity_ct)
else:
rows = zip(date_list, valid_capacity_pc, valid_capacity_ct)
os.makedirs(os.path.dirname(fpath), exist_ok=True)
if time_span == 'APPEND':
of = fsspec.open(fpath, 'a')
with of as f:
writer = csv.writer(f)
for row in rows:
writer.writerow(row)
else:
of = fsspec.open(fpath, 'w')
with of as f:
writer = csv.writer(f)
headings = [
'Observation Date', 'Wet pixel percentage',
'Wet pixel count (n = {0})'.format(masked_all)
]
if include_uncertainty:
headings.append('Invalid pixel count')
writer.writerow(headings)
for row in rows:
writer.writerow(row)
else:
logger.info(f'{str_poly_name} has no new good valid data')
return True
def to_json(self, fname: Union[str, Path], open_kwargs: dict = {}) -> int:
jblob = json.dumps(asdict(self))
with fsspec.open(str(fname), mode='wt', **open_kwargs) as fh:
result = fh.write(jblob)
return result
def download(url, csv_path):
of = fsspec.open(url)
of.fs.download(url, csv_path)
return csv_path
def save_csv(suffix: str, contents: bytes) -> str:
fd, tmpfile = tempfile.mkstemp(prefix="csv_sniffer", suffix=suffix)
os.close(fd)
with fsspec.open(tmpfile, mode="wb", compression="infer") as out:
out.write(contents)
return tmpfile
def bus_peak_frequencies(
gtfs_path: str,
test_date: typing.Optional[datetime.date] = None,
am_peak: typing.Optional[typing.Tuple[int, int]] = None,
pm_peak: typing.Optional[typing.Tuple[int, int]] = None,
) -> geopandas.GeoDataFrame:
"""
Compute AM and PM Peak frequencies for all the lines in a GTFS Feed.
Parameters
==========
gtfs_path: str
The path (or URL) to a GTFS feed.
test_date: datetime.date
The test date for which to compute frequencies. Defaults to February
18th, 2020, an unremarkable weekday February.
am_peak: tuple of integers
The two hours (out of 24) demarcating the AM peak period.
pm_peak: tuple of integers
The two hours (out of 24) demarcating the PM peak period.
"""
# Set default values
test_date = test_date or TEST_DATE
am_peak = am_peak or (6, 9)
pm_peak = pm_peak or (15, 19)
am_duration = am_peak[1] - am_peak[0]
pm_duration = pm_peak[1] - pm_peak[0]
assert am_duration > 0
assert pm_duration > 0
# Download and read the GTFS feed
with fsspec.open(gtfs_path, "rb") as infile:
data = infile.read()
with open(GTFS_FILE, "wb") as outfile:
outfile.write(data)
service_by_date = partridge.read_service_ids_by_date(GTFS_FILE)
feed = partridge.load_geo_feed(GTFS_FILE)
# Get the service for the test date
try:
test_service = next(v for k, v in service_by_date.items()
if k == test_date)
except StopIteration:
raise ValueError(f"Could not find service for {test_date}")
test_trips = feed.trips[feed.trips.service_id.isin(test_service)]
test_stops = feed.stop_times[feed.stop_times.trip_id.isin(
test_trips.trip_id)]
# Get the departure, arrival, and mean time for each trip
trip_timings = test_stops.groupby(test_stops.trip_id).agg({
"departure_time":
min,
"arrival_time":
max
})
trip_timings = trip_timings.assign(
mean_time=trip_timings.departure_time +
(trip_timings.arrival_time - trip_timings.departure_time) / 2.0)
# Find all of the trips that fall within the AM and PM peak times.
am_peak_trips = trip_timings[
(trip_timings.mean_time > am_peak[0] * 60 * 60)
& (trip_timings.mean_time < am_peak[1] * 60 * 60)]
pm_peak_trips = trip_timings[
(trip_timings.mean_time > pm_peak[0] * 60 * 60)
& (trip_timings.mean_time < pm_peak[1] * 60 * 60)]
am_peak_trips = test_trips.merge(
am_peak_trips,
left_on=test_trips.trip_id,
right_index=True,
)
pm_peak_trips = test_trips.merge(
pm_peak_trips,
left_on=test_trips.trip_id,
right_index=True,
)
# Compute the peak frequency
am_peak_frequency = (am_peak_trips.groupby(
[am_peak_trips.route_id,
am_peak_trips.direction_id]).size().to_frame("am_peak_trips"))
am_peak_frequency = am_peak_frequency.assign(
am_peak_frequency=am_duration * 60 / am_peak_frequency.am_peak_trips)
pm_peak_frequency = (pm_peak_trips.groupby(
[pm_peak_trips.route_id,
pm_peak_trips.direction_id]).size().to_frame("pm_peak_trips"))
pm_peak_frequency = pm_peak_frequency.assign(
pm_peak_frequency=pm_duration * 60 / pm_peak_frequency.pm_peak_trips)
peak_frequency = pandas.concat([am_peak_frequency, pm_peak_frequency],
axis=1,
sort=False)
# Add the route short name for easier legibility.
peak_frequency = peak_frequency.join(
feed.routes[["route_id", "route_short_name"]].set_index("route_id"),
how="left",
on="route_id",
)
# Grab the most popular shape as the official one.
route_shapes = (test_trips.groupby("route_id").agg({
"shape_id":
lambda s: s.value_counts().index[0]
}).reset_index().merge(
feed.shapes, how="left",
on="shape_id").set_index("route_id").drop(columns=["shape_id"]))
peak_frequency = peak_frequency.merge(
route_shapes, how="left", right_index=True,
left_index=True).assign(agency=feed.agency.agency_name.iloc[0])
gdf = geopandas.GeoDataFrame(peak_frequency, geometry="geometry")
gdf.crs = f"EPSG:{WGS84}"
return gdf
def write_text_from_href(self, href: str, txt: str, *args: Any,
**kwargs: Any) -> None:
with fsspec.open(href, "w") as destination:
return destination.write(txt)
def open_file(urlpath, mode="rb", compression=None):
return fsspec.open(urlpath,
mode=mode,
compression=compression,
**fsspec_kwargs)
