def get_data(self, company_id):
table_match_dict = robj.processdelta(company_id)
deal_id = tuple(company_id.split('_'))
import read_slt_info as read_slt_info
upObj = read_slt_info.update(deal_id)
docname_docid_lst = upObj.get_documentName_id()
ph_doc_info_dict = {}
doc_ph_info = {}
for (doc_name, doc_id) in docname_docid_lst:
doc_sp_lst = doc_name.split('_')
ph = ''
if len(doc_sp_lst) == 2:
ph = doc_sp_lst[-1]
if len(doc_sp_lst) == 4:
ph = doc_sp_lst[-2] + doc_sp_lst[-1]
ph = ph.replace('AR', 'FY')
ph_doc_info_dict[ph] = (ph, doc_name, doc_id)
doc_ph_info[doc_id] = ph
sorted_ph_lst = report_year_sort.year_sort(ph_doc_info_dict.keys())
len_phs = len(sorted_ph_lst)
doc_sorted_lst = [ph_doc_info_dict[ph][-1] for ph in sorted_ph_lst]
#print sorted_ph_lst
#print doc_sorted_lst
#sys.exit()
new_table_match_dict = {}
for doc_pair, pair_ars in table_match_dict.items():
i1 = doc_sorted_lst.index(doc_pair[0])
i2 = doc_sorted_lst.index(doc_pair[1])
print 'doc_pair: ', doc_pair
if (i1 < i2):
new_table_match_dict[doc_pair] = pair_ars[:]
else:
print 'rev: ', doc_pair, ' == ', i1, i2
new_pair_ars = []
for pair in pair_ars:
new_pair_ars.append((pair[1], pair[0]))
new_table_match_dict[(doc_pair[1],
doc_pair[0])] = new_pair_ars[:]
fname = os.path.join(self.output_path, company_id + '_table_delta.slv')
sh = shelve.open(fname, 'n')
sh['data'] = new_table_match_dict
sh.close()
return
def get_doc_pairs(self, company_id):
deal_id = tuple(company_id.split('_'))
import read_slt_info as read_slt_info
upObj = read_slt_info.update(deal_id)
docname_docid_lst = upObj.get_documentName_id()
ph_doc_info_dict = {}
doc_ph_info = {}
#print docname_docid_lst
#sys.exit()
for (doc_name, doc_id) in docname_docid_lst:
doc_sp_lst = doc_name.split('_')
ph = ''
if len(doc_sp_lst) == 2:
ph = doc_sp_lst[-1]
if len(doc_sp_lst) == 4:
ph = doc_sp_lst[-2] + doc_sp_lst[-1]
ph = ph.replace('AR', 'FY')
ph_doc_info_dict[ph] = doc_id
doc_ph_info[doc_id] = ph
lmdb_folder = os.path.join(self.output_path, company_id)
if not os.path.exists(lmdb_folder):
os.mkdir(lmdb_folder)
sorted_ph_lst = report_year_sort.year_sort(ph_doc_info_dict.keys())
doc_pair_dict = {}
for i, ph in enumerate(sorted_ph_lst):
qh_flg1 = ph[:2]
doc_id1 = ph_doc_info_dict[ph]
for kph in sorted_ph_lst[i + 1:]:
qh_flg2 = kph[:2]
doc_id2 = ph_doc_info_dict[kph]
if qh_flg1 == qh_flg2:
if (doc_id1, doc_id2) not in doc_pair_dict:
doc_pair_dict[(doc_id1, doc_id2)] = 1
break
return doc_pair_dict.keys()
def get_sorted_doc_list(self, company_id):
delta_table_match_dict = self.read_delta_data(company_id)
#print delta_table_match_dict.keys()
#sys.exit()
deal_id = tuple(company_id.split('_'))
import read_slt_info as read_slt_info
upObj = read_slt_info.update(deal_id)
docname_docid_lst = upObj.get_documentName_id()
ph_doc_info_dict = {}
doc_ph_info = {}
for (doc_name, doc_id) in docname_docid_lst:
doc_sp_lst = doc_name.split('_')
ph = ''
if len(doc_sp_lst) == 2:
ph = doc_sp_lst[-1]
if len(doc_sp_lst) == 4:
ph = doc_sp_lst[-2] + doc_sp_lst[-1]
ph = ph.replace('AR', 'FY')
ph_doc_info_dict[ph] = (ph, doc_name, doc_id)
doc_ph_info[doc_id] = ph
lmdb_folder = os.path.join(self.output_path, company_id)
if not os.path.exists(lmdb_folder):
os.mkdir(lmdb_folder)
dfname = os.path.join(lmdb_folder, 'doc_ph_info')
self.lmdb_obj.write_to_lmdb(dfname, doc_ph_info, doc_ph_info.keys())
sorted_ph_lst = report_year_sort.year_sort(ph_doc_info_dict.keys())
len_phs = len(sorted_ph_lst)
doc_sorted_lst = [ph_doc_info_dict[ph][-1] for ph in sorted_ph_lst]
sorted_combination_doc_lst = []
for i, ph in enumerate(sorted_ph_lst):
map_doc_tup = doc_sorted_lst[i]
if (ph[:2] != 'FY') and (i < len(sorted_ph_lst) - 2):
next_ph = sorted_ph_lst[i + 1]
if ph[:1] in ['Q', 'H'] and next_ph[:1] in ['F']: continue
next_doc_tup = doc_sorted_lst[i + 1]
sorted_combination_doc_lst.append((map_doc_tup, next_doc_tup))
elif (ph[:2] == 'FY') and (i < len(sorted_ph_lst) - 2):
next_fy = ph[:2] + str(int(ph[2:]) + 1)
if next_fy not in sorted_ph_lst: continue
next_fy_indx = sorted_ph_lst.index(next_fy)
next_doc_tup = doc_sorted_lst[next_fy_indx]
sorted_combination_doc_lst.append((map_doc_tup, next_doc_tup))
######################################################################
project_id, url_id = company_id.split('_')
norm_res_list = sObj.slt_normresids(project_id, url_id)
doc_page_dict = {}
for doc_tup in norm_res_list:
doc_id, page_number, norm_table_id = doc_tup
if doc_id not in doc_page_dict:
doc_page_dict[doc_id] = {}
if page_number not in doc_page_dict[doc_id]:
doc_page_dict[doc_id][page_number] = []
doc_page_dict[doc_id][page_number].append(norm_table_id)
cache_xml_ids = {}
val_cons_dict = {}
doc_pair_table_pair_dict = {}
doc_id_pairs = delta_table_match_dict.keys()
for doc_pair in doc_id_pairs:
hyp_doc, ref_doc = doc_pair
hyf_ref_lst = delta_table_match_dict[doc_pair]
for (hyp_list, ref_list) in hyf_ref_lst:
pg1 = hyp_list[0].split('_')[-1]
pg2 = ref_list[0].split('_')[-1]
#print doc_pair, pg1, pg2, hyp_list, ref_list
if (pg2 not in doc_page_dict[doc_pair[1]]) or (
pg1 not in doc_page_dict[doc_pair[0]]):
continue
norm_table_id_hyps = doc_page_dict[doc_pair[0]][pg1]
norm_table_id_refs = doc_page_dict[doc_pair[1]][pg2]
selected_hyp = ''
if (len(norm_table_id_hyps) == 1):
selected_hyp = norm_table_id_hyps[0]
else:
n_hyp_list = []
for r in hyp_list:
n_hyp_list += r.split('#')
for norm_table_id_hyp in norm_table_id_hyps:
if (project_id, url_id,
norm_table_id_hyp) in cache_xml_ids:
xmlids1 = cache_xml_ids[(project_id, url_id,
norm_table_id_hyp)]
else:
xmlids1 = self.get_xml_id(project_id, url_id,
norm_table_id_hyp)
cache_xml_ids[(project_id, url_id,
norm_table_id_hyp)] = xmlids1
s1 = sets.Set(xmlids1).intersection(
sets.Set(n_hyp_list))
if list(s1):
selected_hyp = norm_table_id_hyp
break
selected_ref = ''
if (len(norm_table_id_refs) == 1):
selected_ref = norm_table_id_refs[0]
else:
n_ref_list = []
for r in ref_list:
n_ref_list += r.split('#')
for norm_table_id_ref in norm_table_id_refs:
if (project_id, url_id,
norm_table_id_ref) in cache_xml_ids:
xmlids1 = cache_xml_ids[(project_id, url_id,
norm_table_id_ref)]
else:
xmlids1 = self.get_xml_id(project_id, url_id,
norm_table_id_ref)
cache_xml_ids[(project_id, url_id,
norm_table_id_ref)] = xmlids1
s1 = sets.Set(xmlids1).intersection(
sets.Set(n_ref_list))
if list(s1):
selected_ref = norm_table_id_ref
break
#print ' doc_pair: ', doc_pair
#print
if (not selected_hyp) or (not selected_ref):
print 'mmmm Learning Error....'
continue
#sys.exit()
if selected_hyp and selected_ref:
#print [selected_hyp, selected_ref]
if (doc_pair[0],
selected_hyp) not in doc_pair_table_pair_dict:
doc_pair_table_pair_dict[(doc_pair[0],
selected_hyp)] = []
#if selected_hyp not in doc_pair_table_pair_dict[doc_pair]:
# doc_pair_table_pair_dict[doc_pair][selected_hyp] = []
if (doc_pair[1],
selected_ref) not in doc_pair_table_pair_dict[(
doc_pair[0], selected_hyp)]:
doc_pair_table_pair_dict[(doc_pair[0],
selected_hyp)].append(
(doc_pair[1],
selected_ref))
val_cons_dict[(doc_pair[1], selected_ref)] = 1
init_values = []
for k, vs in doc_pair_table_pair_dict.items():
if k not in val_cons_dict:
init_values.append([k])
#print len(init_values)
#sys.exit()
flg = 1
while flg:
flg = 0
new_init_values = []
for init_val in init_values:
last_key = init_val[-1]
if last_key in doc_pair_table_pair_dict:
flg = 1
extended_pos = doc_pair_table_pair_dict[last_key]
print ' extended_pos: ', extended_pos, len(extended_pos)
new_init_val = init_val[:]
for e in extended_pos:
new_init_values.append(init_val[:] + [e])
print len(init_val[:]) + 1, ' === ALL: ', len_phs
else:
new_init_values.append(init_val[:])
init_values = new_init_values[:]
'''
new_init_values = []
for init_value in init_values:
l1 = map(lambda x:x[0]+'#'+x[1], init_value[:])
s1 = sets.Set(l1)
flg = 0
for init_value1 in init_values:
l2 = map(lambda x:x[0]+'#'+x[1], init_value1[:])
s2 = sets.Set(l2)
if (s1 == s2): continue
if s1.issubset(s2):
flg = 1
break
if (flg == 0):
new_init_values.append(init_value)
'''
ar = []
for init_value in init_values:
ar.append((len(init_value), init_value))
ar.sort()
ar.reverse()
print 'Total ar: ', len(ar)
for ar_elm in ar:
print 'Len: ', ar_elm[0], 'ELMS: ', ar_elm[1]
#sys.exit()
ofname = os.path.join(lmdb_folder, 'doc_table_final_chain_pair')
final_pair_dict = {
'sorted_comb_list': sorted_combination_doc_lst,
'doc_table_chain_pair_list': new_init_values
}
self.lmdb_obj.write_to_lmdb(ofname, final_pair_dict,
final_pair_dict.keys())
