Homec4science
Diffusion invenio-infoscience (maint-1.2) Browse

R3600/modules/bibcheck/lib/bibcheck_task.py
b6592965ddb4maint-1.2

Last Change
Raw File

bibcheck_task.py
View Options

# -*- coding: utf-8 -*-
##
## This file is part of Invenio.
## Copyright (C) 2013 CERN.
##
## Invenio is free software; you can redistribute it and/or
## modify it under the terms of the GNU General Public License as
## published by the Free Software Foundation; either version 2 of the
## License, or (at your option) any later version.
##
## Invenio is distributed in the hope that it will be useful, but
## WITHOUT ANY WARRANTY; without even the implied warranty of
## MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
## General Public License for more details.
##
## You should have received a copy of the GNU General Public License
## along with Invenio; if not, write to the Free Software Foundation, Inc.,
## 59 Temple Place, Suite 330, Boston, MA 02111-1307, USA.
"""
BibCheck task
This task will check the records against a defined set of rules.
"""
import sys
import getopt
import os
import traceback
import time
import inspect
import itertools
from ConfigParser import RawConfigParser
from datetime import datetime
from tempfile import mkstemp
from invenio.bibtask import \
task_init, \
task_set_option, \
task_has_option, \
task_get_option, write_message, \
task_update_progress, \
task_sleep_now_if_required, \
get_modified_records_since, \
task_low_level_submission
from invenio.config import \
CFG_VERSION, \
CFG_ETCDIR, \
CFG_PYLIBDIR, \
CFG_SITE_URL, \
CFG_TMPSHAREDDIR, \
CFG_CERN_SITE
from invenio.search_engine import \
perform_request_search, \
search_unit_in_bibxxx
from invenio.bibedit_utils import get_bibrecord
from invenio.bibrecord import record_xml_output, record_add_field
from invenio.pluginutils import PluginContainer
from invenio.intbitset import intbitset
from invenio.dbquery import run_sql
from invenio.bibcatalog import BIBCATALOG_SYSTEM
from invenio.shellutils import split_cli_ids_arg
from invenio.jsonutils import json
CFG_BATCH_SIZE = 100
class RulesParseError(Exception):
""" An exception indicating an error in the rules definition """
def __init__(self, rule_name, error):
Exception.__init__(self, "Invalid rule '%s': %s." % (rule_name,
error))
class AmendableRecord(dict):
""" Class that wraps a record (recstruct) to pass to a plugin """
def __init__(self, record):
dict.__init__(self, record)
self.errors = []
self.amendments = []
self.warnings = []
self.valid = True
self.amended = False
self.holdingpen = False
self.rule = None
self.record_id = self["001"][0][3]
def iterfields(self, fields):
"""
This function accepts a list of marc tags (a 6 character string
containing a 3 character tag, two 1 character indicators and an 1
character subfield code) and returns and yields tuples of marc tags
(without wildcards) and the field value.
Examples:
record.iterfields(["%%%%%%", "%%%%%_"])
--> Iterator of all the field and subfield values.
record.iterfields(["100__a"])
--> The author of record
record.iterfields(["%%%%%u"])
--> All "u" subfields
@param list of marc tags (accepts wildcards)
@yields tuple (position, field_value)
"""
for field in fields:
for res in self.iterfield(field):
yield res
def iterfield(self, field):
""" Like iterfields with only a field """
assert len(field) == 6
field = field.replace("_", " ")
ind1, ind2, code = field[3:]
for tag in self.itertags(field[:3]):
for (local_position, field_obj) in enumerate(self[tag]):
if ind1 in ('%', field_obj[1]) and ind2 in ('%', field_obj[2]):
field_name = tag + field_obj[1] + field_obj[2]
field_name = field_name.replace(' ', '_')
if code == " " and field_obj[3]:
position = field_name + "_", local_position, None
yield position, field_obj[3]
else:
for subfield_position, subfield_tuple in enumerate(field_obj[0]):
subfield_code, value = subfield_tuple
if code in ("%", subfield_code):
position = field_name + subfield_code, local_position, subfield_position
yield position, value
def _query(self, position):
""" Return a position """
tag = position[0].replace("_", " ")
res = self[tag[0:3]]
if position[1] is not None:
res = res[position[1]]
assert res[1] == tag[3] and res[2] == tag[4] # Check indicators
if position[2] is not None:
res = res[0][position[2]]
assert res[0] == tag[5]
return res
def _queryval(self, position):
""" Like _query() but return the value """
if position[2] is None:
return self._query(position)[3]
else:
return self._query(position)[1]
def amend_field(self, position, new_value, message=""):
"""
Changes the value of the field in the specified position to new value
and marks the record as amended.
Optional message to explain what was changed.
"""
tag, localpos, subfieldpos = position
tag = tag.replace("_", " ")
old_value = self._queryval(position)
if new_value != old_value:
if position[2] is None:
fields = self[tag[0:3]]
fields[localpos] = fields[localpos][0:3] + (new_value,)
else:
self._query(position[:2] + (None,))[0][subfieldpos] = (tag[5], new_value)
if message == '':
message = u"Changed field %s from '%s' to '%s'" % (position[0],
old_value.decode('utf-8'), new_value.decode('utf-8'))
self.set_amended(message)
def delete_field(self, position, message=""):
"""
Delete a field or subfield. Returns the deleted field or subfield
"""
if message == "":
message = u"Deleted field %s" % (position[0])
self.set_amended(message)
if position[2] is None:
return self._query(position[:1] + (None, None)).pop(position[1])
else:
return self._query(position[:2] + (None,))[0].pop(position[2])
def add_field(self, tag, value, subfields=None):
""" Add a field """
tag = tag.replace("_", " ")
record_add_field(self, tag[:3], tag[3], tag[4], value, subfields)
self.set_amended("Added field %s" % tag)
def add_subfield(self, position, code, value):
""" Add a subfield to the field in the specified field """
self._query(position[:2] + (None,))[0].append((code, value))
self.set_amended("Added subfield %s='%s' to field %s" % (code, value,
position[0][:5]))
def set_amended(self, message):
""" Mark the record as amended """
write_message("Amended record %s by rule %s: %s" %
(self.record_id, self.rule["name"], message))
self.amendments.append("Rule %s: %s" % (self.rule["name"], message))
self.amended = True
if self.rule["holdingpen"]:
self.holdingpen = True
def set_invalid(self, reason):
""" Mark the record as invalid """
write_message("Record %s marked as invalid by rule %s: %s" %
(CFG_SITE_URL + "/record/%s" % self.record_id, self.rule["name"], reason))
self.errors.append("Rule %s: %s" % (self.rule["name"], reason))
self.valid = False
def warn(self, msg):
""" Add a warning to the record """
self.warnings.append("Rule %s: %s" % (self.rule["name"], msg))
write_message("[WARN] record %s by rule %s: %s" %
(self.record_id, self.rule["name"], msg))
def set_rule(self, rule):
""" Set the current rule the record is been checked against """
self.rule = rule
def itertags(self, tag):
"""
Yields the tags of the record that matching
@param tag: tag with wildcards
@yields tags without wildcards
"""
if "%" in tag:
for key in self.iterkeys():
if ((tag[0] in ("%", key[0])) and
(tag[1] in ("%", key[1])) and
(tag[2] in ("%", key[2]))):
yield key
else:
if tag in self:
yield tag
def is_dummy(self):
return len(list(self.iterfield("001%%_"))) == 1 and \
len(self.keys()) == 1
def task_parse_options(key, val, *_):
""" Must be defined for bibtask to create a task """
if key in ("--all", "-a"):
for rule_name in val.split(","):
reset_rule_last_run(rule_name)
elif key in ("--enable-rules", "-e"):
task_set_option("enabled_rules", set(val.split(",")))
elif key in ("--id", "-i"):
task_set_option("record_ids", intbitset(split_cli_ids_arg(val)))
elif key in ("--queue", "-q"):
task_set_option("queue", val)
elif key in ("--no-tickets", "-t"):
task_set_option("no_tickets", True)
elif key in ("--no-upload", "-b"):
task_set_option("no_upload", True)
elif key in ("--dry-run", "-n"):
task_set_option("no_upload", True)
task_set_option("no_tickets", True)
elif key in ("--config", "-c"):
task_set_option("config", val)
else:
raise StandardError("Error: Unrecognised argument '%s'." % key)
return True
def task_run_core():
"""
Main daemon task.
Returns True when run successfully. False otherwise.
"""
plugins = load_plugins()
rules = load_rules(plugins)
task_set_option('plugins', plugins)
recids_for_rules = get_recids_for_rules(rules)
all_recids = intbitset([])
single_rules = set()
batch_rules = set()
for rule_name, rule_recids in recids_for_rules.iteritems():
all_recids.union_update(rule_recids)
if plugins[rules[rule_name]["check"]]["batch"]:
batch_rules.add(rule_name)
else:
single_rules.add(rule_name)
for batch in iter_batches(all_recids, CFG_BATCH_SIZE):
records_to_upload_holdingpen = []
records_to_upload_replace = []
for rule_name in batch_rules:
rule = rules[rule_name]
rule_recids = recids_for_rules[rule_name]
task_sleep_now_if_required(can_stop_too=True)
records = []
for i, record_id, record in batch:
if record_id in rule_recids:
records.append(record)
if len(records):
check_records(rule, records)
# Then run them trught normal rules
for i, record_id, record in batch:
progress_percent = int(float(i) / len(all_recids) * 100)
task_update_progress("Processing record %s/%s (%i%%)." %
(i, len(all_recids), progress_percent))
write_message("Processing record %s" % record_id)
for rule_name in single_rules:
rule = rules[rule_name]
rule_recids = recids_for_rules[rule_name]
task_sleep_now_if_required(can_stop_too=True)
if record_id in rule_recids:
check_record(rule, record)
if record.amended:
if record.holdingpen:
records_to_upload_holdingpen.append(record)
else:
records_to_upload_replace.append(record)
if not record.valid:
submit_ticket(record, record_id)
upload_amendments(records_to_upload_holdingpen, True)
upload_amendments(records_to_upload_replace, False)
# Update the database with the last time the rules was ran
for rule in rules.keys():
update_rule_last_run(rule)
return True
def submit_ticket(record, record_id):
""" Submit the errors to bibcatalog """
if task_get_option("no_tickets", False):
return
msg = """
Bibcheck found some problems with the record with id %s:
Errors:
%s
Amendments:
%s
Warnings:
%s
Edit this record: %s
"""
msg = msg % (
record_id,
"\n".join(record.errors),
"\n".join(record.amendments),
"\n".join(record.warnings),
"%s/record/%s/edit" % (CFG_SITE_URL, record_id),
)
if isinstance(msg, unicode):
msg = msg.encode("utf-8")
subject = "Bibcheck rule failed in record %s" % record_id
res = BIBCATALOG_SYSTEM.ticket_submit(
subject=subject,
recordid=record_id,
text=subject,
queue=task_get_option("queue", "Bibcheck")
)
write_message("Bibcatalog returned %s" % res)
if res > 0:
BIBCATALOG_SYSTEM.ticket_comment(None, res, msg)
def upload_amendments(records, holdingpen):
""" Upload a modified record """
if task_get_option("no_upload", False) or len(records) == 0:
return
xml = '<collection xmlns="http://www.loc.gov/MARC21/slim">'
for record in records:
xml += record_xml_output(record)
xml += "</collection>"
tmp_file_fd, tmp_file = mkstemp(
suffix='.xml',
prefix="bibcheckfile_%s" % time.strftime("%Y-%m-%d_%H:%M:%S"),
dir=CFG_TMPSHAREDDIR
)
os.write(tmp_file_fd, xml)
os.close(tmp_file_fd)
os.chmod(tmp_file, 0644)
if holdingpen:
flag = "-o"
else:
flag = "-r"
task = task_low_level_submission('bibupload', 'bibcheck', flag, tmp_file)
write_message("Submitted bibupload task %s" % task)
def check_record(rule, record):
"""
Check a record against a rule
"""
plugins = task_get_option("plugins")
record.set_rule(rule)
plugin = plugins[rule["check"]]
if not record.is_dummy():
return plugin["check_record"](record, **rule["checker_params"])
def check_records(rule, records):
"""
Check a set of records against a batch rule
"""
plugins = task_get_option("plugins")
for record in records:
record.set_rule(rule)
plugin = plugins[rule["check"]]
return plugin["check_records"](records, **rule["checker_params"])
def get_rule_lastrun(rule_name):
"""
Get the last time a rule was run, or the oldest representable datetime
if the rule was never ran.
"""
res = run_sql("SELECT last_run FROM bibcheck_rules WHERE name=%s;",
(rule_name,))
if len(res) == 0 or res[0][0] is None:
return datetime(1900, 1, 1)
else:
return res[0][0]
def update_rule_last_run(rule_name):
"""
Set the last time a rule was run to now. This function should be called
after a rule has been ran.
"""
if task_has_option('record_ids') or task_get_option('no_upload', False) \
or task_get_option('no_tickets', False):
return # We don't want to update the database in this case
updated = run_sql("UPDATE bibcheck_rules SET last_run=NOW() WHERE name=%s;",
(rule_name,))
if not updated: # rule not in the database, insert it
run_sql("INSERT INTO bibcheck_rules(name, last_run) VALUES (%s, NOW())",
(rule_name,))
def reset_rule_last_run(rule_name):
"""
Reset the last time a rule was run. This will cause the rule to be
ran on all matching records (not only modified ones)
"""
run_sql("DELETE FROM bibcheck_rules WHERE name=%s", (rule_name,))
def load_plugins():
"""
Will load all the plugins found under the bibcheck_plugins folder.
Returns a list of plugin objects.
"""
plugin_dir = os.path.join(CFG_PYLIBDIR, "invenio/bibcheck_plugins/*.py")
# Load plugins
plugins = PluginContainer(plugin_dir,
plugin_builder=_bibcheck_plugin_builder)
# Check for broken plug-ins
broken = plugins.get_broken_plugins()
if broken:
for plugin, info in broken.items():
print "Failed to load %s:\n" % plugin
print "".join(traceback.format_exception(*info))
enabled = plugins.get_enabled_plugins()
enabled.pop("__init__", None)
return enabled
def load_rule(config, plugins, rule_name):
"""Read rule 'rule_name' from the config file """
checker_params = {}
rule = {
"checker_params": checker_params,
"holdingpen": False,
"name": rule_name
}
def encode(obj):
""" Encode a decoded json object strings """
if isinstance(obj, dict):
return dict([(encode(key), encode(value)) for key, value in
obj.iteritems()])
elif isinstance(obj, list):
return [encode(element) for element in obj]
elif isinstance(obj, unicode):
return obj.encode('utf-8')
else:
return obj
def parse_arg(argument_str, arg_name):
try:
return encode(json.loads(argument_str))
except ValueError:
raise RulesParseError(rule_name, "Invalid value in argument '%s'" %
arg_name)
for key, val in config.items(rule_name):
if key in ("filter_pattern",
"filter_field",
"filter_collection",
"filter_limit"):
rule[key] = val
elif key in ("holdingpen",
"consider_deleted_records"):
rule[key] = val.lower() in ("true", "1", "yes", "on")
elif key == "check":
rule["check"] = val
if val not in plugins:
raise RulesParseError(rule_name, "Invalid checker '%s'" % val)
elif key.startswith("check."):
checker_params[key[len("check."):]] = parse_arg(val, key)
else:
raise RulesParseError(rule_name, "Invalid rule option '%s'" % key)
if "check" not in rule:
raise RulesParseError(rule_name, "Doesn't have a checker")
plugin = plugins[rule["check"]]
if not plugin["mandatory_args"].issubset(checker_params.keys()):
raise RulesParseError(rule_name, "Plugin mandatory argument not specified")
if not plugin["all_args"].issuperset(checker_params.keys()):
raise RulesParseError(rule_name, "Unknown plugin argument")
return rule
def load_rules(plugins):
"""
Load the rules and return a dict with the rules
"""
config = task_get_option("config", "rules.cfg")
filename = os.path.join(CFG_ETCDIR, "bibcheck/", config)
config = RawConfigParser()
config.readfp(open(filename))
rules = {}
rule_names = config.sections()
enabled = task_get_option("enabled_rules", None)
if enabled is not None:
rule_names = enabled.intersection(rule_names)
for rule_name in rule_names:
try:
rules[rule_name] = load_rule(config, plugins, rule_name)
except RulesParseError, ex:
print ex
write_message(ex)
return rules
def get_recids_for_rules(rules):
"""
Generates the final list of record IDs to load.
@param rules dict of rules {rule_name: rule_dict}
@type rules: dict of rules
@return dict {rule_name: array of record IDs}
"""
override_record_ids = task_get_option("record_ids")
recids = {}
for rule_name, rule in rules.iteritems():
if "filter_pattern" in rule:
query = rule["filter_pattern"]
if "filter_collection" in rule:
collections = rule["filter_collection"].split()
else:
collections = None
write_message("Performing given search query: '%s'" % query)
result = intbitset(perform_request_search(
p=query,
of='id',
wl=rule.get('filter_limit', 0),
f=rule.get('filter_field', None),
c=collections
))
else:
result = intbitset(trailing_bits=True)
if override_record_ids is not None:
result.intersection_update(override_record_ids)
else:
last_run = get_rule_lastrun(rule_name)
modified_recids = get_modified_records_since(last_run)
if not "consider_deleted_records" in rule:
modified_recids -= search_unit_in_bibxxx(p='DELETED', f='980__%', type='e')
if CFG_CERN_SITE:
modified_recids -= search_unit_in_bibxxx(p='DUMMY', f='980__%', type='e')
result.intersection_update(modified_recids)
recids[rule_name] = result
return recids
def iter_batches(records, batch_size):
"""
like enumerate_records(), but yield batches of records of size
batch_size instead of records
"""
iterator = enumerate_records(records)
while True:
batch = list(itertools.islice(iterator, batch_size))
if len(batch) > 0:
yield batch
else:
return
def enumerate_records(records):
"""
Given an array of record IDs this function will yield a
triplet of the count (starting from 0), the record ID and
the record object.
@param record: Array of record IDs
@type record: int
@yield: tuple (count, recordId, record structure (dict))
"""
for i, recid in enumerate(records):
record = get_bibrecord(int(recid))
if not record:
write_message("Error: could not load record '%s'." % (recid,))
continue
yield i, int(recid), AmendableRecord(record)
def _bibcheck_plugin_builder(plugin_name, plugin_code):
"""
Custom builder for pluginutils.
@param plugin_name: the name of the plugin.
@type plugin_name: string
@param plugin_code: the code of the module as just read from
filesystem.
@type plugin_code: module
@return: the plugin
"""
if plugin_name == "__init__":
return
plugin = {}
plugin["check_record"] = getattr(plugin_code, "check_record", None)
plugin["check_records"] = getattr(plugin_code, "check_records", None)
plugin["name"] = plugin_name
if (plugin["check_record"] is None) == (plugin["check_records"] is None):
raise Exception("Plugin doesn't implement one check_record method")
plugin["batch"] = plugin["check_records"] is not None
argspec = inspect.getargspec(plugin["check_record"] or plugin["check_records"])
args, defaults = argspec[0], argspec[3]
if len(args) == 0:
msg = "Plugin %s: check_record must accept at least one argument"
raise Exception(msg % plugin_name)
mandatory_args = args[1:len(args)-len(defaults or [])]
plugin["mandatory_args"] = set(mandatory_args)
plugin["all_args"] = set(args[1:])
return plugin
def print_rules():
"""Prints the valid rules to stdout"""
plugins = load_plugins()
for rule_name, rule in load_rules(plugins).items():
print "Rule %s:" % rule_name
if "filter_pattern" in rule:
print " - Filter: %s" % rule["filter_pattern"]
if "filter_collection" in rule:
print " - Filter collection: %s" % rule["filter_collection"]
print " - Checker: %s" % rule["check"]
if len(rule["checker_params"]) > 0:
print " Parameters:"
for param, val in rule["checker_params"].items():
print " %s = %s" % (param, json.dumps(val))
print
def print_plugins():
"""Prints the enabled plugins to stdout"""
all_plugins = load_plugins()
print "Enabled plugins:"
for plugin in all_plugins.values():
print " -%s" % plugin["name"]
optional_args = plugin["all_args"].difference(plugin["mandatory_args"])
if len(plugin["mandatory_args"]):
print " Mandatory args: ", ", ".join(plugin["mandatory_args"])
if len(optional_args):
print " Optional args: ", ", ".join(optional_args)
print
def main():
"""Constructs the BibCheck bibtask."""
usage = """
Scheduled (daemon) options:
-l, --list-plugins List all plugins and exit
-r, --list-rules List all rules and exit
-e, --enable-rules=rules Enable only some rules (comma separated)
-a, --all=rules Run the specified rules in all matching records (not
only modified ones)
-i, --id=ids Run only in the specified record ids or ranges (comma
separated), ignoring all other filters
-q, --queue=queue Create tickets in the specified RT Queue (Default
Bibcheck)
-t, --no-tickets Don't create any ticket in RT. Useful for debugging
-b, --no-upload Don't upload changes to the database
-n, --dry-run Like --no-tickets and --no-upload
-c, --config By default bibcheck reads the file rules.cfg. This
allows to specify a different config file
If any of the options --id, --no-tickets, --no-upload or --dry-run is enabled,
bibcheck won't update the last-run-time of a task in the database.
Examples:
(run a periodical daemon job that checks the rules from rules.cfg)
bibcheck -s1d
(Run bibcheck on records 1, 2, 3, 5, 6, 7, 8, 9 and 10)
bibcheck -i 1,2,3,5-10
(Run only the rule foobar in all the records)
bibcheck -a foobar -e foobar
(Run only the rules foo and bar on modified records)
bibcheck -e foo,bar
"""
try:
opts = getopt.getopt(sys.argv[1:], "lr",
["list-plugins", "list-rules"])[0]
except getopt.GetoptError:
opts = []
for opt, dummy in opts:
if opt in ["-l", "--list-plugins"]:
print_plugins()
return
elif opt in ["-r", "--list-rules"]:
print_rules()
return
# Build and submit the task
task_init(authorization_action='runbibcheck',
authorization_msg="BibCheck Task Submission",
description="",
help_specific_usage=usage,
version="Invenio v%s" % CFG_VERSION,
specific_params=("hvtbnV:e:a:i:q:c:", ["help", "version",
"verbose=", "enable-rules=", "all=", "id=", "queue=",
"no-tickets", "no-upload", "dry-run", "config"]),
task_submit_elaborate_specific_parameter_fnc=task_parse_options,
task_run_fnc=task_run_core)
c4science · Help