41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206 | class DocumentPicker:
"""Pick documents from samediggi.se to be added to the corpus."""
def __init__(self, source_dir):
self.freecorpus = os.getenv("GTFREE")
self.source_dir = source_dir
self.file_dict = {}
self.file_dict.setdefault("sma", [])
self.file_dict.setdefault("sme", [])
self.file_dict.setdefault("smj", [])
self.file_dict.setdefault("swe", [])
self.file_dict.setdefault("none", [])
self.parallel_dict = {}
self.total_file = 0
def classify_files(self):
"""Iterate through all files, classify them according to language"""
for root, dirs, files in os.walk(self.source_dir):
for f in files:
if f.endswith(".html"):
self.total_file += 1
self.classify_file(os.path.join(root, f))
def get_parallel_name(self, file_, a):
dirname = os.path.dirname(file_)
href = a.get("href").replace("http://www.samediggi.se/", "")
if ".." in href:
dirname_parts = dirname.split("/")
href_parts = href.split("/")
dirname = "/".join(dirname_parts[:-1])
href = "/".join(href_parts[1:])
return os.path.join(dirname, href)
def append_parallel(self, file_, a):
if os.path.exists(self.get_parallel_name(file_, a)):
self.parallel_dict[file_].append(self.get_parallel_name(file_, a))
else:
print(
util.lineno(),
self.get_parallel_name(file_, a),
"does not exist",
a.get("title"),
file_,
file=sys.stderr,
)
def get_parallels(self, a, file_):
self.parallel_dict.setdefault(file_, [])
prev = a.getprevious()
next_ = a.getnext()
while prev is not None:
if prev != a:
self.append_parallel(file_, prev)
prev = prev.getprevious()
while next_ is not None:
if next_ != a:
self.append_parallel(file_, next_)
next_ = next_.getnext()
def append_file(self, language, file_):
self.file_dict[language].append(file_)
self.file_dict["none"].remove(file_)
def classify_file(self, file_):
"""Identify the language of the file"""
parser = etree.HTMLParser()
html = etree.parse(file_, parser)
self.file_dict["none"].append(file_)
for img in html.iter("img"):
if img.get("src") is not None and "icon_flag_sme_dim.gif" in img.get("src"):
self.append_file("sme", file_)
self.get_parallels(img.getparent(), file_)
elif img.get("src") is not None and "icon_flag_smj_dim.gif" in img.get(
"src"
):
self.append_file("smj", file_)
self.get_parallels(img.getparent(), file_)
elif img.get("src") is not None and "icon_flag_sma_dim.gif" in img.get(
"src"
):
self.append_file("sma", file_)
self.get_parallels(img.getparent(), file_)
elif img.get("src") is not None and "icon_flag_swe_dim.gif" in img.get(
"src"
):
self.append_file("swe", file_)
self.get_parallels(img.getparent(), file_)
def conclude(self):
total = 0
for key, value in self.file_dict.items():
total += len(self.file_dict[key])
print(key, len(self.file_dict[key]))
print(total, self.total_file)
def check_consistency(self):
"""Check if all files that claim to have parallels actually exist
Remove the parallel file from the list of parallels
"""
for file_ in self.parallel_dict:
for parallel_file in self.parallel_dict[file_]:
try:
self.parallel_dict[parallel_file]
if file_ not in self.parallel_dict[parallel_file]:
self.parallel_dict[file_].remove(parallel_file)
except KeyError:
self.parallel_dict[file_].remove(parallel_file)
def get_goal_name(self, file_, lang):
filename = os.path.basename(file_)
goalfile = os.path.join(
self.freecorpus, "orig", lang, "admin", "sd", "www.samediggi.se", filename
)
return goalfile
def set_metadata(self, file_, lang):
mh = xslsetter.MetadataHandler(
self.get_goal_name(file_, lang) + ".xsl", create=True
)
for key, value in self.set_variables(file_, lang).items():
mh.set_variable(key, value)
mh.write_file()
def set_variables(self, file_, lang):
variables = {}
variables["filename"] = "http://" + file_.replace(".html", "")
variables["license_type"] = "free"
variables["sub_name"] = "Børre Gaup"
variables["sub_email"] = "borre.gaup@uit.no"
variables["mainlang"] = lang
variables["monolingual"] = "1"
if lang != "swe":
variables["translated_from"] = "swe"
if self.parallel_dict[file_]:
variables["parallel_texts"] = "1"
para_langs = ["sma", "sme", "smj", "swe"]
para_langs.remove(lang)
for parallel_file in self.parallel_dict[file_]:
for para_lang in para_langs:
if parallel_file in self.file_dict[para_lang]:
variables["para_" + para_lang] = os.path.basename(parallel_file)
return variables
def move_swe_file(self, file_):
for candidate in self.parallel_dict[file_]:
if candidate in self.file_dict["swe"]:
shutil.copy(candidate, self.get_goal_name(candidate, "swe"))
self.set_metadata(candidate, "swe")
def move_files_set_metadata(self):
for lang in ["sma", "smj", "sme"]:
for file_ in self.file_dict[lang]:
shutil.copy(file_, self.get_goal_name(file_, lang))
self.move_swe_file(file_)
self.set_metadata(file_, lang)
|