def read_label(idx):
gatt = read_yaml("../../data/label/applicability/labels/GATT.yaml")
gatt = cleanse_dict(gatt)
inv_gatt = invert_dict(gatt)
print(len(inv_gatt.keys()))
print(sorted(inv_gatt.keys()))
return inv_gatt[idx]
def rehash_arts_in_text(indices,
yaml_path=
"labels/GATT.yaml"):
"""
From the integer list of labels, change the int into article name
e/g : [0,2] -> ['Article I', 'Article II']
"""
gatt_info = read_yaml(yaml_path)
gatt_dict = cleanse_dict(gatt_info)
gatt_keys_in_int = get_keys(gatt_dict)
in_text = []
for index in indices:
in_text.append(gatt_keys_in_int[index])
return in_text
def filter_panel_eng(list_of_urls, idx):
"""
Filter panel report's urls on given list of urls
:param list_of_urls
:return: one url that of panel report
"""
def _find_panel(urls, ds_idx):
panel_url = None
print("Panel reports: ")
snippet = None
for url in filter_eng(urls):
print("url", url.split("/")[-1])
if 'Q' in url and 'WT' in url:
snippet = url
if 'q' in url and 'WT' in url:
snippet = url
if "{}R.pdf".format(ds_idx) == url.split("/")[-1]:
panel_url = url
break
if "{}R.PDF".format(ds_idx) == url.split("/")[-1]:
panel_url = url
break
if "{}RW.pdf".format(ds_idx) == url.split("/")[-1]:
panel_url = url
break
if "{}R-00.pdf".format(ds_idx) == url.split("/")[-1]:
panel_url = url
break
if "{}R-01.pdf".format(ds_idx) == url.split("/")[-1]:
panel_url = url
break
if not snippet:
snippet = url
print("urlSnippet", snippet)
print("panel_url", panel_url)
return panel_url, snippet
panelURL, urlSNIPPET = _find_panel(list_of_urls, idx)
if not panelURL:
group = None
info = read_yaml("../../data/info.yaml")
links = info['LinkedPanel']
for link in links:
if idx in link:
group = list(link.keys())
print("group", group)
break
if group is not None:
for one in group:
print("linked", one)
panelURL, url_snippet = _find_panel(list_of_urls, one)
print("newly found panelURL", panelURL)
if panelURL:
break
if not panelURL:
panelURL = os.path.join('/'.join(urlSNIPPET.split('/')[:-1]),
'{}R.pdf'.format(idx))
print("final_decision: ", panelURL)
return panelURL
from prep.factual.after_panel.extract import extract_factual_manual, locate_chapter_II, locate_chapter_III
def read_label(idx):
gatt = read_yaml("../../data/label/applicability/labels/GATT.yaml")
gatt = cleanse_dict(gatt)
inv_gatt = invert_dict(gatt)
print(len(inv_gatt.keys()))
print(sorted(inv_gatt.keys()))
return inv_gatt[idx]
if __name__ == "__main__":
pkl_path = "../../data/factual/after_panel/factual.pkl"
info = read_yaml("../../data/info.yaml")
panel_exist = info['Panel']['ds_numb']
mutual_agree = info['Panel']['mutual_agree']
WPF = info['Panel']['WPF']
print(WPF)
LinkedPanel = info['LinkedPanel']
LinkedOmit = info['LinkedOmit']
for idx in panel_exist:
if check_already_exist(pkl_path, idx):
continue
elif idx in mutual_agree:
continue
elif idx in WPF:
print(idx, "in WPF")
continue
elif idx in LinkedOmit:
"""Convert Multi Label Dataset to One-Label"""
import os
import json
from utils.misc.yaml import read_yaml
from utils.misc.dict import get_keys
from utils.misc.json import write_json_line_by_line
from prep.label.cite.parse import cleanse_dict
print(os.getcwd())
multi_label_json = "entire_data.json"
provision_path = "/home/zachary/projects/DeepWTO/data/provision/gatt.yaml"
provision = read_yaml(provision_path)
provision_names = list(provision.keys())
print(provision_names)
print(len(provision_names))
print(provision[provision_names[0]])
gatt = read_yaml("/home/zachary/projects/DeepWTO/data/label/citability/labels"
"/GATT.yaml")
gatt = cleanse_dict(gatt)
gatt_keys = get_keys(gatt)
print(gatt_keys)
art_id = gatt_keys.index("Article I")
print(art_id)
print(provision_names == gatt_keys)
from utils.misc.yaml import read_yaml
read = read_yaml("notable.yaml")
ds_numbs = set()
titles = ['III', 'III:1', 'III:2', 'III:4', 'III:5', 'III:7']
for title in titles:
ds_numbs = ds_numbs.union(set(read[title].split(',')))
ds_numbs = sorted(list(ds_numbs))
print(len(ds_numbs))
print(ds_numbs)
ds_numbs_wo_DS_str = []
for ds_numb in ds_numbs:
if ' ' in ds_numb:
ds_numb = ds_numb[1:]
print(ds_numb)
ds_numbs_wo_DS_str.append(int(ds_numb[2:]))
comparee = read_yaml("../../info.yaml")
comparee = comparee['Panel']['ds_numb']
joint = []
for elem in ds_numbs_wo_DS_str:
if elem in comparee:
joint.append(elem)
if __name__ == "__main__":
print(sorted(set(joint)))
print(len(sorted(set(joint))))
write_yaml(cited, "./stat.yaml")
def sort_yaml(yaml_path):
"""
Read yaml path then sort it by value, then print.
:param yaml_path:
:return: None. only prints.
""
"""
stream = open(yaml_path, "r")
docs = yaml.load_all(stream)
result = dict()
for doc in docs:
for k, v in doc.items():
result[k] = v
sorted_result = sorted(result.items(), key=lambda kv: kv[1], reverse=True)
print(sorted_result)
if __name__ == "__main__":
# main()
# sort_yaml("stat.yaml")
info = read_yaml("../info.yaml")
panel = info['Panel']['ds_numb']
appellate = info['AppellateBody']['ds_numb']
linked = info['LinkedPanel']
print(linked)
def query(article, yaml_path):
data = read_yaml(yaml_path)
print(data[article])
"labels/GATT.yaml"):
"""
From the integer list of labels, change the int into article name
e/g : [0,2] -> ['Article I', 'Article II']
"""
gatt_info = read_yaml(yaml_path)
gatt_dict = cleanse_dict(gatt_info)
gatt_keys_in_int = get_keys(gatt_dict)
in_text = []
for index in indices:
in_text.append(gatt_keys_in_int[index])
return in_text
if __name__ == "__main__":
gatt = read_yaml("labels/GATT.yaml")
gatt = cleanse_dict(gatt)
print(gatt)
gatt_keys = get_keys(gatt)
print(gatt_keys)
for key in gatt_keys:
print(key)
inv_gatt = invert_dict(gatt)
inv_gatt_keys = sorted(list(inv_gatt.keys()))
print(inv_gatt)
print(inv_gatt_keys)
print(len(inv_gatt_keys))
print(rehash_arts_in_text([0, 1]))
# Dump
# path_to_dump = "../../dataset/citability/GATT_523/label.pkl"
from utils.misc.yaml import read_yaml
read = read_yaml("labels/AD.yaml")
if __name__ == "__main__":
new = dict()
keys = read.keys()
for key in keys:
value = []
dss = read[key].split(', ')
for ds in dss:
numb_only = ds[2:]
value.append(numb_only)
new[key] = value
print(new.keys())
print(new['Article XXIV:12'][-1])
pass