def preview_csv(key):
"""
Preview a CSV file
Simply passes the first 25 rows of a dataset's csv result file to the
template renderer.
:param str key: Dataset key
:return: HTML preview
"""
try:
dataset = DataSet(key=key, db=db)
except TypeError:
return error(404, "Dataset not found.")
try:
with dataset.get_results_path().open(encoding="utf-8") as csvfile:
rows = []
reader = csv.reader(csvfile)
while len(rows) < 25:
try:
row = next(reader)
rows.append(row)
except StopIteration:
break
except FileNotFoundError:
abort(404)
return render_template("result-csv-preview.html",
rows=rows,
filename=dataset.get_results_path().name)
def work(self):
"""
Go through all datasources, and if it is configured to automatically
delete old datasets, do so for all qualifying datasets
:return:
"""
for datasource_id in self.all_modules.datasources:
datasource = self.all_modules.datasources[datasource_id]
# default = never expire
if not datasource.get("expire-datasets", None):
continue
cutoff = time.time() - datasource.get("expire-datasets")
datasets = self.db.fetchall(
"SELECT key FROM datasets WHERE key_parent = '' AND parameters::json->>'datasource' = %s AND timestamp < %s",
(datasource_id, cutoff))
# we instantiate the dataset, because its delete() method does all
# the work (e.g. deleting child datasets) for us
for dataset in datasets:
dataset = DataSet(key=dataset["key"], db=self.db)
dataset.delete()
self.log.info(
"Deleting dataset %s/%s (expired per configuration)" %
(datasource, dataset.key))
self.job.finish()
def show_result(key):
"""
Show result page
The page contains dataset details and a download link, but also shows a list
of finished and available processors.
:param key: Result key
:return: Rendered template
"""
try:
dataset = DataSet(key=key, db=db)
except TypeError:
abort(404)
# child datasets are not available via a separate page - redirect to parent
if dataset.key_parent:
genealogy = dataset.get_genealogy()
nav = ",".join([family.key for family in genealogy])
url = "/results/%s/#nav=%s" % (genealogy[0].key, nav)
return redirect(url)
# load list of processors compatible with this dataset
is_processor_running = False
is_favourite = (db.fetchone(
"SELECT COUNT(*) AS num FROM users_favourites WHERE name = %s AND key = %s",
(current_user.get_id(), dataset.key))["num"] > 0)
# if the datasource is configured for it, this dataset may be deleted at some point
datasource = dataset.parameters.get("datasource", "")
if datasource in backend.all_modules.datasources and backend.all_modules.datasources[
datasource].get("expire-datasets", None):
timestamp_expires = dataset.timestamp + int(
backend.all_modules.datasources[datasource].get("expire-datasets"))
else:
timestamp_expires = None
# we can either show this view as a separate page or as a bunch of html
# to be retrieved via XHR
standalone = "processors" not in request.url
template = "result.html" if standalone else "result-details.html"
return render_template(template,
dataset=dataset,
parent_key=dataset.key,
processors=backend.all_modules.processors,
is_processor_running=is_processor_running,
messages=get_flashed_messages(),
is_favourite=is_favourite,
timestamp_expires=timestamp_expires)
def work(self):
"""
Send pg_cancel_backend query to cancel query with given PID
"""
# delete dataset
try:
dataset = DataSet(key=self.job.data["remote_id"], db=self.db)
jobtype = dataset.data["type"]
except TypeError:
# dataset already deleted, apparently
self.job.finish()
return
# now find the job that's tasked with creating this dataset, if it
# exists
try:
job = Job.get_by_remote_ID(remote_id=self.job.data["remote_id"],
jobtype=jobtype,
database=self.db)
except JobNotFoundException:
# no job... dataset already fully finished?
self.job.finish()
return
# ask the manager to interrupt this job
self.manager.request_interrupt(job, self.INTERRUPT_CANCEL)
# done
self.job.finish()
def process(self):
"""
This queues a series of post-processors to run in sequence, with an
overarching dataset to which the results of the last processor in the
sequence are copied. The processor pipeline is then attached to the
overarching dataset so it is clear that all processors were run as part
of that particular preset.
"""
pipeline = self.get_processor_pipeline()
# make sure the last item in the pipeline copies to the preset's dataset
pipeline = pipeline.copy()
pipeline[-1]["parameters"]["attach_to"] = self.dataset.key
# map the linear pipeline to a nested processor parameter set
while len(pipeline) > 1:
last = pipeline.pop()
pipeline[-1]["parameters"]["next"] = [last]
analysis_pipeline = DataSet(parameters=pipeline[0]["parameters"],
type=pipeline[0]["type"],
db=self.db,
parent=self.dataset.key)
# this starts the pipeline
self.queue.add_job(pipeline[0]["type"],
remote_id=analysis_pipeline.key)
def restart_dataset(key):
"""
Run a dataset's query again
Deletes all underlying datasets, marks dataset as unfinished, and queues a
job for it.
:param str key: Dataset key
:return:
"""
try:
dataset = DataSet(key=key, db=db)
except TypeError:
return error(404, message="Dataset not found.")
if current_user.get_id() != dataset.parameters.get(
"user", "") and not current_user.is_admin:
return error(403, message="Not allowed.")
if not dataset.is_finished():
return render_template(
"error.html",
message="This dataset is not finished yet - you cannot re-run it.")
if "type" not in dataset.parameters:
return render_template(
"error.html",
message=
"This is an older dataset that unfortunately lacks the information necessary to properly restart it."
)
for child in dataset.children:
child.delete()
dataset.unfinish()
queue = JobQueue(logger=log, database=db)
queue.add_job(jobtype=dataset.parameters["type"], remote_id=dataset.key)
flash("Dataset queued for re-running.")
return redirect("/results/" + dataset.key + "/")
def process_standalone(processor):
"""
Run a standalone processor
This bypasses the usual 4CAT query-processor structure and allows running
any available processor (see the `/api/get-standalone-processors/`
endpoint) with one API call. The data is returned immediately and not saved
server-side.
Requires authentication.
:param str processor: ID of the processor to run on incoming data
:request-body object data: Data to process, a JSON-formatted list of
objects with each object having at least they keys `post_id`,
`thread_id`, body`, and `author`.
:request-schema data: {
type=object,
properties={
post_id={type=string},
thread_id={type=string},
body={type=string},
author={type=string}
}
}
:request-param str ?access_token: Access token; only required if not
logged in currently.
:return: A JSON object containing the processed data, with a
processor-specific structure.
:return-schema: {
type=object,
additionalProperties={}
}
:return-error 402: If an invalid processor is requested, or if the input is
not properly-formatted JSON.
:return-error 503: If too many other requests are currently being handled,
so that the server does not have the capacity to deal with this request
"""
processors = get_standalone_processors().get_json()
if processor not in processors:
return error(402, error="Processor '%s' is not available" % processor)
if not request.is_json:
return error(402, error="This API endpoint only accepts JSON-formatted data as input")
try:
input = request.get_json(force=True)
except json.JSONDecodeError:
return error(402, error="JSON decoding error")
# check file integrity
required = ("id", "thread_id", "body", "author")
try:
for row in input:
for field in required:
if field not in row:
return error(402, error="Input is valid JSON, but not a list of data objects (missing field '%s')" % field)
except TypeError:
return error(402, error="Input is valid JSON, but not a list of data objects")
if not input:
return error(402, error="Input is empty")
# ok, valid input!
temp_dataset = DataSet(extension="csv", type="standalone", parameters={"user": current_user.get_id(), "after": [processor]}, db=db)
temp_dataset.finish(len(input))
# make sure the file is deleted later, whichever way this request is
# ultimately handled
@after_this_request
def delete_temporary_dataset(response):
temp_dataset.delete() # also deletes children!
return response
# write the input as a csv file so it can be accessed as normal by
# processors
result_file = temp_dataset.get_results_path()
with result_file.open("w") as temp_csv:
writer = csv.DictWriter(temp_csv, fieldnames=required)
writer.writeheader()
for row in input:
writer.writerow({field: row[field] for field in required})
# queue the postprocessor
metadata = processors[processor]
processed = DataSet(extension=metadata["extension"], type=processor, parent=temp_dataset.key, db=db)
queue = JobQueue(database=db, logger=log)
job = queue.add_job(processor, {}, processed.key)
place_in_queue = queue.get_place_in_queue(job)
if place_in_queue > 5:
job.finish()
return error(code=503, error="Your request could not be handled as there are currently %i other jobs of this type in the queue. Please try again later." % place_in_queue)
# wait up to half a minute for the job to be taken up
# if not, tell the user to try again later
start = time.time()
while True:
if time.time() > start + 30:
job.finish()
return error(code=503, error="The server is currently too busy to handle your request. Please try again later.")
if queue.get_place_in_queue(job) != 0:
time.sleep(2)
continue
else:
break
# job currently being processed, wait for it to finish
while True:
try:
job = Job.get_by_remote_ID(job.data["remote_id"], db, processor)
except JobNotFoundException:
break
if not job.is_finished:
time.sleep(2)
else:
break
# job finished, send file - temporary datasets will be cleaned up by
# after_this_request function defined earlier
return send_file(processed.get_results_path(), as_attachment=True)
def show_results(page):
"""
Show results overview
For each result, available analyses are also displayed.
:return: Rendered template
"""
page_size = 20
offset = (page - 1) * page_size
where = ["(key_parent = '' OR key_parent IS NULL)"]
replacements = []
query_filter = request.args.get("filter", "")
depth = request.args.get("depth", "own")
if depth not in ("own", "favourites", "all"):
depth = "own"
if depth == "own":
where.append("parameters::json->>'user' = %s")
replacements.append(current_user.get_id())
if depth == "favourites":
where.append(
"key IN ( SELECT key FROM users_favourites WHERE name = %s )")
replacements.append(current_user.get_id())
if query_filter:
where.append("query LIKE %s")
replacements.append("%" + query_filter + "%")
where = " AND ".join(where)
num_datasets = db.fetchone(
"SELECT COUNT(*) AS num FROM datasets WHERE " + where,
tuple(replacements))["num"]
replacements.append(page_size)
replacements.append(offset)
datasets = db.fetchall(
"SELECT key FROM datasets WHERE " + where +
" ORDER BY timestamp DESC LIMIT %s OFFSET %s", tuple(replacements))
if not datasets and page != 1:
abort(404)
pagination = Pagination(page, page_size, num_datasets)
filtered = []
for dataset in datasets:
filtered.append(DataSet(key=dataset["key"], db=db))
favourites = [
row["key"] for row in db.fetchall(
"SELECT key FROM users_favourites WHERE name = %s", (
current_user.get_id(), ))
]
return render_template("results.html",
filter={"filter": query_filter},
depth=depth,
datasets=filtered,
pagination=pagination,
favourites=favourites)
cli.add_argument("-k", "--key", required=True, help="Dataset key to delete.")
cli.add_argument(
"-q",
"--quiet",
type=bool,
default=False,
help="Whether to skip asking for confirmation. Defaults to false.")
args = cli.parse_args()
if not args.quiet:
confirm = input(
"This will delete the dataset and any child datasets. Are you sure? (y/n)"
)
if confirm.strip().lower() != "y":
sys.exit(0)
logger = Logger()
database = Database(logger=logger, appname="delete-dataset")
# Initialize query
try:
parent = DataSet(key=args.key, db=database)
except TypeError:
print("No dataset found with that key.")
sys.exit(1)
parent.delete()
print(
"Done. Note that running jobs for the datasets above are not stopped; you will have to wait for them to finish on their own."
)
"""
Delete all files from the results folder that are not linked to a query
"""
import glob
import sys
import os
sys.path.insert(0, os.path.abspath(os.path.dirname(__file__)) + "/..")
from common.lib.database import Database
from common.lib.logger import Logger
from common.lib.dataset import DataSet
import config
logger = Logger()
database = Database(logger=logger, appname="result-cleaner")
os.chdir(config.PATH_DATA)
files = glob.glob("*.*")
for file in files:
key = file.split(".")[0].split("-")[-1]
try:
query = DataSet(key=key, db=database)
except TypeError:
print("Not linked to a query: %s" % file)
os.unlink(config.PATH_DATA + "/" + file)
def process(self):
"""
Run 4CAT search query
Gets query details, passes them on to the object's search method, and
writes the results to a CSV file. If that all went well, the query and
job are marked as finished.
"""
query_parameters = self.dataset.get_parameters()
results_file = self.dataset.get_results_path()
self.log.info("Querying: %s" % str(query_parameters))
# Execute the relevant query (string-based, random, countryflag-based)
try:
posts = self.search(query_parameters)
except WorkerInterruptedException:
raise ProcessorInterruptedException(
"Interrupted while collecting data, trying again later.")
# Write posts to csv and update the DataBase status to finished
num_posts = 0
if posts:
self.dataset.update_status("Writing posts to result file")
if not hasattr(self, "extension") or self.extension == "csv":
num_posts = self.items_to_csv(posts, results_file)
elif self.extension == "ndjson":
num_posts = self.items_to_ndjson(posts, results_file)
else:
raise NotImplementedError(
"Datasource query cannot be saved as %s file" %
self.extension)
self.dataset.update_status(
"Query finished, results are available.")
elif posts is not None:
self.dataset.update_status("Query finished, no results found.")
# queue predefined post-processors
if num_posts > 0 and query_parameters.get("next", []):
for next in query_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 query_parameters.get("copy_to", None):
# copy the results to an arbitrary place that was passed
if self.dataset.get_results_path().exists():
# but only if we actually have something to copy
shutil.copyfile(str(self.dataset.get_results_path()),
query_parameters.get("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(query_parameters.get("copy_to"), "w") as empty_file:
empty_file.write("")
self.dataset.finish(num_rows=num_posts)
cli = argparse.ArgumentParser()
cli.add_argument("-i", "--input", required=True, help="csv to import")
cli.add_argument("-u", "--user", default="autologin", help="Username to assign the dataset to")
cli.add_argument("-d", "--datasource", default="custom", help="Datasource to assign to dataset")
args = cli.parse_args()
input = Path(args.input)
if not input.exists():
print("File not found")
sys.exit(1)
with open(input) as i:
reader = csv.DictReader(i)
rows = 0
for row in reader:
rows += 1
required = ("id", "thread_id", "subject", "author", "timestamp", "body")
for field in required:
if field not in reader.fieldnames:
print("Column '%s' missing." % field)
sys.exit(1)
logger = Logger()
new_set = DataSet(
parameters={"user": args.user, "filename": input.name, "time": int(time.time()), "datasource": args.datasource,
"board": "upload"}, type="%s-search" % args.datasource,
db=Database(logger=logger))
shutil.copyfile(input, new_set.get_results_path())
new_set.finish(rows)
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
options = self.get_options(self.dataset.get_parent())
for option, option_settings in options.items():
if option_settings.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()
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?
# the following will be defined automatically upon loading the processor
# there is no need to override manually
filepath = None # do not override
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
options = self.get_options(self.dataset.get_parent())
for option, option_settings in options.items():
if option_settings.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:
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)
@classmethod
def is_filter(cls):
"""
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(
cls,
"category") and cls.category and "filter" in cls.category.lower()
@classmethod
def get_options(cls, parent_dataset=None):
"""
Get processor options
This method by default returns the class's "options" attribute, or an
empty dictionary. It can be redefined by processors that need more
fine-grained options, e.g. in cases where the availability of options
is partially determined by the parent dataset's parameters.
:param DataSet parent_dataset: An object representing the dataset that
the processor would be run on
"""
return cls.options if hasattr(cls, "options") else {}
@abc.abstractmethod
def process(self):
"""
Process data
To be defined by the child processor.
"""
pass
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()