class BasicProcessor(BasicWorker, metaclass=abc.ABCMeta):
"""
Abstract post-processor class
A post-processor takes a finished search query as input and processed its
result in some way, with another result set as output. The input thus is
a CSV file, and the output (usually) as well. In other words, the result of
a post-processor run can be used as input for another post-processor
(though whether and when this is useful is another question).
"""
db = None # database handler
dataset = None # Dataset object representing the dataset to be created
job = None # Job object that requests the execution of this processor
parent = None # Dataset object to be processed, if applicable
source_file = None # path to dataset to be processed, if applicable
description = "No description available" # processor description, shown in web front-end
category = "Other" # processor category, for sorting in web front-end
extension = "csv" # extension of files created by this processor
options = {} # configurable options for this processor
parameters = {} # values for the processor's configurable options
def work(self):
"""
Process a dataset
Loads dataset metadata, sets up the scaffolding for performing some kind
of processing on that dataset, and then processes it. Afterwards, clean
up.
"""
try:
self.dataset = DataSet(key=self.job.data["remote_id"], db=self.db)
except TypeError:
# query has been deleted in the meantime. finish without error,
# as deleting it will have been a conscious choice by a user
self.job.finish()
return
if self.dataset.data.get("key_parent", None):
# search workers never have parents (for now), so we don't need to
# find out what the parent dataset is if it's a search worker
try:
self.parent = DataSet(key=self.dataset.data["key_parent"],
db=self.db)
except TypeError:
# we need to know what the parent dataset was to properly handle the
# analysis
self.log.warning(
"Processor %s queued for orphan query %s: cannot run, cancelling job"
% (self.type, self.dataset.key))
self.job.finish()
return
if not self.parent.is_finished():
# not finished yet - retry after a while
self.job.release(delay=30)
return
self.parent = DataSet(key=self.dataset.data["key_parent"],
db=self.db)
self.source_file = self.parent.get_results_path()
if not self.source_file.exists():
self.dataset.update_status("Finished, no input data found.")
self.log.info("Running post-processor %s on query %s" %
(self.type, self.job.data["remote_id"]))
self.parameters = self.dataset.parameters
self.dataset.update_status("Processing data")
self.dataset.update_version(get_software_version())
if self.interrupted:
return self.abort()
if not self.dataset.is_finished():
try:
self.process()
self.after_process()
except WorkerInterruptedException:
self.abort()
except Exception as e:
frames = traceback.extract_tb(e.__traceback__)
frames = [
frame.filename.split("/").pop() + ":" + str(frame.lineno)
for frame in frames[1:]
]
location = "->".join(frames)
# Not all datasets have parent keys
if len(self.dataset.get_genealogy()) > 1:
parent_key = " (via " + self.dataset.get_genealogy(
)[0].key + ")"
else:
parent_key = ""
raise ProcessorException(
"Processor %s raised %s while processing dataset %s%s in %s:\n %s\n"
% (self.type, e.__class__.__name__, self.dataset.key,
parent_key, location, str(e)))
else:
# dataset already finished, job shouldn't be open anymore
self.log.warning(
"Job %s/%s was queued for a dataset already marked as finished, deleting..."
% (self.job.data["jobtype"], self.job.data["remote_id"]))
self.job.finish()
def after_process(self):
"""
After processing, declare job finished
"""
if self.dataset.data["num_rows"] > 0:
self.dataset.update_status("Dataset saved.")
if not self.dataset.is_finished():
self.dataset.finish()
# see if we have anything else lined up to run next
for next in self.parameters.get("next", []):
next_parameters = next.get("parameters", {})
next_type = next.get("type", "")
available_processors = self.dataset.get_available_processors()
# run it only if the post-processor is actually available for this query
if next_type in available_processors:
next_analysis = DataSet(
parameters=next_parameters,
type=next_type,
db=self.db,
parent=self.dataset.key,
extension=available_processors[next_type]["extension"])
self.queue.add_job(next_type, remote_id=next_analysis.key)
# see if we need to register the result somewhere
if "copy_to" in self.parameters:
# copy the results to an arbitrary place that was passed
if self.dataset.get_results_path().exists():
shutil.copyfile(str(self.dataset.get_results_path()),
self.parameters["copy_to"])
else:
# if copy_to was passed, that means it's important that this
# file exists somewhere, so we create it as an empty file
with open(self.parameters["copy_to"], "w") as empty_file:
empty_file.write("")
# see if this query chain is to be attached to another query
# if so, the full genealogy of this query (minus the original dataset)
# is attached to the given query - this is mostly useful for presets,
# where a chain of processors can be marked as 'underlying' a preset
if "attach_to" in self.parameters:
try:
# copy metadata and results to the surrogate
surrogate = DataSet(key=self.parameters["attach_to"],
db=self.db)
if self.dataset.get_results_path().exists():
shutil.copyfile(str(self.dataset.get_results_path()),
str(surrogate.get_results_path()))
top_parent = self.dataset.get_genealogy()[1]
top_parent.link_parent(surrogate.key)
try:
surrogate.finish(self.dataset.data["num_rows"])
except RuntimeError:
# already finished, could happen (though it shouldn't)
pass
surrogate.update_status(self.dataset.get_status())
except ValueError:
# dataset with key to attach to doesn't exist...
self.log.warning(
"Cannot attach dataset chain containing %s to %s (dataset does not exist)"
% (self.dataset.key, self.parameters["attach_to"]))
self.job.finish()
def abort(self):
"""
Abort dataset creation and clean up so it may be attempted again later
"""
# remove any result files that have been created so far
if self.dataset.get_results_path().exists():
os.unlink(str(self.dataset.get_results_path()))
if self.dataset.get_temporary_path().exists():
shutil.rmtree(str(self.dataset.get_temporary_path()))
# we release instead of finish, since interrupting is just that - the
# job should resume at a later point. Delay resuming by 10 seconds to
# give 4CAT the time to do whatever it wants (though usually this isn't
# needed since restarting also stops the spawning of new workers)
self.dataset.update_status(
"Dataset processing interrupted. Retrying later.")
if self.interrupted == self.INTERRUPT_RETRY:
# retry later - wait at least 10 seconds to give the backend time to shut down
self.job.release(delay=10)
elif self.interrupted == self.INTERRUPT_CANCEL:
# cancel job
self.job.finish()
def iterate_csv_items(self, path):
"""
A generator that iterates through a CSV file
With every iteration, the processor's 'interrupted' flag is checked,
and if set a ProcessorInterruptedException is raised, which by default
is caught and subsequently stops execution gracefully.
:param Path path: Path to csv file to read
:return:
"""
with open(path, encoding="utf-8") as input:
reader = csv.DictReader(input)
for item in reader:
if self.interrupted:
raise ProcessorInterruptedException(
"Processor interrupted while iterating through CSV file"
)
yield item
def write_csv_items_and_finish(self, data):
"""
Write data as csv to results file and finish dataset
Determines result file path using dataset's path determination helper
methods. After writing results, the dataset is marked finished. Will
raise a ProcessorInterruptedException if the interrupted flag for this
processor is set while iterating.
:param data: A list or tuple of dictionaries, all with the same keys
"""
if not (isinstance(data, typing.List)
or isinstance(data, typing.Tuple)) or isinstance(data, str):
raise TypeError(
"write_csv_items requires a list or tuple of dictionaries as argument"
)
if not data:
raise ValueError(
"write_csv_items requires a dictionary with at least one item")
if not isinstance(data[0], dict):
raise TypeError(
"write_csv_items requires a list or tuple of dictionaries as argument"
)
self.dataset.update_status("Writing results file")
with self.dataset.get_results_path().open("w",
encoding="utf-8",
newline='') as results:
writer = csv.DictWriter(results, fieldnames=data[0].keys())
writer.writeheader()
for row in data:
if self.interrupted:
raise ProcessorInterruptedException(
"Interrupted while writing results file")
writer.writerow(row)
self.dataset.update_status("Finished")
self.dataset.finish(len(data))
def is_filter(self):
"""
Is this processor a filter?
Filters do not produce their own dataset but replace the parent dataset
instead.
:todo: Make this a bit more robust than sniffing the processor category
:return bool:
"""
return hasattr(
self, "category"
) and self.category and "filter" in self.category.lower()
@abc.abstractmethod
def process(self):
"""
Process data
To be defined by the child processor.
"""
pass
class BasicProcessor(BasicWorker, metaclass=abc.ABCMeta):
"""
Abstract post-processor class
A post-processor takes a finished search query as input and processed its
result in some way, with another result set as output. The input thus is
a CSV file, and the output (usually) as well. In other words, the result of
a post-processor run can be used as input for another post-processor
(though whether and when this is useful is another question).
"""
db = None # database handler
dataset = None # Dataset object representing the dataset to be created
job = None # Job object that requests the execution of this processor
source_dataset = None # Dataset object to be processed, if applicable
source_file = None # path to dataset to be processed, if applicable
description = "No description available" # processor description, shown in web front-end
category = "Other" # processor category, for sorting in web front-end
extension = "csv" # extension of files created by this processor
options = {} # configurable options for this processor
parameters = {} # values for the processor's configurable options
is_running_in_preset = False # is this processor running 'within' a preset processor?
def work(self):
"""
Process a dataset
Loads dataset metadata, sets up the scaffolding for performing some kind
of processing on that dataset, and then processes it. Afterwards, clean
up.
"""
try:
self.dataset = DataSet(key=self.job.data["remote_id"], db=self.db)
except TypeError:
# query has been deleted in the meantime. finish without error,
# as deleting it will have been a conscious choice by a user
self.job.finish()
return
is_running_in_preset = False
if self.dataset.data.get("key_parent", None):
# search workers never have parents (for now), so we don't need to
# find out what the source_dataset dataset is if it's a search worker
try:
self.source_dataset = DataSet(
key=self.dataset.data["key_parent"], db=self.db)
# for presets, transparently use the *top* dataset as a source_dataset
# since that is where any underlying processors should get
# their data from. However, this should only be done as long as the
# preset is not finished yet, because after that there may be processors
# that run on the final preset result
if self.source_dataset.type.find(
"preset-"
) == 0 and not self.source_dataset.is_finished():
self.is_running_in_preset = True
self.source_dataset = self.source_dataset.get_genealogy(
)[0]
except TypeError:
# we need to know what the source_dataset dataset was to properly handle the
# analysis
self.log.warning(
"Processor %s queued for orphan query %s: cannot run, cancelling job"
% (self.type, self.dataset.key))
self.job.finish()
return
if not self.source_dataset.is_finished(
) and not self.is_running_in_preset:
# not finished yet - retry after a while
# exception for presets, since these *should* be unfinished
# until underlying processors are done
self.job.release(delay=30)
return
self.source_file = self.source_dataset.get_results_path()
if not self.source_file.exists():
self.dataset.update_status("Finished, no input data found.")
self.log.info("Running processor %s on dataset %s" %
(self.type, self.job.data["remote_id"]))
processor_name = self.title if hasattr(self, "title") else self.type
self.dataset.clear_log()
self.dataset.log("Processing '%s' started for dataset %s" %
(processor_name, self.dataset.key))
# start log file
self.parameters = self.dataset.parameters.copy()
self.dataset.update_status("Processing data")
self.dataset.update_version(get_software_version())
# now the parameters have been loaded into memory, clear any sensitive
# ones. This has a side-effect that a processor may not run again
# without starting from scratch, but this is the price of progress
if hasattr(self, "options"):
for option in self.options:
if self.options[option].get("sensitive"):
self.dataset.delete_parameter(option)
if self.interrupted:
self.dataset.log("Processing interrupted, trying again later")
return self.abort()
if not self.dataset.is_finished():
try:
self.process()
self.after_process()
except WorkerInterruptedException as e:
self.dataset.log(
"Processing interrupted (%s), trying again later" % str(e))
self.abort()
except Exception as e:
self.dataset.log("Processor crashed (%s), trying again later" %
str(e))
frames = traceback.extract_tb(e.__traceback__)
frames = [
frame.filename.split("/").pop() + ":" + str(frame.lineno)
for frame in frames[1:]
]
location = "->".join(frames)
# Not all datasets have source_dataset keys
if len(self.dataset.get_genealogy()) > 1:
parent_key = " (via " + self.dataset.get_genealogy(
)[0].key + ")"
else:
parent_key = ""
raise ProcessorException(
"Processor %s raised %s while processing dataset %s%s in %s:\n %s\n"
% (self.type, e.__class__.__name__, self.dataset.key,
parent_key, location, str(e)))
else:
# dataset already finished, job shouldn't be open anymore
self.log.warning(
"Job %s/%s was queued for a dataset already marked as finished, deleting..."
% (self.job.data["jobtype"], self.job.data["remote_id"]))
self.job.finish()
def after_process(self):
"""
After processing, declare job finished
"""
if self.dataset.data["num_rows"] > 0:
self.dataset.update_status("Dataset saved.")
if not self.dataset.is_finished():
self.dataset.finish()
if hasattr(self, "staging_area") and type(
self.staging_area) == Path and self.staging_area.exists():
shutil.rmtree(self.staging_area)
# see if we have anything else lined up to run next
for next in self.parameters.get("next", []):
next_parameters = next.get("parameters", {})
next_type = next.get("type", "")
available_processors = self.dataset.get_available_processors()
# run it only if the post-processor is actually available for this query
if next_type in available_processors:
next_analysis = DataSet(
parameters=next_parameters,
type=next_type,
db=self.db,
parent=self.dataset.key,
extension=available_processors[next_type]["extension"])
self.queue.add_job(next_type, remote_id=next_analysis.key)
else:
self.log.warning(
"Dataset %s (of type %s) wants to run processor %s next, but it is incompatible"
% (self.dataset.key, self.type, next_type))
# see if we need to register the result somewhere
if "copy_to" in self.parameters:
# copy the results to an arbitrary place that was passed
if self.dataset.get_results_path().exists():
shutil.copyfile(str(self.dataset.get_results_path()),
self.parameters["copy_to"])
else:
# if copy_to was passed, that means it's important that this
# file exists somewhere, so we create it as an empty file
with open(self.parameters["copy_to"], "w") as empty_file:
empty_file.write("")
# see if this query chain is to be attached to another query
# if so, the full genealogy of this query (minus the original dataset)
# is attached to the given query - this is mostly useful for presets,
# where a chain of processors can be marked as 'underlying' a preset
if "attach_to" in self.parameters:
try:
# copy metadata and results to the surrogate
surrogate = DataSet(key=self.parameters["attach_to"],
db=self.db)
if self.dataset.get_results_path().exists():
shutil.copyfile(str(self.dataset.get_results_path()),
str(surrogate.get_results_path()))
try:
surrogate.finish(self.dataset.data["num_rows"])
except RuntimeError:
# already finished, could happen (though it shouldn't)
pass
surrogate.update_status(self.dataset.get_status())
except ValueError:
# dataset with key to attach to doesn't exist...
self.log.warning(
"Cannot attach dataset chain containing %s to %s (dataset does not exist)"
% (self.dataset.key, self.parameters["attach_to"]))
self.job.finish()
def abort(self):
"""
Abort dataset creation and clean up so it may be attempted again later
"""
# remove any result files that have been created so far
if self.dataset.get_results_path().exists():
self.dataset.get_results_path().unlink(missing_ok=True)
if self.dataset.get_staging_area().exists():
shutil.rmtree(str(self.dataset.get_staging_area()))
# we release instead of finish, since interrupting is just that - the
# job should resume at a later point. Delay resuming by 10 seconds to
# give 4CAT the time to do whatever it wants (though usually this isn't
# needed since restarting also stops the spawning of new workers)
self.dataset.update_status(
"Dataset processing interrupted. Retrying later.")
if self.interrupted == self.INTERRUPT_RETRY:
# retry later - wait at least 10 seconds to give the backend time to shut down
self.job.release(delay=10)
elif self.interrupted == self.INTERRUPT_CANCEL:
# cancel job
self.job.finish()
def iterate_items(self, path, bypass_map_item=False):
"""
A generator that iterates through a CSV or NDJSON file
With every iteration, the processor's 'interrupted' flag is checked,
and if set a ProcessorInterruptedException is raised, which by default
is caught and subsequently stops execution gracefully.
Processors can define a method called `map_item` that can be used to
map an item from the dataset file before it is processed any further
this is slower than storing the data file in the right format to begin
with but not all data sources allow for easy 'flat' mapping of items,
e.g. tweets are nested objects when retrieved from the twitter API
that are easier to store as a JSON file than as a flat CSV file, and
it would be a shame to throw away that data
There are two file types that can be iterated (currently): CSV files
and NDJSON (newline-delimited JSON) files. In the future, one could
envision adding a pathway to retrieve items from e.g. a MongoDB
collection directly instead of from a static file
:param Path path: Path to file to read
:return Generator: A generator that yields each item as a dictionary
"""
# see if an item mapping function has been defined
# open question if 'source_dataset' shouldn't be an attribute of the dataset
# instead of the processor...
item_mapper = None
if hasattr(self, "source_dataset"
) and self.source_dataset and not bypass_map_item:
parent_processor = self.all_modules.processors.get(
self.source_dataset.type)
if parent_processor:
parent_processor = self.all_modules.load_worker_class(
parent_processor)
if hasattr(parent_processor, "map_item"):
item_mapper = parent_processor.map_item
# go through items one by one, optionally mapping them
if path.suffix.lower() == ".csv":
with path.open(encoding="utf-8") as input:
reader = csv.DictReader(input)
for item in reader:
if self.interrupted:
raise ProcessorInterruptedException(
"Processor interrupted while iterating through CSV file"
)
if item_mapper:
item = item_mapper(item)
yield item
elif path.suffix.lower() == ".ndjson":
# in this format each line in the file is a self-contained JSON
# file
with path.open(encoding="utf-8") as input:
for line in input:
if self.interrupted:
raise ProcessorInterruptedException(
"Processor interrupted while iterating through NDJSON file"
)
item = json.loads(line)
if item_mapper:
item = item_mapper(item)
yield item
else:
raise NotImplementedError("Cannot iterate through %s file" %
path.suffix)
def get_item_keys(self, path=None):
"""
Get item attribute names
It can be useful to know what attributes an item in the dataset is
stored with, e.g. when one wants to produce a new dataset identical
to the source_dataset one but with extra attributes. This method provides
these, as a list.
:param Path path: Path to the dataset file; if left empty, use the
processor's own dataset's path
:return list: List of keys, may be empty if there are no items in the
dataset
:todo: Figure out if this makes more sense as a Dataset method
"""
if not path:
path = self.dataset.get_results_path()
items = self.iterate_items(path)
try:
keys = list(items.__next__().keys())
except StopIteration:
return []
finally:
del items
return keys
def iterate_archive_contents(self, path, staging_area=None):
"""
A generator that iterates through files in an archive
With every iteration, the processor's 'interrupted' flag is checked,
and if set a ProcessorInterruptedException is raised, which by default
is caught and subsequently stops execution gracefully.
Files are temporarily unzipped and deleted after use.
:param Path path: Path to zip file to read
:param Path staging_area: Where to store the files while they're
being worked with. If omitted, a temporary folder is created and
deleted after use
:return: An iterator with a Path item for each file
"""
if not path.exists():
return
if staging_area and (not staging_area.exists()
or not staging_area.is_dir()):
raise RuntimeError("Staging area %s is not a valid folder")
else:
if not hasattr(self, "staging_area") and not staging_area:
self.staging_area = self.dataset.get_staging_area()
staging_area = self.staging_area
with zipfile.ZipFile(path, "r") as archive_file:
archive_contents = sorted(archive_file.namelist())
for archived_file in archive_contents:
if self.interrupted:
if hasattr(self, "staging_area"):
shutil.rmtree(self.staging_area)
raise ProcessorInterruptedException(
"Interrupted while iterating zip file contents")
file_name = archived_file.split("/")[-1]
temp_file = staging_area.joinpath(file_name)
archive_file.extract(file_name, staging_area)
yield temp_file
if hasattr(self, "staging_area"):
temp_file.unlink()
if hasattr(self, "staging_area"):
shutil.rmtree(self.staging_area)
del self.staging_area
def unpack_archive_contents(self, path, staging_area=None):
"""
Unpack all files in an archive to a staging area
With every iteration, the processor's 'interrupted' flag is checked,
and if set a ProcessorInterruptedException is raised, which by default
is caught and subsequently stops execution gracefully.
Files are unzipped to a staging area. The staging area is *not*
cleaned up automatically.
:param Path path: Path to zip file to read
:param Path staging_area: Where to store the files while they're
being worked with. If omitted, a temporary folder is created and
deleted after use
:return Path: A path to the staging area
"""
if not path.exists():
return
if staging_area and (not staging_area.exists()
or not staging_area.is_dir()):
raise RuntimeError("Staging area %s is not a valid folder")
else:
if not hasattr(self, "staging_area"):
self.staging_area = self.dataset.get_staging_area()
staging_area = self.staging_area
paths = []
with zipfile.ZipFile(path, "r") as archive_file:
archive_contents = sorted(archive_file.namelist())
for archived_file in archive_contents:
if self.interrupted:
raise ProcessorInterruptedException(
"Interrupted while iterating zip file contents")
file_name = archived_file.split("/")[-1]
temp_file = staging_area.joinpath(file_name)
archive_file.extract(archived_file, staging_area)
paths.append(temp_file)
return staging_area
def write_csv_items_and_finish(self, data):
"""
Write data as csv to results file and finish dataset
Determines result file path using dataset's path determination helper
methods. After writing results, the dataset is marked finished. Will
raise a ProcessorInterruptedException if the interrupted flag for this
processor is set while iterating.
:param data: A list or tuple of dictionaries, all with the same keys
"""
if not (isinstance(data, typing.List)
or isinstance(data, typing.Tuple)) or isinstance(data, str):
raise TypeError(
"write_csv_items requires a list or tuple of dictionaries as argument"
)
if not data:
raise ValueError(
"write_csv_items requires a dictionary with at least one item")
if not isinstance(data[0], dict):
raise TypeError(
"write_csv_items requires a list or tuple of dictionaries as argument"
)
self.dataset.update_status("Writing results file")
with self.dataset.get_results_path().open("w",
encoding="utf-8",
newline='') as results:
writer = csv.DictWriter(results, fieldnames=data[0].keys())
writer.writeheader()
for row in data:
if self.interrupted:
raise ProcessorInterruptedException(
"Interrupted while writing results file")
writer.writerow(row)
self.dataset.update_status("Finished")
self.dataset.finish(len(data))
def write_archive_and_finish(self,
files,
num_items=None,
compression=zipfile.ZIP_STORED):
"""
Archive a bunch of files into a zip archive and finish processing
:param list|Path files: If a list, all files will be added to the
archive and deleted afterwards. If a folder, all files in the folder
will be added and the folder will be deleted afterwards.
:param int num_items: Items in the dataset. If None, the amount of
files added to the archive will be used.
:param int compression: Type of compression to use. By default, files
are not compressed, to speed up unarchiving.
"""
is_folder = False
if issubclass(type(files), PurePath):
is_folder = files
if not files.exists() or not files.is_dir():
raise RuntimeError(
"Folder %s is not a folder that can be archived" % files)
files = files.glob("*")
# create zip of archive and delete temporary files and folder
self.dataset.update_status("Compressing results into archive")
done = 0
with zipfile.ZipFile(self.dataset.get_results_path(),
"w",
compression=compression) as zip:
for output_path in files:
zip.write(output_path, output_path.name)
output_path.unlink()
done += 1
# delete temporary folder
if is_folder:
shutil.rmtree(is_folder)
self.dataset.update_status("Finished")
if num_items is None:
num_items = done
self.dataset.finish(num_items)
def is_filter(self):
"""
Is this processor a filter?
Filters do not produce their own dataset but replace the source_dataset dataset
instead.
:todo: Make this a bit more robust than sniffing the processor category
:return bool:
"""
return hasattr(
self, "category"
) and self.category and "filter" in self.category.lower()
@abc.abstractmethod
def process(self):
"""
Process data
To be defined by the child processor.
"""
pass