Homec4science
Diffusion invenio-infoscience (iugw2017) Browse

R3600/modules/bibindex/lib/bibindex_engine.py
7b2f8868dcd2iugw2017

Last Change
Raw File

bibindex_engine.py
View Options

# -*- coding: utf-8 -*-
##
## $Id$
## BibIndxes bibliographic data, reference and fulltext indexing utility.
##
## This file is part of CDS Invenio.
## Copyright (C) 2002, 2003, 2004, 2005, 2006, 2007 CERN.
##
## CDS 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.
##
## CDS 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 CDS Invenio; if not, write to the Free Software Foundation, Inc.,
## 59 Temple Place, Suite 330, Boston, MA 02111-1307, USA.
"""
BibIndex indexing engine implementation. See bibindex executable for entry point.
"""
__revision__ = "$Id$"
import marshal
from zlib import compress, decompress
import os
import re
import sys
import time
import Numeric
import urllib2
import tempfile
import traceback
import cStringIO
from invenio.config import \
CFG_BIBINDEX_CHARS_ALPHANUMERIC_SEPARATORS, \
CFG_BIBINDEX_CHARS_PUNCTUATION, \
CFG_BIBINDEX_FULLTEXT_INDEX_LOCAL_FILES_ONLY, \
CFG_BIBINDEX_MIN_WORD_LENGTH, \
CFG_BIBINDEX_REMOVE_HTML_MARKUP, \
CFG_BIBINDEX_STEMMER_DEFAULT_LANGUAGE, \
CFG_MAX_RECID, \
weburl
from invenio.bibindex_engine_config import *
from invenio.search_engine import perform_request_search, strip_accents
from invenio.dbquery import run_sql, escape_string, DatabaseError
from invenio.bibindex_engine_stopwords import is_stopword
from invenio.bibindex_engine_stemmer import stem
from invenio.bibtask import task_init, write_message, get_datetime, \
task_set_option, task_get_option, task_get_task_param, task_update_status, \
task_update_progress
## import optional modules:
try:
import psyco
psyco.bind(get_words_from_phrase)
psyco.bind(WordTable.merge_with_old_recIDs)
psyco.bind(serialize_via_numeric_array)
psyco.bind(serialize_via_marshal)
psyco.bind(deserialize_via_numeric_array)
psyco.bind(deserialize_via_marshal)
except:
pass
## precompile some often-used regexp for speed reasons:
re_subfields = re.compile('\$\$\w')
re_html = re.compile("(?s)<[^>]*>|&#?\w+;")
re_block_punctuation_begin = re.compile(r"^"+CFG_BIBINDEX_CHARS_PUNCTUATION+"+")
re_block_punctuation_end = re.compile(CFG_BIBINDEX_CHARS_PUNCTUATION+"+$")
re_punctuation = re.compile(CFG_BIBINDEX_CHARS_PUNCTUATION)
re_separators = re.compile(CFG_BIBINDEX_CHARS_ALPHANUMERIC_SEPARATORS)
re_datetime_shift = re.compile("([-\+]{0,1})([\d]+)([dhms])")
nb_char_in_line = 50 # for verbose pretty printing
chunksize = 1000 # default size of chunks that the records will be treated by
wordTables = []
base_process_size = 4500 # process base size
_last_word_table = None
## Dictionary merging functions
def intersection(dict1, dict2):
"Returns intersection of the two dictionaries."
int_dict = {}
if len(dict1) > len(dict2):
for e in dict2:
if dict1.has_key(e):
int_dict[e] = 1
else:
for e in dict1:
if dict2.has_key(e):
int_dict[e] = 1
return int_dict
def union(dict1, dict2):
"Returns union of the two dictionaries."
union_dict = {}
for e in dict1.keys():
union_dict[e] = 1
for e in dict2.keys():
union_dict[e] = 1
return union_dict
def diff(dict1, dict2):
"Returns dict1 - dict2."
diff_dict = {}
for e in dict1.keys():
if not dict2.has_key(e):
diff_dict[e] = 1
return diff_dict
def list_union(list1, list2):
"Returns union of the two lists."
union_dict = {}
for e in list1:
union_dict[e] = 1
for e in list2:
union_dict[e] = 1
return union_dict.keys()
## safety function for killing slow DB threads:
def kill_sleepy_mysql_threads(max_threads=CFG_MAX_MYSQL_THREADS, thread_timeout=CFG_MYSQL_THREAD_TIMEOUT):
"""Check the number of DB threads and if there are more than
MAX_THREADS of them, lill all threads that are in a sleeping
state for more than THREAD_TIMEOUT seconds. (This is useful
for working around the the max_connection problem that appears
during indexation in some not-yet-understood cases.) If some
threads are to be killed, write info into the log file.
"""
res = run_sql("SHOW FULL PROCESSLIST")
if len(res) > max_threads:
for row in res:
r_id, dummy, dummy, dummy, r_command, r_time, dummy, dontcare = row
if r_command == "Sleep" and int(r_time) > thread_timeout:
run_sql("KILL %s", (r_id,))
write_message("WARNING: too many DB threads, killing thread %s" % r_id, verbose=1)
return
## MARC-21 tag/field access functions
def get_fieldvalues(recID, tag):
"""Returns list of values of the MARC-21 'tag' fields for the record
'recID'."""
out = []
bibXXx = "bib" + tag[0] + tag[1] + "x"
bibrec_bibXXx = "bibrec_" + bibXXx
query = "SELECT value FROM %s AS b, %s AS bb WHERE bb.id_bibrec=%s AND bb.id_bibxxx=b.id AND tag LIKE '%s'" \
% (bibXXx, bibrec_bibXXx, recID, tag)
res = run_sql(query)
for row in res:
out.append(row[0])
return out
def get_associated_subfield_value(recID, tag, value, associated_subfield_code):
"""Return list of ASSOCIATED_SUBFIELD_CODE, if exists, for record
RECID and TAG of value VALUE. Used by fulltext indexer only.
Note: TAG must be 6 characters long (tag+ind1+ind2+sfcode),
otherwise en empty string is returned.
FIXME: what if many tag values have the same value but different
associated_subfield_code? Better use bibrecord library for this.
"""
out = ""
if len(tag) != 6:
return out
bibXXx = "bib" + tag[0] + tag[1] + "x"
bibrec_bibXXx = "bibrec_" + bibXXx
query = """SELECT bb.field_number, b.tag, b.value FROM %s AS b, %s AS bb
WHERE bb.id_bibrec=%s AND bb.id_bibxxx=b.id AND tag LIKE '%s%%'""" % \
(bibXXx, bibrec_bibXXx, recID, tag[:-1])
res = run_sql(query)
field_number = -1
for row in res:
if row[1] == tag and row[2] == value:
field_number = row[0]
if field_number > 0:
for row in res:
if row[0] == field_number and row[1] == tag[:-1] + associated_subfield_code:
out = row[2]
break
return out
def get_field_tags(field):
"""Returns a list of MARC tags for the field code 'field'.
Returns empty list in case of error.
Example: field='author', output=['100__%','700__%']."""
out = []
query = """SELECT t.value FROM tag AS t, field_tag AS ft, field AS f
WHERE f.code='%s' AND ft.id_field=f.id AND t.id=ft.id_tag
ORDER BY ft.score DESC""" % field
res = run_sql(query)
for row in res:
out.append(row[0])
return out
## Fulltext word extraction functions
def get_fulltext_urls_from_html_page(htmlpagebody):
"""Parses htmlpagebody data looking for url_directs referring to
probable fulltexts.
Returns an array of (ext,url_direct) to fulltexts.
Note: it looks for file format extensions as defined by global
'CONV_PROGRAMS'structure.
"""
out = []
for ext in CONV_PROGRAMS.keys():
expr = re.compile( r"\"(http://[\w]+\.+[\w]+[^\"'><]*\." + \
ext + r")\"")
match = expr.search(htmlpagebody)
if match:
out.append([ext, match.group(1)])
else: # FIXME: workaround for getfile, should use bibdoc tables
expr_getfile = re.compile(r"\"(http://.*getfile\.py\?.*format=" + ext + r"&version=.*)\"")
match = expr_getfile.search(htmlpagebody)
if match:
out.append([ext, match.group(1)])
return out
def get_words_from_fulltext(url_direct_or_indirect,
split=str.split,
lower=str.lower,
force_file_extension=None):
"""Returns all the words contained in the document specified by
URL_DIRECT_OR_INDIRECT with the words being split by various
SRE_SEPARATORS regexp set earlier. If FORCE_FILE_EXTENSION is
set (e.g. to "pdf", then treat URL_DIRECT_OR_INDIRECT as a PDF
file. (This is interesting to index Indico for example.) Note
also that URL_DIRECT_OR_INDIRECT may be either a direct URL to
the fulltext file or an URL to a setlink-like page body that
presents the links to be indexed. In the latter case the
URL_DIRECT_OR_INDIRECT is parsed to extract actual direct URLs
to fulltext documents, for all knows file extensions as
specified by global CONV_PROGRAMS config variable.
"""
if CFG_BIBINDEX_FULLTEXT_INDEX_LOCAL_FILES_ONLY and \
url_direct_or_indirect.find(weburl) < 0:
return []
write_message("... reading fulltext files from %s started" % url_direct_or_indirect, verbose=2)
fulltext_urls = None
if not force_file_extension:
url_direct = None
fulltext_urls = None
# check for direct link in url
url_direct_or_indirect_ext = lower(split(url_direct_or_indirect,".")[-1])
if url_direct_or_indirect_ext in CONV_PROGRAMS.keys():
fulltext_urls = [(url_direct_or_indirect_ext, url_direct_or_indirect)]
# Indirect url. Try to fetch the real fulltext(s)
if not fulltext_urls:
# read "setlink" data
try:
htmlpagebody = urllib2.urlopen(url_direct_or_indirect).read()
except:
sys.stderr.write("Error: Cannot read %s.\n" % url_direct_or_indirect)
return []
fulltext_urls = get_fulltext_urls_from_html_page(htmlpagebody)
write_message("... fulltext_urls = %s" % fulltext_urls, verbose=9)
else:
fulltext_urls = [[force_file_extension, url_direct_or_indirect]]
words = {}
# process as many urls as they were found:
for (ext, url_direct) in fulltext_urls:
write_message(".... processing %s from %s started" % (ext, url_direct), verbose=2)
# sanity check:
if not url_direct:
break
# read fulltext file:
try:
url = urllib2.urlopen(url_direct)
except:
sys.stderr.write("Error: Cannot read %s.\n" % url_direct)
break # try other fulltext files...
tmp_fd, tmp_name = tempfile.mkstemp('invenio.tmp')
data_chunk = url.read(8*1024)
while data_chunk:
os.write(tmp_fd, data_chunk)
data_chunk = url.read(8*1024)
os.close(tmp_fd)
# try all available conversion programs according to their order:
bingo = 0
for conv_program in CONV_PROGRAMS[ext]:
if os.path.exists(conv_program):
# intelligence on how to run various conversion programs:
cmd = "" # wil keep command to run
bingo = 0 # had we success?
if os.path.basename(conv_program) == "pdftotext":
cmd = "%s -enc UTF-8 %s %s.txt" % (conv_program, tmp_name, tmp_name)
elif os.path.basename(conv_program) == "pstotext":
if ext == "ps.gz":
# is there gzip available?
if os.path.exists(CONV_PROGRAMS_HELPERS["gz"]):
cmd = "%s -cd %s | %s > %s.txt" \
% (CONV_PROGRAMS_HELPERS["gz"], tmp_name, conv_program, tmp_name)
else:
cmd = "%s %s > %s.txt" \
% (conv_program, tmp_name, tmp_name)
elif os.path.basename(conv_program) == "ps2ascii":
if ext == "ps.gz":
# is there gzip available?
if os.path.exists(CONV_PROGRAMS_HELPERS["gz"]):
cmd = "%s -cd %s | %s > %s.txt"\
% (CONV_PROGRAMS_HELPERS["gz"], tmp_name,
conv_program, tmp_name)
else:
cmd = "%s %s %s.txt" \
% (conv_program, tmp_name, tmp_name)
elif os.path.basename(conv_program) == "antiword":
cmd = "%s %s > %s.txt" % (conv_program, tmp_name, tmp_name)
elif os.path.basename(conv_program) == "catdoc":
cmd = "%s %s > %s.txt" % (conv_program, tmp_name, tmp_name)
elif os.path.basename(conv_program) == "wvText":
cmd = "%s %s %s.txt" % (conv_program, tmp_name, tmp_name)
elif os.path.basename(conv_program) == "ppthtml":
# is there html2text available?
if os.path.exists(CONV_PROGRAMS_HELPERS["html"]):
cmd = "%s %s | %s > %s.txt"\
% (conv_program, tmp_name,
CONV_PROGRAMS_HELPERS["html"], tmp_name)
else:
cmd = "%s %s > %s.txt" \
% (conv_program, tmp_name, tmp_name)
elif os.path.basename(conv_program) == "xlhtml":
# is there html2text available?
if os.path.exists(CONV_PROGRAMS_HELPERS["html"]):
cmd = "%s %s | %s > %s.txt" % \
(conv_program, tmp_name,
CONV_PROGRAMS_HELPERS["html"], tmp_name)
else:
cmd = "%s %s > %s.txt" % \
(conv_program, tmp_name, tmp_name)
else:
sys.stderr.write("Error: Do not know how to handle %s conversion program.\n" % conv_program)
# try to run it:
try:
write_message("..... launching %s" % cmd, verbose=9)
errcode = os.system(cmd)
if errcode == 0 and os.path.exists("%s.txt" % tmp_name):
bingo = 1
break # bingo!
else:
write_message("Error while running %s for %s.\n" % (cmd, url_direct), sys.stderr)
except:
write_message("Error running %s for %s.\n" % (cmd, url_direct), sys.stderr)
# were we successful?
if bingo:
tmp_name_txt_file = open("%s.txt" % tmp_name)
for phrase in tmp_name_txt_file.xreadlines():
for word in get_words_from_phrase(phrase):
if not words.has_key(word):
words[word] = 1
tmp_name_txt_file.close()
else:
write_message("No conversion success for %s.\n" % (url_direct), sys.stderr)
# delete temp files (they might not exist):
try:
os.unlink(tmp_name)
os.unlink(tmp_name + ".txt")
except StandardError:
write_message("Error: Could not delete file. It didn't exist", sys.stderr)
write_message(".... processing %s from %s ended" % (ext, url_direct), verbose=2)
write_message("... reading fulltext files from %s ended" % url_direct_or_indirect, verbose=2)
return words.keys()
# tagToFunctions mapping. It offers an indirection level necesary for
# indexing fulltext. The default is get_words_from_phrase
tagToWordsFunctions = {'8564_u':get_words_from_fulltext}
def get_words_from_phrase(phrase, split=str.split):
"""Return list of words found in PHRASE. Note that the phrase is
split into groups depending on the alphanumeric characters and
punctuation characters definition present in the config file.
"""
words = {}
if CFG_BIBINDEX_REMOVE_HTML_MARKUP and phrase.find("</") > -1:
phrase = re_html.sub(' ', phrase)
phrase = phrase.lower()
# 1st split phrase into blocks according to whitespace
for block in split(strip_accents(phrase)):
# 2nd remove leading/trailing punctuation and add block:
block = re_block_punctuation_begin.sub("", block)
block = re_block_punctuation_end.sub("", block)
if block:
block = apply_stemming_and_stopwords_and_length_check(block)
if block:
words[block] = 1
# 3rd break each block into subblocks according to punctuation and add subblocks:
for subblock in re_punctuation.split(block):
subblock = apply_stemming_and_stopwords_and_length_check(subblock)
if subblock:
words[subblock] = 1
# 4th break each subblock into alphanumeric groups and add groups:
for alphanumeric_group in re_separators.split(subblock):
alphanumeric_group = apply_stemming_and_stopwords_and_length_check(alphanumeric_group)
if alphanumeric_group:
words[alphanumeric_group] = 1
return words.keys()
def apply_stemming_and_stopwords_and_length_check(word):
"""Return WORD after applying stemming and stopword and length checks.
See the config file in order to influence these.
"""
# stem word, when configured so:
if CFG_BIBINDEX_STEMMER_DEFAULT_LANGUAGE != "":
word = stem(word, CFG_BIBINDEX_STEMMER_DEFAULT_LANGUAGE)
# now check against stopwords:
if is_stopword(word):
return ""
# finally check the word length:
if len(word) < CFG_BIBINDEX_MIN_WORD_LENGTH:
return ""
return word
def remove_subfields(s):
"Removes subfields from string, e.g. 'foo $$c bar' becomes 'foo bar'."
return re_subfields.sub(' ', s)
def get_index_id(indexname):
"""Returns the words/phrase index id for INDEXNAME.
Returns empty string in case there is no words table for this index.
Example: field='author', output=4."""
out = 0
query = """SELECT w.id FROM idxINDEX AS w
WHERE w.name='%s' LIMIT 1""" % indexname
res = run_sql(query, None, 1)
if res:
out = res[0][0]
return out
def get_index_tags(indexname):
"""Returns the list of tags that are indexed inside INDEXNAME.
Returns empty list in case there are no tags indexed in this index.
Note: uses get_field_tags() defined before.
Example: field='author', output=['100__%', '700__%']."""
out = []
query = """SELECT f.code FROM idxINDEX AS w, idxINDEX_field AS wf,
field AS f WHERE w.name='%s' AND w.id=wf.id_idxINDEX
AND f.id=wf.id_field""" % indexname
res = run_sql(query)
for row in res:
out.extend(get_field_tags(row[0]))
return out
def get_all_indexes():
"""Returns the list of the names of all defined words indexes.
Returns empty list in case there are no tags indexed in this index.
Example: output=['global', 'author']."""
out = []
query = """SELECT name FROM idxINDEX"""
res = run_sql(query)
for row in res:
out.append(row[0])
return out
def split_ranges(parse_string):
"""Parse a string a return the list or ranges."""
recIDs = []
ranges = parse_string.split(",")
for arange in ranges:
tmp_recIDs = arange.split("-")
if len(tmp_recIDs)==1:
recIDs.append([int(tmp_recIDs[0]), int(tmp_recIDs[0])])
else:
if int(tmp_recIDs[0]) > int(tmp_recIDs[1]): # sanity check
tmp = tmp_recIDs[0]
tmp_recIDs[0] = tmp_recIDs[1]
tmp_recIDs[1] = tmp
recIDs.append([int(tmp_recIDs[0]), int(tmp_recIDs[1])])
return recIDs
def get_word_tables(tables):
global wordTables
if tables:
indexes = tables.split(",")
for index in indexes:
index_id = get_index_id(index)
if index_id:
wordTables.append({"idxWORD%02dF" % index_id: \
get_index_tags(index)})
else:
write_message("Error: There is no %s words table." % index, sys.stderr)
else:
for index in get_all_indexes():
index_id = get_index_id(index)
wordTables.append({"idxWORD%02dF" % index_id: \
get_index_tags(index)})
return wordTables
def get_date_range(var):
"Returns the two dates contained as a low,high tuple"
limits = var.split(",")
if len(limits)==1:
low = get_datetime(limits[0])
return low, None
if len(limits)==2:
low = get_datetime(limits[0])
high = get_datetime(limits[1])
return low, high
return None, None
def create_range_list(res):
"""Creates a range list from a recID select query result contained
in res. The result is expected to have ascending numerical order."""
if not res:
return []
row = res[0]
if not row:
return []
else:
range_list = [[row[0], row[0]]]
for row in res[1:]:
row_id = row[0]
if row_id == range_list[-1][1] + 1:
range_list[-1][1] = row_id
else:
range_list.append([row_id, row_id])
return range_list
def beautify_range_list(range_list):
"""Returns a non overlapping, maximal range list"""
ret_list = []
for new in range_list:
found = 0
for old in ret_list:
if new[0] <= old[0] <= new[1] + 1 or new[0] - 1 <= old[1] <= new[1]:
old[0] = min(old[0], new[0])
old[1] = max(old[1], new[1])
found = 1
break
if not found:
ret_list.append(new)
return ret_list
def serialize_via_numeric_array(arr):
"""Serialize Numeric array into a compressed string."""
return compress(Numeric.dumps(arr))
def deserialize_via_numeric_array(string):
"""Decompress and deserialize string into a Numeric array."""
return Numeric.loads(decompress(string))
def serialize_via_marshal(obj):
"""Serialize Python object via marshal into a compressed string."""
return escape_string(compress(marshal.dumps(obj)))
def deserialize_via_marshal(string):
"""Decompress and deserialize string into a Python object via marshal."""
return marshal.loads(decompress(string))
class WordTable:
"A class to hold the words table."
def __init__(self, tablename, fields_to_index, separators="[^\s]"):
"Creates words table instance."
self.tablename = tablename
self.recIDs_in_mem = []
self.fields_to_index = fields_to_index
self.separators = separators
self.value = {}
def get_field(self, recID, tag):
"""Returns list of values of the MARC-21 'tag' fields for the
record 'recID'."""
out = []
bibXXx = "bib" + tag[0] + tag[1] + "x"
bibrec_bibXXx = "bibrec_" + bibXXx
query = """SELECT value FROM %s AS b, %s AS bb
WHERE bb.id_bibrec=%s AND bb.id_bibxxx=b.id
AND tag LIKE '%s'""" % (bibXXx, bibrec_bibXXx, recID, tag);
res = run_sql(query)
for row in res:
out.append(row[0])
return out
def clean(self):
"Cleans the words table."
self.value = {}
def put_into_db(self, mode="normal"):
"""Updates the current words table in the corresponding DB
idxFOO table. Mode 'normal' means normal execution,
mode 'emergency' means words index reverting to old state.
"""
write_message("%s %s wordtable flush started" % (self.tablename, mode), verbose=2)
write_message('...updating %d words into %s started' % \
(len(self.value), self.tablename), verbose=2)
task_update_progress("%s flushed %d/%d words" % (self.tablename, 0, len(self.value)))
self.recIDs_in_mem = beautify_range_list(self.recIDs_in_mem)
if mode == "normal":
for group in self.recIDs_in_mem:
query = """UPDATE %sR SET type='TEMPORARY' WHERE id_bibrec
BETWEEN '%d' AND '%d' AND type='CURRENT'""" % \
(self.tablename[:-1], group[0], group[1])
write_message(query, verbose=9)
run_sql(query)
nb_words_total = len(self.value)
nb_words_report = int(nb_words_total/10.0)
nb_words_done = 0
for word in self.value.keys():
self.put_word_into_db(word)
nb_words_done += 1
if nb_words_report != 0 and ((nb_words_done % nb_words_report) == 0):
write_message('......processed %d/%d words' % (nb_words_done, nb_words_total))
task_update_progress("%s flushed %d/%d words" % (self.tablename, nb_words_done, nb_words_total))
write_message('...updating %d words into %s ended' % \
(nb_words_total, self.tablename), verbose=9)
write_message('...updating reverse table %sR started' % self.tablename[:-1])
if mode == "normal":
for group in self.recIDs_in_mem:
query = """UPDATE %sR SET type='CURRENT' WHERE id_bibrec
BETWEEN '%d' AND '%d' AND type='FUTURE'""" % \
(self.tablename[:-1], group[0], group[1])
write_message(query, verbose=9)
run_sql(query)
query = """DELETE FROM %sR WHERE id_bibrec
BETWEEN '%d' AND '%d' AND type='TEMPORARY'""" % \
(self.tablename[:-1], group[0], group[1])
write_message(query, verbose=9)
run_sql(query)
write_message('End of updating wordTable into %s' % self.tablename, verbose=9)
elif mode == "emergency":
for group in self.recIDs_in_mem:
query = """UPDATE %sR SET type='CURRENT' WHERE id_bibrec
BETWEEN '%d' AND '%d' AND type='TEMPORARY'""" % \
(self.tablename[:-1], group[0], group[1])
write_message(query, verbose=9)
run_sql(query)
query = """DELETE FROM %sR WHERE id_bibrec
BETWEEN '%d' AND '%d' AND type='FUTURE'""" % \
(self.tablename[:-1], group[0], group[1])
write_message(query, verbose=9)
run_sql(query)
write_message('End of emergency flushing wordTable into %s' % self.tablename, verbose=9)
write_message('...updating reverse table %sR ended' % self.tablename[:-1])
self.clean()
self.recIDs_in_mem = []
write_message("%s %s wordtable flush ended" % (self.tablename, mode))
task_update_progress("%s flush ended" % (self.tablename))
def load_old_recIDs(self, word):
"""Load existing hitlist for the word from the database index files."""
query = "SELECT hitlist FROM %s WHERE term=%%s" % self.tablename
res = run_sql(query, (word,))
if res:
return deserialize_via_numeric_array(res[0][0])
else:
return None
def merge_with_old_recIDs(self,word,set):
"""Merge the system numbers stored in memory (hash of recIDs with value +1 or -1
according to whether to add/delete them) with those stored in the database index
and received in set universe of recIDs for the given word.
Return 0 in case no change was done to SET, return 1 in case SET was changed.
"""
set_changed_p = 0
for recID,sign in self.value[word].items():
if sign == -1 and set[recID]==1:
# delete recID if existent in set and if marked as to be deleted
set[recID] = 0
set_changed_p = 1
elif set[recID] == 0:
# add recID if not existent in set and if marked as to be added
set[recID] = 1
set_changed_p = 1
return set_changed_p
def put_word_into_db(self, word):
"""Flush a single word to the database and delete it from memory"""
set = self.load_old_recIDs(word)
if set: # merge the word recIDs found in memory:
if self.merge_with_old_recIDs(word,set) == 0:
# nothing to update:
write_message("......... unchanged hitlist for ``%s''" % word, verbose=9)
pass
else:
# yes there were some new words:
write_message("......... updating hitlist for ``%s''" % word, verbose=9)
run_sql("UPDATE %s SET hitlist=%%s WHERE term=%%s" % self.tablename,
(serialize_via_numeric_array(set), word))
else: # the word is new, will create new set:
set = Numeric.zeros(CFG_MAX_RECID+1, Numeric.Int0)
Numeric.put(set, self.value[word].keys(), 1)
write_message("......... inserting hitlist for ``%s''" % word, verbose=9)
run_sql("INSERT INTO %s (term, hitlist) VALUES (%%s, %%s)" % self.tablename,
(word, serialize_via_numeric_array(set)))
if not set: # never store empty words
run_sql("DELETE from %s WHERE term=%%s" % self.tablename,
(word,))
del self.value[word]
def display(self):
"Displays the word table."
keys = self.value.keys()
keys.sort()
for k in keys:
write_message("%s: %s" % (k, self.value[k]))
def count(self):
"Returns the number of words in the table."
return len(self.value)
def info(self):
"Prints some information on the words table."
write_message("The words table contains %d words." % self.count())
def lookup_words(self, word=""):
"Lookup word from the words table."
if not word:
done = 0
while not done:
try:
word = raw_input("Enter word: ")
done = 1
except (EOFError, KeyboardInterrupt):
return
if self.value.has_key(word):
write_message("The word '%s' is found %d times." \
% (word, len(self.value[word])))
else:
write_message("The word '%s' does not exist in the word file."\
% word)
def update_last_updated(self, starting_time=None):
"""Update last_updated column of the index table in the database.
Puts starting time there so that if the task was interrupted for record download,
the records will be reindexed next time."""
if starting_time is None:
return None
write_message("updating last_updated to %s...", starting_time, verbose=9)
return run_sql("UPDATE idxINDEX SET last_updated=%s WHERE id=%s",
(starting_time, self.tablename[-3:-1],))
def add_recIDs(self, recIDs, opt_flush):
"""Fetches records which id in the recIDs range list and adds
them to the wordTable. The recIDs range list is of the form:
[[i1_low,i1_high],[i2_low,i2_high], ..., [iN_low,iN_high]].
"""
global chunksize, _last_word_table
flush_count = 0
records_done = 0
records_to_go = 0
for arange in recIDs:
records_to_go = records_to_go + arange[1] - arange[0] + 1
time_started = time.time() # will measure profile time
for arange in recIDs:
i_low = arange[0]
chunksize_count = 0
while i_low <= arange[1]:
# calculate chunk group of recIDs and treat it:
i_high = min(i_low+opt_flush-flush_count-1,arange[1])
i_high = min(i_low+chunksize-chunksize_count-1, i_high)
try:
self.chk_recID_range(i_low, i_high)
except StandardError, e:
write_message("Exception caught: %s" % e, sys.stderr)
if task_get_option('verbose') >= 9:
traceback.print_tb(sys.exc_info()[2])
task_update_status("ERROR")
self.put_into_db()
sys.exit(1)
write_message("%s adding records #%d-#%d started" % \
(self.tablename, i_low, i_high))
if CFG_CHECK_MYSQL_THREADS:
kill_sleepy_mysql_threads()
task_update_progress("%s adding recs %d-%d" % (self.tablename, i_low, i_high))
self.del_recID_range(i_low, i_high)
just_processed = self.add_recID_range(i_low, i_high)
flush_count = flush_count + i_high - i_low + 1
chunksize_count = chunksize_count + i_high - i_low + 1
records_done = records_done + just_processed
write_message("%s adding records #%d-#%d ended " % \
(self.tablename, i_low, i_high))
if chunksize_count >= chunksize:
chunksize_count = 0
# flush if necessary:
if flush_count >= opt_flush:
self.put_into_db()
self.clean()
write_message("%s backing up" % (self.tablename))
flush_count = 0
self.log_progress(time_started,records_done,records_to_go)
# iterate:
i_low = i_high + 1
if flush_count > 0:
self.put_into_db()
self.log_progress(time_started,records_done,records_to_go)
def add_recIDs_by_date(self, dates, opt_flush):
"""Add records that were modified between DATES[0] and DATES[1].
If DATES is not set, then add records that were modified since
the last update of the index.
"""
if not dates:
table_id = self.tablename[-3:-1]
query = """SELECT last_updated FROM idxINDEX WHERE id='%s'
""" % table_id
res = run_sql(query)
if not res:
return
if not res[0][0]:
dates = ("0000-00-00", None)
else:
dates = (res[0][0], None)
if dates[1] is None:
res = run_sql("""SELECT b.id FROM bibrec AS b
WHERE b.modification_date >= %s ORDER BY b.id ASC""",
(dates[0],))
elif dates[0] is None:
res = run_sql("""SELECT b.id FROM bibrec AS b
WHERE b.modification_date <= %s ORDER BY b.id ASC""",
(dates[1],))
else:
res = run_sql("""SELECT b.id FROM bibrec AS b
WHERE b.modification_date >= %s AND
b.modification_date <= %s ORDER BY b.id ASC""",
(dates[0], dates[1]))
alist = create_range_list(res)
if not alist:
write_message( "No new records added. %s is up to date" % self.tablename)
else:
self.add_recIDs(alist, opt_flush)
def add_recID_range(self, recID1, recID2):
empty_list_string = serialize_via_marshal([])
wlist = {}
self.recIDs_in_mem.append([recID1,recID2])
# secondly fetch all needed tags:
for tag in self.fields_to_index:
if tag in tagToWordsFunctions.keys():
get_words_function = tagToWordsFunctions[tag]
else:
get_words_function = get_words_from_phrase
bibXXx = "bib" + tag[0] + tag[1] + "x"
bibrec_bibXXx = "bibrec_" + bibXXx
query = """SELECT bb.id_bibrec,b.value FROM %s AS b, %s AS bb
WHERE bb.id_bibrec BETWEEN %d AND %d
AND bb.id_bibxxx=b.id AND tag LIKE '%s'""" % (bibXXx, bibrec_bibXXx, recID1, recID2, tag)
res = run_sql(query)
for row in res:
recID,phrase = row
if not wlist.has_key(recID): wlist[recID] = []
if tag == "8564_u":
# Special treatment for fulltext indexing. 8564
# $$u contains URL, and $$y link name. If $$y is
# actually a file name, that is if it ends with
# something like .pdf or .ppt, then $$u is treated
# as direct URL to the PDF file, and is indexed as
# such. This is useful to index Indico files.
# FIXME: this is a quick fix only. We should
# rather download all 856 $$u files and analyze
# content in order to decide how to index them
# (directly for Indico, indirectly for Setlink).
filename = get_associated_subfield_value(recID,'8564_u', phrase, 'y')
filename_extension = filename.split('.')[-1].lower()
if filename_extension in CONV_PROGRAMS.keys():
new_words = get_words_function(phrase, force_file_extension=filename_extension) # ,self.separators
else:
new_words = get_words_function(phrase) # ,self.separators
else:
new_words = get_words_function(phrase) # ,self.separators
wlist[recID] = list_union(new_words,wlist[recID])
# were there some words for these recIDs found?
if len(wlist) == 0: return 0
recIDs = wlist.keys()
for recID in recIDs:
# was this record marked as deleted?
if "DELETED" in self.get_field(recID, "980__c"):
wlist[recID] = []
write_message("... record %d was declared deleted, removing its word list" % recID, verbose=9)
write_message("... record %d, termlist: %s" % (recID, wlist[recID]), verbose=9)
# Using cStringIO for speed.
query_factory = cStringIO.StringIO()
qwrite = query_factory.write
qwrite( "INSERT INTO %sR (id_bibrec,termlist,type) VALUES" % self.tablename[:-1])
qwrite( "('" )
qwrite( str(recIDs[0]) )
qwrite( "','" )
qwrite( serialize_via_marshal(wlist[recIDs[0]]) )
qwrite( "','FUTURE')" )
for recID in recIDs[1:]:
qwrite(",('")
qwrite(str(recID))
qwrite("','")
qwrite(serialize_via_marshal(wlist[recID]))
qwrite("','FUTURE')")
query = query_factory.getvalue()
query_factory.close()
run_sql(query)
query_factory = cStringIO.StringIO()
qwrite = query_factory.write
qwrite("INSERT INTO %sR (id_bibrec,termlist,type) VALUES" % self.tablename[:-1])
qwrite("('")
qwrite(str(recIDs[0]))
qwrite("','")
qwrite(serialize_via_marshal(wlist[recIDs[0]]))
qwrite("','CURRENT')")
for recID in recIDs[1:]:
qwrite( ",('" )
qwrite( str(recID) )
qwrite( "','" )
qwrite( empty_list_string )
qwrite( "','CURRENT')" )
query = query_factory.getvalue()
query_factory.close()
try:
run_sql(query)
except DatabaseError:
# ok, we tried to add an existent record. No problem
pass
put = self.put
for recID in recIDs:
for w in wlist[recID]:
put(recID, w, 1)
return len(recIDs)
def log_progress(self, start, done, todo):
"""Calculate progress and store it.
start: start time,
done: records processed,
todo: total number of records"""
time_elapsed = time.time() - start
# consistency check
if time_elapsed == 0 or done > todo:
return
time_recs_per_min = done/(time_elapsed/60.0)
write_message("%d records took %.1f seconds to complete.(%1.f recs/min)"\
% (done, time_elapsed, time_recs_per_min))
if time_recs_per_min:
write_message("Estimated runtime: %.1f minutes" % \
((todo-done)/time_recs_per_min))
def put(self, recID, word, sign):
"Adds/deletes a word to the word list."
try:
word = word[:50].lower()
if self.value.has_key(word):
# the word 'word' exist already: update sign
self.value[word][recID] = sign
else:
self.value[word] = {recID: sign}
except:
write_message("Error: Cannot put word %s with sign %d for recID %s." % (word, sign, recID))
def del_recIDs(self, recIDs):
"""Fetches records which id in the recIDs range list and adds
them to the wordTable. The recIDs range list is of the form:
[[i1_low,i1_high],[i2_low,i2_high], ..., [iN_low,iN_high]].
"""
count = 0
for arange in recIDs:
self.del_recID_range(arange[0],arange[1])
count = count + arange[1] - arange[0]
self.put_into_db()
def del_recID_range(self, low, high):
"""Deletes records with 'recID' system number between low
and high from memory words index table."""
write_message("%s fetching existing words for records #%d-#%d started" % \
(self.tablename, low, high), verbose=3)
self.recIDs_in_mem.append([low,high])
query = """SELECT id_bibrec,termlist FROM %sR as bb WHERE bb.id_bibrec
BETWEEN '%d' AND '%d'""" % (self.tablename[:-1], low, high)
recID_rows = run_sql(query)
for recID_row in recID_rows:
recID = recID_row[0]
wlist = deserialize_via_marshal(recID_row[1])
for word in wlist:
self.put(recID, word, -1)
write_message("%s fetching existing words for records #%d-#%d ended" % \
(self.tablename, low, high), verbose=3)
def report_on_table_consistency(self):
"""Check reverse words index tables (e.g. idxWORD01R) for
interesting states such as 'TEMPORARY' state.
Prints small report (no of words, no of bad words).
"""
# find number of words:
query = """SELECT COUNT(*) FROM %s""" % (self.tablename)
res = run_sql(query, None, 1)
if res:
nb_words = res[0][0]
else:
nb_words = 0
# find number of records:
query = """SELECT COUNT(DISTINCT(id_bibrec)) FROM %sR""" % (self.tablename[:-1])
res = run_sql(query, None, 1)
if res:
nb_records = res[0][0]
else:
nb_records = 0
# report stats:
write_message("%s contains %d words from %d records" % (self.tablename, nb_words, nb_records))
# find possible bad states in reverse tables:
query = """SELECT COUNT(DISTINCT(id_bibrec)) FROM %sR WHERE type <> 'CURRENT'""" % (self.tablename[:-1])
res = run_sql(query)
if res:
nb_bad_records = res[0][0]
else:
nb_bad_records = 999999999
if nb_bad_records:
write_message("EMERGENCY: %s needs to repair %d of %d records" % \
(self.tablename, nb_bad_records, nb_records))
else:
write_message("%s is in consistent state" % (self.tablename))
return nb_bad_records
def repair(self, opt_flush):
"""Repair the whole table"""
# find possible bad states in reverse tables:
query = """SELECT COUNT(DISTINCT(id_bibrec)) FROM %sR WHERE type <> 'CURRENT'""" % (self.tablename[:-1])
res = run_sql(query, None, 1)
if res:
nb_bad_records = res[0][0]
else:
nb_bad_records = 0
if nb_bad_records == 0:
return
query = """SELECT id_bibrec FROM %sR WHERE type <> 'CURRENT' ORDER BY id_bibrec""" \
% (self.tablename[:-1])
res = run_sql(query)
recIDs = create_range_list(res)
flush_count = 0
records_done = 0
records_to_go = 0
for arange in recIDs:
records_to_go = records_to_go + arange[1] - arange[0] + 1
time_started = time.time() # will measure profile time
for arange in recIDs:
i_low = arange[0]
chunksize_count = 0
while i_low <= arange[1]:
# calculate chunk group of recIDs and treat it:
i_high = min(i_low+opt_flush-flush_count-1,arange[1])
i_high = min(i_low+chunksize-chunksize_count-1, i_high)
try:
self.fix_recID_range(i_low, i_high)
except StandardError, e:
write_message("Exception caught: %s" % e, sys.stderr)
if task_get_option['verbose'] >= 9:
traceback.print_tb(sys.exc_info()[2])
task_update_status("ERROR")
self.put_into_db()
sys.exit(1)
flush_count = flush_count + i_high - i_low + 1
chunksize_count = chunksize_count + i_high - i_low + 1
records_done = records_done + i_high - i_low + 1
if chunksize_count >= chunksize:
chunksize_count = 0
# flush if necessary:
if flush_count >= opt_flush:
self.put_into_db("emergency")
self.clean()
flush_count = 0
self.log_progress(time_started,records_done,records_to_go)
# iterate:
i_low = i_high + 1
if flush_count > 0:
self.put_into_db("emergency")
self.log_progress(time_started,records_done,records_to_go)
write_message("%s inconsistencies repaired." % self.tablename)
def chk_recID_range(self, low, high):
"""Check if the reverse index table is in proper state"""
## check db
query = """SELECT COUNT(*) FROM %sR WHERE type <> 'CURRENT'
AND id_bibrec BETWEEN '%d' AND '%d'""" % (self.tablename[:-1], low, high)
res = run_sql(query, None, 1)
if res[0][0]==0:
write_message("%s for %d-%d is in consistent state"%(self.tablename,low,high))
return # okay, words table is consistent
## inconsistency detected!
write_message("EMERGENCY: %s inconsistencies detected..." % self.tablename)
write_message("""EMERGENCY: Errors found. You should check consistency of the %s - %sR tables.\nRunning 'bibindex --repair' is recommended.""" \
% (self.tablename, self.tablename[:-1]))
raise StandardError
def fix_recID_range(self, low, high):
"""Try to fix reverse index database consistency (e.g. table idxWORD01R) in the low,high doc-id range.
Possible states for a recID follow:
CUR TMP FUT: very bad things have happened: warn!
CUR TMP : very bad things have happened: warn!
CUR FUT: delete FUT (crash before flushing)
CUR : database is ok
TMP FUT: add TMP to memory and del FUT from memory
flush (revert to old state)
TMP : very bad things have happened: warn!
FUT: very bad things have happended: warn!
"""
state = {}
query = "SELECT id_bibrec,type FROM %sR WHERE id_bibrec BETWEEN '%d' AND '%d'"\
% (self.tablename[:-1], low, high)
res = run_sql(query)
for row in res:
if not state.has_key(row[0]):
state[row[0]]=[]
state[row[0]].append(row[1])
ok = 1 # will hold info on whether we will be able to repair
for recID in state.keys():
if not 'TEMPORARY' in state[recID]:
if 'FUTURE' in state[recID]:
if 'CURRENT' not in state[recID]:
write_message("EMERGENCY: Record %d is in inconsistent state. Can't repair it." % recID)
ok = 0
else:
write_message("EMERGENCY: Inconsistency in record %d detected" % recID)
query = """DELETE FROM %sR
WHERE id_bibrec='%d'""" % (self.tablename[:-1], recID)
run_sql(query)
write_message("EMERGENCY: Inconsistency in record %d repaired." % recID)
else:
if 'FUTURE' in state[recID] and not 'CURRENT' in state[recID]:
self.recIDs_in_mem.append([recID,recID])
# Get the words file
query = """SELECT type,termlist FROM %sR
WHERE id_bibrec='%d'""" % (self.tablename[:-1], recID)
write_message(query, verbose=9)
res = run_sql(query)
for row in res:
wlist = deserialize_via_marshal(row[1])
write_message("Words are %s " % wlist, verbose=9)
if row[0] == 'TEMPORARY':
sign = 1
else:
sign = -1
for word in wlist:
self.put(recID, word, sign)
else:
write_message("EMERGENCY: %s for %d is in inconsistent state. Couldn't repair it." % (self.tablename, recID))
ok = 0
if not ok:
write_message("""EMERGENCY: Unrepairable errors found. You should check consistency
of the %s - %sR tables. Deleting affected entries from these tables
is recommended.""" % (self.tablename, self.tablename[:-1]))
raise StandardError
def test_fulltext_indexing():
"""Tests fulltext indexing programs on PDF, PS, DOC, PPT,
XLS. Prints list of words and word table on the screen. Does not
integrate anything into the database. Useful when debugging
problems with fulltext indexing: call this function instead of main().
"""
print get_words_from_fulltext("http://doc.cern.ch/cgi-bin/setlink?base=atlnot&categ=Communication&id=com-indet-2002-012") # protected URL
print get_words_from_fulltext("http://doc.cern.ch/cgi-bin/setlink?base=agenda&categ=a00388&id=a00388s2t7") # XLS
print get_words_from_fulltext("http://doc.cern.ch/cgi-bin/setlink?base=agenda&categ=a02883&id=a02883s1t6/transparencies") # PPT
print get_words_from_fulltext("http://doc.cern.ch/cgi-bin/setlink?base=agenda&categ=a99149&id=a99149s1t10/transparencies") # DOC
print get_words_from_fulltext("http://doc.cern.ch/cgi-bin/setlink?base=preprint&categ=cern&id=lhc-project-report-601") # PDF
sys.exit(0)
def test_word_separators(phrase="hep-th/0101001"):
"""Tests word separating policy on various input."""
print "%s:" % phrase
for word in get_words_from_phrase(phrase):
print "\t-> %s" % word
def main():
"""Main that construct all the bibtask."""
task_set_option('cmd', 'add')
task_set_option('id', [])
task_set_option("modified", [])
task_set_option("collection", [])
task_set_option("maxmem", 0)
task_set_option("flush", 10000)
task_set_option("windex", get_word_tables(None))
task_init(authorization_action='runbibindex',
authorization_msg="BibIndex Task Submission",
description="""Examples:
\t%s -a -i 234-250,293,300-500 -u admin@localhost
\t%s -a -w author,fulltext -M 8192 -v3
\t%s -d -m +4d -A on --flush=10000\n""" % ((sys.argv[0],) * 3), help_specific_usage=""" Indexing options:
-a, --add\t\tadd or update words for selected records
-d, --del\t\tdelete words for selected records
-i, --id=low[-high]\t\tselect according to doc recID
-m, --modified=from[,to]\tselect according to modification date
-c, --collection=c1[,c2]\tselect according to collection
Repairing options:
-k, --check\t\tcheck consistency for all records in the table(s)
-r, --repair\t\ttry to repair all records in the table(s)
Specific options:
-w, --windex=w1[,w2]\tword/phrase indexes to consider (all)
-M, --maxmem=XXX\tmaximum memory usage in kB (no limit)
-f, --flush=NNN\t\tfull consistent table flush after NNN records (10000)
""",
specific_params=("adi:m:c:w:krM:f:", [
"add",
"del",
"id=",
"modified=",
"collection=",
"windex=",
"check",
"repair",
"maxmem=",
"flush=",
]),
task_stop_helper_fnc=task_stop_table_close_fnc,
task_submit_elaborate_specific_parameter_fnc=task_submit_elaborate_specific_parameter,
task_run_fnc=task_run_core)
def task_submit_elaborate_specific_parameter(key, value, opts, args):
""" Given the string key it checks it's meaning, eventually using the
value. Usually it fills some key in the options dict.
It must return True if it has elaborated the key, False, if it doesn't
know that key.
eg:
if key in ['-n', '--number']:
self.options['number'] = value
return True
return False
"""
if key in ("-a", "--add"):
task_set_option("cmd", "add")
if ("-x","") in opts or ("--del","") in opts:
raise StandardError, "Can not have --add and --del at the same time!"
elif key in ("-k", "--check"):
task_set_option("cmd", "check")
elif key in ("-r", "--repair"):
task_set_option("cmd", "repair")
elif key in ("-d", "--del"):
task_set_option("cmd", "del")
elif key in ("-i", "--id"):
task_set_option('id', task_get_option('id') + split_ranges(value))
elif key in ("-m", "--modified"):
task_set_option("modified", get_date_range(value))
elif key in ("-c", "--collection"):
task_set_option("collection", value)
elif key in ("-w", "--windex"):
task_set_option("windex", get_word_tables(value))
elif key in ("-M", "--maxmem"):
task_set_option("maxmem", int(value))
if task_get_option("maxmem") < base_process_size + 1000:
raise StandardError, "Memory usage should be higher than %d kB" % \
(base_process_size + 1000)
elif key in ("-f", "--flush"):
task_set_option("flush", int(value))
else:
return False
return True
def task_stop_table_close_fnc():
""" Close tables to STOP. """
global _last_word_table
if _last_word_table:
_last_word_table.put_into_db()
def task_run_core():
"""Runs the task by fetching arguments from the BibSched task queue. This is
what BibSched will be invoking via daemon call.
The task prints Fibonacci numbers for up to NUM on the stdout, and some
messages on stderr.
Return 1 in case of success and 0 in case of failure."""
global _last_word_table
if task_get_option("cmd") == "check":
for table in task_get_option("windex"):
wordTable = WordTable(table.keys()[0], table.values()[0])
_last_word_table = wordTable
wordTable.report_on_table_consistency()
_last_word_table = None
return True
for table in task_get_option("windex"):
wordTable = WordTable(table.keys()[0], table.values()[0])
_last_word_table = wordTable
wordTable.report_on_table_consistency()
try:
if task_get_option("cmd") == "del":
if task_get_option("id"):
wordTable.del_recIDs(task_get_option("id"))
elif task_get_option("collection"):
l_of_colls = task_get_option("collection").split(",")
recIDs = perform_request_search(c=l_of_colls)
recIDs_range = []
for recID in recIDs:
recIDs_range.append([recID,recID])
wordTable.del_recIDs(recIDs_range)
else:
write_message("Missing IDs of records to delete from index %s." % wordTable.tablename,
sys.stderr)
raise StandardError
elif task_get_option("cmd") == "add":
if task_get_option("id"):
wordTable.add_recIDs(task_get_option("id"), task_get_option("flush"))
elif task_get_option("collection"):
l_of_colls = task_get_option("collection").split(",")
recIDs = perform_request_search(c=l_of_colls)
recIDs_range = []
for recID in recIDs:
recIDs_range.append([recID,recID])
wordTable.add_recIDs(recIDs_range, task_get_option("flush"))
else:
wordTable.add_recIDs_by_date(task_get_option("modified"), task_get_option("flush"))
# only update last_updated if run via automatic mode:
wordTable.update_last_updated(task_get_task_param('task_starting_time'))
elif task_get_option("cmd") == "repair":
wordTable.repair(task_get_option("flush"))
else:
write_message("Invalid command found processing %s" % \
wordTable.tablename, sys.stderr)
raise StandardError
except StandardError, e:
write_message("Exception caught: %s" % e, sys.stderr)
if task_get_option("verbose") >= 9:
traceback.print_tb(sys.exc_info()[2])
task_update_status("ERROR")
if _last_word_table:
_last_word_table.put_into_db()
sys.exit(1)
wordTable.report_on_table_consistency()
_last_word_table = None
return True
### okay, here we go:
if __name__ == '__main__':
main()
c4science · Help